Upload
others
View
7
Download
0
Embed Size (px)
Citation preview
Advanced Electronics for
Fourier-Transform Ion Cyclotron
Resonance Mass Spectrometry
by
Tzu-Yung Lin
A thesis submitted in partial fulfilment of the requirements for the degree of
Doctor of Philosophy in Engineering
School of Engineering
November 2012
For my grandpa, Mong.
Acknowledgements
There are many people who have helped me along the way toward the comple-
tion of this work. Especially, I would like to thank my supervisors, Prof. Peter
B. O’Connor (Department of Chemistry) and Prof. Roger J. Green (School of
Engineering), for their patient guidance, support, and their valuable time. They
offered a unique opportunity for me to look at the mass spectrometry instrumen-
tation problems from an engineering point of view. Furthermore, the relocation
with Prof. O’Connor extends my experience to the magnificent cultures of the
British Isles and continental Europe.
Meanwhile, Prof. Ronald W. Knepper (Department of Electrical and Com-
puter Engineering) and Prof. Catherine E. Costello (Cardiovascular Proteomics
Center) from Boston University, Massachusetts, USA, provided useful sugges-
tions and support, especially during the time when I was conducting research
within Boston University. Useful discussions from other Boston University mem-
bers, such as Dr. Konstantine Aizikov, Dr. Raman Mathur, Dr. Cheng Lin, Dr.
Weidong Cui, Dr. Chunxiang Yao, Dr. Xiaojuan Li, and Dr. Nadezda Sargaeva
iii
are also acknowledged.
My appreciation is also extended to staff and student members here at the
University of Warwick, Coventry, UK. Dr. David P. A. Kilgour helped with
the generation of the data used to plot the Mathieu equation stability diagram,
and shared much knowledge about quadrupole/ion trap mass spectrometry. Dr.
Mark P. Barrow and Dr. Alex W. Colburn provided useful information and sug-
gestions about the research field of mass spectrometry and the related electronics.
Huilin Li helped with the interpretation of the IR-ECD spectra, and allowed me
to participate in and to learn from some of her proteomics research projects.
Rod Wesson kindly suggested some design ideas for the quadrupole power sup-
ply. He also helped with some of the circuit drawings and the making of the
power supply printed circuit boards. Ian Griffith helped with the manufactur-
ing of the preamplifier boards. Robert Day provided insight for the preamplifier
stabilization. There were also many helpful discussions between the members of
the Ion Cyclotron Resonance Laboratory. They are Rebecca H. Wills, Yulin Qi,
Pilar Perez Hurtado, Andrea F. Lopez-Clavijo, Andrew Soulby, Juan Wei, and
Samantha L. Benson.
Some of the conversations during the conferences were inspiring too. Espe-
cially, I would like to thank Prof. Ron M. A. Heeren, Prof. I. Jon Amster and Dr.
Don Rempel for their helpful discussions in the BMSS (British Mass Spectrometry
Society) 3-Day Meeting, the EFTMS (European Fourier Transform Mass Spec-
trometry) Conference, and the ASMS (American Society for Mass Spectrometry)
iv
Conference on Mass Spectrometry and Allied Topics.
Finally, I’d like to emphasize on my special thanks to my family and my
friends, especially my grandparents, my parents, my brothers, and Joanna, with
their great support throughout my PhD experience. The financial support from
Department of Chemistry, Warwick Centre for Analytical Science (EPSRC (Engi-
neering and Physical Sciences Research Council): EP/F034210/1), School of En-
gineering, and NIH (National Institutes of Health) (NIH/NCRR-P41 RR10888;
NIH/NIGMS-R01GM078293) are also gratefully appreciated.
v
Abstract
With the development of mass spectrometry (MS) instruments starting in the
late 19th century, more and more research emphasis has been put on MS related
subjects, especially the instrumentation and its applications. Instrumentation
research has led modern mass spectrometers into a new era where the MS per-
formance, such as resolving power and mass accuracy, is close to its theoretical
limit. Such advanced performance releases more opportunities for scientists to
conduct analytical research that could not be performed before.
This thesis reviews general MS history and some of the important mile-
stones, followed by introductions to ion cyclotron resonance (ICR) technique
and quadrupole operation. Existing electronic designs, such as Fourier-transform
ion cyclotron resonance (FT-ICR) preamplifiers (for ion signal detection) and
radio-frequency (RF) oscillators (for ion transportation/filtering) are reviewed.
Then the potential scope for improvement is discussed.
Two new FT-ICR preamplifiers are reported; both preamplifiers operate at
room temperature. The first preamplifier uses an operational amplifier (op amp)
vi
in a transimpedance configuration. When a 18-kΩ feedback resistor is used, this
preamplifier delivers a transimpedance of about 85 dBΩ, and an input current
noise spectral density of around 1 pA/√
Hz. The total power consumption of
this circuit is around 310 mW when tested on the bench. This preamplifier has a
bandwidth of ∼3 kHz to 10 MHz, which corresponds to the mass-to-charge ratio,
m/z, of approximately 18 to 61k at 12 T for FT-ICR MS. The transimpedance
and the bandwidth can be adjusted by replacing passive components such as the
feedback resistor and capacitor. The feedback and bandwidth limitation of the
circuit is also discussed. When using an 0402 type surface mount resistor, the
maximum possible transimpedance, without sacrificing its bandwidth, is approx-
imated to 5.3 MΩ. Under this condition, the preamplifier is estimated to be able
to detect ∼110 charges.
The second preamplifier employs a single-transistor design using a different
feedback arrangement, a T-shaped feedback network. Such a feedback system
allows ∼100-fold less feedback resistance at a given transimpedance, hence pre-
serving bandwidth, which is beneficial to applications demanding high gain. The
single-transistor preamplifier yields a low power consumption of ∼5.7 mW, and
a transimpedance of 80 dBΩ in the frequency range between 1 kHz and 1 MHz
(m/z of around 180 to 180k for a 12-T FT-ICR system). In trading noise per-
formance for higher transimpedance, an alternative preamplifier design has also
been presented with a transimpedance of 120 dBΩ in the same frequency range.
The previously reported room-temperature FT-ICR preamplifier had a volt-
vii
age gain of about 25, a bandwidth of around 1 MHz when bench tested, and
a voltage noise spectral density of ∼7.4 nV/√
Hz. The bandwidth performance
when connecting this preamplifier to an ICR cell has not been reported. However,
from the transimpedance theory, the transimpedance preamplifiers reported in
this work will have a bandwidth wider by a factor of the open-loop gain of the
amplifier.
In a separate development, an oscillator is proposed as a power supply for
a quadrupole mass filter in a mass spectrometer system. It targets a stabilized
output frequency, and a feedback control for output amplitude stabilization. The
newly designed circuit has a very stable output frequency at 1 MHz, with a fre-
quency tolerance of 15 ppm specified by the crystal oscillator datasheet. Within
this circuit, an automatic gain control (AGC) unit is built for output amplitude
stabilisation. A new transformer design is also proposed. The dimension of the
quadrupole being used as a mass filter will be determined in the future. This
circuit (in particular the transformer and the quadrupole connection/mounting
device) will be finalised after the design of the quadrupole.
Finally, this thesis concludes with a discussion between the gain and the noise
performance of an FT-ICR preamplifier. A brief analysis about the correlation
between the gain, cyclotron frequency, and input capacitance is performed. Fu-
ture work is also suggested for extending this research.
viii
List of Awards, Publications, &
Presentations
Awards
Thales Sponsorship for ICTON 2012 for Joint Best Research Student Paper
from the University of Warwick, July 2012.
School of Engineering Travel Bursary, University of Warwick, February 2012.
BMSS Travel Grants (twice), British Mass Spectrometry Society,
August 2011, & February 2012.
Patent
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Transimpedance am-
plifier. University of Warwick. UK Patent Application GB1205775.8, 30 March
2012.
Journal/Conference Papers
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. A low noise single-
transistor transimpedance preamplifier for Fourier-transform mass spectrometry
ix
using a T feedback network. Review of Scientific Instruments. 83(9):094102,
2012.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Use of optical receiver
circuit design techniques in a transimpedance preamplifier for high performance
mass spectrometry. In 14th International Conference on Transparent Optical
Networks (ICTON), pages 1–4. Coventry, UK, 2–5 July 2012.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. A gain and bandwidth
enhanced transimpedance preamplifier for Fourier-transform ion cyclotron res-
onance mass spectrometry. Review of Scientific Instruments, 82(12):124101,
2011.
Huilin Li, Tzu-Yung Lin, Steve L. Van Orden, Yao Zhao, Mark P. Barrow,
Ana M. Pizarro, Yulin Qi, Peter J. Sadler, and Peter B. O’Connor. Use of
top-down and bottom-up Fourier transform ion cyclotron resonance mass spec-
trometry for mapping calmodulin sites modified by platinum anticancer drugs.
Analytical Chemistry, 83(24):9507–9515, 2011.
Huilin Li, Yao Zhao, Hazel I. A. Phillips, Yulin Qi, Tzu-Yung Lin, Peter J.
Sadler, and Peter B. O’Connor. Mass spectrometry evidence for cisplatin as a
protein cross-linking reagent. Analytical Chemistry, 83(13):5369–5376, 2011.
Konstantin Aizikov, Donald F. Smith, David A. Chargin, Sergei Ivanov, Tzu-
Yung Lin, Ron M. A. Heeren, and Peter B. O’Connor. Vacuum compatible
sample positioning device for matrix assisted laser desorption/ionization Fourier
transform ion cyclotron resonance mass spectrometry imaging. Review of Sci-
entific Instruments, 82(5):054102, 2011.
x
Selected Oral Presentations
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Novel Electronics
for Fourier-Transform Ion Cyclotron Resonance Mass Spectrometry Mass Spec-
trometry Group, Max-Planck-Institut fur Kohlenforschung Mulheim, Germany,
4 October 2012.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. An improved low-
noise preamplifier design for Fourier-transform ion cyclotron resonance mass
spectrometry. Environmental Molecular Sciences Laboratory, Pacific Northwest
National Laboratory. Richland, WA, USA, 5 September 2012.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Advanced front-end
electronics for Fourier-transform ion cyclotron resonance mass spectrometry.
Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory.
Tallahassee, FL, USA, 27 August 2012.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Low noise front-end
electronics for Fourier-transform ion cyclotron resonance mass spectrometry.
In Department of Chemistry Postgraduate Research Symposium. University of
Warwick. Coventry, UK, 30 May 2012.
Selected Poster Presentations
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. An improved low-
noise preamplifier design for Fourier-transform ion cyclotron resonance mass
spectrometry. In 60th ASMS Conference on Mass Spectrometry and Allied Top-
ics. Vancouver, BC, Canada, 20–24 May 2012.
xi
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. Advanced front-end
electronics for Fourier-transform ion cyclotron resonance mass spectrometry. In
10th European Fourier Transform Mass Spectrometry Conference. Coventry,
UK, 1–5 April 2012.
Tzu-Yung Lin, Roger J. Green, David P. A. Kilgour, and Peter B. O’Connor.
Circuit design of a high power rf oscillator for multipole ion guide mass filtering.
In 32nd BMSS 3-Day Meeting. Cardiff, UK, 11–14 September 2011.
Tzu-Yung Lin, Roger J. Green, and Peter B. O’Connor. A new preamplifier
design for Fourier-transform ion cyclotron resonance mass spectrometry. In 59th
ASMS Conference on Mass Spectrometry and Allied Topics. Denver, CO, USA,
5–9 June 2011.
Tzu-Yung Lin, Raman Mathur, Cheng Lin, Konstantin Aizikov, Ronald W.
Knepper, and Peter B. O’Connor. An amplitude and frequency stabilized high
power oscillator for mass filtering and multipole ion guides. In 57th ASMS
Conference on Mass Spectrometry and Allied Topics. Philadelphia, PA, USA,
31 May–4 June 2009.
Konstantin Aizikov, Jason J. Cournoyer, Cheng Lin, Tzu-Yung Lin,
Nadezda P. Sargaeva, and Peter B. O’connor. Experimental evidence of ion
cyclotron resonance frequency modulations induced by inhomogeneities of the
trapping electric field. In 56th ASMS Conference on Mass Spectrometry and
Allied Topics. Denver, CO, USA, 1–5 June 2008.
xii
Contents
Acknowledgements iii
Abstract vi
List of Awards, Publications, & Presentations ix
List of Tables xvii
List of Figures xviii
List of Abbreviations xxiii
1 Introduction 1
1.1 FT-ICR Preamplifier Problems to Be Solved . . . . . . . . . . . . 2
1.2 Quadrupole Power Supply Problems in Mass Filtering . . . . . . . 3
1.3 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 Fourier-Transform Ion Cyclotron Resonance
Mass Spectrometry 8
2.1 Introduction to Mass Spectrometry . . . . . . . . . . . . . . . . . 9
2.1.1 Brief History & Milestones . . . . . . . . . . . . . . . . . . 9
2.1.2 Mass Spectrometry Composition . . . . . . . . . . . . . . 14
2.2 Ion Cyclotron Resonance Technique . . . . . . . . . . . . . . . . . 16
2.2.1 ICR Cell Operation . . . . . . . . . . . . . . . . . . . . . . 18
2.2.2 In-Cell Ion Motion . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Signal Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3.1 Ion Detector . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.3.2 Data System . . . . . . . . . . . . . . . . . . . . . . . . . 25
xiii
2.4 Advantages of FT-ICR MS . . . . . . . . . . . . . . . . . . . . . . 25
2.4.1 Resolving Power . . . . . . . . . . . . . . . . . . . . . . . 27
2.4.2 Mass Accuracy . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4.3 Flexibility . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.5 Front-End Electronics . . . . . . . . . . . . . . . . . . . . . . . . 33
2.5.1 Signal & Gain . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.5.2 Amplifier Noise . . . . . . . . . . . . . . . . . . . . . . . . 37
2.5.3 Existing FT-ICR Preamplifier Designs . . . . . . . . . . . 40
2.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3 Quadrupole Ion Guide & Mass Filter 48
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2 Theory of the Quadrupole Operation . . . . . . . . . . . . . . . . 51
3.2.1 Quadrupolar Potential . . . . . . . . . . . . . . . . . . . . 51
3.2.2 Mathieu Equation . . . . . . . . . . . . . . . . . . . . . . . 52
3.2.3 Stability Diagram . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Mass Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.4 Existing Power Supply Designs . . . . . . . . . . . . . . . . . . . 57
3.5 Quadrupole Power Supply Problems in Mass Filtering . . . . . . . 65
3.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4 Test Equipment & Software Programs 70
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.2 NI PXI Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.3 NI LabVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.3.1 I-V Characteristics . . . . . . . . . . . . . . . . . . . . . . 73
4.3.2 AC Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.4 Noise Performance . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.5 Other Software/Hardware Used . . . . . . . . . . . . . . . . . . . 81
4.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
5 Transimpedance Preamplifier Using an Operational Amplifier 83
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
5.2 Transimpedance Amplifier . . . . . . . . . . . . . . . . . . . . . . 86
5.2.1 Bandwidth Extension . . . . . . . . . . . . . . . . . . . . . 86
5.2.2 Cyclotron Frequency Correlation . . . . . . . . . . . . . . 89
5.3 Transimpedance Preamplifier Circuit Design . . . . . . . . . . . . 92
xiv
5.3.1 Main Stage . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.3.2 Input Stage . . . . . . . . . . . . . . . . . . . . . . . . . . 97
5.3.3 Printed Circuit Board . . . . . . . . . . . . . . . . . . . . 98
5.4 Computer Simulation . . . . . . . . . . . . . . . . . . . . . . . . . 98
5.5 Transimpedance Preamplifier Testing Results . . . . . . . . . . . 100
5.5.1 Frequency Response . . . . . . . . . . . . . . . . . . . . . 100
5.5.2 Noise Performance . . . . . . . . . . . . . . . . . . . . . . 104
5.6 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.6.1 Bandwidth & Feedback Impedance . . . . . . . . . . . . . 105
5.6.2 Noise & Feedback Impedance . . . . . . . . . . . . . . . . 106
5.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
6 Single-Transistor Transimpedance Preamplifier Using
a T Feedback Network 109
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
6.2 Transimpedance Amplifier Transfer Function . . . . . . . . . . . . 111
6.2.1 Single-Resistor Feedback . . . . . . . . . . . . . . . . . . . 111
6.2.2 T Feedback Network . . . . . . . . . . . . . . . . . . . . . 114
6.3 Single-Transistor Preamplifier Circuit Design . . . . . . . . . . . . 116
6.3.1 Common-Source Amplifier . . . . . . . . . . . . . . . . . . 116
6.3.2 Feedback Loop Configuration . . . . . . . . . . . . . . . . 121
6.3.3 Printed Circuit Board . . . . . . . . . . . . . . . . . . . . 125
6.4 Circuit Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.5 Results & Discussions . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
7 Radio-Frequency Oscillator for a Quadrupole Mass Filter 135
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
7.1.1 Component Selection . . . . . . . . . . . . . . . . . . . . . 137
7.1.2 Circuit Testing . . . . . . . . . . . . . . . . . . . . . . . . 138
7.2 RF Oscillator Design . . . . . . . . . . . . . . . . . . . . . . . . . 139
7.2.1 Crystal Oscillator . . . . . . . . . . . . . . . . . . . . . . . 139
7.2.2 Bandpass Filter . . . . . . . . . . . . . . . . . . . . . . . . 142
7.2.3 Gain Control Scheme . . . . . . . . . . . . . . . . . . . . . 142
7.2.4 RF Oscillator Printed Circuit Board . . . . . . . . . . . . 144
7.2.5 Testing Results . . . . . . . . . . . . . . . . . . . . . . . . 145
xv
7.3 Precision Rectifier Design . . . . . . . . . . . . . . . . . . . . . . 149
7.3.1 Precision Rectifier Printed Circuit Board . . . . . . . . . . 151
7.4 Power Amplifier Design . . . . . . . . . . . . . . . . . . . . . . . . 152
7.4.1 Power Amplifier Board . . . . . . . . . . . . . . . . . . . . 154
7.5 Transformer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
7.5.1 Testing Results (Power Amplifier & Transformer) . . . . . 157
7.5.2 Future Transformer Modification . . . . . . . . . . . . . . 159
7.6 Conclusion & Future Work . . . . . . . . . . . . . . . . . . . . . . 160
8 Conclusion 162
8.1 FT-ICR Preamplifier . . . . . . . . . . . . . . . . . . . . . . . . . 163
8.1.1 Preamplifier Using an Operational Amplifier . . . . . . . . 163
8.1.2 Single-Transistor Preamplifier Using a T Feedback Network 163
8.1.3 Cyclotron Frequency Correlation . . . . . . . . . . . . . . 164
8.2 Power Supply for a Quadrupole Mass Filter . . . . . . . . . . . . 165
9 Future Work 166
9.1 Preamplifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
9.1.1 T Feedback Network . . . . . . . . . . . . . . . . . . . . . 167
9.1.2 Preamplifier Noise Performance . . . . . . . . . . . . . . . 168
9.1.3 Cyclotron Frequency Correlation . . . . . . . . . . . . . . 169
9.1.4 System Test . . . . . . . . . . . . . . . . . . . . . . . . . . 169
9.2 Power Supply for a Quadrupole Mass Filter . . . . . . . . . . . . 170
References 171
A Appendix 190
A.1 Datasheet: JFET BF862 . . . . . . . . . . . . . . . . . . . . . . . 191
A.2 Datasheet: Operational Amplifier AD8099 . . . . . . . . . . . . . 201
A.3 Datasheet: Operational Amplifier LT6205/06/07 . . . . . . . . . . 225
A.4 Datasheet: Bipolar Power Transistor MJE18008 . . . . . . . . . . 244
xvi
List of Tables
2.1 Summary of the existing FT-ICR preamplifiers. . . . . . . . . . . 43
3.1 Summary of the existing RF power supplies. . . . . . . . . . . . . 65
xvii
List of Figures
2.1 Sir Joseph J. Thomson’s apparatus for measuring the charge-to-
mass ratio of cathode rays. . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Replica of Francis W. Aston’s third mass spectrometer. . . . . . . 11
2.3 Apparatus for the multiple acceleration of ions. . . . . . . . . . . 12
2.4 The composition of a modern mass spectrometer. . . . . . . . . . 15
2.5 Components inside the mass analyser, ion detector, and data sys-
tem illustrated in Fig. 2.4. . . . . . . . . . . . . . . . . . . . . . . 15
2.6 The cyclotron principle. . . . . . . . . . . . . . . . . . . . . . . . 17
2.7 The operation of a cylindrical ICR cell. . . . . . . . . . . . . . . . 19
2.8 Three types of ion motions in an ICR cell. . . . . . . . . . . . . . 20
2.9 Ion trajectory in a cubic Penning trap . . . . . . . . . . . . . . . 22
2.10 A typical schematic of the detection signal processing chain for an
FT-ICR system. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.11 The signal processing procedure of an FT-ICR data system. . . . 26
2.12 The Bruker 12-T solariX FT-ICR mass spectrometer in the Ion
Cyclotron Resonance Laboratory, University of Warwick. . . . . . 27
2.13 Full width at half maximum (FWHM) ∆m. . . . . . . . . . . . . 28
2.14 Tuning mix m/z 922 peak with ∼1-minute transient and ∼5.8M
resolving power using a 12-T FT-ICR system. . . . . . . . . . . . 29
2.15 Overlapped isotopic distributions of two ions being identified by
an FT-ICR mass spectrometer. . . . . . . . . . . . . . . . . . . . 31
xviii
2.16 IR-ECD spectra of the Substance P m/z 449.9 peak. . . . . . . . 32
2.17 Different ICR cell configurations. . . . . . . . . . . . . . . . . . . 34
2.18 Gain and noise distribution in a three-stage amplifier system. . . . 35
2.19 Equivalent circuits of the thermal noise for a resistor R. . . . . . . 39
2.20 The noise model for an amplifier. . . . . . . . . . . . . . . . . . . 40
2.21 A conventional FT-ICR preamplifier. . . . . . . . . . . . . . . . . 42
2.22 The room-temperature FT-ICR amplifier system designed by Mathur
and co-workers in 2007. . . . . . . . . . . . . . . . . . . . . . . . . 44
2.23 The cryogenic FT-ICR preamplifier reported by Mathur and co-
workers in 2008. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.1 Three major stages of a quadrupole ion guide power supply, and
the schematic of a quadrupole ion guide illustrating the electrical
connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2 The Mathieu stability diagram. . . . . . . . . . . . . . . . . . . . 54
3.3 The first stability region in a stability diagram. . . . . . . . . . . 55
3.4 Schematic of the simplified RF source designed by Jones and An-
derson. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.5 Schematic of the RF oscillator designed by Mathur and O’Connor. 59
3.6 Principle schematic of the RF power supply designed by Cermak . 60
3.7 (a) (b) Schematic of the frequency stabilized RF generator for ion
traps designed by Chang and Mitchell. . . . . . . . . . . . . . . . 61
3.7 (c) Schematic of the clamp circuit of the frequency stabilized RF
generator for ion traps designed by Chang and Mitchell. . . . . . 62
3.8 Detailed schematic of the RF circuit designed by Robbins et al. . 63
3.9 The ion trap driving circuit using CMOS inverters, designed by
Jau et al. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.10 The feedback scheme with common-base setup for oscillation in
the 2002 design by O’Connor et al. . . . . . . . . . . . . . . . . . 67
3.11 The 215-V regulator and the 15-V DC voltage supply in the 2002
design by O’Connor et al. . . . . . . . . . . . . . . . . . . . . . . 68
xix
4.1 NI PXI platform and DC power supplies for circuit test. . . . . . 72
4.2 Schematic of the transistor I-V characteristic testing circuit. . . . 74
4.3 Front panel of the LabVIEW program for obtaining I-V charac-
teristics of a transistor. . . . . . . . . . . . . . . . . . . . . . . . . 75
4.4 Block diagram of the LabVIEW program for obtaining I-V char-
acteristics of a transistor. . . . . . . . . . . . . . . . . . . . . . . . 76
4.5 Front panel of the LabVIEW program for the AC analysis. . . . . 78
4.6 Block diagram of the LabVIEW program for the AC analysis. . . 79
5.1 Components inside the ion detector of a modern mass spectrometer. 84
5.2 Two types of preamplifier systems. . . . . . . . . . . . . . . . . . 88
5.3 Noise power spectral densities of four types of transistors. . . . . . 93
5.4 Schematic of the transimpedance preamplifier using an operational
amplifier AD8099. . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.5 Average noise level of chip resistors. . . . . . . . . . . . . . . . . . 95
5.6 Single layer printed circuit board of the AD8099 preamplifier. . . 99
5.7 SPICE simulations of the AD8099 preamplifier transimpedance. . 101
5.8 Voltage gain frequency response of the main (AD8099) stage. . . . 102
5.9 Transimpedance frequency response of the preamplifier system in
three feedback conditions. . . . . . . . . . . . . . . . . . . . . . . 103
5.10 One of the input/output waveforms fetched by the NI PXI system. 104
5.11 Measured input current noise spectral density with 18 kΩ tran-
simpedance. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
6.1 Transimpedance amplifiers with two feedback arrangements. . . . 113
6.2 Schematic of the single-transistor transimpedance preamplifier us-
ing a single-resistor feedback. . . . . . . . . . . . . . . . . . . . . 118
6.3 Common-source amplifier using the JFET BF862: (a) SPICE sim-
ulated correlations between the source resistance and voltage gain/
drain current; (b) measured BF862 current-voltage characteristic. 120
6.4 (a) Schematic of the single-transistor transimpedance preamplifier
using a T feedback network. . . . . . . . . . . . . . . . . . . . . . 123
xx
6.4 (b) Measured transimpedance frequency response of the single-
transistor transimpedance preamplifier using a T feedback network.124
6.5 Printed circuit board for single-transistor preamplifiers and AD8099
voltage amplifiers. . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.6 (a) Schematic of the AD8099 preamplifier with a T feedback net-
work, and the AD8099 voltage amplifier (VA). . . . . . . . . . . . 128
6.6 (b) Measured transimpedance of the AD8099 preamplifier with
three feedback systems. . . . . . . . . . . . . . . . . . . . . . . . . 129
6.6 (c) SPICE simulation of the AD8099 preamplifier frequency re-
sponse in different permutations of the presence of two 8-pF feed-
back capacitors in the T feedback network. . . . . . . . . . . . . . 130
7.1 Block diagram of the newly designed RF power supply for a quadrupole
mass filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
7.2 (a) Block diagram of the schematic shown in Fig. 7.2b. . . . . . . 140
7.2 (b) Schematic of the fixed frequency RF oscillating source with
output amplitude control. . . . . . . . . . . . . . . . . . . . . . . 141
7.3 The frequency response simulation of the Deliyannis-Friend band-
pass filter in Fig. 7.2. . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.4 A transformer in which its secondary coil is canter-tapped. . . . . 145
7.5 Printed circuit board of of the fixed frequency RF oscillating source
with output amplitude control. . . . . . . . . . . . . . . . . . . . 146
7.6 The oscilloscope screen snapshots of the output waveforms and
their FFT results from both the crystal oscillator and the bandpass
filter. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
7.7 Measured RF oscillator output peak-to-peak amplitude and the
power consumption correlated to the applied feedback voltage. . . 148
7.8 Schematic of the precision rectifier. . . . . . . . . . . . . . . . . . 150
7.9 Two-layer printed circuit board of the precision rectifier. . . . . . 151
7.10 Schematic of the power amplifier stage. . . . . . . . . . . . . . . . 153
7.11 Power amplifier board with components mounted. . . . . . . . . . 156
xxi
7.12 Winding of a bifilar coil with current flowing in the same direction. 157
7.13 Photo of the hand-wound air-core bifilar transformer. . . . . . . . 158
7.14 The oscilloscope screen snapshot of the power amplifier output. . 159
7.15 Cross-section view of a Litz wire . . . . . . . . . . . . . . . . . . . 160
xxii
List of Abbreviations
AC . . . . . . Alternating Current
ADC . . . . . . Analogue-to-Digital Converter
AGC . . . . . . Automatic Gain Control
BJT . . . . . . Bipolar Junction Transistor
CAD . . . . . . Collision-Activated Dissociation
CID . . . . . . Collision-Induced Dissociation
CMOS . . . . . . Complementary Metal-Oxide-Semiconductor
ESI . . . . . . Electrospray Ionization
FET . . . . . . Field-Effect Transistor
FFT . . . . . . Fast Fourier Transform
FT-ICR . . . . . . Fourier-Transform Ion Cyclotron Resonance
FTMS . . . . . . Fourier-Transform Mass Spectrometry
FWHM . . . . . . Full Width at Half Maximum
GC . . . . . . Gas Chromatography
xxiii
ICR . . . . . . Ion Cyclotron Resonance
IR . . . . . . Infrared
IRMPD . . . . . . Infrared Multiphoton Dissociation
JFET . . . . . . Junction Field-Effect Transistor
LC . . . . . . Liquid Chromatography
LMCO . . . . . . Low-Mass Cut Off
MALDI . . . . . . Matrix-Assisted Laser Desorption/Ionization
MOSFET . . . . . . Metal-Oxide-Semiconductor Field-Effect Transistor
MS . . . . . . Mass Spectrometry
MS/MS . . . . . . Tandem Mass Spectrometry
NS-CAD . . . . . . Nozzle-Skimmer CAD
Op Amp . . . . . . Operational Amplifier
PCB . . . . . . Printed Circuit Board
PCI . . . . . . Peripheral Component Interconnect
PXI . . . . . . PCI eXtensions for Instrumentation
RF . . . . . . Radio Frequency
SNR . . . . . . Signal-to-Noise Ratio
SORI-CAD . . . . . . Sustained Off-Resonance Irradiation CAD
SPICE . . . . . . Simulation Program with Integrated Circuit Emphasis
TTL . . . . . . Transistor-Transistor Logic
UVPD . . . . . . Ultraviolet Photodissociation
VA . . . . . . Voltage Amplifier
xxiv
CHAPTER 1
Introduction
The late 19th century marks the start of the mass spectrometry (MS) instrument
development. MS instrumentation research has led to a continuous improvement
in mass spectrometer performance, especially the resolving power and mass accu-
racy. This advanced performance allows scientists to conduct analytical research
in new areas that could not be performed before.
This thesis focuses on the electronics for Fourier-transform ion cyclotron res-
onance (FT-ICR) MS, in hope that the new designs proposed here contribute to
the MS community, resulting in better mass spectrometer designs in the future,
for solving more complicated analytical problems. In this work, existing FT-ICR
preamplifiers (for ion signal detection) and radio-frequency (RF) oscillators (for
ion transportation/filtering) are reviewed. Then the potential scope for future
improvement is suggested.
1
1. Introduction 2
1.1 FT-ICR Preamplifier Problems to Be Solved
For improving FT-ICR performance, single-charge detection is proposed to be
one of the solutions to minimise the space charge issue in an ion cyclotron res-
onance (ICR) cell. The sensitivity and the signal-to-noise performance of an
FT-ICR system has to be improved to achieve the goal of single-charge detec-
tion. More advanced ICR cell designs are also introduced for improving FT-ICR
performance. In such a case, more parasitic capacitance will be seen by the
input node of the preamplifier, and such capacitance at input node limits the
bandwidth. Therefore, an FT-ICR preamplifier with improved bandwidth, noise
performance, and signal sensitivity is essential.
Among the exiting FT-ICR preamplifier designs reviewed in Section 2.5.3, the
signal current from the FT-ICR mass analyser, an ICR cell, is converted into a
voltage using a large input resistor, which unfortunately acts as a major noise
source at the input node. Meanwhile, the parasitic capacitance from the ICR
cell and cabling shunts such a large input resistor, causing the bandwidth to be
limited. The parasitic capacitance from the ICR cell in an FT-ICR system can
vary from around 10 pF to over 100 pF (Kaiser et al., 2011a), depending on
the cell dimensions, feedthroughs, and cabling. The preamplifier circuit reported
by Mathur and co-workers in 2008 used a 10-MΩ input resistor (Mathur et al.,
2008). A 100-pF capacitance shunting a 10-MΩ resistance causes a 1/RC corner
at ∼160 Hz. However, a 12-T FT-ICR system demands a bandwidth of around
1 MHz.
1. Introduction 3
With the newly designed preamplifiers reported in Chapters 5 and 6, this work
presents solutions to preserve preamplifier bandwidth using both transimpedance
and the T feedback techniques. Components with excellent noise performance are
used for improving the signal-to-noise ratio (SNR) of the system, hence improving
the FT-ICR sensitivity.
The design constraints of an FT-ICR preamplifier, such as the gain, sensi-
tivity, noise performance, and cyclotron frequency correlation are also studied
in this work. In the future, it is planned to use these new designs to test the
potential of performing single-charge detection at room temperature.
1.2 Quadrupole Power Supply Problems in Mass Filtering
A quadrupole mass filter is a commonly used ion filtering device in MS. When
operated in the RF-only mode (details will be discussed in Section 3.2), it can
be used as an ion guide to transfer ions in a mass spectrometer. Such a device is
also commonly used in an FT-ICR system. Among the existing quadrupole power
supply designs reviewed in Section 3.4, many of the ion guide power supplies use
a similar feedback scheme to start the oscillation. Those systems oscillate at
the resonant frequency of the output stage. Such frequency is determined by
the equivalent output reactance, which depends on the output transformer size,
tuning capacitors, cabling, and ion guide dimensions. As the impedance varies
with operation conditions, the output frequency is unstable. Meanwhile, the
output amplitude of most of those circuits is also instable. For instance, the
1. Introduction 4
design reported by Mathur and co-workers (Mathur and O’Connor, 2006) has
been tested to have a more than 1% output amplitude variation in less than one
minute.
As discussed later in Chapter 3, to drive a mass filter with 0.1-Da resolution,
RF oscillator output amplitude and frequency variations of less than 5.0× 10−4
are necessary. In Chapter 7, a new oscillator design is proposed and tested. Such
an oscillator targets a stabilized output frequency, and a feedback control for
output amplitude stabilisation. The newly designed circuit has a very stable
output frequency at 1 MHz, with a frequency tolerance of 15 ppm specified by
the crystal oscillator datasheet. For output amplitude stabilization, an automatic
gain control (AGC) unit is built in this circuit. It is believed that the new design
will produce a stable RF power supply for driving a quadrupole mass filter.
1.3 Thesis Outline
This thesis consists of three major parts. The first part of this thesis comprises
Chapters 1, 2, 3, and 4, covering the research questions, theories, and the testing
methods used for this work. Chapter 2 reviews the general MS related history and
some of the important milestones, followed by the introduction of the composition
of a modern mass spectrometer. Then, the theories of the ICR technique, one
of the mass analysing methods, is reviewed. This is followed by the introduction
of the FT-ICR signal processing method. To understand the electronic detection
limit for an ICR cell in an FT-ICR system, the nature of the ion signal and
1. Introduction 5
electronic noise have been studied. Then the existing preamplifiers for ICR signal
detection are reviewed. Finally, the potential scope for improvement is suggested
by the end of this chapter.
In Chapter 3, the theory of operating a quadrupole ion guide is provided,
followed by an introduction to the Mathieu equation, stability diagram, and
quadrupole mass filtering theory. Existing quadrupole power supply designs are
reviewed. Then the problems of building a power supply for a quadrupole mass
filter are discussed in this chapter, whilst a new power supply design will be
reported in Chapter 7.
Chapter 4 reports the testing equipment and the computer softwares used in
this work, followed by the presentation of the methods used to test the designed
circuits reported in the second part of this thesis, including Chapters 5, 6, and 7.
The second part of this dissertation reports the new electronic designs, in-
cluding two FT-ICR preamplifiers, and an oscillator for a quadrupole mass filter.
In Chapter 5, a preamplifier using an operational amplifier (op amp) in a tran-
simpedance configuration is reported. This chapter starts with a brief review
of the transimpedance technique and the input capacitance tolerance of a tran-
simpedance amplifier. Then the correlation between the cyclotron frequency of
the signal form an ICR cell and the transimpedance (gain) of the preamplifier is
reviewed to understand the preamplifier design constraints. This is followed by
the presentations of the newly designed preamplifier and its printed circuit board
(PCB) for testing. The preamplifier has been computer simulated and tested on
1. Introduction 6
the bench. The tested frequency response and the noise performance are shown.
This chapter concludes with a discussion between the feedback impedance, band-
width, noise performance, and the estimated numbers of ions to be detected in a
12-T FT-ICR system.
Chapter 6 begins with a discussion of the theories of two feedback arrange-
ments, a single-resistor feedback and a T-shaped feedback network, for tran-
simpedance amplifiers. Then a single-transistor transimpedance preamplifier de-
sign is proposed to extend further the noise performance. Biasing conditions,
input/output impedance, and over-all transimpedance of such a design is stud-
ied. A PCB for testing purpose has been manufactured. Then, this is followed
by the bench testing report of the proposed T feedback network and the single-
transistor transimpedance preamplifier. This chapter concludes with a discussion
of the gain and noise performance of a preamplifier, and a suggestion of possible
constructing elements for a T feedback network for circuit optimization.
Chapter 7 presents a RF oscillator as a quadrupole mass filter power sup-
ply, which has a stabilised output frequency, and a feedback control for output
amplitude stabilisation. This chapter first introduces the building blocks of the
new quadrupole mass filter power supply, and is followed by the reports of the
electronic details of each building block. Three PCBs have been built for testing
the following parts of this power supply: a RF oscillator, a bandpass filter, a gain
control circuit, a power amplifier, and a feedback control circuit. Then a trans-
former design is suggested, followed by the discussion of the correlation between
1. Introduction 7
resonant frequency and impedance of the output stage.
Finally, the third part of this thesis contains the conclusion, suggested future
works, reference and appendices. Chapter 8 and Chapter 9 present the conclu-
sion and the suggested future works for extending this research, respectively. In
particular, a discussion between the preamplifier gain and noise performance is
presented. A brief analysis of the correlation between the gain, cyclotron fre-
quency for FT-ICR MS, and input capacitance is also performed.
In the reference, the page number labeled in the parentheses indicates where
the citation is mentioned in this thesis. In the appendices, datasheets of the key
components of this work are included.
CHAPTER 2
Fourier-Transform Ion Cyclotron
Resonance Mass Spectrometry
This chapter first reviews the general MS related history and some of the im-
portant milestones, followed by the introduction of the composition of a modern
mass spectrometer. In Section 2.2, the theories of the ion cyclotron resonance
(ICR) technique, one of the mass analysing methods, is reviewed. This is followed
by the introduction of the method for FT-ICR signal processing. Section 2.4 in-
troduces the advantages of an FT-ICR mass spectrometer. The nature of the ion
signal and the electronic noise have been studied to understand the electronic
detection limit for an ICR cell in an FT-ICR system. Then the existing pream-
plifiers for ICR signal detection is reviewed, followed by the suggested potential
scope for improvement.
8
2. FT-ICR MS 9
2.1 Introduction to Mass Spectrometry
After positively or negatively charged ions (from samples which are introduced
either directly, or from a gas chromatography (GC)/liquid chromatography (LC)
system) are generated and sent into a mass spectrometer, a mass spectrometer
separates those charged ions according to their mass-to-charge ratios, m/z, and
records the m/z with relative abundances. Because of its selectivity, specificity,
and sensitivity to a given sample substance, mass spectrometry (MS) is generally
recognized as one of the essential microanalytical tools for determining elemental
composition or structural information in chemical, biological, or other areas of
research.
2.1.1 Brief History & Milestones
The development of a mass spectrometer can be traced back to the late 19th cen-
tury, when Sir Joseph J. Thomson used a vacuum tube to measure the charge-to-
mass ratio of cathode rays, and was awarded the Nobel Prize in Physics in 1906
”in recognition of the great merits of his theoretical and experimental investiga-
tions on the conduction of electricity by gases.”1 However, Thomson’s curiosity
about the electrical discharge behaviors originated from the discovery of ”Kanal-
strahlen” (canal rays) by Eugene Goldstein at Berlin Observatory (Watson and
Sparkman, 2008). Figure 2.1 (Beynon and Morgan, 1978) shows his apparatus
for such measurement, in which the letter ”M” means the magnet that generates
1See http://www.nobelprize.org/nobel prizes/physics/laureates/1906/ for information aboutthe Nobel Prize in Physics 1906, accessed 10 August 2012.
2. FT-ICR MS 10
a magnetic field perpendicular to the plane of Fig. 2.1.6
Fig. 3. Thomson’s apparatus for measuring the -=tio of charge-to-mass for cathode rays. The deflection produced on the collimated beam issuing from the discharge tube by a magnetic Geld was balanced by that from an electric fieid.
The ratio of the charge to mass e/me * and the velocities u of the cathode rays could now be measured assuming that they consisted of e stream of electrified particles, and this was achieved before the turn of the century by several independent researchers. In J.J. Thomson’s method [26,27] illus- trated in Fig. 3; the beam of the cathode rays issuing through a perforated anode was further coliimated by passing it through a metal slit at anode potential_ An electric field was applied between the two plates A and B thus producing a force in the plane of Fig. 3 in a direction at right angles to the direction of motion of the rays. A magnetic field in a direction perpendicular to the plane of Fig. 3 was produced by an electromagnet with its poles at right angks to the plates A and B and this produced a force, also in the plane of Fig. 3, but in the direction opposite to that of the electric field.
Assume that the individual particles follow one another at equal intervals of time 6r and that they are separated by distances 6s equal to u - 6t. The current along the path of the cathode rays is the charge conveyed past any point of the path per second and is thus equal to e/&t where e is the charge on the particles. By Laplace’s equation the force on a length 6s of this cur- rent when placed in a magnetic field W is R(e/st))ss sin 8 where 8 is the angle between the magnetic field and the direction of the cathode rays. Substitut- ing for 6s one obtains the result that
Force due to magnetic field = He u sin 8 (1)
If E is the value of the electric field between the plates A and B then
Force due to electric field = Ee (2)
If the values of magnetic and electric field strengths are adjusted in such a man- ner that the position at which the beam of cathode rays strikes a fluorescent screen, C, is tie same as the position before the fields are applied, then the two forces of eqns. (I) and (2) are equal and opposite. In Thomson’s experi-
l The symboIs and units are those recpmmended in “Quantities, Units and Symbols”, a report. by the Symbols Committee of the Royal Society (1971).
:-
Figure 2.1: Sir Joseph J. Thomson’s apparatus for measuring the charge-to-mass ratio of cathode rays (Beynon and Morgan, 1978).
The discovery and research of isotopes are also widely recognized as milestones
in the mass spectrometry history. In particular, Frederick Soddy and Francis W.
Aston received their Nobel Prizes in Chemistry in 1921 and 1922, respectively,
for Soddy’s ”contributions to our knowledge of the chemistry of radioactive sub-
stances, and his investigations into the origin and nature of isotopes,”2 and for
Aston’s ”discovery, by means of his mass spectrograph, of isotopes, in a large
number of non-radioactive elements, and for his enunciation of the whole-number
rule.”3 Figure 2.2 (Watson and Sparkman, 2008) shows the mass spectrometer
designed by Aston.
The oil drop experiment performed in 1909 leads Robert A. Millikan to the
Nobel Prize in Physics in 1923 for his ”work on the elementary charge of electric-
ity and on the photoelectric effect.”4 Some believe that the oil drop experiment
can be considered the first example of the electrospray ionization (ESI) method,
2See http://www.nobelprize.org/nobel prizes/chemistry/laureates/1921/ for informationabout the Nobel Prize in Chemistry 1921, accessed 10 August 2012.
3See http://www.nobelprize.org/nobel prizes/chemistry/laureates/1922/ for informationabout the Nobel Prize in Chemistry 1922, accessed 10 August 2012.
4See http://www.nobelprize.org/nobel prizes/physics/laureates/1923/ for information aboutthe Nobel Prize in Physics 1923, accessed 10 August 2012.
2. FT-ICR MS 11
Figure 2.2: Replica of Francis W. Aston’s third mass spectrometer, commis-sioned by the American Society for Mass Spectrometry (Watson and Sparkman,2008).
which is now a widely used ionization method for MS. A U.S. patent was granted
to Ernest O. Lawrence in 1934 for the invention of the cyclotron (Lawrence, 1934),
and later in 1939 the Nobel Prize in Physics was awarded to Lawrence ”for the
invention and development of the cyclotron and for results obtained with it, espe-
cially with regard to artificial radioactive elements.”5 Figure 2.3 (Lawrence and
Livingston, 1932) shows the ion acceleration apparatus developed by Lawrence.
The cyclotron concept was later adapted, and in 1974 the mass analysis method of
Fourier-transform ion cyclotron resonance (FT-ICR) was invented by Comisarow
and Marshall (Comisarow and Marshall, 1974a).
Although it was not until 1989 that the next MS related Nobel laureates
of Hans G. Dehmelt and Wolfgang Paul were recognized (one half awarded in
Physics) ”for the development of the ion trap technique,”6 the interests toward
the MS related topics remained active in between the Nobel ”gap years” (Wat-
5See http://www.nobelprize.org/nobel prizes/physics/laureates/1939/ for information aboutthe Nobel Prize in Physics 1939, accessed 10 August 2012.
6See http://www.nobelprize.org/nobel prizes/physics/laureates/1989/ for information aboutthe Nobel Prize in Physics 1989, accessed 10 August 2012.
2. FT-ICR MS 12
(a) Photo of the vacuum tube withcover removed.
26 E. O. LAWRENCE AND M. S, LIVINGSTON
Copper r'o y ja ssJ'e a/s
00
Yacuu yn +um p
IpI'I Jpp'ppptl
r/////.
/
iI
//
/
+T—: /~0 v. Z~sT
/i/IPm/JI/ - /2 yp//P
Deflecdi n y po/en Pi o IE/ec 7 romefer
R
Fig. 2. Diagram of apparatus for the multiple acceleration of ions.
p,
rp
~::,";-,Nj+::--'g:,";:Ip
I4 I
1m i~„'HI
i ~ p& II
I LI4
II
Fig. 3. Tube for the multiple acceleration of light ions—with cover removed.
(b) Apparatus diagram.
Figure 2.3: Apparatus for the multiple acceleration of ions, invented byLawrence (Lawrence and Livingston, 1932).
2. FT-ICR MS 13
son and Sparkman, 2008). One of the evidences can be found in the MS review
paper written by Hipple and Shepherd and published in the journal of Analytical
Chemistry in 1949 (Hipple and Shepherd, 1949), in which 176 references were
cited. Later in 1998, another MS review paper in Analytical Chemistry writ-
ten by Burlingame and co-workers became a 70-page report with 1409 citations
(Burlingame et al., 1998).
Later, in 1996 Robert F. Curl Jr., Sir Harold W. Kroto, and Richard E.
Smalley were awarded the Nobel Prize in Chemistry in 1996 ”for their discovery
of fullerenes,”7 where the C60 signal was first recorded by a time-of-flight (TOF)
mass spectrometer in 1985 (Kroto, 1997). The Nobel Prize in Chemistry 2002
was received ”for the development of methods for identification and structure
analyses of biological macromolecules.”8 Whilst one half of the prize was for the
development of the nuclear magnetic resonance (NMR) spectroscopy method, the
other half went jointly to John B. Fenn and Koichi Tanaka ”for their development
of soft desorption ionization methods for mass spectrometric analyses of biological
macromolecules,”8 in which the ionization techniques of ESI and matrix-assisted
laser desorption/ionization (MALDI) were developed.
Nowadays, there are many good references about the history of MS. The re-
view sections and book sections/chapters mentioned in this thesis can be excellent
starting points for researchers to investigate further.
7See http://www.nobelprize.org/nobel prizes/chemistry/laureates/1996/ for informationabout the Nobel Prize in Chemistry 1996, accessed 10 August 2012.
8See http://www.nobelprize.org/nobel prizes/chemistry/laureates/2002/ for informationabout the Nobel Prize in Chemistry 2002, accessed 10 August 2012.
2. FT-ICR MS 14
2.1.2 Mass Spectrometry Composition
”A modern mass spectrometer is constructed from elements which approach the
state-of-the-art in solid-state electronics, vacuum systems, magnet design, pre-
cision machining, and computerized data acquisition and processing” (Ligon,
1979). In general, a mass spectrometer is composed of an inlet system for sample
introduction, an ion source to create charged ions, a mass analyser to measure
the mass-to-charge ratio, m/z, of the charged sample, a signal detector, and a
data processing system. Usually the ion source, mass analyser, and ion detector
are located in a vacuum chamber, as shown in Fig. 2.4.
The components inside the mass analyser, ion detector, and data system are
listed in Fig. 2.5. Generally, in an FT-ICR system, a mass analyser contains an
ion transferring/filtering/accumulating system and an ICR cell. A quadrupole
ion filter can be used here for ion filtering. The operation of a quadrupole ion
filter will be introduced in Chapter 3, and a proposed new electronic device for
running a quadrupole ion filter will be presented in Chapter 7. An ion detec-
tor consists of electronic devices for processing the analogue signal coming from
the mass analyser. The signal is amplified and filtered here, before being sent
into an analogue-to-digital converter (ADC) in the data system. The preampli-
fier in an ion detector system is believed to be one of the key components for
improving signal-to-noise performance electronically. Later in this chapter, the
operation theory of an ICR cell and the signal processing in the ion detector will
be discussed, and newly designed preamplifiers will be presented in Chapter 5
2. FT-ICR MS 15
Sample
Inlet
Ion
Source
Data
System
Vacuum
Pumps
vacuum chamber
Figure 2.4: The composition of a modern mass spectrometer.
Data
System
• Ion transferring/filtering/accumulating
• Ion cyclotron resonance cell
• Preamplifier
• Instrumentation amplifiers & filters
• Analogue-to-digital converter
• Computer
Figure 2.5: Components inside the mass analyser, ion detector, and data systemillustrated in Fig. 2.4. In particular, this thesis will cover the designs of thepreamplifier (Chapters 5 & 6) and the power supply for ion filtering (Chapter 7).
2. FT-ICR MS 16
and Chapter 6.
Depending on the type of analysers, there are different types of mass spec-
trometers, such as sector instruments, ion traps, time-of-flight (TOF) instru-
ments, and Fourier-transform mass spectrometers. Each of them offers different
performance features in terms of sensitivity, speed, resolving power, and mass
accuracy. Among them, the FT-ICR MS, one type of the Fourier-transform mass
spectrometry (FTMS) (Amster, 1996), offers excellent flexibility, the highest re-
solving power, and the best mass accuracy. As a result, the FT-ICR MS has
become the instrument of choice for proteomics (Li et al., 2011b; Li et al., 2011a;
Lourette et al., 2010; Cui et al., 2011), biological imaging (Aizikov et al., 2011;
McDonnell et al., 2010; Smith et al., 2011; Taban et al., 2007), petroleum (Hsu
et al., 2011), and environmental research (Barrow et al., 2010; Headley et al.,
2011). The FT-ICR MS approach also shows promise for archeological dating
(Perez Hurtado and O’Connor, 2012). Here, the mass analyser of an FT-ICR
mass spectrometer will be discussed in Section 2.2, whilst the ion detector will
be mentioned in Section 2.3.
2.2 Ion Cyclotron Resonance Technique
Positioned within an ultra high vacuum chamber (< 10−9 mbars) with a ho-
mogeneous magnetic field, an ion cyclotron resonance (ICR) cell is the main
analysing component of FT-ICR MS. The detecting technique of ion cyclotron
resonance (ICR) was introduced by Comisarow and Marshall in 1974 at Univer-
2. FT-ICR MS 17
sity of British Columbia, Vancouver, Canada (Comisarow and Marshall, 1974a;
Comisarow and Marshall, 1974b). The operation of the ICR can be derived from
the cyclotron principle recognized by Lawrence (Lawrence and Livingston, 1932),
and is illustrated in Fig. 2.6 (Comisarow, 1993). In a homogeneous magnetic field
B, charged ion with ionic charge q, ionic mass m, and velocity v, can be con-
strained to orbit circularly with a characteristic angular cyclotron frequency ωcyc
(Comisarow and Marshall, 1976; Comisarow, 1993), where
ωcyc = 2πfcyc =qB
m. (2.1)
Note that here fcyc is the cyclotron frequency. The correlations between the
cyclotron frequency and the ion mass in a 3-T magnetic field is also shown in
Fig. 2.6.
172 M.B. Comisarow / Fundamental aspects of FT-ICR
C Y C L O T R O N P R I N C I P L E
• ° • • •
• ° ° • •
B • • • , •
• • • • •
• • ° ° •
F), ,~ F = q v x B
co- qB rn f
B = 3 Tesla
1 MHz 200 kHz 100 kHz 50 kHz f (Hz)
i I I I I I I t ~ I I I r I I
0 500 1000 m (Da)
Fig. 1. Cyclotron principle. A charged particle of mass m and charge q is constrained to a circular orbit by a homogeneous magnetic field B. The orbit has a characteristic frequency to, called the cyclotron frequency, given by eq. (1).
field is created within an apparatus called an ICR cell [2], by applying an alternating voltage to a pair of opposing ICR cell plates. When the frequency of the alternating electric field matches the natural cyclotron frequency of a particular ion mass m, the translational ICR motion of the ion will be excited. This is shown in fig. 2. As the ion absorbs energy from the alternating electric field, its velocity increases and all ions of a given mass are accelerated together as they follow the spiral path shown in the figure. ICR detection, that is, detection of the presence of excited ICR motion, is accomplished by connecting a resistor across a set of orthogonal plates and detecting the current induced in the circuit by the excited ICR motion [3]. It can be shown [3] that this current is given by eq. (2), whose parameters are N, the number of ions in the experiment, q, the ion charge, r, the ion orbital radius, B, the magnetic field, m, the ion mass, and d, the ICR cell spacing:
Nq2rB sin ~ot. (2) Is(t) = md
In conventional ICR instruments, excitation and detection are simultaneously accomplished, one mass at a time, with tuned electronic circuitry such as a marginal oscillator [4] and a scanned magnetic field.
Figure 2.6: The cyclotron principle (Comisarow, 1993).
2. FT-ICR MS 18
2.2.1 ICR Cell Operation
Figure 2.7 shows the operation of a cylindrical ICR cell. Ions are first trans-
ported and trapped in the cell, as shown in Fig. 2.7(a). At this stage, ions may
oscillate with incoherent, low, thermal amplitude. Then as Fig. 2.7(b) indicates,
an oscillating electric field is applied to the excitation plates to excite the ions
into a higher rotating radius for detection. The ions which have their angular
cyclotron frequency ωcyc the same as the excitation frequency will be ”irradiated”
to a larger, coherent cyclotron orbit closer to the detection plates. The excita-
tion signal will be a sweep of frequency to excite all ions of interests to their
rotation orbit. The ICR cell detects such ion rotating motion in the magnetic
field (perpendicular to the plane of Fig. 2.7) by electrostatic induction, and the
introduced ”image current” will be picked up by the front-end electronics con-
nected to the detection plates, as shown in Fig. 2.7(c). Note that the rotating
ions may collide with neutral background air molecules, resulting in the loss of
their kinetic energy. Then the radius of their rotating orbit is reduced. As the
ions move further away from the detection plates, the intensity of the induced
signal current is decreased. Therefore, an ultra high vacuum condition is required
to delay such a signal decay due to the background air pressure in an ICR cell.
2.2.2 In-Cell Ion Motion
There are three types of ion motions that ions trapped in an ICR cell undergo,
as shown in Fig. 2.8 (Schmid et al., 2000). The ion motion with frequency fcyc
2. FT-ICR MS 19
(a) (b)
+
_
(c)
Figure 2.7: The operation of a cylindrical ICR cell.
characterized by Eq. (2.1) is called the cyclotron motion. In the ICR cell, ideally
ions are trapped in the center of the cell, but in reality, undergo harmonic oscil-
lations along the injection-axis (z-direction). Such a oscillating motion is called
the trapping motion or z-motion. Meanwhile, ICR frequencies are a function not
only of the magnetic force, but also of the electrostatic trapping field, which is
ideally hyperbolic. The motion caused by this radial component of the trapping
field is called the magnetron motion. In reality, with the ICR cell typically used in
commercial instruments, the electrostatic trapping field is a good approximation
of a hyperbolic field only near the center of the cell. Thus, with those different
ion motions and the imperfect electric fields typically used, the observed ICR
frequencies are modulated as the ions orbit in the cell.
To describe such frequency quantitatively, one should start with the Lorentz
force described in Fig. 2.6, in which the force F , the velocity v, and the magnetic
2. FT-ICR MS 20
one of the trapping plates is an orifice through whichions can enter the cell (Fig. 1). Ions are trapped andstored for up to hours in the cell by applying a smallvoltage to the trapping plates (usually +/− 1–2 Volts,depending on the polarity of the ions). At the sametime the excitation plates and the detection plates areheld at ground potential. The function of the remainingtwo plates are described below.
The third component is the ultrahigh vacuum sys-tem. Although all types of mass spectrometer requirevacuum, FTICR-MS is most sensitive in this respect.As mentioned above, ions can be stored for long pe-riods in the analyzer cell. Long dwell times of ions inthe cell are only possible in a very high vacuum be-cause residual gas molecules from air disturb the mo-tion of the ions and so shorten the time for analysis anddetection. To perform ultrahigh mass resolution analy-sis, a pressure of 10−9 to 10−10 mbars is necessary inFTICR-mass spectrometry. Such a vacuum is providedby turbo molecular pumps or cryogenic pumps. In thecase of ESI, in which solutes are ionized under atmo-spheric pressure, several pump stages are required toovercome the enormous pressure difference from ionintroduction to ion detection.
Ion Motion
Ions describe three discrete forms of motion in theanalyzer cell: trapping motion, cyclotron motion, andmagnetron motion (Fig. 2). After injection into thecell, ions undergo harmonic oscillations in the electricfield between the trapping plates (trapping motion).Simultaneously, ions in the analyzer cell are exposedto the strong magnetic field and undergo stable cyclic
motion in a plane perpendicularly to the magneticfield, the so-called cyclotron motion. Movement ofions parallel to the magnetic field is not influenced bythis field. Each ion rotates with its typical frequency inrespect to its mass-to-charge-ratio (m/q), the so-calledcyclotron frequency. This frequency lies in the rangeof several kilohertz to several megahertz. The thirdmotion, called magnetron motion, which is centeredaround the cell axis, arises from the combination of themagnetic and the electric field. The combination ofthese three motions leads to a complex three-dimensional movement of ions in the analyzer cell.
Cyclotron Motion
The cyclic motion of ions due to a strong magneticfield is the basis of FTICR-MS. Here we explain in asimplified manner the principle of the cyclotron mo-tion.
In a magnetic fieldB, ions of chargeq and velocityv experience the Lorentz forceFL, [Eq. (1)] perpen-dicular to both the ions velocity and the magnetic fieldlines.
FL = q ? v ^ B (1)
The Lorentz force is directed inward and is coun-terbalanced by the centrifugal forceFZ, which is di-rected outward (Fig. 3) and defined by the ion massm,the ions velocityvxy in the x-y-plane and the orbitalradiusr [Eq. (2)].
FZ =m ? vxy
2
r(2)
Figure 3. Cyclotron motion of ions in the magnetic field induced by thecounterbalance of the centrifugal forceFz and Lorentz forceFL.Figure 2. Scheme of the three ion motions in the analzyer cell.
SCHMID ET AL.: FTICR-MASS SPECTROMETRY FOR HIGH-RESOLUTION ANALYSIS 151
Figure 2.8: Three types of ion motions in an ICR cell (Schmid et al., 2000).
field B are perpendicular to each other and have the relation of
F = mdv
dt= qv ×B , (2.2)
where m is the mass and q is the ionic charge of the ion. Then the magnitude of
such force F can be a function of the angular velocity ω and the rotating radius
r where
F = mω2r = qBωr . (2.3)
In an ICR cell, the ions are trapped using a electrostatic trapping potential
Vtrap, which can be applied to two end electrodes that are positioned at z = ±az/2
from the cell center along the z-axis. Such potential causes a radical force qE
(E is the electric field) that opposites the Lorentz force described in Eq. (2.3).
2. FT-ICR MS 21
Therefore the force becomes (Marshall et al., 1998)
F = mω2r = qBωr − qVtrapα
az2r , (2.4)
where α is the geometrical constant depending on the ICR cell. Then by rear-
ranging Eq. (2.4), the following equation can be obtained
ω2 − qBω
m+qVtrapα
maz2= 0 . (2.5)
The solution, as shown in Eq. (2.6), to the quadratic Eq. (2.5) describes the
perturbations applied to the cyclotron frequency ωcyc mentioned in Eq. (2.1).
ω± =ωcyc
2±√
(ωcyc
2)2
− ωz2
2, (2.6)
where ω+ is called the reduced cyclotron frequency, ω− is the magnetron frequency
of the magnetron motion shown in Fig. 2.8, and
ωz =
√2qVtrapα
maz2(2.7)
is the trapping oscillation frequency of the trapping motion. Figure 2.9 (Marshall
et al., 1998) is an example of the ion trajectory with cyclotron motion (indicated
by vc), magnetron motion (indicated by vm), and trapping motion (indicated by
vT ).
2. FT-ICR MS 22
n MARSHALL, HENDRICKSON, AND JACKSON
FIGURE 13. Ion motion in a 2 in. cubic Penning trap in a perfectly homogeneous magnetic field of 3Tfor an ion of m/z 2,300, for 10V trapping voltage. The three natural motional frequencies and amplitudeshave relative magnitudes given by: v/ Å 4.25 vz, vz Å 8.5 v0, r0 Å 4zmax Å 8r/ (Schweikhard et al.,1995). Note that the field produced by a cubic trap is not perfectly quadrupolar as manifested by the shapeof the magnetron orbit (not a perfect circle) in the xy plane. The magnetic field points in the negative zdirection.
C. Mass Calibrationv0 Å
vc
20
√Svc
2 D2
0 v2z
2(‘‘Magnetron’’ frequency) (31b)
Equation (31a) shows that the imposition of a quadrupolard.c. trapping potential reduces the ion cyclotron orbitalfrequency, because the radially outward-directed magnetic
in which field effectively reduces the magnetic field strength. In theabsence of electric space charge and trapping potentials,measurement of the frequency of a single ion of knownvz Å
√2qVtrapa
ma2 mass would serve to calibrate the magnetic field strength,from which the m/z values of all other ions in the mass(‘‘Trapping’’ oscillation frequency, in S.I. units) (31c)spectrum could be computed from Eq. (4a). However, theintroduction of a trapping potential changes the ICR mass/
andfrequency relation to the form shown in Eq. (31a), fromwhich Eq. (32) can be derived (Ledford et al., 1984). Aand B are constants obtained by fitting a particular set ofvc Å
qB0
m(‘‘Unperturbed’’ cyclotron frequency, S.I. units).
ICR mass spectral peak frequencies for ions of at leasttwo known m/z values to Eq. (32).(4a)
The three natural ion motional modes (cyclotron rota- m
zÅ A
n// B
n2/
. (32)tion, magnetron rotation, and trapping oscillation) areshown in Fig. 13, and their relative frequencies as a func-tion of ion m/z and d.c. trapping potential are shown in Strictly speaking, Eqs. (31) and (32) should be valid
only in the single-ion limit (i.e., no Coulomb interactionsFig. 14 (Schweikhard et al., 1995). The magnetron andtrapping frequencies are usually much less than the cyclo- between ions). In practice, however, FT-ICR MS experi-
ments are usually performed with sufficiently few ions sotron frequency, and generally are not detected (except assmall sidebands when the ion trap is misaligned with the that the space charge perturbation is small; moreover, the
perturbation affects calibrant and analyte ions. That’s basi-magnet axis and/or the ion motional amplitudes approachthe dimensions of the trap) (Allemann et al., 1981; Mar- cally why ICR mass calibration works so well, even when
many ions are present. ‘‘Internal’’ calibration (i.e., cali-shall and Grosshans, 1991; Mitchell et al., 1989).
14
8218 329/ 8218$$0329 09-14-98 10:53:33 msra W: MSR
Figure 2.9: Ion trajectory in a cubic Penning trap, where vc indicates thecyclotron motion, vm indicates the magnetron motion, and vT indicates thetrapping motion (Marshall et al., 1998).
2.3 Signal Processing
2.3.1 Ion Detector
The ion detector shown in Fig. 2.4 takes the responsibility of sensing the electronic
signals induced from a mass analyser. Figure 2.10 (Mathur and O’Connor, 2009)
illustrates a typical schematic of the detection signal processing chain for an
FT-ICR system. Before the digitized signal is sent into a computer for further
analysis, the analogue signal is processed by a preamplifier, instrument amplifiers,
filters, and an ADC.
A preamplifier, which is usually mounted inside the ultra high vacuum cham-
ber as close as possible to the ICR cell, is the key front-end electronic component
in FT-ICR MS. The preamplifier detects the image current (Comisarow, 1978)
induced by the excited ion packets rotating at a coherent orbit in the cell. These
2. FT-ICR MS 23
EXPERIMENTAL
A custom-made MALDI-FTMS system24 was used for all the
experiments which has a ultra-low-noise BUSM detection
preamplifier.12 A saturated solution of C60 in toluene was
spotted onto a stainless steel plate. Spectra were generated
using a single shot of a 355 nm Nd:YAG laser (Continuum
Inc., Santa Clara, CA, USA) at 50–150mJ/pulse. A pair of
hexapoles driven by radio-frequency (RF) oscillators25,26
were used to transfer the ions to the ICR cell. After 200 ms of
thermal stabilization, the ions were resonantly excited into
coherent cyclotron orbits by the application of a broadband
RF sweep. An RF sweep voltage of 80Vp–p was applied for
8 ms and swept from 150 to 3000 Da. 256K samples were
taken from the amplified ICR signal at a rate of 1 MHz (total
transient length of 0.262 s). The digitized data was fast
Fourier transformed without apodization. The standard
detection scheme for an FTMS experiment is shown in Fig. 1.
Simulated spectra were generated using a MATLAB script
(The MathWorks Inc., Natick, MA, USA). Three undamped
sinusoidal functions were used to model the first three
isotopes of C60. The instantaneous signal, f(t), at any time, t,
can be found by superposition principle, as shown in Eqn.
(2).9,21
fðtÞ ¼ A1 sinð2 p fc1 tÞ þ A2 sinð2 p fc2 tÞ
þ A3 sinð2 p fc3 tÞ (2)
Here, A1, A2, A3 and fc1, fc2, fc3 are the abundances and
cyclotron frequencies for the first three most abundant
isotopes of C60, respectively. Substituting the values for 7
Tesla FTICRMS, we obtain:
fðtÞ ¼ sinð2 p pi 149141 tÞ þ :649
sinð2 pi 148934 tþ :206
sinð2 pi 148728 tÞ (3)
Note that, this is a simplified model of induced ion current
in ICR; sufficient to analyze the artifacts arising due to signal
processing which is the primary goal of this article. A
detailed signal model can be found in an article by Grosshans
et al.11 The density functional theory (DFT) routine defined in
MATLAB was used to obtain the spectral information from
the time-domain data. The frequency spectrum was
converted into the mass-to-charge domain using calibration
constants for a 7 Tesla magnetic field.15,16,27
RESULTS AND DISCUSSION
In these sets of experiments, C60 was chosen as the
compound of interest. The intermodulation of time-domain
signals due to the first three isotopes of C60 is shown in
Fig. 2(a). The induced image signals from the three isotopes
interfere constructively and destructively with each other
causing the generation of the beat pattern. The spacing
between these beats is determined by the frequencies of the
neighboring isotopes.28 The Fourier transform of the time-
domain signal from Fig. 2(a) is shown in Fig. 2(b), where fc,
149 kHz, is the cyclotron frequency of the C60 ion at 7 T. A
simulated mass spectrum and experimental mass spectrum
of C60 are shown in Figs. 2(c) and 2(d), respectively. As the
physical model of the electric field inhomogeneity is not
included here, the odd harmonics which are present in real
spectra are not seen in Fig. 2(b), but are apparent in the real
spectra (data not shown). Figure 2 represents a nearly
ideal detection case in FTICRMS.
In FTICRMS, the detection amplifier sends the signal to the
ADC. In a case when the detection amplifier is highly
sensitive (high gain & low noise), saturation of the amplifier
or overloading of the ADC can occur, which generates
additional spectral artifacts. Figure 3(a) shows the C60 beat
pattern with a direct current (DC) offset causing the
saturation/clipping of part of the negative signal. The DC
offset could be caused by the temperature drift in electronic
components of the amplifier, or imbalance of the two FFTs
(mismatch).12 The frequency spectrum of the simulated
transient is shown in Fig. 3(b). The spectral components due
to C60 are present, fc; however, the so-called ’clipping’ of the
time-domain signal generates the harmonics of the cyclotron
frequency within the spectrum. Ideally, ICR detection is a
differential scheme and no even harmonics should appear in
the resulting mass spectra. However, in this case the
clipping, which is caused by the imbalance in the ICR
detection circuit, gives rise to the even harmonics.
The first harmonic ( fc), second harmonic (2fc) and the third
harmonic (3fc) at 149 kHz, 298 kHz, and 447 kHz,
respectively, are seen in Fig. 3(b). The frequency of the
fourth harmonic is more than the Nyquist frequency of
500 kHz; this results in the undersampling (aliasing), and a
Figure 1. FTMS detection scheme.
Copyright # 2009 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2009; 23: 523–529
DOI: 10.1002/rcm
524 R. Mathur and P. B. O’Connor
Figure 2.10: A typical schematic of the detection signal processing chain for anFT-ICR system (Mathur and O’Connor, 2009).
currents are generally on the scale of tens of femtoamps to a few hundred pi-
coamps, depending of the number of ions trapped, and the radius of rotating
orbit. The root mean square (RMS) signal current (Comisarow, 1978) from an
ICR cell can be given by Eq. (2.8a).
Is (r.m.s) =Nq2B√
2m× r
d(2.8a)
=1× (1.60× 10−19)2 × 12√2× 1000× 1.66× 10−27
× 1
4
= 3.27× 10−14 ' 33 (fA/charge), (2.8b)
where N is the number of excited ions, q is the ionic charge, B is the magnetic
field strength, m is the ion mass, r is the ion rotation orbital radius, and d is the
cell diameter (spacing). For a singly-charged, 1000-Da ion rotating at the orbit
radius of1
4the cell diameter in a 12-T magnetic field, namely, N = 1, B = 12 T,
m = 1000, andr
d=
1
4, Eq. (2.8b) calculates a RMS current signal of about
33 fA/charge.
2. FT-ICR MS 24
The amplified signal will be further processed in a signal processing chain
involving second stage amplifiers, band pass filters, and an ADC, in air outside
the vacuum system. The digitized signal then can be analysed by a computer.
The total gain within such a signal chain should be carefully designed so that the
amplified signal fits in the dynamic range of the ADC. Insufficient gain sacrifices
the detection sensitivity, whereas overloaded gain causes saturation of the ADC,
which results in the generation of artifact peaks by the Fourier transform (Mathur
and O’Connor, 2009). Considering the worst-case scenario of single-charge de-
tection, 1-bit change at the least significant bit of the ADC output is needed to
detect this current signal of a single ion given by Eq. (2.8b). If the ADC has a
resolution of 16 bits with an input range of ±1 V, the minimum total gain, which
is the transimpedance AT in this case, of the full amplifier must be
AT =2× 1
216
3.27× 10−14= 9.3× 108 (Ω). (2.9)
Note this calculation assumes a noiseless signal, but upon digitization, noise
allows detection of periodic signals less than 1 bit change (Marshall and Verdun,
1990). With such an over-all transimpedance, theoretically the maximum signal
current can be detected without saturation at the ADC is around 2.1 nA, namely,
about 66000 1000-Da ions (at the orbit radius of1
4the cell diameter in a 12-T
magnetic field) being detected.
2. FT-ICR MS 25
2.3.2 Data System
The induced image current from the ICR cell is recorded in the time domain
after being sent to a data system computer. The data system Fourier transforms
the recorded signal into the frequency domain. The frequency information is fur-
ther calibrated to yield the mass spectrum in the mass-to-charge m/z domain.
Such a procedure is illustrated in Fig. 2.11 using an tandem mass spectrome-
try (MS/MS), methods to gain different structural information, spectrum of the
Substance P m/z 449.9 peak, recorded by the Bruker (Billerica, Massachusetts,
USA) 12-T solariX FT-ICR mass spectrometer shown in Fig. 2.12.
Equation (2.6) can be revised to have a form of
fobs ≈ a(m
z)−1 + bVtrap + cVtrap
2(m
z) , (2.10)
as reported by Li and co-workers (Li et al., 1994; Zhang et al., 2005). The
Eq. (2.10) presents one of the commonly used calibration functions for the FT-
ICR MS.
2.4 Advantages of FT-ICR MS
The FT-ICR MS is currently the mass spectrometer with the highest resolving
power and mass accuracy. Another greatest advantage of the FT-ICR MS is its
flexibility to be equipped with different ion sources, inlet systems, and MS/MS
methods.
2. FT-ICR MS 26
Fast Fourier Transform
Calibration
Time Domain
Frequency Domain
m/z Domain
1000 600 200 Frequency (kHz)
Time (s) 0.06 0.12 0.18 0
I (a
.u.)
0.0
x105
-1.0
1.0
m/z
x108
2.0
1.0
0.0
Figure 2.11: The signal processing procedure of an FT-ICR data system. First,the transient data are recorded, and then Fourier transformed into frequencydomain. After calibration, the mass spectrum in the mass-to-charge m/z domainis shown.
2. FT-ICR MS 27
Figure 2.12: The Bruker 12-T solariX FT-ICR mass spectrometer in the IonCyclotron Resonance Laboratory, University of Warwick.
2.4.1 Resolving Power
A mass spectrometer with higher resolution means the ability to obtain better
separation of peaks in a given mass spectrum. Usually such ability is described as
the resolving power, R.P. (Marshall et al., 1998; Watson and Sparkman, 2008),
R.P. =m
∆m, (2.11)
where m is the mass, and ∆m is the full width at half maximum (FWHM) of
the peak, as illustrated by Fig. 2.13 (Watson and Sparkman, 2008). In Fig. 2.13
the peak of m/z = 2000 has a ∆m of 0.5, so that the resolving power R.P. =
2000/0.5 = 4000.
Nowadays, a 12-T FT-ICR system can record spectra with resolving power
routinely > 1M. The maximum resolving power RFT−ICR that an FT-ICR system
2. FT-ICR MS 28
Figure 2.13: Full width at half maximum (FWHM) ∆m (Watson and Spark-man, 2008) [modified].
can achieve for a data set is dependent on the transient duration ttran and the
cyclotron frequency fcyc, where
RFT−ICR ≥fcyc × ttran
2. (2.12)
Figure 2.149 shows a spectrum of a tuning mix10 m/z 922 peak with ∼5.8M
resolving power, obtained from a ∼1-minute transient after apodization, using
the 12-T FT-ICR system shown in Fig. 2.12.
9The mass spectra reported in Fig. 2.14 and Fig. 2.16 were prepared using the BrukerDaltonics DataAnalysis Version 4.0 SP 4 and the FTMS Processing tool build 17 from BrukerCorporation (Billerica, Massachusetts, USA).
10The tuning mix was purchased from Agilent Technologies (Santa Clara, California, USA).
2. FT-ICR MS 29
0 5 10 15 20 25 30 35 40 45 50 55 t (s)
x10 4
-1.0
-0.5
0.0
0.5
1.0
I (a.u.)
917 922 927 m/z
x10 8
0.0
1.0
2.0
3.0
922.05
923.05
922.049102
922.042 922.046 922.05 922.054
922.049118
922.042 922.044 922.046 922.048 922.05 922.052 922.054 922.056 m/z
x10 8
0.5
1.0
• FWHM(m/z): 0.000160.
• Resolution: 5,760,000.
• Apodization: Sine-Bell.
• 12T solariX.
• Narrowband m/z: 922.
• Window: 20.
(a)
(b)
(c)
Figure 2.14: Tuning mix m/z 922 peak with (a) ∼1-minute transient and∼5.8M resolving power using a 12-T FT-ICR system, obtained using a 12-TsolariX FT-ICR mass spectrometer, with Narrowband mode (center mass 922;mass range 20) and Q1 isolation of mass 922 with isolation resolution of 20.
2. FT-ICR MS 30
2.4.2 Mass Accuracy
Mass accuracy is the measurement to yield the error between the measured m/z
and the true m/z in a given spectrum. Mass accuracy M.A. can be calculated in
parts per million (ppm) or parts per billion (ppb) as
M.A. =Mobs −Mtrue
Mtrue
× 106 (ppm) =Mobs −Mtrue
Mtrue
× 109 (ppb), (2.13)
where Mobs is the experimentally observed, and Mtrue is the true mass values,
respectively. Recently, the FT-ICR mass accuracy of around 200 ppb has been
reported (Smith et al., 2012; Savory et al., 2011). With the improvement of
electric field homogeneity (such as the introduction of a better cell design), the
reduction of the space charge effect (to trap and detect one ion at a time), and
the advanced method of mass calibration, mass accuracy can be pushed close to
its theoretical value.
Ultra high resolving power and mass accuracy are required in different areas
of research. For instance, Fig. 2.15 demonstrates the overlapped isotopic distri-
butions of two ions (one labeled with the dots and the other labeled with crosses)
being identified by an FT-ICR mass spectrometer (Li et al., 2011b).
2.4.3 Flexibility
An FT-ICR mass spectrometer has the flexibility to be coupled with different ion
sources/inlet systems, and to perform many MS/MS techniques, such as electron-
2. FT-ICR MS 31
Figure 2.15: Overlapped isotopic distributions of two ions (one labeled withthe dots and the other labeled with crosses) being identified by an FT-ICR massspectrometer (Li et al., 2011b).
capture dissociation (ECD) (Zubarev et al., 1998; Zubarev et al., 2002), sustained
off-resonance irradiation collision-activated/-induced dissociation (SORI-CAD/-
CID) (Gauthier et al., 1991; Flora et al., 2001), infrared multiphoton dissociation
(IRMPD) (Little et al., 1994), ultraviolet photodissociation (UVPD) (Ly and
Julian, 2009), or double resonance (Comisarow et al., 1978; Lin et al., 2006),
etc., for obtaining detailed structural information. The FT-ICR system shown in
Fig. 2.12 has 5 different ion sources (APCI, APPI, EI/CI, ESI, & MALDI), and
6 MS/MS method classes (ECT, ETD, IRMPD, CAD, NS-CAD, & SORI-CAD),
and has the potential of being equipped with more sources and MS/MS methods.
Figure 2.169 illustrates an example of injecting an infrared (IR) laser to the
trapped precursor ions during an ECD process to increase the ECD efficiency.
The first spectrum at the top indicates the isolated Substance P m/z 449.9 peak.
2. FT-ICR MS 32
The second spectrum from the top shows the spectrum when IR laser is intro-
duced but the energy is low enough not to break any covalent bond. The third
spectrum from the top is the normal ECD spectrum of this Substance P m/z
449.9 peak. With the assistance of the ”IR laser heating” during the ECD pro-
cess, the spectrum at the bottom illustrates an increased product ion intensity.
Slide courtesy of Huilin Li, Department of Chemistry, University of Warwick
200 400 600 800 1000 1200 m/z
[M+3H]3+ isolation_449.9
20
60
100
[%]
b102+
[M+3H-NH3]3+
IRMPD(18%)_449.9
20
60
100 [M+3H]3+
b102+ [M+3H-NH3]3+
[M+3H]3+ ECD_449.9
20
60
100
c4 c8
2+
c92+
c1 b10
2+ [M+2H] 2+
c6 c10
2+
c5
c7 c8 c9
[M+2H] 2+
IR(18%)-ECD_449.9
20
60
100 [M+3H]3+
b102+
[M+3H-NH3]3+
c1 c4 c82+
c92+
c102+
c5
c5-H2O
c6 c7 c8 c9 c10
Figure 2.16: IR-ECD spectra of the Substance P m/z 449.9 peak to show theflexibility of the FT-ICR MS flexibility to perform ECD whilst injecting IR laser.
2. FT-ICR MS 33
2.5 Front-End Electronics 11
In the 80s and the 90s, different ICR cell designs were conceived (as shown in
Fig. 2.17) to reduce the perturbation caused by the imperfect electrostatic trap-
ping potential to improve performance (Caravatti and Allemann, 1991; Marshall
et al., 1998). In recent years, cell design research remains active; a few modern
designs have been reported to optimize the in-cell electric field (Brustkern et al.,
2008; Tolmachev et al., 2008; Kim et al., 2008; Weisbrod et al., 2010; Misharin
et al., 2010; Kaiser et al., 2011b; Nikolaev et al., 2011) to push the FT-ICR MS
performance close to its theoretical limit.
With imperfect cells, ion packets experience an imperfect, non-hyperbolic
electric field when rotating within an ICR cell at a ’higher’ orbital radius. The
space charge forces arising from the Coulombic interaction further dynamically
perturb the electric fields experienced by the ions and thus affect both frequency
and stability (peak width) of the peaks in the spectra thus limiting the mass
accuracy. These effects can also result in a rapid loss of coherence in the transient
which limits the resolving power (Aizikov et al., 2009). To avoid such phenomena,
fewer ions are sent into the cell and excited to a ’lower’ rotating orbit for detection.
As a consequence, at the detection plates of the cell, the induced current due to
the ion motions is smaller, resulting in a weaker signal being presented to the
11This section is partially reproduced from the following two journal articles, ”A Gain andBandwidth Enhanced Transimpedance Preamplifier for Fourier-Transform Ion Cyclotron Res-onance Mass Spectrometry,” Review of Scientific Instruments, in 2011 (Lin et al., 2011),and ”A Low Noise Single-Transistor Transimpedance Preamplifier for Fourier-Transform MassSpectrometry Using a T Feedback Network,” Review of Scientific Instruments, in 2012 (Linet al., 2012).
2. FT-ICR MS 34
front-end electronics.n MARSHALL, HENDRICKSON, AND JACKSON
FIGURE 12. ICR ion trap configurations. E Å Excitation; D Å Detection; T Å End Cap (‘‘Trapping’’).(a) cubic (Comisarow, 1981; Comisarow, 1980); (b) cylindrical (Comisarow and Marshall, 1976; Elkindet al., 1988; Kofel et al., 1986; Lee et al., 1980); (c) end caps segmented to linearize excitation potential(‘‘infinity’’ trap) (Caravatti and Allemann, 1991); (d) and (e) open-ended, without or with capacitive rfcoupling between the three sections (Beu and Laude, 1992; Beu and Laude, 1992; Gabrielse et al., 1989);(f) dual (Littlejohn and Ghaderi, 1986); and (g) ‘‘matrix-shimmed’’ (Guan and Marshall, 1995).
whose corners just touch the inner walls of the comparable not (in the absence of an additional switching network)dipolar detection, and requires that quadrupolar excitationcylindrical trap. [We have chosen a cubic trap inscribed
(rather than circumscribed) relative to a cylinder, because be performed with 2-plate rather than 4-plate excitation(Jackson et al., 1997).both traps each have the largest possible equal diameter,
and thus are equally separated from the inner (cylindrical) From Eq. (26), it is straightforward to solve the equa-tion of ion z-motionwall of the enveloping vacuum chamber.] The hyperbolic
trap is near-perfect for trapping potential, but is very non-linear for dipolar excitation. Although capacitive coupling Axial Force Å m
d2z
dt2Å 0qÇF(x, y, z) (27)
of the three-cylinder open trap improves its dipolar excita-tion linearity only slightly for ions in the z Å 0 midplane,
to obtain an ion z-position that oscillates sinusoidally withthe capacitively coupled design provides near-linear exci-time,tation from one end of the trap to the other, and, therefore,
virtually eliminates ‘‘z-ejection.’’ Also, although bquad for z(t) Å z(0)cos(2pnzt) (28a)the matrix-shimmed trap is not as large as for the cylindri-cal traps, it closely approximates the ‘‘ideal’’ bquad of 8/3
nz Å1
2p
√2qVtrapa
ma2(S.I. units) (28b)
(i.e., the limiting value for an infinitely extended tetragonaltrap). The ‘‘infinity’’ trap improves dipolar excitation, but
12
8218 329/ 8218$$0329 09-14-98 10:53:33 msra W: MSR
Figure 2.17: Different ICR cell configurations, where D indicates the detection,E indicates the excitation, and T indicates the trapping plates (Marshall et al.,1998).
2.5.1 Signal & Gain
As suggested by Eq. (2.8b), the image current signal from an ICR cell can be
theoretically as low as 33 fA (for a singly-charged, 1000-Da ion rotating at the
orbit radius of1
4the cell diameter in a 12-T magnetic field) for the goal of single-
charge detection. To have a 1-bit change at the least significant bit of the output
of a 16-bit ADC with input range of ±1 V, a transimpedance of 9.3 × 108 Ω
2. FT-ICR MS 35
(∼180 dBΩ) is needed for the electronics between the ICR cell and the ADC.
Such a large gain can be achieved by introducing several amplifying stages in
series, including a preamplifier at the first stage, and several instrumentation
amplifiers at the later stages. Figure 2.18 shows such a signal chain with three
amplifying stages. The first stage amplifier provides transimpedance AT for the
input signal current s and the noise current n0 coming from the ICR cell. The
transimpedance preamplifier converts the input current into a voltage output for
the following second and third stage voltage amplifiers. After the preamplifier,
the signal is further amplified by voltage gains G2 and G3 at the second and third
stage, respectively.
n1
A T G2 G3
n2 n3
s+n0 out
sn_3stage_amp.pdf
Figure 2.18: Gain and noise distribution in a three-stage amplifier system.
However, each amplifying stage adds its own intrinsic noise onto the signal.
In Fig. 2.18 such noises are referred back to the inputs of the first, second, and
third stage amplifiers and are represented by n1, n2, and n3 respectively. Note
that since the noise n1, n2, and n3 are the referred noise of each amplifier, noise
n1 is current noise whilst the noise n2 and n3 are noise values in voltage. Here, for
the interpretation convenience, in the following calculations the amplified output
signal S, the output noise N , and all other symbols (AT , G2 and G3; s, n0, n1,
2. FT-ICR MS 36
n2, and n3) indicate only the magnitudes (without any dimension or unit) of each
of them, respectively.
Here, after the signal processing chain, the output contains not only the am-
plified signal, S, where
S = sATG2G3 , (2.14)
but also the noise, N , where
N = (n0 + n1)ATG2G3 + n2G2G3 + n3G3 . (2.15)
The gains G2 and G3 are generally greater than 1. With large transimpedance
AT , where AT n2 and AT n3, the signal-to-noise ratio (SNR), S/N , at
the output can be obtained by dividing Eqn. (2.14) by Eqn. (2.15), and can be
approximated as:
S/N =sATG2G3
(n0 + n1)ATG2G3 + n2G2G3 + n3G3
=s
n0 + n1 +n2
AT
+n3
ATG2
' s
n0 + n1
. (2.16)
To conclude, by having significant gain at the first amplifying stage, the con-
tribution of the electronic noise is limited to the first stage, namely, the intrinsic
noise of the preamplifier and detection components. To electronically improve the
signal-to-noise performance, it is required that the first stage amplifier has not
only as much gain as possible within the bandwidth of interest, but also minimal
2. FT-ICR MS 37
noise (n1), so that the preamplifier design is crucial for best performance.
2.5.2 Amplifier Noise
Reducing the noise from a preamplifier can be a major difficulty for a circuit
designer. Typically, noise consists of all of the voltages and currents which ac-
company a signal of interest, and includes (Letzter and Webster, 1970):
• Johnson-Nyquist noise (thermal noise). Thermal noise is the electronic
noise caused by electron’s thermal agitation inside any electrical conduc-
tor, first measured and explained by John B. Johnson and Harry Nyquist
(Johnson, 1928; Nyquist, 1928).
• Noise from electronic components or amplifiers (such as shot noise or flicker
noise). Shot noise is caused by the random fluctuation of current (due to
the discrete nature of charges) at semiconductor junctions. It was first
introduced in 1918 by Walter H. Schottky who studied the current fluctu-
ations in vacuum tubes (Schottky, 1918). Flicker noise was first measured
by John B. Johnson (Johnson, 1925) and thereafter Walter H. Schottky
provided a theoretical explanation about such noise12 (Schottky, 1926).
• Environmental noise. It includes the interference from the lightning, au-
tomotive ignition, structure vibration, etc. Environmental noise can be
limited to a negligible level when proper grounding/shielding is provided.
12See http://arxiv.org/abs/physics/0204033v1 for information about the history of flickernoise, accessed 15 August 2012.
2. FT-ICR MS 38
• Statistical fluctuations. It is the noise from the quantization nature of all
measurements.
As proper grounding and shielding to the amplifier circuitry minimise the in-
terference from the environmental noise, and the noise of statistical fluctuations
rarely becomes a problem, the other two types of noise, thermal noise and am-
plifier noise, are the major noise sources to be minimised for improving noise
performance in a given amplifier design (Letzter and Webster, 1970).
Caused by the thermal agitation of electrons inside conductors, thermal noise
can be described as the noise power Pn,
Pn = kBT∆f , (2.17)
which is a function of Boltzmann’s constant kB, the absolute temperature T ,
and the bandwidth ∆f . To describe the thermal noise output from a resistor,
the equivalent circuit of a noisy resistor can be modeled by a noiseless resistor
R, either coupling in series with a noise voltage source en, or shunting a noise
current source in, shown in Fig. 2.19, where
en =√
4kBRT∆f , (2.18)
and
in =
√4kBT∆f
R. (2.19)
2. FT-ICR MS 39
Note that here the en and in are the RMS voltage and current, respectively.
0 1 2 3 4 5 6
A
B
C
D
E
R
en
inR
Figure 2.19: Equivalent circuits of the thermal noise for a resistor R.
Shot noise and flicker noise are considered electronic noise sources and are
commonly seen in an active device. Usually, flicker noise dominates thermal and
shot noise from the frequency of direct current (DC) to ∼100 Hz (Letzter and
Webster, 1970). To characterize the noise performance of an amplifier, Letzter
and Webster suggested a model comprising a noiseless amplifier with both voltage
and current noise sources, ena and ina, respectively, connected to the input, and
a noiseless source resistor, R, with its noise voltage source, en, coupled in series
to the input (Letzter and Webster, 1970), as shown in Fig. 2.20.
From Eq. (2.17), thermal noise power can be decreased by cooling (T ), or
reducing the bandwidth (∆f). To design and cool a preamplifier to work in
the cryogenic temperature of ∼4 K can significantly reduce the thermal noise by
about 10 fold. At a given temperature (such as at room temperature), modifying
the resistance value in an amplifier system can also reduce the thermal noise volt-
2. FT-ICR MS 40
Figure 2.20: The noise model for an amplifier suggested by Letzter and Webster(Letzter and Webster, 1970), including a noiseless amplifier with gain A, anequivalent amplifier noise current generator ina, an equivalent amplifier noisevoltage generator ena, and a noiseless source (the input signal source) resistor Rwith its noise voltage generator en.
age. To characterize such a change in an amplifier system illustrated by Fig. 2.20,
the corresponding modification to either the noise current generator ina, or the
noise voltage generator ena, or both, need to be evaluated carefully. Additionally,
limiting the noise generated by the active components can be another approach
to improve the signal-to-noise performance of a preamplifier. It can be done by
reducing the number of active components being used, and by choosing ultra low
noise components.
2.5.3 Existing FT-ICR Preamplifier Designs
As discussed earlier, in terms of sensitivity, the front-end electronics, especially
the preamplifier, in an FT-ICR system plays a crucial role. The potential im-
provement scope for such preamplifiers remains significant. Conventional FT-ICR
preamplifiers can provide a signal-to-noise limitation that requires at least 30–100
2. FT-ICR MS 41
ions to achieve a signal-to-noise of 3 (Kaur and O’Connor, 2004; Limbach et al.,
1993). To obtain the best system performance, the design goal for a preamplifier
is generally to boost its gain to the maximum and to limit its noise to the mini-
mum. New preamplifier designs may allow single-charge detection, which would
maximize the potential dynamic range of FT-ICR instruments.
Figure 2.21 shows a conventional FT-ICR preamplifier, consisting of a large
input resistor and a unity-gain stable operational amplifier (op amp) OPA627,13
which has a gain-bandwidth product of 16 MHz and a reported noise of 4.5 nV/√
Hz.
The preamplifier is mounted close to the ICR cell inside a vacuum chamber to
limit the capacitance from cabling, as shown in Fig. 2.21a. Figure 2.21b shows
the preamplifier circuit board. Two sets of preamplifiers are sitting on a ceramic
board. Each set is responsible for the signal from one of the two detection plates
on the ICR cell. The schematic of such a preamplifier is shown in Fig. 2.21c.
The current signal I from the ICR cell is converted to voltage by the large input
resistor R, and then buffered by a unity-gain voltage follower. The capacitance
C indicates the effective parasitic capacitance at the op amp input.
Recently, efforts have been made toward the enhancement of the performance
of preamplifiers for FT-ICR MS. In 2007, Mathur et al. reported a room tem-
perature differential preamplifier (Mathur et al., 2007), which was based on the
Jefferts and Walls’ design (Jefferts and Walls, 1989) updated with modern com-
ponents, with similar configuration shown in Fig. 2.21c. The amplifier system
13See http://www.ti.com/product/opa627 for information about op amp OPA627, accessed 15August 2012.
2. FT-ICR MS 42
(a) The ICR cell and the mounted preamplifier board.
(b) Top view of the preamplifier board.
(c) Schematic of the preamplifier.
Figure 2.21: A conventional FT-ICR preamplifier.
2. FT-ICR MS 43
(a preamplifier, shown in Fig. 2.22a, and an instrumentation amplifier, shown in
Fig. 2.22b) designed by Mathur et al. figures a voltage noise spectral density of
7.4 nV/√
Hz at 100 kHz, and a total gain of about 3500 (around 25 at the first
stage) between the frequency of 10 kHz and 1 MHz.
In 2002, O’Connor proposed an idea of holding the vacuum system of the FT-
ICR MS inside a helium cooled cold bore at the temperature of 4.2 K (O’Connor,
2002). Significantly reducing the thermal noise, the preamplifier design (shown
in Fig. 2.23) reported in 2008 (using gallium arsenide (GaAs) field-effect tran-
sistors (FETs) to avoid the semiconductor mobility ”freezing out” at cryogenic
temperature) showed about 20 times improvement in SNR, and had a voltage
gain of 250 with 3-dB frequency of 850 kHz (Mathur et al., 2008). Later in 2011,
Ivanov and co-workers designed another cryogenic amplifier using a silicon ger-
manium (SiGe) transistor, with a reported input voltage noise spectral density of
about 35 pV/√
Hz (Ivanov et al., 2011). However, the high cost of maintaining
liquid helium to preserve the 4-K environment prevents such a cryogenic FT-ICR
system from being popularized.
DesignerOperatingTemperature
VoltageGain
BandwidthNoiseSpectralDensity
conventional room unity 16 MHz 4.5 nV/√
Hz
(Mathur et al., 2007) room 25 1 MHz 7.4 nV/√
Hz
(Mathur et al., 2008) cryogenic 250 850 kHz –
Table 2.1: Summary of the existing FT-ICR preamplifiers.
2. FT-ICR MS 44
to approximately 50 ohms, primarily due to the balancingpotentiometer at the emitter of Q5 and Q6. The constantcurrent sources, Q7–Q9, kept the base-emitter voltage ofQ5 and Q6 equal. This ensures that the current gain(CMRR/differential gain) is identical for each signal line.The voltage gain versus frequency characteristic plot
of the BUSM amplifier is shown in Figure 4. The 3dBhigh frequency roll-off of the transimpedance amplifieris at 2.81 MHz, which corresponds to approximately 30
Da on a 7 tesla FTMS. The midband voltage gain of theamplifier is 3500.The electrical noise performance of the BUSM transim-
pedance amplifier was compared with that of a commercial amplifier (Figure 5). The commercial amplifierconsists of a unity gain source follower, using OPA637, asa preamplifier followed by a high gain instrumentationamplifier. The midband gain of the commercial amplifieris 1000. The BUSM transimpedance amplifier was pow-
Figure 2. Schematic of the BUSM low noise wideband transimpedance preamplifier.
Figure 3. Schematic of the BUSM instrumentation amplifier.
2238 MATHUR ET AL. J Am Soc Mass Spectrom 2007, 18, 2233–2241
(a) Schematic of the preamplifier.
to approximately 50 ohms, primarily due to the balancingpotentiometer at the emitter of Q5 and Q6. The constantcurrent sources, Q7–Q9, kept the base-emitter voltage ofQ5 and Q6 equal. This ensures that the current gain(CMRR/differential gain) is identical for each signal line.The voltage gain versus frequency characteristic plot
of the BUSM amplifier is shown in Figure 4. The 3dBhigh frequency roll-off of the transimpedance amplifieris at 2.81 MHz, which corresponds to approximately 30
Da on a 7 tesla FTMS. The midband voltage gain of theamplifier is 3500.The electrical noise performance of the BUSM transim-
pedance amplifier was compared with that of a commercial amplifier (Figure 5). The commercial amplifierconsists of a unity gain source follower, using OPA637, asa preamplifier followed by a high gain instrumentationamplifier. The midband gain of the commercial amplifieris 1000. The BUSM transimpedance amplifier was pow-
Figure 2. Schematic of the BUSM low noise wideband transimpedance preamplifier.
Figure 3. Schematic of the BUSM instrumentation amplifier.
2238 MATHUR ET AL. J Am Soc Mass Spectrom 2007, 18, 2233–2241
(b) Schematic of the instrumentation amplifier.
Figure 2.22: The room-temperature FT-ICR amplifier system designed byMathur and co-workers in 2007 (Mathur et al., 2007).
2. FT-ICR MS 45
MATHUR et al.: LOW-NOISE BROADBAND CRYOGENIC PREAMPLIFIER 1785
Fig. 4. Preamplifier circuit schematic.
occurring at 700 KHz. Due to the difficulty of accurate noisemeasurements, it is critical to have a stable, well shielded testcircuit which was possible in this case at 77 K, but not possibleat 4 K due to cracking of the coax cable used—which allowedRFI noise to leak into the measurement. A better noise mea-surement at 4 K may be possible in the future, but was deemedunnecessary in this case as the mass spectra itself (Fig. 9) is anoise measurement.
B. Preamplifier Performance in a High-Field SuperconductingElectromagnet
In the ICR detection preamplifier, it is optimal to mount thepreamplifier in close proximity to the detection plates to min-imize the lead wire capacitance. However, the high magneticfield in which the ICR cell is placed may adversely affect theperformance of the FETs. Most of the commercial JFETs whichwere tested during the room-temperature preamplifier designdid not operate well in the magnetic field due to the Hall effect[23]. A general solution is to align the channel of the FETs tothe magnetic field; however, the transverse package of the GaAsMESFETs selected here prohibits such alignment.
For this reason, functionality tests were performed on thecommon source amplifier in Fig. 3 by placing the preamplifier in
the presence of the high magnetic field. The preamplifier was in-serted inside the room—temperature bore of the 7-T supercon-ducting magnet. The gain bandwidth plot of GaAs MESFETs(FSU01LG) was obtained using a function generator and an os-cilloscope for the amplifier, once, inside the magnet, and then,outside the magnet. The Bode plot thus obtained is shown inFig. 8. It was noticed that there was minor affect on the gainof the preamplifier when placed in the 7-T field. The preampli-fier was also rotated at different angles with respect to the mag-netic field in the bore of the magnet, which only resulted in aslight change (less than 0.2 V) in the bias voltage and transcon-ductance of the FETs. This variation (reduction) in drain cur-rent is attributed to the change in magnetoresistance of the FETchannel [23]. The Lorentz force acting on the charge carrierscauses them to undergo motion, leading to a drain cur-rent which is dependent on the applied magnetic field. The fielddependence generally causes a decrease in transconductance insilicon-based JFETs, which was observed in previous room tem-perature preamplifier designs (unpublished), in agreement withthe literature [24]. However, the GaAs MESFETs tested here re-tained most of their intrinsic gain even in the presence of 7- and14-T magnetic field. Verification of the cryogenic preamplifierin a 14-T magnetic field at 4 K was done by conducting severalFTICRMS experiments, as discussed below.
Figure 2.23: The cryogenic FT-ICR preamplifier reported by Mathur and co-workers in 2008 (Mathur et al., 2008).
2. FT-ICR MS 46
Among the designs mentioned above, the signal current is converted into a
voltage by a large input resistor, which acted as a major noise source at the
input node, before being further processed by the following voltage amplifiers.
As single-charge detection is one of the solutions to avoid space charge issues in an
ICR cell, a preamplifier with improved signal sensitivity and noise performance
is essential for such a goal.
The parasitic capacitance from the ICR cell and cabling shunts such a large
input resistor, which limits the bandwidth. The parasitic capacitance from the
ICR cell in an FT-ICR system can vary from around 10 pF to over 100 pF (Kaiser
et al., 2011a), depending on the cell dimensions, feedthroughs, and cabling. The
preamplifier circuit reported by Mathur and co-workers in 2008 used a 10-MΩ
input resistor (Mathur et al., 2008). A 100-pF capacitance shunting a 10-MΩ
resistance causes a 1/RC corner at ∼160 Hz. However, a 12-T FT-ICR system
demands a bandwidth of at least 1 MHz. Meanwhile, more complicated modern
ICR cell designs introduce more input capacitance to the preamplifier. As a
result, a preamplifier with an enhanced tolerance to the input capacitance is
demanded for further FT-ICR systems.
2.6 Conclusion
This chapter first reviews the general MS related history and the theory of the
FT-ICR operation. The nature of the ion signal and the electronic noise have
also been studied to understand the electronic detection limit for an ICR cell
2. FT-ICR MS 47
in an FT-ICR system. In particular, the cyclotron frequency equation and the
calculated theoretical current signal intensity from a 12-T FT-ICR system serve
as important references for designing the MS ion detector and data system. Then
the existing preamplifiers for ICR signal detection is reviewed, followed by the
suggested potential scope for improvement. New FT-ICR preamplifier designs
and their testing results will be reported later in Chapter 5 (a transimpedance
preamplifier using an op amp) and Chapter 6 (a single-transistor transimpedance
preamplifier using a T feedback network).
CHAPTER 3
Quadrupole Ion Guide & Mass Filter
This chapter introduces the theory of operating a quadrupole ion guide is pro-
vided, followed by the introduction of the Mathieu equation, stability diagram,
and quadrupole mass filtering theory. Then this chapter reviews the existing
quadrupole power supply designs. The problems of building a power supply for
a quadrupole mass filter are discussed in this chapter, whilst a new power supply
design will be presented in Chapter 7.
3.1 Introduction
The linear quadrupole has been used for ion transportation and mass filtering
in scientific apparatus since the late 50s (Paul, 1990; Douglas, 2009), and is
still widely used in mass spectrometers. The history of quadrupoles and the
theory of the motion of charged particles in radio-frequency (RF) fields have been
comprehensively reviewed in many excellent papers and books, in particular, by
Gerlich (Teloy and Gerlich, 1974; Gerlich, 1992), Dawson (Dawson, 1976), Paul
48
3. Quadrupole Ion Guide & Mass Filter 49
(Paul, 1990), March (March, 1997) and Todd (March and Todd, 2009), and
Douglas (Douglas, 2009) and co-workers (Douglas et al., 2005).
A quadrupole comprises ideally hyperbolic but commonly cylindrical elec-
trodes in a square formation, as shown in Fig. 3.1. Two pairs of rods are con-
nected with opposite-polarity RF signal applied electrically, and establish a two-
dimensional field in the x-y plane. Ions oscillate in the x-y plane whilst traveling
along the z direction (March and Todd, 2005).
+ ( U + V cosωt )
– ( U + V cosωt ) –
–
+ +
2r
–
+ +
– x
y z
+ +
+
Transformer/ Matching Circuit
Power Amplifier
RF Oscillator
Figure 3.1: Three major stages of a quadrupole ion guide power supply, andthe schematic of a quadrupole ion guide illustrating the electrical connections.
To operate a quadrupole with radius r (the inscribed circle tangential to the
rods’ surfaces), opposite polarity RF signals ±φ(t) with amplitude V , angular
frequency ω, and a superimposed direct current (DC) offset U , are applied to the
adjacent rods of a quadrupole, where
φ(t) = ±(U + V cosωt) . (3.1)
3. Quadrupole Ion Guide & Mass Filter 50
The RF signals mentioned in Eq. (3.1) are generated by a ”power supply” which
comprises three major stages/parts, as illustrated by Fig. 3.1:
• RF Oscillator (First Stage): the oscillating source. In some circuits, such
as the power supply reported by O’Connor and co-workers (O’Connor et al.,
2002), this is a feedback system involving the next stage power amplifier. By
feeding the output the amplifier back to the input, the circuit will oscillate
at a self-tuned resonant frequency.
• Power Amplifier (Second Stage): the main driving stage to amplify the
signal generated by the source to drive the quadrupole ion guide.
• Transformer/Matching Circuit (Third Stage): the output stage after the
power amplifier. It is usually either a transformer to convert the output
voltage of the power amplifier to a higher value, or a matching circuit for
resonant frequency tuning or impedance matching, or both. If a DC offset
is to applied to the quadrupole, the offset voltage can be introduced in this
stage.
Due to the strict requirement of the output frequency and amplitude tolerances,
to design a power supply for a quadrupole mass filter requires more attention
than for a ion guide. The details of such design requirements will be discussed
later.
3. Quadrupole Ion Guide & Mass Filter 51
3.2 Theory of the Quadrupole Operation
The theory of operating a quadrupole device is widely reviewed. The basic con-
cepts will be summarized here, based on the interpretation in the book written
by March and Todd (March and Todd, 2005).
3.2.1 Quadrupolar Potential
To simplify the following derivation, the assumption has to be made that there is
only one gas-phase ion travels in a quadrupole with hyperbolic electrodes, infinite
rod length, and complete absence of background gas. The potential in the electric
field of a quadrupolar device in the Cartesian co-ordinates φx,y,z has a general
form of
φx,y,z =φ0
2r2(λx2 + σy2 + γz2) , (3.2)
where λ, σ, and γ are weighting constants for the x, y, and z coordinates, re-
spectively; φ0 is the net potential applied to the single ion in the quadrupole,
where
φ0 = φx−pair − φy−pair = 2(U + V cosωt) . (3.3)
To satisfy the Laplace condition of
∇2φx,y,z = 0 , (3.4)
3. Quadrupole Ion Guide & Mass Filter 52
the numbers of
λ = −σ = 1 ; γ = 0 (3.5)
can be assigned for the two-dimensional devices. As a result, the quadrupolar
potential at point (x, y) can be expressed as
φx,y =φ0
2r2(x2 − y2) . (3.6)
3.2.2 Mathieu Equation
According to Eq. (3.3), if only the component of motion in the x-direction is
considered, the force acting on this ion at the point (x, 0), Fx, is
Fx = md2x
dt2= −ze(dφx,y
dx)y=0 = −zeφ0x
r2, (3.7)
where z is the number of charges on the ion, e is the electron charge, m is the
mass of the ion, and the negative sign suggests that the force acting on the ion
is in the opposite direction of the increasing x. By substituting Eq. (3.3) into
Eq. (3.7), the equation can be expanded to
d2x
dt2+ (
2zeU
mr2+
2zeV cosωt
mr2)x = 0 , (3.8)
3. Quadrupole Ion Guide & Mass Filter 53
which is a canonical form of the Mathieu Equation
(d2u
dξ2) + (au − 2qucos2ξ)u = 0 . (3.9)
By substituting u = x and ξ = ωt/2, the dimensionless stability parameters
ax and qx become
ax =8zeU
mr2ω2and qx =
−4zeV
mr2ω2, (3.10)
where U , V , ω, and r are the previously mentioned DC voltage, RF signal am-
plitude, angular frequency, and the inscribed circle radius, respectively. Because
of the conditions given by Eq. (3.5), the derivation of the force on the traveling
ion in the y-direction can be obtained as ay = −ax and qy = −qx.
3.2.3 Stability Diagram
As a result, the ion motion (in either x or y direction) in such electric fields
can be described by the Mathieu equation with parameters au and qu, where u
represents the co-ordinate axis x or y depending on the geometrical direction
to be considered. The solutions to the Mathieu equation can be interpreted in
terms of ion trajectory stability in the stability diagram, as shown in Fig. 3.2
(de Hoffmann and Stroobant, 2007). A stable ion trajectory can be obtained
when the parameters au and qu fall into one of the stability regions, where the
traveling ion is stable in both x and y-directions, of the Mathieu equation. It is
3. Quadrupole Ion Guide & Mass Filter 54
shown in Fig. 3.2 that there are a few regions that are stable along both x and
y-directions, such areas A, B, C, and D.
chap02 JWBK172-Hoffmann August 7, 2007 0:9
94 2 MASS ANALYSERS
A
A
B
C
D
au
au
au
qu
qu
qu
Stable along xStable along y
Figure 2.7Stability areas for an ion alongx or y (above) and along x andy (below); u represents either xor y. The four stability areas arelabelled A to D and are circled.The area A is that used commonlyin mass spectrometers and is en-larged. The direct potential partis shown for positive U (shaded)or negative U. From now on wewill consider the positive area. Re-produced (modified) from MarchR.E. and Hughes R.J., QuadrupoleStorage Mass Spectrometry, Wiley,New York, 1989, with permission.
proportional triangles. Figure 2.8 represents in a U, V diagram the areas A obtained withdifferent masses.
We can see in this diagram that scanning along a line maintaining the U/V ratio constantallows the successive detection of the different masses. So long as the line keeps goingthrough stability areas, then the higher the slope, the better the resolution.
If U = 0, there is no direct current and the resolution becomes zero. However, the valueof V imposes a minimum on stable masses. Thus, if V is increased from 0 to V so as toreach slightly beyond the stability area m1, all of the ions with masses equal to or lowerthan m1 will have an unstable trajectory, and all of those with masses above m1 will have astable trajectory.
In practice, the highest detectable m/z ratio is about 4000 Th, and the resolution hoversaround 3000. Thus, beyond 3000 u the isotope clusters are no longer clearly resolved. As
Figure 3.2: The Mathieu stability diagram (de Hoffmann and Stroobant, 2007).
It is common to run a quadrupole in the first stability region (area A in
Fig. 3.2), though to run the quadrupole device in other stable regions was also
reported (Hiroki et al., 1991). Figure 3.3 illustrates the first stable region. The
area confined by q-axis and both blue and red solid lines is the stable region.
When running a quadrupole as an ion guide, in the so-called ”RF-only” mode,
the DC voltage U is set to 0 to operate the quadrupole so that ions are posi-
tioned along the q-axis, which allows any ion to be transported stably in a given
quadrupole system as long as its mass-to-charge ratio (m/z ) satisfies the low-mass
cut off (LMCO) condition of qu ≤ 0.908. For instance, a quadrupole with size r
3. Quadrupole Ion Guide & Mass Filter 55
of 5 mm operating at the frequency of 1 MHz and the supplied RF amplitude V
of 200 V has a LMCO of 86 Da for a singly-charged ion.
0 . 0 0 . 2 0 . 4 0 . 6 0 . 8 1 . 00 . 0
0 . 1
0 . 2
0 . 3
a u
q u
S c a n l i n e
( 0 . 9 0 8 , 0 )
( 0 . 7 0 6 , 0 . 2 3 7 )
Figure 3.3: The first stability region in a stability diagram.
3.3 Mass Filter
To run a quadrupole as a mass filter, a DC voltage (U in Eq. (3.1)) is super-
imposed on the RF signal supplying the system. In general, the quadrupole is
tuned and run along certain ”scan line” (the dotted line shown in Fig. 3.3). On
the scan line, heavier ions lie closer to the origin. With a fixed running frequency,
the slope of the scan line can be adjusted by changing the ratio of DC voltage
U and RF amplitude V . When the parameters of the system are set so that the
3. Quadrupole Ion Guide & Mass Filter 56
scan line is inside the stable region but very close to the tip, where qu = 0.706
and au = 0.237, only ions within a very narrow window of m/z values can be
transported. The closer to the tip, the higher the potential resolving power of
the mass filter. To tune the system electronically closer to this tip of the stabil-
ity diagram, signals of very stable values of U , V , and ω have to be generated
to supply the quadrupole mass filter. This requirement becomes more stringent
as one increases the masses of the ions one wishes to resolve. Specifically, the
variation of the output amplitude and frequency of a mass filter power supply
has to be limited to narrow the filtering window. The requirements for precise
mass selection was described by Austin and co-workers in a quadrupole mass
spectrometry book edited by Dawson in 1976 (Austin et al., 1976) as
∆m
m=
∆V
V− 2∆ω
ω. (3.11)
It was also stated by Austin et al. (Austin et al., 1976) that over the operating
range of 0–200 Da, a mass stability of better than 0.1 Da can be achieved if ∆ω/ω
and ∆V/V are below 2.5×10−4 and 5.0×10−4, respectively. As a result, most of
the reported quadrupole power supplies are capable of driving an ion guide, but
only a few of them have the potential to drive a mass filter with 0.1-Da resolution.
3. Quadrupole Ion Guide & Mass Filter 57
3.4 Existing Power Supply Designs
In 1976, some basic building blocks, such as a rectifier, a crystal oscillator, a RF
output circuit, for a mass filter power supply were given by Austin et al. in the
book edited by Dawson in 1976 (Austin et al., 1976). Such circuit intended to
supply a 4-MHz signal up to 1000 V.
Since the late 90s, a series of power supplies for ion guides have been designed
for mass spectrometry instrumentation. In 1997, Jones et al. reported a simple
RF power supply for ion guides (Jones et al., 1997), using a pair of 6146B trans-
mitter vacuum tubes in a push-pull configuration. The operating frequency was
set by the output impedance, which was a combination of the output tank coil
and the total shunting capacitance, and could be tuned up to ∼30 MHz. The
output could be switched off by a transistor-transistor logic (TTL) signal, and
the RF amplitude could be adjusted from 50 to 600 V by computer or manually,
while the maximum power dissipation was ∼140 W. Such design was further im-
proved by Jones and Anderson in 2000, as shown in Fig. 3.4 (Jones and Anderson,
2000), with reduced complexity, size, and cost.
In 2002, O’Connor et al. reported a high voltage RF oscillator for driving
multipole ion guides (O’Connor et al., 2002). This oscillator was a modification
based on Jones and Anderson’s design, to (a) replace the vacuum tubes by the
bipolar junction transistors (BJTs) 2SC5392, (b) introduce a tightly coupled air-
core transformer to separate the DC offset from the power supply voltage, while
providing feedback signal, and (c) include an automatic gain control (AGC) to
3. Quadrupole Ion Guide & Mass Filter 58
To enable rf output, the cathodes of the tubes are held atground potential. When the A–B jumper~lower left, Fig. 1!is in the A position, the rf generator is controlled by anexternal TTL input. When the input is high~5 V!, Q1 isturned on, holding the cathodes at ground potential, and en-abling rf output. When the TTL signal is low~0 V!, Q1 turnsoff and allows the cathodes to float, shutting off the rf output.Q1 must be a high voltage transistor, because the cathodesfloat to near the plate voltage whenQ1 is off. If external
control of the rf output is not needed, then the cathodes canbe permanently grounded, and the B position of the A–Bjumper is provided for this purpose.
In the oscillator section, the voltage oscillates betweenground and the applied plate voltage. For ion guiding/trapping applications, the rf must oscillate symmetricallyabout an externally provided ‘‘float’’ voltage. The requireddc decoupling is provided by the ‘‘output circuit’’ indicatedby a second dotted rectangle in Fig. 1.C4 andC5 provide alow-impedance path for radio frequencies, but block theplate dc voltage. In the current design, the coupling capaci-tors ~C4 and C5! are fixed value ceramic capacitors. Thevalue of these capacitors is not critical, as long as they areequal and large compared toCLOAD ~typically tens of pico-farads!. Because capacitors are typically low-precision com-ponents, it may be necessary to add small capacitors in par-allel to C4 or C5 to balance rf output~i.e., make theamplitude of the two phases identical!. The float voltage issupplied by an external power supply, and is coupled to therf output connectors through a pair of rf chokes~L1 andL2!.C1 andC2 protect the float supply by shunting to groundany rf that passes the rf chokes.
In a push–pull oscillator like this, the rf amplitude isapproximately equal to the applied plate high voltage, allow-ing rf amplitude to be simply controlled or programmed bythe external high voltage supply. The 6146B tubes are ratedat 700 V—sufficient for most ion guiding/trapping purposes.The threshold for oscillation~i.e., the useful lower limit of rfamplitude! is controlled by losses in the circuit and load, andwill depend somewhat on construction details. For a typicalion guide system with capacitance in the 100 pF range, thethreshold is generally less than 50 V. The maximum powerdissipation for the generator is 140 W, and the dissipation ismostly in the form of heating in the tubes,L tank, and otherinternal components. The unit shown here draws approxi-mately 100 mA from the high voltage supply~positive po-larity! when operating at 600 V with no external load, i.e., 60W dissipation. The additional load from a well constructedion guide/trap is less than 10 W at 600 V. As a consequence,it is possible to add additional output circuits to allow asingle generator to drive several independently floatable ionguide segments. We routinely drive two ion guide segmentswith a single generator.2–5 Of course, as additional load ca-pacitance is added, the operating frequency will drop unlessCtune is reduced to compensate for it. For the componentvalues given in Table I, the operating frequency with noexternal load is;7 MHz, dropping to;5 MHz when drivinga typical octapole ion guide.
To illustrate our component layout, Figs. 2 and 3 showfront and bottom views of the rf generator. The generator isconstructed around a small aluminum chassis (20 cm311 cm34 cm), with tubes andL tank on top, and most othercomponents in the bottom. Not shown are covers, con-structed of perforated aluminum sheet, that completely sur-round the unit. Note that dangerously high dc and rf voltagesare present in the generator in operation. For safety and toreduce rf emissions, it is important that the unit be enclosed.
Figure 2 showsL tank, hand wound on a plastic coil form,the two tubes,Q1 ~mounted on the base at rear!, and a small
FIG. 1. Schematic of the simplified rf source.
TABLE I. Typical component values.
Ref. Value Type Notes
C1 0.1mF at 1 kV disk ceramic noncritical rf bypassingC2 0.1mF at 1 kV disk ceramic noncritical rf bypassingC3 0.1mF at 1 kV disk ceramic noncritical rf bypassingC4 1000 pF at 3 kV disk ceramic aC5 1000 pF at 3 kV disk ceramic aCtune 44 pF at 2 kV disk ceramic for setting frequencyC7 0.1mF at 1 kV disk ceramic noncritical rf bypassingC8 10 pF at 500 V micaC9 10 pF at 500 V micaR1 18 V, 1 W carbon for ON/OFF keying dampingR2 25 K, 10 W wire wound screen voltageR3 25 K, 10 W wire wound screen voltageR4 36 K, 1/2 W carbon grid biasR5 36 K, 1/2 W carbon grid biasL1 2.5 mH 60 mA rf choke current rating 60 mA or moreL2 2.5 mH 60 mA rf choke current rating 60 mA or moreL3 2.5 mH 250 mA rf choke current rating 250 mA or moreL tank 30 mH hand wound see the textQ1 Motorola BU208 1500 VCE must be a high voltage part
aC4 andC5 couple rf to the output. To get the best rf balance, small value3 kV capacitors can be added in parallel withC4 or C5 to compensate forvariations in capacitor values.
4336 Rev. Sci. Instrum., Vol. 71, No. 11, November 2000 R. M. Jones and S. L. Anderson
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
Figure 3.4: Schematic of the simplified RF source designed by Jones andAnderson (Jones and Anderson, 2000).
linearly correlate the output RF amplitude of 0–500 V with a reference DC volt-
age of 0–10 V. The output frequency was tunable from 500 kHz to 1.5 MHz by
changing the impedance of the matching components. In their report, a sim-
ple regulating circuit was also provided for an unregulated power supply. Later
in 2006, Mathur and O’Connor implemented a similar oscillator, in which the
BJT BUH51 replaced 2SC5392, on a printed circuit board (PCB) (Mathur and
O’Connor, 2006). The circuit of such a design is shown in Fig. 3.5 (Mathur and
3. Quadrupole Ion Guide & Mass Filter 59
O’Connor, 2006). Mathur and O’Connor further studied the PCB design con-
straints, such as track spacing and width, heat dissipation, parasitic impedance,
and electromagnetic interference. The details and the PCB files can be down-
loaded from the Internet.14
varied substantially across the mass range, 1 ms at m /z200 Da but 3.5 ms at m /z 8.5 kDa for an ion with 20 eV ofkinetic energy. The broadband excitation voltage and thesweep rate also had to be adjusted for different ions to gen-erate a coherent, detectable ion packet.19
The mass spectrum MS of Ub is shown in Fig. 5. Thepeaks corresponding to the matrix adducts were also ob-served and are assigned in Fig. 5a. The time of flight of Ub
ions in the 134 cm long hexapole was approximately 3.5 msand the excitation voltage on the ICR cell of 100 Vp-p wasswept for 8 ms from m /z 1 kDa to 10 kDa. In an attempt toremove the matrix adducts, the ions were heated using apseudocontinuous CO2 laser Synrad model 48-2, Mukilteo,
FIG. 1. Schematic of the rf oscillator circuit. The connections of the transformer coils at points a-a, b-b, c-c, and d-d are elaborated in Fig. 4.
FIG. 2. Color online Radio frequency oscillator printed circuit board.FIG. 3. Color online Radio frequency oscillator with the transformer coil,power supply, and the cooling fan enclosed in a shielded box.
114101-5 rf oscillator design Rev. Sci. Instrum. 77, 114101 2006
Downloaded 27 Apr 2011 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
Figure 3.5: Schematic of the RF oscillator designed by Mathur and O’Connor(Mathur and O’Connor, 2006).
In 2005, Cermak designed a compact RF power supply that could be run
between the frequency range of 4 to 8 MHz, depending on the output trans-
former and capacitors (Cermak, 2005). In this design, two power metal-oxide-
14See http://warwick.ac.uk/oconnorgroup/research/rfoscillator/ for information about the RFoscillator PCB designed by Mathur and O’Connor in 2006, accessed 10 August 2012.
3. Quadrupole Ion Guide & Mass Filter 60
semiconductor field-effect transistors (MOSFETs) were used as the main power
amplification stage, which was driven by externally synchronized oscillators de-
rived monostable flip-flops and buffers, as shown in Fig. 3.6 (Cermak, 2005). A
stable operation, when the amplitude was 200 V with ∼50 W power consumption,
was reported.
plied modern electronic components, which resulted in aninexpensive, compact construction with high output powerand efficiency.
II. DESIGN
The principle schematic of the generator is shown in Fig.1, and an outline of typical waveforms is depicted in Fig. 2.The circuitry is similar to those used in resonant switchingpower supply units. The crucial difference between our de-sign and commercial power supply units is our higher avail-able operating frequency. The radio-frequency output voltageof the generator is generated by the transformer TR1. Thetransformer is driven by a symmetrical push–pull powerstage consisting of two power metal–oxide–semiconductorfield effect transistorssMOSFETsd T1 and T2. The supplyvoltage for the power stages+Vd is provided by an externalpower supply unitsPSUd. The control signals for switchingthe power stage are derived from two monostable flip-flopcircuits smonoflopsd and two drivers from output signals ofan externally synchronized oscillator. The width of the con-trol pulses together with the magnitude of the supply voltagedetermines the amount of energy that is accumulated in theoutput transformer during each half cycle of the output sig-nal, thereby influencing the amplitude of the generator’s out-put voltage. The synchronization of the oscillator by the out-put voltage ensures that the device operates at approximatelythe resonance frequency of the output resonance circuit con-sisting of the inductance L and the capacity of the connectedtrap. When the resonance frequency shifts, for instance dueto changes in temperature, the generator follows thesechanges and still operates with optimal efficiency. The outputvoltage can be shifted by applying an auxiliary dc voltageVo via an inductance Lo that is connected to the center of theresonant circuit and which acts as a low pass filter.
The output transformer TR1 directly supplies the trap-ping device, which behaves as the capacitive part of the out-put resonance circuit. The design of the transformer is crucialfor the function of the whole generatorssee Sec. IIId. Usu-ally, the inductance of the secondary coil should be substan-
tially higher than the inductance L of the connected resonantcircuit. The conversion ratio of the transformer determinesthe effective resistive load of the resonant circuit during theswitch-on time of the power stage. When a sufficiently highconversion ratio is selected, the generator will not affect the
FIG. 1. Principle schematic of the generator.
FIG. 2. Outline of the generator waveforms. The diagram shows the tem-poral dependencies of important signals from the generator’s electronicssfrom top to bottomd: the sinusoidal output signals, the output signals of thedelay elements gained by delaying the digitized generator’s output signal,output signal of the oscillator externally synchronized by the delay elements,output signals of the monostable flip-flopssmonoflopsd derived from theoscillator output and driving the power MOSFETs. The lowest traces showthe signals on the drains of the MOSFETs. These signals swing around thesupply voltage +V and reach approximately the ground potential during theswitch-on time of each particular MOSFET. Note that the real shape of thewaveforms on the drains depends on the properties of the MOSFET itself, ofthe output transformer, and of the output resonance circuit and may signifi-cantly differ from the form shownssee Fig. 5 and corresponding textd.
063302-3 RF power supply for guides and traps Rev. Sci. Instrum. 76, 063302 ~2005!
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
Figure 3.6: Principle schematic of the RF power supply designed by Cermak(Cermak, 2005).
In 2006, Chang and Mitchell reported a frequency stabilized RF generator to
drive ion traps. Vacuum tubes 6146W were used, and the oscillation frequency
was phase locked to an external reference oscillator (Chang and Mitchell, 2006).
With the presence of an amplitude gain control unit, the circuit, as shown in
Fig. 3.7, was set to run at the frequency of ∼800 kHz, with amplitude of 8–
400 V.
In 2008, Robbins and co-workers designed a computer-controlled, variable-
frequency power supply that allowed an output RF amplitude of 5–500 V over
3. Quadrupole Ion Guide & Mass Filter 61
transistor-transistor logic TTL level signal OSCCTL tooscillate or to clamp. This signal is used to control analogswitches which connect the operational amplifier inputs toground during clamping.
A unity voltage gain current buffer drives the high volt-age power transistor Q1. The current gain varies from tran-sistor to transistor, and the resistance for R13 that places thetransistor response at its largest gradient must be individually
FIG. 2. Schematic of the amplitude gain control AGC circuit.
FIG. 1. Principle schematic of the rfoscillator design.
063101-2 B. T. Chang and T. B. Mitchell Rev. Sci. Instrum. 77, 063101 2006
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
(a) Principle schematic.
transistor-transistor logic TTL level signal OSCCTL tooscillate or to clamp. This signal is used to control analogswitches which connect the operational amplifier inputs toground during clamping.
A unity voltage gain current buffer drives the high volt-age power transistor Q1. The current gain varies from tran-sistor to transistor, and the resistance for R13 that places thetransistor response at its largest gradient must be individually
FIG. 2. Schematic of the amplitude gain control AGC circuit.
FIG. 1. Principle schematic of the rfoscillator design.
063101-2 B. T. Chang and T. B. Mitchell Rev. Sci. Instrum. 77, 063101 2006
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
(b) Gain control unit.
Figure 3.7: Schematic of the frequency stabilized RF generator for ion trapsdesigned by Chang and Mitchell (Chang and Mitchell, 2006)
3. Quadrupole Ion Guide & Mass Filter 62
determined. Likewise, the capacitances C4 and C5 in Fig. 1may have to be adjusted to take into account tube variabil-ity. Typically, the buffer output voltage is about 6 V in thisregime. Values of R13 below 220 ohms must be avoided toprevent damaging the power transistor with excessive cur-rent.
A schematic of the clamp circuit is given in Fig. 3. Weuse a solid state switch U3 to short the LC tank on the sec-ondary side of the transformer T1. This specific design limitsus to output amplitudes of 400 V due to the voltage limitsof the selected switch.
Fast and repeatable clamping requires that one beginsthe clamp at a specific phase of the output wave form. This isachieved with a “synchronizer” subcircuit whose location isindicated in Fig. 3. A TTL transition from the OscCtl signalarms a level detector which begins clamping after it is trig-gered by a signal derived from the rf output. There is arandom time delay between zero and a wave period from theOscCtl transition to the clamp start pulse produced by this/approach, unless one works in the frequency-locked modeand synchronizes the OscCtl transition with the referenceoscillator.
III. PERFORMANCE
The performance of the rf oscillator has been tested witha hand-wound transformer designed to produce oscillationsin the 1 MHz region. The air-core transformer consists of athree layer secondary winding overlaid upon a one layer pri-mary winding. The diameter of the cylindrical 8.5 in. long
nylon form upon which the transformer is wound is 2 in. Theprimary is made of 32 turns of AWG16 wire and is centertapped. The secondary is made of 96 turns of AWG16 wire,three layers each of 32 turns; each spaced over the 8.5 in.length. The secondary is also center tapped. There is also asingle turn sensing loop consisting of insulated hookup wirewrapped over the center.
The primary center tap is capacitively coupled toground, and is driven by the dc high voltage power supply.The two 6146 W vacuum tube anodes drive the primaryends. Two load capacitances of about 200 pF are attached tothe secondary from each side to the center tap. With theseloads, the free-running resonant frequency is 800 kHz,close to the estimated fLC. We operate in an amplitude rangeof 8–400 V, the upper range of which requires a dc powersupply of 200 V capable of 150 mA. With the clampingcircuitry disconnected, the maximum amplitude is set by the700 V limits of the tubes and the choice of turns ratio in theoutput transformer.
The performance of the clamping circuitry is shown inFig. 4. Two separate clamped waveforms were recorded witha Tektronics TDS5052B oscilloscope with different voltageresolutions. The 330 V amplitude oscillations are reduced toa level of 0.5 V within two wave periods.
Figure 5 shows the frequency spectrum of a 170 V am-plitude output, as calculated from a fast Fourier transformwith a Hanning window. The spectral purity is similar to thatof the transistor oscillator of Ref. 6, although the fundamen-tal width is narrower, and the high frequency noise is moreclosely correlated with harmonics.
During free-run operation of the generator, the oscillatorfrequency f free drifts due to temperature variations, with a1 °C increase producing a 150 Hz f free decrease. After afew hours of continuous operation the unit reaches thermalequilibrium, and variations in f free are observed to be slowlyvarying minutes and within a 50 Hz range.
During frequency-locked operation of the generator, areference signal generated by a commercial generator is con-nected to the grid of tube U1. The generator output thenphase locks on to the input reference frequency f ref, providedthe difference f lock= f ref− f free is small enough. The rangeover which frequency lock occurs is measured to be linear
FIG. 3. Schematic of one of the clamp control circuits.
FIG. 4. Two wave forms showing the oscillator output when clamped.
063101-3 Frequency stabilized rf generator Rev. Sci. Instrum. 77, 063101 2006
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
(c) Clamp circuit.
Figure 3.7: Schematic of the frequency stabilized RF generator for ion trapsdesigned by Chang and Mitchell (Chang and Mitchell, 2006).
3. Quadrupole Ion Guide & Mass Filter 63
the frequency range of 350–750 kHz (Robbins et al., 2008). In such system,
the reference waveform was produced by a computer-controlled waveform gener-
ator, as shown in the detailed schematic in Fig. 3.8 (Robbins et al., 2008). At
the transformer output stage, a computer-controlled stepper motor was used to
change the shaft angular position of an air-gap variable tuning capacitor to match
the output impedance to a resonant frequency assigned by the computer.
connected to pin 5 COSC. The output frequency is con-trolled by an input current applied to pin 10 IIN. Internalcircuits shape the triangular waves to produce 2 V peak-to-peak sine waves. In the current configuration, frequency ac-curacy is approximately 10 kHz, which is sufficient for therf-only ion guide application. Higher frequency accuracy canbe obtained through the use of a crystal reference and aphase locked loop.
Following the signal path, the sine wave output of the
MAX038 is then sent through a 1.9 MHz low pass filterPLP-1.9, Minicircuits, Brooklyn, NY to remove unwantedhigh frequency components. The signal is then sent through a20:1 resistive divider to attenuate it for the dynamic rangeconsiderations of subsequent stages. After passing through abuffer stage OP27, Analog Devices, Norwood, MA the sig-nal, now a 50 mV sine waveform, is sent to a variable gainamplifier AD603, Analog Devices, Norwood, MA. TheAD603 is a linear-in-decibel amplifier whose gain is con-
FIG. 1. Simplified schematic of the rf circuit showing all of the functional elements.
FIG. 2. Detailed schematic of the rf circuit showing the pinouts and connections of the major components.
034702-3 Multipole ion guides power supply Rev. Sci. Instrum. 79, 034702 2008
Downloaded 20 Jun 2012 to 137.205.23.69. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
Figure 3.8: Detailed schematic of the RF circuit designed by Robbins et al.(Robbins et al., 2008).
In 2011, Jau et al. reported a low power RF oscillator using complementary
metal-oxide-semiconductor (CMOS) logic gates, which was utilized on a 2×2 cm
PCB for driving a small ion trap (2× 2× 10 mm) (Jau et al., 2011). This design
delivered frequencies from 0.1 to 10 MHz, and the output RF amplitude was
tested up to 400 V while the DC voltage supply to the system could be lower
than 7 V. The circuit is shown in Fig. 3.9 (Jau et al., 2011).
3. Quadrupole Ion Guide & Mass Filter 64
023118-2 Jau et al. Rev. Sci. Instrum. 82, 023118 (2011)
VRF
(1) (2)
0
U1A
1 2
U2D
9 8
U2A
1 2R1
0.5 -- 10 k
U1D
9 8
U1B
3 4
U2E
11 10
OutputTo RF electrodes
12
U2B
3 4
U1E
11 10
U1C
5 6
C10.5 -- 2 pF
12
U2F
13 12
U2C
5 6
L
10uH -- 1mH1 2
U1F
13 12
C0
12
FIG. 1. (Color online) The RF driving circuit is made of 12 CMOS inverters (two 74AC04 ICs) with an LC loop. A better performance can be achieved byusing four NC7NZ04 ICs (integrated circuits).
being the active components. By using the basic oscillationtheory, we have designed a simple, efficient RF driving cir-cuit as shown in Fig. 1.
A. Circuit characteristics
In the frequency range of our interest, the finite Q of theLC resonator is generally due to loss in the inductor. Thelosses due to the wire resistance, skin effect, and the hystere-sis can be well approximated as an effective resistor in serieswith an ideal inductor. The loss due to the eddy current is usu-ally modeled as an effective resistor parallel to the inductor.In both configurations for the effective resistor, the resonancevoltage across the capacitor has nearly a 90 phase lag to thedriving voltage across the LC . To achieve Barkhausen sta-bility criterion for an oscillator, we would like the gain of theclosed loop [from (1) to (2) and then back to (1) in Fig. 1] to begreater than one and the transient phase in the loop to be zeroor multiples of 2π at the LC resonance. For the ideal logicinverters with infinitesimal switching time, zero propagationdelay, and no input capacitance, the correct phase conditioncan be obtained if R1C1ω0 (Ag + 1). Here, Ag is the to-tal voltage gain of the first three inverters in series. Since theoscillation starts from a small perturbation in the loop to theinput of the inverter, the inverters are working at linear regimeat beginning. Once the oscillation builds up, the inverters cansaturate their outputs, and an additional phase, which can beincorporated into the propagation delay, can be introduced tothe loop. We will discuss the influence of this additional phasein Subsections III B. In practice, without losing the closedloop gain, we would like to lower C1 and increase R1 as muchas possible to reduce the power consumption from R1. For anideal circuit, the oscillation frequency is simply the LC reso-nant frequency,
f0 = 1
2π√
LC, (3)
where C = C0 + C1 + CL . Here, CL is the load capacitancefrom the connection of the output. If there is no limitation ofthe output current from the inverters, the output peak-to-peakRF voltage, VRFp−p = 2VRF, is determined by the Q of the
LC and the supply voltage VS of the circuit. Assuming thatthe sinusoidal driving signal across the LC circuit has peak-to-peak voltage VO , the voltage across the capacitor, VRFp−p,is boosted up by a factor of Q, and we find
VRFp−p = QVO = 4
πQVS . (4)
At the output of the inverters, the voltage is oscillating in asquare waveform between 0 and VS . So VO = 4VS/π , whichis the amplitude of the first harmonic of the square wave.
B. Effects of nonideal inverters
There are several effects of using realistic inverters. Forexample, the output impedance of the inverters can degradethe Q of the entire LC resonant circuit and also shift the res-onant frequency. In reality, the inverters have limited gain andcannot output infinite current. Therefore, the maximum out-put RF voltage is given by
VRFmax = IOmax
2π f0C, (5)
where IOmax is the maximum current that the CMOS inverterscan provide. Higher capacitive loads require higher IOmax tomaintain the maximum RF voltage.
To increase the circuit gain and amend the low currentoutputs of the CMOS logic gates, we use three inverters inseries to enhance the loop gain and nine inverters in paral-lel to raise the maximum output current. The 12 inverters inFig. 1 can be achieved by using two 74AC04 chips or fourNC7NZ04 chips (Fairchild Semiconductor Co.) or any otherswith similar characteristics. According to the specifications ofthese logic gates, the circuit of Fig. 1 is guaranteed to deliverthe peak current more than 200 mA to the LC circuit, and themeasured equivalent open-loop gain at VCC = 3.3 V from (1)to (2) is ∼140 dB for 74AC04. Practical logic gates have de-lays in signal propagation, which influence the RF oscillationespecially at higher frequency. From the manufacturer data,at room temperature, VCC = 3.3 V, and 50 pF capacitive load,the typical value of total propagation delay τD from (1) to (2)is 18 ns for 74AC04 and 11.6 ns for NC7NZ04. The propaga-tion delay is usually longer with heavier loads. This leads to
Downloaded 25 May 2011 to 137.205.124.72. Redistribution subject to AIP license or copyright; see http://rsi.aip.org/about/rights_and_permissions
Figure 3.9: The ion trap driving circuit using CMOS inverters, designed by Jauet al. (Jau et al., 2011).
As described by Eq. (3.10), the dimension of a quadrupole (r, the inscribed
circle radius) plays a role when one tries to tune the operation of a quadrupole
system into the stability regions of the stability diagram. Therefore, the spec-
ifications (in particular, the output frequency and amplitude) of the reviewed
existing power supply designs are different. The power supply circuits mentioned
above are summarised in Table 3.1.
Apart from the power supply circuits, a zero-method control circuit to regulate
the DC and RF voltages was reported by Tsukakoshi et al. in 2000 (Tsukakoshi
et al., 2000). In 2007, Franceschi et al. reported a matching network, with
capability to provide a DC offset, to match the output impedance of a commercial
RF generator to an ion guide system with high Q (Franceschi et al., 2007).
Another LC coupling network was reported by Canterbury et al. in 2010 for high
field asymmetric waveform ion mobility spectrometry (Canterbury et al., 2010).
3. Quadrupole Ion Guide & Mass Filter 65
Designer / Output Output MainYear Reported Frequency Amplitude Component
(Hz) (V)
(Austin et al., 1976) 4M 1000 vacuum tube
(Jones et al., 1997) 30M 50–600 vacuum tube(Jones and Anderson, 2000)
(O’Connor et al., 2002) 0.5–1.5M 0–500 BJT(Mathur and O’Connor, 2006)
(Cermak, 2005) 4–8M 200 MOSFET
(Chang and Mitchell, 2006) 800k 8–400 vacuum tube
(Robbins et al., 2008) 350–750k 5–500powerop amp
(Jau et al., 2011) 0.1–10M 400CMOSlogic gate
Table 3.1: Summary of the existing RF power supplies.
3.5 Quadrupole Power Supply Problems in Mass Filtering
Among the designs mentioned in Section 3.4, transformers are commonly used at
the output stage of the RF oscillator circuitry. Potentially, the output RF ampli-
tude can be modified by changing the turns ratio of the transformer. However, the
impedance of a transformer changes with its dimensions and the number of turns
of the coil. Quadrupole ion guides are capacitive loads. The resonant frequency
of the oscillator output is determined by the equivalent output inductance and
capacitance, which depend on the transformer size, tuning capacitors, cabling,
and the ion guide dimensions and resultant capacitance. When the operating
3. Quadrupole Ion Guide & Mass Filter 66
frequency is fixed, increasing the output-to-input turns ratio of the output trans-
former may theoretically increase the output amplitude, but the offset resonant
frequency due to the output impedance change could result in a dramatic am-
plitude decrease. It was commonly reported that the output RF amplitude was
changed after connecting the oscillator circuit to the quadrupole. The impedance
mismatch is believed to be the main reason causing such amplitude loss. On the
other hand, if the transformer is part of the resonance circuit, such impedance
change will modify the output frequency. Therefore, if in the RF oscillator stage
(the first stage, as shown in the diagram in Fig. 3.1) of a power supply, a LC
resonance circuit is used for generating the RF signal, such a power supply can
only drive an ion guide, not a mass filter.
Many of the ion guide power supply designs previously reported (Jones et al.,
1997; Jones and Anderson, 2000; O’Connor et al., 2002; Mathur and O’Connor,
2006) used a similar feedback scheme to start the oscillation. Those oscillators
operated with a self-tuned resonant frequency, which was set by the output trans-
former and the shunting capacitance. For instance, Fig. 3.10 (O’Connor et al.,
2002) shows the basic building block of the differential common-base power am-
plifier of the design by O’Connor and co-workers. The output transformer forms
part of the resonance circuit, and also generates feedback signal for the oscil-
lation. As a result, the oscillation frequency of such a circuit depends on the
impedance of the LC tank circuit, which is a combination of the impedance of
the transformer, the tuning capacitor, the quadrupole load to be connected, and
3. Quadrupole Ion Guide & Mass Filter 67
other stray impedance, at the output. However, in practice, a change in the
operating temperature will alter the capacitance of the components, causing a
resonant frequency drift.
isolating capacitors to remove the DC voltage used bythe transistors. A center tap on the secondary coil of thetransformer allows DC biasing of the multipoles. Thetransistors are operated in a common-base configura-tion in order to preserve as much of the transistorbandwidth as possible, whereas the more usual com-mon-emitter configuration would normally reduce the
bandwidth. The transistor selected for use is the Pana-sonic 2SC5392 (Digikey, Thief River Falls, MN), rated tohave a 3 dB cutoff frequency of 20 MHz in the common-base configuration.Feedback required to achieve oscillation comes from
a small feedback coil and pair of 200 series resistorsconnected between the transistor emitters. The seriesresistor is needed to convert the feedback voltage into afeedback current. The amplitude of the oscillation is setby the DC value of the emitter current, which isdetermined by the bias voltage to the emitter resistors.The feedback signal is 180° out of phase with theoscillator so that, once oscillation begins, the amplitudeof oscillation increases until the emitter current to theemitter receiving the positive voltage swing is reducedto zero. Since the two emitters are in series with thefeedback coil, the cutoff of emitter current to onetransistor also serves to limit the current to the otherone, preventing a further increase in oscillation ampli-tude. The reader will note that this limiting mechanismdoes not depend on saturation of the collector voltage,which would significantly alter the transistor character-istics and probably shift the oscillation frequency awayfrom that of the tank circuit. Small signal transistoroscillators often do depend on transistor saturation tolimit oscillation amplitude, but this is only satisfactorywhen transistors being used exhibit negligible sideeffects due to saturation.The RF amplitude is controlled by adjustment of the
emitter current. When the negative bias voltage to theemitters is increased, the feedback amplitude requiredto cut off emitter current will also increase, allowing theoscillation amplitude to increase. Thus, the voltagedeveloped across the tank circuit is proportional to theemitter bias voltage.
Coil Design
Devising an acceptable coil design was a major part ofthis oscillator project. Many coils were wound beforesettling on the present design. To maintain as high a Q(meaning low loss) as possible, relatively large wiresizes were used; the tank circuit is wound from 18-gauge (1.024 mm diameter) copper magnet wire and thetransistor drive coil and feedback coil use 16-gaugecopper magnet wire (1.291 mm diameter).A high coupling constant between the primary and
the secondary coil of the transformer is crucial. Withoutit, the transistors, which are hooked to the primary coil,will tend to oscillate at a frequency other than that setby the inductance (provided by the secondary coil) andthe capacitance of the tank circuit. It was found to bedifficult to adequately couple the transistor winding toa relatively long center-tapped tank coil because eachtransistor tended to drive only its own half of the tankcoil rather than the entire coil. Splitting the secondarycoil into two identical windings, one on top of the other,with the primary coil on top of both, proved to give thebest coupling. The use of overlapping tank coils also
Figure 1. The basic common-base transistor oscillator circuitcoupled to a traditional LC tank circuit on the multipole rods.
Figure 2. The complete diagram for the oscillator and automaticgain control (AGC) feedback circuit. Unless otherwise noted,resistance values are in ohms, and capacitance values are inmicrofarads.
1371J Am Soc Mass Spectrom 2002, 13, 1370–1375 RF OSCILLATOR FOR MULTIPOLE ION GUIDES
Figure 3.10: The feedback scheme with common-base setup for oscillation inthe 2002 design by O’Connor et al. (O’Connor et al., 2002).
Meanwhile, in Mathur and O’Connor’s design (Mathur and O’Connor, 2006),
an AGC unit was built to sense the output at the transformer/matching circuit
third stage to modify the gain at the power amplifier second stage. The AGC
unit seemed to stabilize the output amplitude. However, a Zener diode was
used in the regulator circuit (shown in Fig. 3.11) reported in 2002 (O’Connor
et al., 2002) to provide the +15-V DC voltage in the 2006 circuit (Mathur and
O’Connor, 2006). The output voltage after a Zener diode is a function of the
biasing current, which changes according to the load shunting this Zener diode.
3. Quadrupole Ion Guide & Mass Filter 68
Therefore, such a supply voltage drift changes the DC conditions of not only the
operational amplifiers (op amps) in both the AGC unit and the regulator circuit,
but also the amplifying transistors. Such instabilities cause a constant, more
than 1% change to the output amplitude.
reduces the required number of turns in the tank coilbecause of the increased inductance of overlappingcoils.The final coil design uses a 5.08 cm diameter poly(vinyl
chloride) PVC coil form with each half of the secondarytank coil consisting of 30 turns of 18-gauge magnet wire.The primary transistor winding consists of 20 turns of16-gauge magnet wire, center-tapped, essentially com-pletely covering the tank coil windings. Three layers of0.004 inch (0.1 mm) thick Mylar (DuPont, Wilmington,DE) serve to insulate the tank coil halves from each otherand the transistor coil from the outer tank coil half. Thefeedback coil consists of 6 turns of 16-gaugemagnet insidethe coil form, with the turns spread to equal the length ofthe other coils. There is also a single turn sensing loopused by the feedback (AGC) circuit; this is merely a singleloop of insulated hookup wire placed over the center ofthe other windings.
Transistors and Power Dissipation
The experimentally determined full power require-ments at maximum (1000 Vp-p) amplitude is about 40 Wat 500 kHz and 20 W at 1.5 MHz, with half the powerbeing dissipated in the tank circuit and the other half inthe transistors. The transistor chosen for the oscillator isthe Panasonic 2SC5392. This is a 500 V 20 MHz transis-tor in a TO-220 case with a 25 W dissipation at a 25 °Ccase temperature. However, actual dissipation is deter-mined by thermal design so that, with a good heat sinkand fan, 10 W of heat dissipation would be considereda reasonably conservative estimate. Since the two oscil-lator transistors would have to dissipate up to 20 W (10W each), it was decided to use parallel transistors, fourtotal, to limit the dissipation of each transistor to 5 W.This is a sufficient derating to allow continuous opera-tion at 20 W with a minimum chance of prematuretransistor failure.This transistor also has a relatively low collector-base
capacitance for a power transistor of about 100 pF.Because of the 3:1 turns ratio between the tank coil andthe driver coil, this capacitance will effectively be only11 pF across half the tank coil, thereby not significantlyinfluencing the tuning of the tank circuit. However,connecting the pairs of transistors in parallel, as isdiscussed above, also doubles the effective transistorcapacitance seen by the tank coil to 22 pF, but this valueis still acceptable.The 3:1 turns ratio between tank coil and driver coil
means that with the maximum 1000 Vp-p across the tankcoil, only 333 Vp-p will appear across each transistor.With a transistor supply voltage of 215 V, the collectorwill swing from 49 to 381 V. Thus, collector saturation isavoided.
Emitter Feedback
Due primarily to the decision to use parallel transistors,the coupling of the feedback coil to the emitters in-
volves using multiple resistors and decoupling capac-itors to guarantee an equal power distribution amongthe four transistors, as seen in Figure 2. The readerwill also note also that there are two additionalresistors that effectively connect the two emitterpairs. These resistors serve to limit the amplitude ofthe reverse emitter-base voltage when the oscillationamplitude is limiting. Before limiting occurs, theemitter–emitter impedance is very low so these resis-tors have almost no effect on circuit operation. How-ever, when limiting begins, the impedance of thetransistor that is cut off becomes rather high andmight allow the reverse emitter-base voltage to ex-ceed the maximum rating of the transistor, were it notfor the shunting effect of these resistors.
Automatic Gain Control
AGC is provided by comparing the rectified signal fromthe sensor loop with the desired signal using a NationalLM6134 op-amp (Digikey, Thief River Falls, MN) sothat a 0–10 Vdc will linearly correlate with a 0–1000 Vp-p
output. The difference between the measured and de-sired signals is amplified by the op-amp and used todrive a Darlington-connected transistor pair to providethe emitter bias for the oscillator transistors. The gainand frequency response of the amplifier was chosen toguarantee a stable AGC circuit.
Power Supply
The supply voltage used by the transistors of theoscillator is 215 V, 200 V between the collectors andbases and 15 V to power the op-amp and Darlingtondriver. It might be possible to use an unregulatedsupply and depend upon the AGC circuit to maintainthe proper oscillation amplitude in the presence of avarying supply voltage, but the cost and complexity ofregulating the supply voltage is minimal, so a regulatoris included as a part of the design. Figure 3 shows thepower supply diagram.
Figure 3. The power supply for the oscillator. Unless otherwisenoted, resistance values are in ohms, and capacitance values are inmicrofarads.
1372 O’CONNOR ET AL. J Am Soc Mass Spectrom 2002, 13, 1370–1375
Figure 3.11: The 215-V regulator and the 15-V DC voltage supply in the 2002design by O’Connor et al. (O’Connor et al., 2002).
As a result, Mathur and O’Connor’s oscillator can be a good power supply to
operate an ion guide, but for driving a mass filter with narrow mass window, the
output frequency and amplitude stabilities have to be improved. In Chapter 7,
a new oscillator design is proposed and partially tested. It is believed that the
new design can be a RF power supply for driving a quadrupole mass filter.
3.6 Conclusion
In this chapter, the theory of operating a quadrupole as a ion guide or a mass
filter is introduced. A stable ion trajectory can be obtained when the stability
3. Quadrupole Ion Guide & Mass Filter 69
parameters are tuned to be inside the stability regions.
Existing power supplies for a quadrupole system are reviewed. Most of them
operate at a frequency below 10 MHz and has an output amplitude less than
500 V, and are suitable for ion transportation. When a quadrupole is used for
mass filtering, very stable output frequency (ω) and amplitude (V ) have to be
generated by the power supply. In particular, it is preferred to have ∆ω/ω and
∆V/V below 2.5×10−4 and 5.0×10−4, respectively, for a 0.1-Da resolution. The
design of a new RF power supply for driving a quadrupole mass filter will be
proposed later in Chapter 7.
CHAPTER 4
Test Equipment & Software Programs
Test automation was widely used to test the circuits in this thesis. This chapter
presents the testing equipment, computer softwares, and testing methods used to
test the designed circuits reported in the following chapters (Chapters 5, 6, and
7).
4.1 Introduction
The PCI (Peripheral Component Interconnect) eXtensions for Instrumentation
(PXI) platform15 and the control software LabVIEW from National Instruments
(Austin, Texas) were utilized to test the performance of the designed circuits.
It allowed faster circuit test execution and reporting, and a larger number of
sampling points for an more accurate results after averaging.
Apart from the National Instruments (NI) system, the direct current (DC)
power supply TTi EL301R from Thurlby Thandar Instruments Ltd. (Hunting-
15See http://www.ni.com/white-paper/4811/en for information about the PXI system, ac-cessed 15 August 2012.
70
4. Test Equipment & Software Programs 71
don, UK), and the lead-acid battery LC-R067R2P from Panasonic Corporation
(Osaka, Japan) were used as power supplies. The spectrum analyser IFR A-7550
from Aeroflex Inc. (Plainview, New York, USA) was used for noise analysis.
The oscilloscope Tektronix (Beaverton, Oregon, USA) DPO2014 was also used
both to monitor, and to perform the fast Fourier transform (FFT) on the output
waveforms. Simulation Program with Integrated Circuit Emphasis (SPICE) sim-
ulation was carried out by using the computer simulation software NI Multisim
v11.0.2. The circuit schematics reported in Chapters 5 and 6 were drawn also
using NI Multisim.
4.2 NI PXI Platform
The NI PXI system used for this report includes a PXI-5122 oscilloscope, a PXI-
5421 arbitrary waveform generator, a PXI-6733 analogue output card, a PXI-8336
control card, and a PXI-1042 chassis. Figure 4.1 shows this PXI system and two
DC power supplies. The 2-channel NI PXI-5122 oscilloscope provides a sampling
rate of 100 MS/s and a 14-bit resolution with 100-MHz bandwidth.16 The NI PXI-
6733 analogue output card has a output voltage range between -10 and +10 V
and a current driving ability of 5 mA.17 The combination of the PXI-5122 and
the PXI-6733 can be very useful for testing the I-V characteristics of a transistor.
NI PXI-5421 arbitrary waveform generator can generate any arbitrary waveform
16See http://sine.ni.com/nips/cds/view/p/lang/en/nid/12615 for information about the NIPXI-5122 oscilloscope, accessed 20 August 2012.
17See http://sine.ni.com/nips/cds/view/p/lang/en/nid/11311 for information about the NIPXI-6733 analogue output card, accessed 20 August 2012.
4. Test Equipment & Software Programs 72
between the frequency range of DC and 43 MHz.18 The frequency response of
a amplifier system can be measured by setting up an alternating current (AC)
analysis using both PXI-5122 and PXI-5421. The controlling and data collecting
programs used with the PXI system will be discussed in the next section.
Figure 4.1: NI PXI platform and DC power supplies for circuit test.
18See http://sine.ni.com/nips/cds/view/p/lang/en/nid/12714 for information about the NIPXI-5421 arbitrary waveform generator, accessed 20 August 2012.
4. Test Equipment & Software Programs 73
4.3 NI LabVIEW
The software NI LabVIEW 2009 (Service Pack 1) was used to control and to
collect data from the PXI system shown in Fig. 4.1.
4.3.1 I-V Characteristics
The I-V characteristics of transistors reported in the later chapter was measured
using the NI PXI-5122 oscilloscope (for measuring DC voltages) and PXI-6733
analogue output card (for supplying DC voltages). Figure 4.2 illustrates the
schematic of the circuit used for transistor I-V characteristic testing. The drain
node of the transistor being tested is coupled to an output channel of the PXI-
6733 card through a resistor R. The gate node of the transistor is coupled directly
to another output channel of the PXI-6733 card. The voltage values at both the
drain node and the gate node of the transistor (VDS and VGS, respectively) are
monitored by the PXI-5122 oscilloscope.
The LabVIEW program utilized for control and data collection is shown in
both Fig. 4.3 (front panel) and Fig. 4.4 (block diagram). The parameters at
the front panel control both PXI-5122 and PXI-6733 cards. The program scans
both voltages applied to nodes ’DC 1’ and ’DC 2’ shown in Fig. 4.2 according to
the settings at the front panel. The window ’Measurements’ shown in Fig. 4.3
is not used, since only one measurement is recorded for each voltage step. The
recorded voltage information will be converted to data sets of gate-source volt-
age VGS, drain-source voltage VDS, and drain current ID, for plotting the I-V
4. Test Equipment & Software Programs 74
0 1 2 3 4 5 6
A
B
C
D
E
R VD
DC 1
GDC 2 V
Figure 4.2: Schematic of the transistor I-V characteristic testing circuit.
characteristics.
4.3.2 AC Analysis
The AC analyses reported in the following chapters were carried out with the
NI PXI-5122 oscilloscope and PXI-5421 arbitrary waveform generator. The DC
voltages needed were supplied by the DC power supplies. The LabVIEW program
for reporting frequency responses is shown in both Fig. 4.5 (front panel) and
Fig. 4.6 (block diagram).
This AC analysis program uses PXI-5421 arbitrary waveform generator to
send out specified waveforms. The frequency of the waveform is scanned ac-
cording to the specified starting/stop frequency and frequency steps. Multiple
measurements are taken for each frequency step, according to the parameter ’no.
of measurements’ at the front panel shown in Fig. 4.5. Peak-to-peak voltage
4. Test Equipment & Software Programs 75
Characteristics JFET.vi
C:\LabVIEW Data\Characteristics JFET.vi
Last modified on 13/01/2011 at 00:04
Printed on 15/07/2012 at 16:48
Page 2
Front Panel
Channel Parameters
PXI1Slot5
PXI1Slot3
PXI1Slot4
Select Device
10.00
Maximum Value
-10.00
Minimum Value
0
0
0
0
Output Array (Channel 0)
4096
Min. Record Length
100.00
Max Input Freq
0,1Input Channel2.85778scalar result
3.43785mean
0.785048stdev
2.69292min
4.61755max
820num in stats
0
Measurements
Voltage Average
Scalar Measurement
STOP
-1.00Max.Time (s)
Clear Stats
1.000MMin. Sample Rate
0.5
-0.0
0.1
0.2
0.3
0.4
2.047m-2.048m
Enforce Realtime
20.00
Vertical Range
C:\LabVIEW Data\temp\
Output Path
temp
File Name
PXI1Slot2
Resource Name reset device
0.05
Vg Increment
0.01
Vg max voltage
6
Active ao# for Vg
7
Active ao# for Vcc
0.1
Vcc Increment
10
Vcc max voltage
9.5
-0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
Vds
100 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5
Vgs = -0.500
Vgs = -0.450
Vgs = -0.400
Characteristic
-0.5
Vg initial voltage
0
Current Vcc Value
0
Current Vg Value
status
0
code
source
error out
Figure 4.3: Front panel of the LabVIEW program for obtaining I-V character-istics of a transistor.
4. Test Equipment & Software Programs 76
Chara
cteri
stic
sJF
ET.v
i
C:\
LabVIE
WD
ata
\Chara
cte
rist
ics
JFET.v
i
Last
modifie
don
13/0
1/2
011
at
00:0
4
Pri
nte
don
15/0
7/2
012
at
16:4
8
Page
3
Blo
ck
Dia
gra
m
Min
imum
Valu
e
Maxim
um
Valu
e
Sele
ctD
evic
e
Outp
ut
Arr
ay
(Channel0)
.txt
C:\
LabV
IEW
Data
\Pre
Am
p2010_LabV
IEW
\
%.5
f
Ou
tpu
tP
ath
File
Nam
e
Vg
Incr
em
ent
reso
urc
enam
e
rese
tdevic
e
Analo
gD
BL
1C
han
1Sam
p
-1
0
AO
Volt
ag
e
8
Vg
max
volt
age
Vg
Volt
age
Arr
ay
1
Analo
gD
BL
1C
han
1Sam
p
0
200
vert
icalra
nge
Enfo
rce
Realt
ime
Acq
uis
itio
nG
raph
min
sam
ple
rate
Tru
e
Cle
ar
Sta
ts?
Max.T
ime
(s)
Norm
al
acq
uis
itio
nT
ype
scala
rm
easu
rem
ent
1D
Clu
ster
Sta
tist
ics
1D
Clu
ster
Measu
rem
en
ts
0 1
DC
1
pro
be
att
enuati
on
min
reco
rdle
ngth
Analo
gD
BL
1C
han
1Sam
p
stop
2statu
s0
Curr
ent
Vcc
Valu
e
1
100
Curr
ent
Vg
Valu
e
Vgs
=
1
dt
t0Y
Vcc
Incr
em
ent
00:0
0:0
0.0
00
DD
/MM
/YYYY
NI_
ChannelN
am
e
3
AO
Volt
ag
e
Acti
ve
ao#
for
Vg
Act
ive
ao#
for
Vcc
AO
Volt
ag
e
Vcc
Incr
em
ent
0
Vcc
max
volt
age
Vcc
Volt
age
Arr
ay
1
input
channel
Max
Input
Fre
q
1m
ega
oh
m
Analo
gD
BL
1C
han
1Sam
p
0
0
Analo
gD
BL
1C
han
1Sam
p
0
0
Vg
init
ialvolt
age
Chara
cterist
ic
err
or
out
Fals
e
Figure 4.4: Block diagram of the LabVIEW program for obtaining I-V charac-teristics of a transistor.
4. Test Equipment & Software Programs 77
amplitude is measured by the PXI-5122 oscilloscope. After all the measurements
are completed, the mean, standard deviation, and minimum/maximum values of
each frequency step will be recorded. Since the memory size of the computer
used was big enough, the maximum sampling rate of 100 MS/s was always used
for both channels of the PXI-5122 oscilloscope. Twenty points were measured
and averaged to obtain both the input current going into the node Iin and the
output voltage measured at node Vout in the schematic figures shown in the later
chapters.
The transimpedance measured by the NI PXI system is reported using the
commonly used form of Bode magnitude plots. The magnitude axis of such a
plot is often reported in decibel scale. Since the transimpedance (gain) of a tran-
simpedance amplifier has a unit of V/A = Ω, in a Bode plot, the transimpedance
is reported using dBΩ, where dBΩ is defined as X (dBΩ) = 20 × log[Y (Ω)], in
which X and Y are the transimpedance in dBΩ and Ω, respectively (Lin et al.,
2012).
For the frequency response reported in Chapter 5, the peak-to-peak amplitude
of the testing input sinusoidal current is 0.12 mA, whereas in Chapter 6 the peak-
to-peak amplitude of the testing input sinusoidal current is 120 µA for circuits
with estimated overall transimpedance below 20 kΩ, and 20 µA for circuits with
estimated transimpedance above 1 MΩ.
4. Test Equipment & Software Programs 78
AC
Analy
sis
Fgen
&niS
cope
-fo
rPre
Am
p(a
lldata
).vi
C:\
LabVIE
WD
ata
\Pre
Am
p2011cs_
LabVIE
W\A
CAnaly
sis
Fgen
&niS
cope
-fo
rPre
Am
p(a
lldata
).vi
Last
modifie
don
11/0
1/2
012
at
17:2
9
Pri
nte
don
15/0
7/2
012
at
16:4
2
Page
2
Fro
nt
Panel
PXI1
Slo
t2
Reso
urc
eN
am
e
0,1
Input
Channel
1000
Min
.R
ecord
Length
2.4
0
Vert
icalR
ange
Enfo
rce
Realtim
e
0.0
-0.0
0.0
4.9
94u
-4.9
96u
100.0
00M
Min
.Sam
ple
Rate
10.0
00
Low
Ref.
50.0
00
Mid
Ref.
90.0
00
Hig
hR
ef.
Perc
enta
ge
Refe
rence
Level
Unit
s
Wavefo
rmM
easu
rem
ent
Vert
icaland
Channel
Tri
ggeri
ng
Tim
ing
Edge
Tri
gger
Type
5.0
0M
ax.T
ime
0.0
0Tri
gger
Level
STO
P
statu
s
0
code
sourc
e
scope
err
or
out
Negative
Posi
tive
Tri
gger
Slo
pe
DC
Tri
gger
Coupling
Voltage
Am
plitu
de
Sca
lar
Measu
rem
ent
Edge
trig
geri
ng
para
mete
rs
0.0
70329
scala
rre
sult
0.0
708463
mean
0.0
0975734
stdev
0.0
445989
min
0.0
824219
max
20
num
inst
ats
0
Measu
rem
ents
Channel
1Tri
gger
Sourc
e
AC
Vert
icalCoupling
rese
tdevic
e
1.0
Pro
be
Att
enuation
20
no
.o
fm
ea
su
re
me
nts
PXI1
Slo
t6
Re
so
urce
Na
me
0.0
0
Sta
rtP
ha
se
Nois
e
Sin
e
Square
Tri
angle
Ram
pU
p
Ram
pD
ow
n
DC
Wa
ve
form
6.0
-6.0
-4.0-2
.00.0
2.0 4
.0
0.0
0
DC
Off
se
t(V
)
10.0
0.0
2.0
4.0
6.0
8.0
0.0
6
Am
pli
tu
de
(V
)
10.0
M0.0
2.0
M
4.0
M6.0
M
8.0
M
11.0
k
Fre
qu
en
cy
(H
z)
0
Confi
gure
dC
hannels
statu
s
0
code
sourc
e
fun
gen
err
or
out
1 0.7
20187
0.7
19773
0 0 0
Outp
ut
Arr
ay
(Channel
0)
0.1
15686
1 0.1
16235
0 0 0
Outp
ut
Arr
ay
(Channel
1)
0.1
15384
0.1
15344
0.1
15384
0.1
15344
0.1
15088
0.1
16235
0.1
15982
0.1
16182
0.1
15982
0.1
15928
0.1
15674
0.1
16182
0.1
1615
0.1
1615
0.1
16587
0.1
16182
0.1
16238
0.7
19773
0.7
20475
0.7
19811
0.7
19858
0.7
19184
0.7
17865
0.7
17008
0.7
16372
0.7
13646
0.7
08742
0.6
99595
0.6
78103
0.0
0144637
0.0
00897892
0.0
010506
0.0
0108005
0.0
00938725
0.0
0108462
0.0
00847866
0.0
011017
0.0
0109257
0.0
0119826
0.0
0117982
0.0
0104897
0.7
16792
0.7
1884
0.7
17329
0.7
17763
0.7
18032
0.7
15878
0.7
15878
0.7
14662
0.7
12376
0.7
07259
0.6
97357
0.6
76536
0.7
21963
0.7
22234
0.7
21693
0.7
21693
0.7
20882
0.7
2007
0.7
18718
0.7
18573
0.7
15744
0.7
12136
0.7
01471
0.6
80322
11000
15743.8
22533.3
32250.9
46159.1
66065.3
94556.1
135334
193697
277229
396784
567898
0.1
15686
0.1
15687
0.1
15755
0.1
15682
0.1
15656
0.1
1549
0.1
15557
0.1
15665
0.1
15755
0.1
15807
0.1
15836
0.1
15516
0.0
00238651
0.0
00190353
0.0
00235001
0.0
00243509
0.0
00189561
0.0
00173968
0.0
00250048
0.0
00252952
0.0
00236028
0.0
00269304
0.0
00261644
0.0
00323324
0.1
15226
0.1
15451
0.1
1525
0.1
15226
0.1
15198
0.1
15165
0.1
15165
0 0
Mean
Valu
es
(Channel1,
Ch0
[mean,
stdev,
min
,m
ax])
20
Fre
qS
tep
s
1M
Sto
pF
re
qu
en
cy
40.0
0M
Max
Input
Fre
q
1m
ega
ohm
Input
Impedance
C:\
LabV
IEW
Data
\Pre
Am
p2011cs_
LabV
IEW
\
Outp
ut
Path
tem
p
File
Nam
e
50
ohm
s
Im
pe
da
nce
1E+
10
multip
lier
for
reco
rdle
ngth
Figure 4.5: Front panel of the LabVIEW program for the AC analysis.
4. Test Equipment & Software Programs 79
AC
Analy
sis
Fgen
&niS
cope
-fo
rPre
Am
p(a
lldata
).vi
C:\
LabVIE
WD
ata
\Pre
Am
p2011cs_
LabVIE
W\A
CAnaly
sis
Fgen
&niS
cope
-fo
rPre
Am
p(a
lldata
).vi
Last
modifie
don
11/0
1/2
012
at
17:2
9
Pri
nte
don
15/0
7/2
012
at
16:4
2
Page
3
Blo
ck
Dia
gra
m
scerr
or
out
Resourc
eN
am
e
111
10
2
Configure
dC
hannels
frequency
(Hz) A
mplitu
de
(V)
DC
Off
set
(V)
Wavefo
rm
Sta
rtPhase
34
Sta
ndard
Functi
on
fgerr
or
out
Outp
ut
Arr
ay
(Channel0)
Outp
ut
Arr
ay
(Channel1)
Mean
Valu
es
(Channel1,
Ch0
[mean,
stdev,
min
,m
ax])
vert
icalra
nge
Enfo
rce
Realtim
e
niS
cope
Chan.
Base
dM
idR
ef
Chan.
Base
dLow
Ref
Chan
Base
dH
igh
Ref
Ref.
LevelU
nit
s
Acti
ve
Channel
Acquis
itio
nG
raph
min
sam
ple
rate
low
refe
rence
mid
refe
rence
hig
hre
fere
nce
Refe
rence
LevelU
nit
s
trig
ger
type
Max.T
ime
Imm
edia
teR
ef
Tri
gger
"Im
media
te"
50
stop
statu
s
scerr
or
out
Update
Err
or
Out
(to
show
warn
ings)
Norm
al
acquis
itio
nType
scala
rm
easu
rem
ent
23
4
5
6
7
8
1D
Clu
ster
Sta
tist
ics
1D
Clu
ster
Measu
rem
ents
vert
icalcoupling
pro
be
att
enuati
on
no.
of
measu
rem
ents
statu
s
Am
plitu
de
(V)
pro
be
att
enuati
on
40
0 1
67
3
statu
s
3
100
pro
be
att
enuati
on
4
input
channel
100M
Max
Input
Fre
q
input
impedance
min
record
length
1
16777116
multip
lier
for
record
length
Fre
qSte
ps
Sto
pFre
quency
9 Fre
qSte
ps
.txt
C:\
LabV
IEW
Data
\Pre
Am
p2011cs_LabV
IEW
\%
.7f
Outp
ut
Path
File
Nam
e
Impedance
reso
urc
enam
e
rese
tdevic
e
pro
be
att
enuati
on
Figure 4.6: Block diagram of the LabVIEW program for the AC analysis.
4. Test Equipment & Software Programs 80
4.4 Noise Performance 19
It had been reported that the RF noise from the switching power supply may
affect the noise performance of a preamplifier (Mathur et al., 2007; Wu et al.,
2010). The lead-acid batteries mentioned above were introduced to compare the
noise performance. Noise performance tests were conducted at an unshielded en-
vironment using the preamplifier reported in Chapter 5. Such test concluded that
no difference in noise performance was observed, when using either the switching
power supplies or the lead-acid batteries as the power supplies to the preampli-
fier used. It could be the results of the proper grounding and bypass/decoupling
capacitors used.
The noise performance reported in the following chapters was measured when
the amplifier input was floating and a DC blocking capacitor was coupled between
the output and the spectrum analyser. The noise power was measured in dBm.
The noise power in dBm, Pn(dBm), can be calculated from the noise power in
watts, Pn(W ), by
Pn(dBm) = 10× log(Pn(W )
1× 10−3) . (4.1)
On the contrary, the measured noise power in dBm can be converted back to
Watts, and can be referred to the output voltage Vout by
Pn(W ) = 10(Pn(dBm)−30
10) =
(Vout)2
Rload
, (4.2)
19This section is partially reproduced from the journal article, ”A Gain and BandwidthEnhanced Transimpedance Preamplifier for Fourier-Transform Ion Cyclotron Resonance MassSpectrometry,” Review of Scientific Instruments, in 2011 (Lin et al., 2011).
4. Test Equipment & Software Programs 81
where Rload is the load resistance of the spectrum analyser. The transimpedance
AT of the preamplifier and the resolution bandwidth (Res) of the spectrum anal-
yser have to be considered to further correlate the output voltage Vout of the
preamplifier back to the input current spectral density in, where
in =
√Pn(W ) ×Rload
(Res)× (AT )2(pA/
√Hz) . (4.3)
By using Eq. (4.2) and Eq. (4.3), the input current noise spectral density in
(in pA/√
Hz) data were calculated from the measured noise power Pn(dBm) (in
dBm) and the transimpedance AT data collected using the AC analysis technique
mentioned in Section 4.3.2.
4.5 Other Software/Hardware Used
Apart from the NI LabVIEW and Multisim mentioned previously, other software
programs were used to assist other related works.
In Chapters 5 and 6, the printed circuit board (PCB) layouts were designed
using the computer-aided design software, Altium Designer Build 8.4 (Service
Pack 4) from Altium Limited (Sydney, Australia). The PCBs used in both chap-
ters were manufactured in-house using the PCB prototyping plotter ProtoMat
S62 from LPKF Laser & Electronics AG (Garbsen, Germany).
The software Proteus 7.10 from Labcenter Electronics (Grassington, UK) was
used for some of the circuit drawings and all of the PCB designs reported in
4. Test Equipment & Software Programs 82
Chapter 7. The designed PCBs reported in Chapter 7 were etched, and populated
in-house.
4.6 Conclusion
In this chapter, the methods used to test the designed circuits reported in the fol-
lowing chapters are presented. The NI PXI system allows a reliable and efficient
testing of circuits. The LabVIEW programmes used to control the PXI cards
and to fetch data have a great flexibility for defining circuit testing conditions.
When testing the noise performance of a circuit, proper shielding and power sup-
ply decoupling should be provided. As of the spectrum analyser, a more recent
model with better sensitivity may be necessary when testing the noise behaviour
of a circuit with excellent noise performance.
CHAPTER 5
Transimpedance Preamplifier Using an
Operational Amplifier
This chapter reports a preamplifier using an operational amplifier in a tran-
simpedance configuration, and is partially reproduced from the journal article,
”A Gain and Bandwidth Enhanced Transimpedance Preamplifier for Fourier-
Transform Ion Cyclotron Resonance Mass Spectrometry,” Review of Scientific
Instruments, in 2011 (Lin et al., 2011).
An ion detector is one of the components in a modern mass spectrometer, as
discussed in Section 2.1.2. Usually, an ion detector consists of electronic devices
for processing the analogue signal coming from the mass analyser, as shown in
Fig. 5.1. The signal is amplified and filtered here, before being sent into an
analogue-to-digital converter (ADC) in the data system. The preamplifier in an
ion detector system is believed to be one of the key components for improving
signal-to-noise performance electronically. Newly designed preamplifiers will be
83
5. Transimpedance Preamplifier Using an Operational Amplifier 84
presented in this chapter and Chapter 6.
+
_
Preamplifier Instrumentation
amplifiers Filters
Figure 5.1: Components inside the ion detector of a modern mass spectrom-eter (the composition of a modern mass spectrometer is illustrated in Fig. 2.4).Chapters 5 & 6 report new preamplifier designs for a 12-T FT-ICR system.
This chapter begins with a brief review of the transimpedance technique and
the input capacitance tolerance of a transimpedance amplifier. Then the corre-
lation between the cyclotron frequency of the signal form an ICR cell and the
transimpedance (gain) of the preamplifier is reviewed to understand the tran-
simpedance amplifier design constraints. The newly designed preamplifier and
its printed circuit board (PCB) for testing are presented. The preamplifier is
SPICE simulated and is tested on the bench for its frequency response and its
noise performance. Then this chapter concludes with a discussion between the
feedback impedance, bandwidth, noise performance, and the estimated numbers
of ions that can be detected in a 12-T FT-ICR system.
5.1 Introduction
It was discussed in Chapter 2, that in a signal processing chain, the noise per-
formance is dominated by the noise from the signal source and from the first
5. Transimpedance Preamplifier Using an Operational Amplifier 85
stage ’signal processor,’ the preamplifier, if the preamplifier is designed with sig-
nificant gain. Similar to the photodiodes in optical communication systems, the
signal analyser of a Fourier-transform ion cyclotron resonance (FT-ICR) mass
spectrometer, an ICR cell, is a capacitive device. Electrostatically induced image
currents are expected at the input of the preamplifier. The parasitic capacitance
from the cell can vary from around 10 pF to over 100 pF (Kaiser et al., 2011a),
depending on the cell dimensions, feedthroughs, and cabling. A high capacitance
at the preamplifier input limits the bandwidth, causing potential signal intensity
loss.
The nature of the ion signal from an FT-ICR mass spectrometer and the
electronic noise were studied and reported in Chapter 2 to further understand
the electronic detection limit. A new transimpedance preamplifier was designed,
computer simulated, built, and tested. The preamplifier design featured its en-
hanced tolerance of the capacitance of the detection device, lower intrinsic noise,
and larger flat mid-band gain (input current noise spectral density of around
1 pA/√
Hz when the transimpedance is about 85 dBΩ).
The designed preamplifier has a bandwidth of ∼3 kHz to 10 MHz, which
corresponds to the mass-to-charge ratio, m/z, of approximately 18 to 61k for
a 12-T FT-ICR system. The transimpedance and the bandwidth can be easily
adjusted by replacing passive components. The feedback limitation of the circuit
will be discussed in this chapter.
5. Transimpedance Preamplifier Using an Operational Amplifier 86
5.2 Transimpedance Amplifier
Since the signal to be detected in an FT-ICR system is a current due to the
motion of ions, a transimpedance amplifier to convert the current input to a
voltage output for further amplification is needed. As reviewed in Section 2.5.3,
conventional preamplifiers convert the signal current into voltage by using an
input resistor and a voltage amplifier. Such a current-voltage conversion can be
also performed by a transimpedance amplifier.
5.2.1 Bandwidth Extension
A transimpedance amplifier is a widely used current-voltage converting solution
for many applications such as optical communication (Green et al., 2008a; Chen
et al., 2005; El-Diwany et al., 1981). The negative-feedback transimpedance
technique has the advantages of reducing the effective input load capacitance to
the amplifier and extending the bandwidth by a factor equal to the open-loop
gain of the amplifier (Green and McNeill, 1989; Hullett and Moustakas, 1981).
Therefore, it appears to be ideal for detecting ion signals from an ICR cell, which
in general acts like a current source (Comisarow, 1978) with large capacitance
(from ∼10 pF to over ∼100 pF, as mentioned earlier) depending on cabling and
the size of the cell.
A typical FT-ICR preamplifier can be modeled from a circuit schematic such
as Fig. 5.2a, where C is the total source capacitance from the ideal alternating
current (AC) current source I, which represents the current signal from the ICR
5. Transimpedance Preamplifier Using an Operational Amplifier 87
cell. The operational amplifier (op amp) with open-loop gain G has an input
resistor R, which is responsible for transforming the input current into voltage.
The current-to-voltage transfer function H(ω) can be derived as
H(ω) =R
1 + jω(RC), (5.1)
and the 3-dB bandwidth ω3dB is given by
ω3dB =1
RC. (5.2)
If the configuration of such an amplifier is replaced by a transimpedance amplifier
with negative feedback, the circuit becomes what is shown in Fig. 5.2b. The
transfer function becomes
H(ω) =R
1 + jωR(C/G). (5.3)
Namely, the transimpedance technique effectively reduces the input capacitance
by a factor of G, hence the 3-dB frequency is then extended to
ω3dB =G
RC. (5.4)
With such bandwidth increment, the gain-bandwidth product is further pushed
to higher frequency range and therefore higher gain can be introduced to the am-
5. Transimpedance Preamplifier Using an Operational Amplifier 88
(a) A unity gain operational amplifier with an input resistor R to convertcurrent signal into voltage.
(b) A transimpedance amplifier with a feedback resistor R.
Figure 5.2: Two types of preamplifier systems. The ideal current source Irepresents the signal from the ICR cell. The capacitor C is the combination ofthe parasitic capacitance of the cell and cabling.
5. Transimpedance Preamplifier Using an Operational Amplifier 89
plifier system with the same 3-dB bandwidth. Equivalently, the tolerance to the
parasitic capacitance of the cell and cabling is stronger. Therefore, the preampli-
fier can be located further away from the cell to isolate it from possible electrical
perturbation induced by the high magnetic field. Such a stronger tolerance makes
the transimpedance an ideal choice when a higher magnetic field, such as the 21-
T magnet (Painter et al., 2006; Xian et al., 2012), or the technique of excitation
and detection on the same ICR plate (Chen et al., 2012)20 is introduced in the
near future.
5.2.2 Cyclotron Frequency Correlation
The cyclotron frequency from the ion signal in an ICR cell is described in Chap-
ter 2. It can be easily seen by rewriting Eq. (2.1) that the mass-to-charge ratio
is inversely proportional to the cyclotron frequency ωcyc,
m
q=
B
ωcyc
, (5.5)
where m is the ion mass, q is the ionic charge of the ion, and B indicates the
magnetic field strength. Since the image current, Is, induced by the rotating ions
in an ICR cell is modeled by Eq. (2.8a), substituting the q/m in Eq. (2.8a) by
Eq. (5.5) yields
Is (r.m.s) =Nqr√
2dωcyc , (5.6)
20The method of excitation and detection on the same ICR plate introduced by Chen andco-workers uses a few protection diodes at the input node of the ICR preamplifier. Such diodesincrease the capacitance seen by the preamplifier input.
5. Transimpedance Preamplifier Using an Operational Amplifier 90
where N is the number of excited ions, r is the ion rotation orbital radius, and d
is the cell diameter (spacing). As described, the intensity of the induced image
current in an ICR cell is a function of the cyclotron frequency. Such dependency
can be either calibrated by the signal processing computer, or can be eliminated
by the front-end circuitry under certain conditions.
When a transimpedance preamplifier is introduced as the front-end electronics
solution, from Eq. (5.3), the magnitude of the transfer function H(ω) at the
cyclotron frequency ωcyc can be calculated as
|H(ωcyc)| =R√
1 + [ωcycR(C/G)]2. (5.7)
Recall that the magnitude of the output voltage |Vout| after the transimpedance
preamplifier is |Is||H(ω)|. Here, if ωcycR(C/G) is much greater than 1, namely,
the magnitude of the feedback resistance is much greater than the magnitude of
the reactance of the effective input capacitance,
R >>1
ωcyc(C/G), (5.8)
Eq. (5.7) can be approximated as
|H(ωcyc)| '1
ωcyc(C/G). (5.9)
5. Transimpedance Preamplifier Using an Operational Amplifier 91
From Eq. (5.6) and Eq. (5.9), the magnitude of the output voltage becomes
|Vout| = |Is||H(ωcyc)| 'Nqr√
2d(C/G), (5.10)
which is independent of the cyclotron frequency. On the contrary, if the mag-
nitude of the feedback resistance is much smaller than the magnitude of the
reactance of the effective input capacitance,
R <<1
ωcyc(C/G), (5.11)
Eq. (5.7) can be approximated as
|H(ωcyc)| ' R , (5.12)
and the magnitude of the output voltage becomes
|Vout| = |Is||H(ωcyc)| 'Nqr√
2dRωcyc . (5.13)
In such a case, calibration will be necessary for maintaining the constancy of the
gain for each peak in a spectrum.
The gain G for a generic op amp is typically greater than 104. Assuming an
input capacitance of 10 pF and the cyclotron frequency fcyc of signals between
10 kHz and 10 MHz (mass-to-charge ratio, m/z, of ∼18 to 18k for a 12-T FT-ICR
5. Transimpedance Preamplifier Using an Operational Amplifier 92
system), the reactance magnitude of the effective input capacitance1
ωcyc(C/G)
can be between 16M and 16G. As it is not practical to have a feedback resistance
much larger than 16 GΩ, for simplicity, the condition described by Eq. (5.11)
shall be fulfilled when designing a transimpedance preamplifier using an op amp
for an FT-ICR system in which the input capacitance is ∼10 pF.
5.3 Transimpedance Preamplifier Circuit Design
The basic design of the transimpedance preamplifier uses a junction field-effect
transistor (JFET) input stage and an op amp main stage for gain. It was reported
that metal-oxide-semiconductor field-effect transistor (MOSFET) often has the
smallest high-frequency noise, whilst JFET usually has the best low-frequency
noise behavior (Fabris and Manfredi, 2002). An example is shown in Fig. 5.3.
In comparison with the N-type JFET, the reported N-type MOSFET has better
noise performance at the frequency range higher than 1 MHz. As in this ap-
plication the frequency of interest mostly falls in the range between 1 kHz and
1 MHz, a JFET device can be one of the best front-end candidates with large
input impedance.
The schematic of the preamplifier circuit is shown in Fig. 5.4. The overall
feedback is controlled by a single resistor (R1 in Fig. 5.4), which determines the
transimpedance of the preamplifier. The input node ’Iin’ of the preamplifier is
direct current (DC) coupled to the detection plate of the ICR cell. Therefore,
the same resistor in the feedback loop also biases the detection plates of the ICR
5. Transimpedance Preamplifier Using an Operational Amplifier 93
1980 IEEE TRANSACTIONS ON NUCLEAR SCIENCE, VOL. 49, NO. 4, AUGUST 2002
Fig. 2. Conceptual behavior of theS(f) power density as a function offrequency. a) N-channel MOSFET belonging to a submicron CMOS monolithicprocess. b) P-channel MOSFET belonging to a submicron CMOS monolithicprocess. c) N-channel JFET belonging to a classical JFET monolithic process.
JFET-CMOS processes like, for instance DMILL. However,for the purpose of the following analysis, an adequate approx-imation is the one which represents as ,where is a coefficient ordinarily of the order of unity, which,however, may become consistently larger than 1 in shortchannel MOSFETs operating in strong inversion.
Equations (1) and (2) also highlight the difference in thelow-frequency noise mechanism of the two types of device. InMOSFETs the low-frequency region of the noise spectrum isgoverned by noise [33]. The term which appears in (2) todescribe the low-frequency noise in a JFET is approximatedas the tail of a Lorentzian distribution. Such an approximationis valid for low noise JFETs in most of practical applicationsand fails only in devices exposed to a substantial dose ofradiation.
A conceptual comparison of the relative merits of N andP-channel MOSFETS belonging to the same process and anN-channel JFET, all assumed to be of identicalvalues, isprovided by the model plots of Fig. 2. The MOSFETs behavioris typical of a submicron process with gate oxide thickness inthe 5-nm region. The JFET behaves as a device belonging to aclassical low-noise monolithic process with a few micron gatelength.
The plots aim at pointing out that, of the three considered de-vices, the N-channel MOSFET has the smallest high-frequencynoise, while usually the JFET has the best behavior in thelow-frequency region. However, as compared to MOSFETS ofthe previous generations, the new MOSFET processes featurea considerable improvement in their noise governedlow-frequency behavior. To understand the reasons for this, itshould be remembered that the coefficient of noise ina MOSFET can be expressed as
(3)
where depends on the quality of the gate oxide and on theheight of the barrier at the junction Si/SiOand is thegate-to-channel capacitance per unit area.
Fig. 3. Frequency dependence of theS (f),S (f) spectral power densitiesfor devices belonging to different monolithic processes. a) N-channel MOSFETbelonging to a CMOS process featuringL = 0:25 m, t = 5 nm, actualgate lengthL = 0:5 m, drain currentI = 0:5 mA, input capacitanceC =6 pF. b) P-channel MOSFET belonging to the same CMOS process as above,actual gate lengthL = 0:5 m, drain currentI = 0:5 mA, input capacitanceC = 6 pF. c) N-channel JFET belonging to an all N-JFET monolithic process,actual gate lengthL = 5 m, input capacitanceC = 10 pF, drain currentI = 5 mA. d) P-channel JFET belonging to DMILL BiCMOS JFET process,actual gate lengthL = 1 m, input capacitanceC = 9 pF, drain currentI = 1 mA.
The reduction in the gate oxide thickness and a likely im-provement in the oxide quality have resulted in a substantiallylower noise in the most advanced MOSFET processes.Still, the P-channel features less noise than the N-channelMOSFET, as pointed out in Fig. 2. The ratiodepends on the process and can vary from a minimum of a fewunits to a maximum of about 30.
The expressions of ENCare given by (1) for the MOSFETand by (2) for a JFET [41]
ENC (4)
ENC (5)
In the previous equations, is the sum of all open-loopcapacitances shunting the preamplifier input, including thedetector capacitance , the feedback capacitance in thecharge-sensitive loop and strays. is the input capacitance ofthe front-end device.
The actual power densities of the series noise relevant tofour devices, all of gate width in the 2000- m region areplotted as functions of frequency in Fig. 3. The devices are: anN-channel and a P-channel MOSFET belonging to the sameCMOS process, featuring a minimum gate length
m and a gate oxide thickness nm, an N-channelJFET part of a monolithic low-noise JFET technology andthe P-channel JFET belonging to the DMILL process. Thefrequency spectra of Fig. 3 were made at equal ratios between
Figure 5.3: Noise power spectral densities of four types of transistors: a) aN-type MOSFET, b) a P-type MOSFET, c) a N-type JFET, and d) a P-typeJFET (Fabris and Manfredi, 2002).
cell, in this case, to a virtual ground potential, which is necessary to preserve ion
trajectories in the cell.
In order to limit the intrinsic noise from the passive components, low noise
surface mount ceramic capacitors and surface mount thin-film chip resistors from
Panasonic Corporation (Osaka, Japan) are selected to define the biasing con-
ditions of the preamplifier system and to provide the transimpedance in the
feedback loop. Figure 5.5 shows the average noise level of chip resistors reported
in the Surface Mount Resistors Technical Guide Ver.3 from Panasonic.21 As re-
ported, a thin-film chip resistor has a lower noise level than a thick-film chip
21See http://industrial.panasonic.com/www-data/pdf/AOA0000/AOA0000PE36.pdf for theSurface Mount Resistors Technical Guide from Panasonic, accessed 30 August 2012.
5. Transimpedance Preamplifier Using an Operational Amplifier 94
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
A A
B B
C C
D D
E E
F F
G G
U1
AD8099
3
2
4
7
6
18
5
VDD
6V
C10
10nF
R6
3.3kΩ
Vout
C9
4.7µF
C7
10nF
C8
4.7µF
VSS
-6V
C2
220nF
VSS
-6VR4180Ω
R31kΩ
C410nF
C34.7µF
R1
18kΩ
R21kΩ
IinQ1
BF862
C6
4.7pF
VDD
6V
C1
0.85pF
C54.7µF
R5820Ω
R8
2kΩKey=A
50%
R915kΩ
VDD
6V
R715kΩ
VSS
-6V
Figure 5.4: Schematic of the transimpedance preamplifier using an operationalamplifier AD8099.
5. Transimpedance Preamplifier Using an Operational Amplifier 95
resistor.
TECHNICAL GUIDE Ver.3
11
(3). Storage in places outside the temperature range of 5deg to 35deg and humidity
range of 45% to 85%RH. (4). Storage over a year after our delivery (This item also applies to the case where the
storage method specified in item (1) to (3) has been followed.
9. Technique 9.1. Circuit design
If using surface mount resistors, pay attention the following performance. (Primarily for
film-chip resistors.)
9.1.1. Resistor noise
In general, resistor noise is calculated from the following formula.
Resistor noise = thermal fuse + 1 / f noise
It is thermal noise that depends on shake of speed distribution by clash of carrier and
grid. It is 1 / f-noise that factor, which controlled electric current, shakes in some cause
and, as the result, it arises from density of carrier and modulation of electric current. It is
thought to be in proportion to a reciprocal of frequency.
Regarding thick-film chip resistor, it is formed resistance value by connect resistance.
Therefore, 1 / f-noise shall be primarily noise, and it is calculated from the following
formula.
Noise Index (dB) = A –10 Log (w l t)
A: Resistive material, value by manufacturing condition
w • l • t: W-dimensions, L-dimensions, t-dimensions
Noise level average of chip resistor by shape is shown in Fig.14.
Fig.14. Noise level average of chip resistor
-40
-50
-30
-20
-10
0
+10
+20
2G3G,6G8G14
12,1W
6Y3Y
10Ω 100Ω 1kΩ 10kΩ 100kΩ 1MΩ
電流雑音特性(QUANTECH 315C)
ノイ
ズIN
DE
X(
dB
)
厚膜抵抗器
薄膜抵抗器
Thick-film chip resistors
2G: 1005
3G: 1608
6G: 2012
8G: 3216
14: 3225
12: 4532
1W: 6432
Thin-film chip resistors
Current noise characteristic (QUANTECH315C) N
oise
IND
EX
(dB
)
+20
+10
0
-10
-20
-30
-40
-50 10 100 1K 10K 100K 1M (ohm)
8G
2G
3G,6G
14
12,1W
Thick-film resistor
Thin-film resistor
3Y
6Y
Figure 5.5: Average noise level of chip resistors, reported in the Surface MountResistors Technical Guide Ver. 3, Panasonic.21
5.3.1 Main Stage
The main stage is formed by an ultralow distortion op amp AD8099 from Analog
Devices, Inc. (Norwood, Massachusetts),22 biasing resistors R3–R5, bias adjust-
ment resistors R7–R9, feedback resistor R6 and capacitor C6, DC blocking ca-
pacitor C2, and bypass capacitors C3–C5 and C7–C10. AD8099’s ultralow noise
input voltage spectral density of typically 0.95 nV/√Hz (specified at 100 kHz),
ultralow distortion (-92 dBc at 10 MHz), and wide bandwidth (700 MHz at the
gain of 2) are specifically utilized in this preamplifier system. The AD8099’s
gain-bandwidth product of 3.8 GHz makes this op amp a good candidate for
22 See Appendix A.2 for information about the op amp AD8099.
5. Transimpedance Preamplifier Using an Operational Amplifier 96
having a large gain at around 10 MHz. Pin 5 of the op amp is not connected, as
suggested by the datasheet22 when the gain is set to around 20. Pin 1 and pin 6
are internally connected inside the surface mount package to increase the routing
flexibility on a PCB.
The AD8099 datasheet reports that AD8099 can operate stably when the gain
is less than 20. As the gain-bandwidth product of this op amp is 3.8 GHz and
only a 10-MHz bandwidth is required for this application, it is planed to set the
gain of this main stage close to 20. Therefore, resistors R4 and R6 are used to
set the AC signal gain of this main stage to 19 (R6/R4 + 1). On the contrary,
one can change the resistance values of R4 and R6 if a gain of greater than 20 is
needed, given that this op amp AD8099 is tested stable when gain is over 20. If
a gain of greater than 20 causes a stability issue to this op amp, the introduction
of multiple gain stages (such as connecting op amps in a cascade configuration),
or other op amps should be considered in this application. The 3-dB bandwidth
f3dB (high frequency cut-off), which can be calculated by Eq. (5.14), is further
limited by the capacitor C6 together with the feedback resistor R6 to 10 MHz.
f3dB =1
2πRC. (5.14)
The DC biasing conditions of both AD8099 inputs are balanced by R3 to-
gether with R4 and R5, and can be trimmed by the trimmer resistor R8 to
maintain the output DC level at 0 V. In this application, pin 8 (the disable pin)
is connected to +6V DC power rail so that the op amp remains on whenever the
5. Transimpedance Preamplifier Using an Operational Amplifier 97
power to the preamplifier system is on. Potentially, connecting this disable pin
with a control signal can be a good solution when isolation between the ICR cell
and the amplifying circuitry is necessary.
5.3.2 Input Stage
The large input impedance of the JFET BF862 from NXP Semiconductors N.V.
(Eindhoven, The Netherlands)23 makes such a transistor an ideal input stage,
as the biasing and feedback circuitry at the input node of the op amp AD8099
lowers the input impedance of this op amp. The ultralow noise (noise input
voltage spectral density of typically 0.8 nV/√Hz at 100 kHz as specified on
the datasheet)23 characteristics of BF862 also limits the possible intrinsic noise
added by this first stage buffer into the preamplifier system. According to the
datasheet, the transition frequency of this JFET is 715 MHz. Such a bandwidth
specification is suitable for this 12-T FT-ICR preamplifier application, where a
bandwidth of less than 10 MHz is needed. Meanwhile, this BF862 was tested to be
vacuum compatible and was proven stable when operates under a 7-T magnetic
field. Such characteristics allow this JFET to be a good front-end choice, as the
front-end device may be placed inside the vacuum chamber where high magnetic
field exists.
The BF862 input stage can be configured either as a source follower or a
common-source amplifier. A common-source amplifier provides more voltage gain
23 See Appendix A.1 for information about the JFET BF862.
5. Transimpedance Preamplifier Using an Operational Amplifier 98
to increase the open-loop gain (G in Eq. (5.4)) to extend the 3-dB bandwidth,
but also results in an unwanted 180-degree phase shift before unity gain, causing
the preamplifier to oscillate. Therefore, this BF862 input stage is designed to be
a source follower with an 1-kΩ source resistor (R2 in Fig. 5.4) to set the drain
current at around 6.5 mA.
5.3.3 Printed Circuit Board
The preamplifier circuit was built on a single-layer printed circuit board (PCB).
The populated PCB (sized ∼65× 60 mm) is shown in Fig. 5.6. On the PCB, all
of the components are located inside a ground ring to shield them from environ-
mental noise. Bypass capacitors are placed as close as possible to the AD8099
chip, as suggested by the AD8099 datasheet, for optimum distortion and power
supply rejection performance. Note that this board is for bench testing purpose.
When placing such a PCB into a vacuum chamber close to the ICR cell in an
FT-ICR system, a smaller-sized, two-layer board with ground plane should be
considered to limit the noise coupling and parasitic capacitance.
5.4 Computer Simulation
To test the designed bandwidth and noise performance, SPICE simulation has
been carried out using NI Multisim. Initially, the designed bandwidth is at the
range of 1 kHz to 10 MHz, which corresponds to an output mass-to-charge ratio,
m/z, of roughly 18 to 180k for a 12-T FT-ICR mass spectrometry (MS) system.
5. Transimpedance Preamplifier Using an Operational Amplifier 99
Figure 5.6: Single layer printed circuit board of the AD8099 preamplifier (sized∼65× 60 mm) with components fully populated.
Such a wide-band design goal offers enough buffer when the parasitic capacitance
narrows the bandwidth. Use of a bandpass filtering configuration minimizes the
artifacts, such as the signal aliasing caused by high frequency signal fold-over, and
the offset in the signal (Mathur and O’Connor, 2009), in FT-ICR mass spectra.
In this design, the low frequency cutoff is defined by the capacitor C2 and the
effective resistance (∼1 kΩ) in series with it, whereas the high frequency 3-dB
point is defined by the feedback resistor R1 together with the capacitor C1. Both
frequency poles can be estimated by using Eq. (5.14).
The AC analysis simulations are performed to show the transimpedance with
different feedback resistance and capacitance values. Two feedback resistors,
180 Ω and 18 kΩ, are chosen to show the gain variance. Figure 5.7a shows the
results when feedback capacitor C1 is not connected to the system. In the simula-
5. Transimpedance Preamplifier Using an Operational Amplifier 100
tion program, the parasitic capacitance of the 0805 sized surface mount feedback
resistor R1 is set to 85 fF. The AC analysis simulation when C1 is 0.85 pF is
shown in Fig. 5.7b. As expected, the low frequency cut-off is independent of the
feedback impedance and is slightly below 1 kHz for both cases. When the feed-
back capacitance of C1 is fixed at 0.85 pF, from Eq. (5.14) the 3-dB bandwidth
occurs at around 10 MHz and 1.0 GHz when R1 is 18 kΩ and 180 Ω, respectively.
The simulation results support such estimations.
In Fig. 5.7a, the peakings around 20 MHz and 200 MHz of the 18-kΩ and
180-Ω curves, respectively, are the results of the phase shift caused by the LC
resonance of the op amp output impedance, and can be eliminated by introducing
the 3-dB poles before the peakings. As shown by the 18-kΩ curve in Fig. 5.7b,
the 0.85-pF feedback capacitor and the 18-kΩ feedback resistor cause a pole at
around 10 MHz, thus the transimpedance drops before the peak at 20 MHz. For
the 180-Ω curve, the same feedback capacitance of 0.85 pF generates a 3-dB pole
at about 1.0 GHz, and the 200-MHz peak remains.
5.5 Transimpedance Preamplifier Testing Results
5.5.1 Frequency Response
The voltage gain of the main stage and the overall transimpedance performance
have been measured. The voltage gain of the op amp AD8099 was tested when
the first stage (BF862 and its source resistor R2 in Fig. 5.4) was not mounted to
5. Transimpedance Preamplifier Using an Operational Amplifier 101
1 1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M 1 G- 1 0
01 02 03 04 05 06 07 08 09 0
1 0 0
Trans
impe
danc
e (dB
(V/A))
F r e q u e n c y ( H z )
1 8 0 o h m f e e d b a c k 1 8 k o h m f e e d b a c k
(a) Frequency response without the feedback capacitor C1 in Fig. 5.4.
1 1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M 1 G- 1 0
01 02 03 04 05 06 07 08 09 0
1 0 0
Trans
impe
danc
e (dB
(V/A))
F r e q u e n c y ( H z )
1 8 0 o h m f e e d b a c k 1 8 k o h m f e e d b a c k
(b) Frequency response with 0.85-pF feedback capacitor.
Figure 5.7: SPICE simulations of the AD8099 preamplifier transimpedance.
5. Transimpedance Preamplifier Using an Operational Amplifier 102
the PCB, to ensure that the performance of this main stage matches the designed
conditions of providing a voltage gain of 19 between 1 kHz and 10 MHz. The
test signal was fed into the system via the DC blocking capacitor C2, and the
output was measured directly from the output of AD8099. The result shown in
Fig. 5.8 demonstrates an agreement with the designed voltage gain of 19 (about
25 dB) and designed 3-dB bandwidth of 10 MHz.
1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M0
5
1 0
1 5
2 0
2 5
3 0
Gain
(dB)
F r e q u e n c y ( H z )
A D 8 0 9 9 V o l t a g e G a i n
Figure 5.8: Voltage gain frequency response of the main (AD8099) stage.
Figure 5.9 shows the transimpedance of the entire preamplifier system in three
conditions when different feedback resistors and capacitor are soldered onto the
system: (i) only a 180-Ω resistor, (ii) only an 18-kΩ resistor, and (iii) an 18-kΩ
resistor together with a 0.8-pF capacitor. It can be seen that the peaking of the
18-kΩ single-resistor feedback curve agrees with the SPICE simulation results
5. Transimpedance Preamplifier Using an Operational Amplifier 103
shown in Fig. 5.7. Shunting a 0.8-pF capacitor with the 18-kΩ feedback is a
solution to avoid this peak.
1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M2 0
3 0
4 0
5 0
6 0
7 0
8 0
9 0
Trans
impe
danc
e (dB
(V/A))
F r e q u e n c y ( H z )
1 8 0 o h m f e e d b a c k 1 8 k o h m f e e d b a c k 1 8 k o h m & 0 . 8 p F f e e d b a c k
Figure 5.9: Transimpedance frequency response of the preamplifier system inthree feedback conditions: (i) only a 180-Ω resistor, (ii) only an 18-kΩ resistor,and (iii) an 18-kΩ resistor together with a 0.8-pF capacitor.
Figure 5.10 illustrates one of the input/output waveforms fetched by the NI
PXI system. It corresponds to the transimpedance measured at 10 kHz reported
in Fig. 5.9 (iii), in which the preamplifier feedback system is an 18-kΩ resistor
shunting a 0.8-pF capacitor. The large green line indicates the output waveform,
whilst the red line represents the input waveform. The peak-to-peak amplitude
of the measured output voltage is around 1.5 V, whilst the peak-to-peak input
current amplitude is about 0.1 mA.
5. Transimpedance Preamplifier Using an Operational Amplifier 104
Figure 5.10: One of the input/output waveforms fetched by the NI PXI system.This is the waveform of the preamplifier system reported in Fig. 5.9 (iii), in whichthe preamplifier feedback system is an 18-kΩ resistor shunting a 0.8-pF capacitor.The testing frequency is 10 kHz, whilst the output voltage (green line) is ∼1.5 V(peak-to-peak) and input current (red line) is ∼0.1 mA (peak-to-peak).
5.5.2 Noise Performance
The noise performance was tested when the input node was floating. The total
output power was recorded, converted into input current noise spectral density
by Eq. (4.2) and Eq. (4.3), and then plotted in Fig.5.11. As reported, at the
frequency of 1 MHz, the noise spectral density is around ∼1 pA/√
Hz. The
measured noise performance of this preamplifier system agrees with the general
noise characteristics of an op amp, where the noise spectral density curve is the
combination of low frequency 1/f noise and high frequency white noise.
5. Transimpedance Preamplifier Using an Operational Amplifier 105
1 0 k 1 0 0 k 1 M 1 0 M02468
1 01 21 41 61 82 0
Curre
nt sp
ectra
l den
sity (p
A/sqrt
(Hz))
F r e q u e n c y ( H z )
I n p u t c u r r e n t n o i s e s p e c t r a l d e n s i t y
Figure 5.11: Measured input current noise spectral density with 18 kΩ tran-simpedance.
5.6 Discussions
5.6.1 Bandwidth & Feedback Impedance
The feedback resistor sets the transimpedance of the preamplifier system. Nev-
ertheless, the parasitic capacitance within the feedback resistor itself limits the
maximum value of the feedback resistance. Theoretically, the resistance of the
feedback resistor (R1 in Fig. 5.4) shall be as large as possible due to a required
large transimpedance (assuming that the condition described by Eq. (5.11) is
ignored here). This limits the capacitance of C1 to a range that may be similar
5. Transimpedance Preamplifier Using an Operational Amplifier 106
to the parasitic capacitance contributed by the resistor R1. In reality, when R1
is large enough, the feedback capacitor C1 may not be necessary to meet the
required high frequency cutoff. For example, a ∼200 kΩ 0805 surface mount
package resistor, which can have the parasitic capacitance of about 80 fF, will
set the 3-dB point (according to Eq. (5.14)) at around 10 MHz. When a smaller
package is used, such parasitic capacitance can be limited to a lower value. For
instance, a 0402 surface mount resistor has a typical parasitic capacitance value
of around 30 fF . As a result, by introducing a resistor with 0402 package, the
transimpedance of this system can be pushed to 530 kΩ for a 10-MHz bandwidth,
or to 5.3 MΩ for 1 MHz.
5.6.2 Noise & Feedback Impedance
The measured noise power is independent of the feedback resistance, namely, the
transimpedance of this preamplifier. If a 5.3 MΩ 0402 surface mount resistor is
used as the feedback whilst narrowing the bandwidth to 1 MHz, the equivalent
input current noise spectral density will become 3.7 fA/√
Hz, rather than the
reported ∼1 pA/√
Hz (shown in Fig. 5.11) at 1 MHz. Meanwhile, the bandwidth
can be easily adjusted by changing the capacitance of C1 (for high frequency
cut-off) and C2 (for low frequency cut-off) in Fig. 5.4. When the bandwidth is
narrowed to 1.0 MHz (commonly used m/z region of having a low mass cut-off at
∼180 m/z for a 12-T FT-ICR MS system), with a 5.3 MΩ feedback resistor the
electronic noise generated by this preamplifier referred back to the input current
5. Transimpedance Preamplifier Using an Operational Amplifier 107
will become roughly 3.7 pA.
The detection bandwidth can be traded for sensitivity by narrowing the detec-
tion window, which can be done by introducing filtering technique or by changing
the preamplifier bandwidth. At threshold, the signal current has to be at least
equal to the noise level in order to be detected, which means from Eqn. (2.8),
around 110 singly-charged, 1000-Dalton ions (assuming the rotating orbit radius
of1
4the cell diameter, in a 12-T magnetic field) to be in the ICR cell for detec-
tion. In lieu of detecting the 1-MHz bandwidth in one detection using a 1-MHz
detection window, a narrower 100-kHz window can be introduced to finish the
same task in 10 detections. In which case, with a 5.3 MΩ feedback resistor
∼ 35 charges (= 3.7× 10−15 ×√
100× 103 ÷ 33× 10−15) can be detected in one
scan. In reality, techniques such as multiple acquisition can be introduced to
allow a weaker signal than noise.
5.7 Conclusion
As the key front-end electronic component, the preamplifier plays a critical role
in pushing the unmatched FT-ICR MS performance to the limit. Following the
studies of the ICR signal model from Comisarow (Comisarow, 1978), the gain
and noise distribution in a multistage amplifier system, and the electronic noise,
this chapter reports a preamplifier with transimpedance configuration to have not
only stronger tolerance to the intrinsic capacitance of the cell, but also higher
designed gain as a result of the bandwidth increment.
5. Transimpedance Preamplifier Using an Operational Amplifier 108
The improved preamplifier has a good flexibility of adjusting its transimpedance
and bandwidth, whilst maintaining flat mid-band gain at the frequency of inter-
est. The total power consumption of this circuit is around 310 mW when tested
on the bench. With the chosen 18-kΩ feedback resistor, 0.8-pF feedback capac-
itor, and 220-nF DC blocking capacitor, the transimpedance of the preamplifier
is around 85 dBΩ between 3 kHz and 10 MHz; the input current spectral den-
sity is about 1 pA/√
Hz at 1 MHz. When using a 0805 type feedback resistor,
this preamplifier has been tested stable of providing a transimpedance between
45 and 85 dBΩ (depending on the feedback resistance value used), whist main-
taining a bandwidth of 10 MHz. When using a 0402 type feedback resistor, this
preamplifier is estimated to provide a transimpedance up to 5.3 MΩ for a 1-MHz
bandwidth. In the near future, this preamplifier will be further mounted onto an
FT-ICR MS system to test the performance.
CHAPTER 6
Single-Transistor Transimpedance
Preamplifier Using a T Feedback
Network
This chapter is partially reproduced from the journal article, ”A Low Noise Single-
Transistor Transimpedance Preamplifier for Fourier-Transform Mass Spectrome-
try Using a T Feedback Network,” Review of Scientific Instruments, in 2012 (Lin
et al., 2012).
This chapter starts with a discussion of the theories and the comparison of two
different feedback arrangements, a single-resistor feedback and a T-shaped feed-
back network, for transimpedance amplifiers. Then it reports a single-transistor
transimpedance preamplifier design to push the noise performance further. This
is followed by the study of the biasing conditions, input/output impedance, and
over-all transimpedance of such a design. A PCB for testing purpose is man-
109
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 110
ufactured. Then this is followed by the bench testing reports of the proposed
T feedback network (using the transimpedance amplifier reported in Chapter 5)
and the single-transistor transimpedance preamplifier. This chapter concludes
with the discussion of the gain and noise performance of a preamplifier, and a
suggestion of possible constructing elements for a T feedback network for circuit
optimization.
6.1 Introduction
In the previous chapter, a transimpedance preamplifier was designed and tested
for a Fourier-transform ion cyclotron resonance (FT-ICR) system. With the abil-
ity of effective input capacitance reduction over existing voltage amplifier designs,
a transimpedance preamplifier can potentially provide an ideal preamplifier solu-
tion for many mass spectrometry systems, including FT-ICR mass spectrometry
(MS). Here, efforts have been taken to further lower the noise generated by the ac-
tive components of an operational amplifier (op amp) in a preamplifier. One way
is to introduce a very low noise first-stage and to use only one active component.
Consequently, a novel single-transistor transimpedance preamplifier with a
lower power consumption is introduced. A low noise, high input impedance
JFET, BF862 from NXP Semiconductors N.V. (Eindhoven, The Netherlands),24
is used as the main amplification stage of this transimpedance preamplifier. The
noise generated by the active components in the previous design reported in
24See Appendix A.1 for information about the JFET BF862.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 111
Chapter 5 is now limited. Only one transistor, which is the low noise JFET,
BF862, with the equivalent noise input voltage of typically 0.8 nV/√
Hz specified
on the datasheet, in the new design is responsible for such noise. Furthermore, a
T-shaped feedback network is introduced as both the feedback and the gate bias-
ing solutions to avoid a large gate biasing resistor, and to increase the bandwidth
flexibility. The T feedback network is studied using the previously reported
AD8099 preamplifier. Such a feedback system allows ∼100-fold less feedback
resistance at a given transimpedance, hence preserving bandwidth, which is ben-
eficial to applications demanding high gain.
6.2 Transimpedance Amplifier Transfer Function
6.2.1 Single-Resistor Feedback
A basic transimpedance amplifier is constructed out of an op amp with a feedback
resistor, as shown in Fig. 6.1a. An op amp senses the voltage difference between
its non-inverting input (V+) and inverting input (V−), and amplifies it with its
open-loop gain (A). Consequently, the voltage at output becomes
Vout = A× (V+ − V−). (6.1)
An ideal op amp has a few particular characteristics, including infinite input
impedance and infinite open-loop gain. In reality, the open-loop gain, A, is a
finite large value of typically greater than 104. When analysing the op amp
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 112
circuit, it is common to assume that the input impedance is large so that the
current flowing into either the inverting or non-inverting input is negligible.
With the hypotheses mentioned above, for a basic transimpedance amplifier
with open-loop gain A, a feedback resistor Rf , a ideal current signal source Iin,
and a grounded non-inverting input node (V+ = 0), as shown in Fig. 6.1a, the
voltage at the inverting input, V−, can be rewritten as
V− =−VoutA
. (6.2)
There is no current flowing into the inverting input, so the input Iin will flow
completely through the feedback resistor Rf ,
Iin =V− − Vout
Rf
. (6.3)
From Eq. (6.2) and Eq. (6.3), the transfer function, or the transimpedance,
Tsingle−resistor of this negative-feedback op amp circuit can be rewritten as
Tsingle−resistor =VoutIin
=−Rf
1 +1
A
' −Rf . (6.4)
Thus the closed-loop gain of this transimpedance amplifier system is independent
of the op amp characteristics (its open-loop gain A), but is a function of the
characteristics of the passive components (the feedback resistance Rf in this
case) used to ’close’ the loop.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 113
(a) Transimpedance amplifier with single-resistor feedback.
(b) Transimpedance amplifier with T-shaped feedback network.
Figure 6.1: Transimpedance amplifiers with two feedback arrangements (Iinindicates an ideal current signal source).
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 114
6.2.2 T Feedback Network
The single feedback resistor can be replaced by a three-resistor T-shaped feedback
network (Barros, 1982; Fish and Katz, 1977) consisting of two resistors connected
in series, and a third resistor coupled between the junction node of the two series
resistors and a reference potential (ground in this case), as shown in Fig.6.1b.
Then the circuit analysis shall be started with assuming that the voltage at the
joint node of the T network, where the resistors R1, R2, and R3 are connected, is
Vx. The current flowing into the resistor R1 is still Iin, and can be expressed as
Iin =V− − VxR1
. (6.5)
By substituting Eq. (6.2) into Eq. (6.5), Vx can be expressed as
Vx = V− −R1Iin =−VoutA−R1Iin . (6.6)
At the joint node of the T network, the current flowing from the resistor R1 is
equal to the current flowing into both R2 and R3, namely,
Iin =Vx − 0
R3
+Vx − Vout
R2
. (6.7)
By substituting Vx from Eq. (6.6), Iin can be rewritten as
Iin =
−VoutA−R1Iin
R3
+
−VoutA−R1Iin − VoutR2
, (6.8)
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 115
and so
(1 +R1
R3
+R1
R2
)Iin = −[1
A(
1
R2
+1
R3
) +1
R2
]Vout . (6.9)
If it is chosen that R2 = 100R3, the equivalent resistance of R2 ‖ R3 will be
about the same as R3 ([R2 ‖ R3] = [100R3 ‖ R3] =100
101R3 ' R3). Since typically
A > 104,
A
R2
(R2 ‖ R3) ' AR3
R2
=A
100 1 . (6.10)
Then the transimpedance of the system becomes
TT−network =VoutIin
= −1 +
R1
R3
+R1
R2
1
A(R2 ‖ R3)+
1
R2
= −R1 +R2 +
R1R2
R3
1 +1
A
R2
(R2 ‖ R3)
(6.11)
' −(R1 +R2 +R1R2
R3
) . (6.12)
As a result, by replacing the single feedback resistor with the T feedback net-
work, the resistance of the feedback system can be reduced without sacrificing the
overall transimpedance in a negative-feedback transimpedance amplifier system.
For instance, using the designed resistance values of 47 kΩ , 18 kΩ , and 180 Ω
for R1, R2, and R3, respectively, the transimpedance becomes about 4.8 MΩ.
Thus, with roughly the same theoretical transimpedance, the resistance values of
the feedback resistors can be dropped by about 100 fold. Since the bandwidth
of a transimpedance amplifier is a function of the feedback impedance, reducing
the feedback resistance increases the bandwidth. Despite different technologies
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 116
being introduced for amplifier bandwidth extension (Green, 1986; Green et al.,
2008a; Analui and Hajimiri, 2004; Mohan et al., 2000; Chien and Chan, 1999),
the T feedback network is undoubtedly another simple approach to preserve the
bandwidth without sacrificing the transimpedance. Note that here, the resistors
R1, R2 and R3 can be replaced by complex impedances (such as Z1, Z2, and
Z3) in order to obtain transimpedance characteristics which vary with frequency
according to particular requirements in a given application.
6.3 Single-Transistor Preamplifier Circuit Design
6.3.1 Common-Source Amplifier
The first attempt to design a single-transistor transimpedance preamplifier was
made by adjusting the impedance values of passive components, to adjust the
DC conditions and to fulfill the requirement of having a large transimpedance
between the frequency of 1 kHz and 1 MHz, which corresponds to the mass-to-
charge ratio, m/z, between ∼180 and ∼180k in a 12-T FT-ICR system.
Voltage Gain
A common-source JFET amplifier with source degeneration is usually formed by
a single JFET, a drain resistor RD, a gate resistor RG, and a source resistor RS
(Q1, R4, R3, and R6, respectively, in Fig. 6.2). The drain current is modulated by
the JFET, according to the input voltage signal applied to the gate, and voltage
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 117
output is expected after such modulated current is converted into voltage by the
drain resistor (or the load resistor), RD. The voltage gain Av of a common-source
amplifier can be written as
Av = − gmRout
1 + gmRS
' − gmRD
1 + gmRS
, (6.13)
where gm is the transconductance of the transistor, and Rout is the equivalent
output resistance, which can be approximated as RD, when any resistance shunt-
ing the drain resistor RD is significantly larger than the resistance of RD. Note
that datasheets often specify the transconductance gm as the forward transfer
admittance yfs, where ’y’ means the admittance, ’f’ stands for forward trans-
fer, and ’s’ indicates the common-source configuration (Deshpande, 2008), and
yfs = dId/dVgs, the ratio of the drain current change over the change of the
gate-source voltage.
Biasing Condition
In particular, the forward transfer admittance, yfs, of the JFET BF862 increases
with its drain current, as suggested by the datasheet. Reduction of the resistance
of the resistor R6 results in increase of the drain current, causing a larger transfer
admittance. Although this larger transfer admittance leads to larger gain, it
sacrifices the advantages of not only the low power consumption characteristic
(due to the low drain current) to lower the heat dissipation, but also the large
voltage gain variation between the DC and AC signals to make this design a better
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 118
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
A A
B B
C C
D D
E E
F F
G G
VDD
6V
R41kΩ
R6470Ω
C54.7µF
R54.7Ω
Q1
BF862
C310nF
C24.7µF
C1220nF
R1
18kΩ
R31MΩ
C4
220nF
Vout
Iin
Figure 6.2: Schematic of the single-transistor transimpedance preamplifier com-prised by a feedback loop (consisting of a feedback resistor R1 and a DC blockingcapacitor C1), and a common-source JFET amplifier with source degeneration.
high-pass filter. Figure 6.3a shows the computer simulated correlations between
the voltage gain, drain current ID, and power consumption of this common-
source JFET amplifier (the circuit of Fig. 6.2 without the feedback loop) with
the permutation of five different source resistors (R6 in Fig. 6.2). As expected,
the AC voltage gain is higher when a smaller source resistor is selected, at the
cost of both higher DC drain current, and less DC and AC voltage gain difference.
The 470-Ω source resistor has been selected for this common-source JFET am-
plifier. An observed drain current of around 1.0 mA when testing this common-
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 119
source amplifier agreed with the SPICE simulation reported in Fig. 6.3a. Since
this low DC drain current was below typical operating current of between 10 and
25 mA suggested by the datasheet, the characteristic curve was measured (shown
in Fig. 6.3b) to test that the JFET was working in its saturation region without
being pinched off.
From the measured drain current of ∼1.0 mA, the DC drain voltage VD of
5.0 V and the source voltage VS of around 0.47 V can be calculated. Therefore,
the operation point of this design is slightly above the red (rhombus) line when
VDS is around 4.5 V in Fig. 6.3b. It can be noticed that this operation point is
very close to the x-axis, where the JFET is off. However, the AC input signal from
an ICR cell of a Fourier-transform mass spectrometer is far below the threshold to
turn this transistor off. Since a 33 fA/charge current for a singly-charged 1000-Da
ion in a 12-T FT-ICR system can be estimated by Eq. (2.8b), and typically less
than 106 ions are expected for detection, a maximum of 33 nA current (assuming
ions are singly charged) may be fed into the preamplifier. The calculated input
voltage from the signal can be as high as 33 mV when using this common-source
amplifier (as shown in Fig. 6.2, but without the feedback loop consisting of the
capacitors C1, C4, and the resistor R1) with a 1 MΩ gate resistor. Such input
condition makes the operation of this JFET still above the Vgs = −0.5 V trace
in Fig. 6.3b. It can be concluded that this biasing condition matches the design
goal of being benefit from both the low power consumption (∼5.7 mW) and the
better low-frequency filtering (DC and AC voltage gain difference of ∼16 dB, as
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 120
1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M0
1 0
2 0
3 0
Volta
ge G
ain (d
B)
F r e q u e n c y ( H z )
R S = 4 7 0 Ω , I D = 0 . 9 5 m A R S = 3 6 0 Ω , I D = 1 . 2 m A R S = 2 6 0 Ω , I D = 1 . 6 m A R S = 1 5 0 Ω , I D = 2 . 4 m A R S = 4 7 Ω , I D = 5 . 1 m A
(a) SPICE simulated correlations between the source resistance RS (R6 inFig. 6.2) and voltage gain/drain current ID of the BF862 common-sourceamplifier (as shown in Fig. 6.2, but without the feedback loop).
0 1 2 3 4 50
2
4
6
8
1 0
I D (mA)
V D S ( V )
V g s = 0 V V g s = - 0 . 1 V V g s = - 0 . 2 V V g s = - 0 . 3 V V g s = - 0 . 4 V V g s = - 0 . 5 V
(b) JFET BF862 current-voltage characteristic curve (measured by theNI PXI system) to test the BF862 for very low DC drain currents. Theoperation point of this design will be slightly above the red (rhombus)line when VDS is around 4.5 V.
Figure 6.3: Common-source amplifier using the JFET BF862.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 121
shown in Fig. 6.3a).
6.3.2 Feedback Loop Configuration
When a feedback resistor (R1 in Fig. 6.2) is added into this common-source
amplifier to modify this circuit into a transimpedance amplifier, not only the
DC blocking capacitor C1 is needed to separate the DC potentials of the gate
and the drain nodes, but there are loading issues to be carefully considered. It
can be seen clearly from Eq. (6.13) that the drain resistance RD plays a role
when determining the voltage gain Av. With the feedback loop coupled between
the input and the output, the feedback resistor R1 becomes part of shunting
resistance to RD. As a result, there will be a minimum threshold resistance for
R1 to keep the voltage gain Av at a reasonable scale for a given application.
Input/Output Impedance
At the input node (the gate of the JFET), it is expected that the majority of the
input signal current flows into the the feedback resistor R1 instead of the gate
biasing resistor R3. Assuming the input resistance of the JFET is too large (in
comparison to R1 and R3) to be considered, the ratio of R3 over R1 should be
so large that the input current grounded via R3 can be negligible. For a larger
transimpedance, a larger feedback resistor R1 is desired, resulting in a desired
much larger R3 to push the signal into the feedback loop. Thus there is also a
maximum threshold resistance of R1 or R3, such that the JFET input resistance
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 122
can be neglected. Therefore, to push the transimpedance of this single-transistor
preamplifier, namely, to further increase the feedback resistance value, such gate
biasing resistor has to be removed.
An effort has been made to re-locate the biasing gate resistor into the feedback
network. Consequently, the single feedback resistor of R1 in Fig. 6.2 is replaced by
a three-resistor, T-shaped feedback system, consisting of R1, R2, and R3 shown
in Fig. 6.4a. The schematic shown in Fig. 6.4a represents the low noise single-
transistor transimpedance preamplifier with a T feedback network. Without a
gate resistor, the gate can still be properly biased to a fixed potential (ground
in this case) through both resistors of R1 and R3. Note that the input of the
preamplifier is DC coupled to the detection plates of the ICR cell, and thereby
provides the ground potential needed for stable ion trajectories.
Moreover, the output impedance of this Single-transistor transimpedance
preamplifier needs to be matched to the impedance of a 50-Ω transmission line
in an FTMS system. A voltage amplifier (VA), such as the AD8099 VA shown
in Fig. 6.6a, can be introduced for this purpose.
Effective Transimpedance Using a T feedback network
An overall transimpedance of 4.8 MΩ can be estimated by Eq. (6.12). However, by
using a common-source amplifier instead of an op amp as the main amplification
stage, the assumption of the open-loop gain A > 104 is no longer valid, causing the
actual transimpedance [calculated by Eq. (6.11)] smaller than the estimated value
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 123
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
A A
B B
C C
D D
E E
F F
G G
VDD
6V
R41kΩ
R6470Ω
C54.7µF
R54.7Ω
Q1
BF862
C310nF
C24.7µF
C4
220nF
R1
47kΩ
Vout
R3180Ω
R2
18kΩ
C1
220nF
Iin
(a) Schematic of the single-transistor transimpedance preamplifier using a T feedback network.
Figure 6.4: Single-transistor transimpedance preamplifier.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 124
1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M5 0
6 0
7 0
8 0
Trans
impe
danc
e (dB
Ω)
F r e q u e n c y ( H z )
B F 8 6 2 p r e a m p l i f i e r w i t h 4 7 0 Ω s o u r c e r e s i s t o r B F 8 6 2 p r e a m p l i f i e r w i t h 4 7 Ω s o u r c e r e s i s t o r
(b) Measured transimpedance frequency response of the single-transistor transimpedance pream-plifier using a T feedback network (schematic shown in Fig. 6.4a), where two resistance valuesof the source resistor R6 (47 Ω and 470 Ω) are tested to show the variation of transimpedanceaccording to the biasing condition change.
Figure 6.4: Single-transistor transimpedance preamplifier.
from Eq. (6.12). Such results can be proved later by the measured transimpedance
data.
In a single-feedback-resistor transimpedance amplifier system (as shown in
Fig. 6.1a), the bandwidth can be limited by adding a capacitor in parallel with
the feedback resistor Rf . The same technique can be used in a transimpedance
amplifier system with the T feedback network. In Fig. 6.1b, such a capacitor
can be added in parallel with either R1 or R2. Figure 6.6c shows the simulated
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 125
3-dB transimpedance bandwidth corners using two feedback capacitors in the T
feedback network with several different permutations (shunting neither, one, or
two of the consecutive resistors with an 8 pF capacitor). The AD8099 Pream-
plifier with the T feedback network shown in Fig. 6.6a was simulated. When no
capacitor was connected onto the T feedback network, the 3-dB frequency corner
was provided by the parasitic capacitance shunting the feedback system. When
a 8 pF capacitor shunts either the resistor R11 or R12, a lower frequency 3-dB
corner was expected. When both consecutive resistors (R11 and R12 in Fig. 6.6a)
were shunted by a 8 pF capacitor, not only an earlier 3-dB corner was expected,
but also the transimpedance after the 3-dB corner dropped more steeply with
frequency.
6.3.3 Printed Circuit Board
A single-layer PCB was designed and manufactured in-house. All components
were located inside a ground ring to assist shielding from environmental noise.
Low noise surface mount thin-film chip resistors from Panasonic Corporation
(Osaka, Japan) were selected to limit the intrinsic noise from the passive compo-
nents.
The populated PCB (sized∼ 100×60mm2, where the single-transistor pream-
plifier occupies the areas of about 20× 15 mm2), is shown in the Supplementary
Fig. 6.5. This PCB is for bench testing purpose. To test the FT-ICR performance
with this newly designed amplifier, a smaller sized board should be designed, so
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 126
that the stray impedance can be limited. The PCB manufactured for the previ-
ous design reported in Chapter 5 is also used here to test the performance of the
T feedback network.
Figure 6.5: Printed circuit board of four sets of amplifier circuits: (i) the single-transistor preamplifier with a single-resistor feedback (inside the blue circle onthe top-left corner); (ii) the single-transistor preamplifier with the T feedbacknetwork (inside the red circle on the right close to the center); (iii) two sets ofAD8099 voltage amplifiers.
6.4 Circuit Testing
The T feedback network was tested using the previously reported AD8099 pream-
plifier reported in Chapter 5. Figure 6.6a illustrates the schematic of the AD8099
preamplifier with the T feedback network. A voltage amplifier using the same op
amp AD8099 is also shown in Fig. 6.6a. In this report, when testing the T feed-
back network using the AD8099 Preamplifier, the resistance values were always
R1 = 47 kΩ, R2 = 18 kΩ, and R3 = 180 Ω, with a 0.8-pF capacitor C1 shunting
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 127
R2. When testing the newly designed single-transistor transimpedance pream-
plifier with the T feedback network, the 0.8-pF capacitor was not connected.
Since the AD8099 is a high speed op amp with gain-bandwidth product of about
3.8 GHz, the existence of the 0.8-pF capacitor (C1 in Fig. 6.6a) is to limit the
bandwidth to around 1 MHz, which corresponds to the mass-to-charge ratio, m/z
of ∼180 in a 12-T FT-ICR system. For the JFET BF862, its bandwidth char-
acteristic will limit the 3-dB corner to around 1 MHz at a transimpedance of
∼80 dBΩ. To compare the performance of the T feedback network with a single
resistor feedback, a feedback system with only a 4.7-MΩ resistor (in such case,
the capacitor C1 and the resistor R1–R3 were replaced by a 4.7 MΩ resistor,
coupling the Iin and Vout node in Fig. 6.6a) is also tested.
6.5 Results & Discussions
Figure 6.6b shows the frequency response using the AD8099 preamplifier in three
feedback conditions: (i) an 18-kΩ resistor in parallel with a 0.8-pF capacitor
(reported in Chapter 5) (ii) a single 4.7-MΩ feedback resistor, and (iii) the T
feedback network as shown in Fig. 6.6a. The change of the DC blocking capacitor
C2 could be also noticed. C2 was 220 nF in the previous design when the 18-
kΩ curve was measured. In order to set the low corner frequency at around
1 kHz (m/z of 180k in a 12-T system), the 220-nF C2 is replaced by a 4.7-
µF capacitor. A narrower bandwidth for the single 4.7-MΩ feedback curve can
be noticed. It is caused by the parasitic capacitance of this 4.7-MΩ resistor.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 128
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
A A
B B
C C
D D
E E
F F
G G
U1
AD8099
3
2
4
7
6
18
5
6V
R6
3.3kΩ
Vout
-6V
C2
4.7µF
in
-6V
R7180ΩR5
1kΩ
R1
47kΩ
R41kΩ
I Q1
BF862
C3
4.7pF6V
C44.7µF
R8820Ω
R2
18kΩ
R3180Ω
C1
0.8pF
AD_in+
(a) Schematic of the AD8099 Preamplifier with a T feedback network. When the transistor Q1,capacitors C1, C2, and resistors R1–R4 are disconnected, the rest of the circuit forms an AD8099VA, which can be used to match the output impedance of the single-transistor preamplifier to a50-Ω transmission line in an FTMS system.
Figure 6.6: The AD8099 preamplifier with a T feedback network.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 129
1 0 1 0 0 1 k 1 0 k 1 0 0 k 1 M 1 0 M 1 0 0 M4 0
5 0
6 0
7 0
8 0
9 0
1 0 0
1 1 0
1 2 0
1 3 0
Trans
impe
danc
e (dB
Ω)
F r e q u e n c y ( H z )
4 . 7 M Ω f e e d b a c k T - f e e d b a c k ( e q u i v . 4 . 8 M Ω ) 1 8 k Ω & 0 . 8 p F f e e d b a c k
(b) Measured transimpedance of the AD8099 preamplifier with three feedback systems: (i) asingle 18-kΩ feedback resistor shunting a 0.8-pF capacitor; (ii) a single 4.7-MΩ feedback resistor;(iii) the T feedback network as shown by the inset.
Figure 6.6: The AD8099 preamplifier with a T feedback network.
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 130
1 0 0 k 1 M 1 0 M 1 0 0 M
5 0
6 0
7 0
8 0
9 0
1 0 0
1 1 0
1 2 0
1 3 0
1 4 0
Trans
impe
danc
e (dB
Ω)
F r e q u e n c y ( H z )
N o c a p . i n t h e T - f e e d b a c k O n l y C 1 1 c o n n e c t e d O n l y C 1 2 c o n n e c t e d T - f e e d b a c k w i t h b o t h C 1 1 & C 1 2
(c) SPICE simulation of the AD8099 preamplifier frequency response in differentpermutations of the presence of the two 8-pF feedback capacitors (C11 and C12shown by the inset) in the T feedback network.
Figure 6.6: The AD8099 preamplifier with a T feedback network.
Despite the bandwidth variation, in Fig. 6.6b the measured transimpedance of
the preamplifier using the T feedback network (consisting of 47-kΩ, 18-kΩ, 180-Ω
resistors, and a 0.8-pF capacitor) shows a good agreement with the preamplifier
using a single 4.7-MΩ feedback resistor.
The transimpedance frequency response of the single-transistor transimpedance
preamplifier with the T feedback network is measured when a 470-Ω or a 47-Ω
source resistor is used. As shown in Fig. 6.4b, the use of a 470-Ω source resistor
shows a flat transimpedance of a few dBΩ less than the 47-Ω source-resistance
preamplifier. However there is better filtering at low frequency for the 470-Ω
source-resistance curve. As the DC drain current is about 0.95 mA and 5.1 mA
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 131
for the source resistance values of 470 Ω and 47 Ω, respectively, with the sup-
plied voltage of 6 V, the power consumption when using a 47 Ω source resistor
is about 25-mW more than, or around 5-fold of, the power used in a 470-Ω
source-resistance preamplifier. More consumed power means more heat dissipa-
tion. Undesired heating can cause instability, so minimizing preamplifier power
consumption in MS applications is important, especially when the preamplifier
is located in vacuum close to the mass analyzing device for minimal parasitic
capacitance from cabling.
The noise performance of the newly designed single-transistor transimpedance
preamplifier with the T feedback network is not showed, since the noise spectrum
is below the sensitivity of the spectrum analyser, IFR A-7550 (mentioned in Ch-
pater 4), utilized to measured the noise performance. In such a single-transistor
preamplifier design, this T feedback network replaces not only the feedback re-
sistor, but also the biasing resistor (for biasing both the input (gate) terminal
of the transistor and the detection plates of the ICR cell). Thus the amplifier
noise cannot be simply modeled with a simple current or voltage noise source
following a standard noise analysis procedures. Instead, the method suggested
by Letzter and Webster (Letzter and Webster, 1970) must be employed for each
resistor in the network, but calculating the ”referred input noise current” and
”referred input noise voltage” will be difficult as both the noise and the gain
are affected. To predict the noise behavior in this case calls for further careful
research. To test the resulting signal-to-noise behavior, a noise analysis system
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 132
with sensitivity better than the spectrum analyzer used for this report will be
necessary.
The T feedback network provides more bandwidth for a given transimpedance,
since the parasitic capacitance shunts a reduced resistance in this feedback sys-
tem. The T feedback network also increases the transimpedance and the flex-
ibility to obtain transimpedance characteristics for a given application. The
single-transistor transimpedance preamplifier with a T feedback network pro-
vides a low-noise preamplification solution, but less gain in comparison to an op
amp based preamplifier, such as AD8099 preamplifier. Recall that the signal-
to-noise performance of the complete system can be maximized electronically by
not only reducing the noise but also increasing the gain of a preamplifier. An
optimized balance between the preamplifier gain and noise characteristics needs
to be obtained for the best signal-to-noise performance in a given system.
6.6 Conclusion
A new low noise, low power consumption preamplifier for an FT-ICR mass spec-
trometer using a single-transistor transimpedance preamplifier with a T feedback
network has been designed, manufactured and tested. The characteristics of using
a resistive T feedback network is studied by using an op amp. The introduction
of the T feedback network in an op amp based transimpedance system allows
∼100-fold less resistance (under the circumstance that the resistance value of R2
in Fig. 6.1b is 100 times the resistance value of R3) for a given transimpedance,
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 133
hence more bandwidth can be preserved.
In response to the need of the ultra low noise, in this preamplifier design,
a very low noise common-source JFET amplifier is used as the main amplifica-
tion stage. The very large open-loop gain offered by an op amp is traded with
better noise performance and lower power consumption (∼5.7 mW). In such a
case, the measured transimpedance of around 80 dBΩ between the frequencies of
about 1 kHz and 1 MHz satisfies the need of a 12-T FT-ICR mass spectrometer
with corresponding m/z of approximately 180 to 180k. Alternatively, using the
AD8099 Preamplifier with the T feedback network, the transimpedance can be
around 120 dBΩ with roughly the same bandwidth. To use the single-transistor
transimpedance preamplifier with a T feedback network as the front-end elec-
tronics for the FT-ICR MS, the reported AD8099 VA can be a buffer candidate
to match the output impedance to the impedance of a signal transmission cable.
Although here the preamplifier is designed for the application of FT-ICR MS,
a similar technique can be introduced to other mass spectrometers such as the
Orbitrap (Makarov, 2000), time-of-flight (TOF), and ion trap systems. Each of
the three constructing elements in a T feedback network can be in any form of
the combinations of resistive, capacitive, or inductive elements to create any com-
plex impedance to have optimized frequency and transimpedance characteristics
for the requirements of any particular application. Since this design can be an
ideal front-end amplification solution to be applied to any charge/current detect-
ing device, it can be introduced to other areas of applications, such as nuclear
6. Single-JFET Transimpedance Preamplifier Using a T Feedback 134
magnetic resonance spectroscopy (Appelt et al., 2006), optical communication
systems (Green et al., 2008b), or charge-coupled devices (CCDs), etc.
CHAPTER 7
Radio-Frequency Oscillator for a
Quadrupole Mass Filter
A mass analyser is one of the components in a modern mass spectrometer, as
discussed in Section 2.1.2. Generally, in an FT-ICR system, a mass analyser
contains an ion transferring/filtering/accumulating system and an ion cyclotron
resonance (ICR) cell. A quadrupole ion filter can be used here for ion filtering
before sample ions being transferred into the ICR cell for detection. The opera-
tion of a quadrupole ion filter has been introduced in Chapter 3, and a proposed
new electronic device for driving a quadrupole ion filter will be presented in this
chapter.
This chapter reports a radio-frequency (RF) oscillator as a power supply for
operating a quadrupole mass filter. The design reported here has a stabilised
output frequency, and a feedback control for output amplitude stabilisation. This
RF oscillator is designed to operate at room temperature. It can provide two out-
135
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 136
of-phase sinusoidal waveforms at the frequency of 1 MHz and the amplitude of
500 V to the quadrupole filter rods. A 200-V DC power supply will be used to
supply this power amplifier. The chapter first introduces the building blocks of
the new quadrupole mass filter power supply, and is followed by the report of the
electronic details of each building block. Three PCBs are built for testing the
RF oscillator, bandpass filter, gain control circuit, power amplifier, and feedback
control circuits reported here. Then a transformer design is suggested, and is
followed by the discussion of the correlation between the resonant frequency,
and the impedance of the output stage. Future works for this project are also
suggested at the end of this chapter.
7.1 Introduction
The theory of operating a quadrupole mass filter has been discussed in Chapter 3.
The existing power supplies to quadrupole devices have also been briefly reviewed
in Chapter 3. Here, efforts are made to design a new RF power supply for driving
a quadrupole mass filter. The new design, as illustrated by the block diagram in
Fig. 7.1, includes:
• a newly designed RF oscillator (the first stage in the mass filter power sup-
ply diagram shown in Fig. 3.1) with a new built-in automatic gain control
(AGC) unit for output amplitude stabilization;
• a power amplifier (the second stage) modification from the design by Mathur
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 137
and O’Connor in 2006 (Mathur and O’Connor, 2006);
• a redesigned rectifier to generate feedback signal (based on the signal am-
plitude on the quadrupole rod) for the AGC;
• a proposed air-core transformer (the third stage) with better balance, and
more cross-section area to carry the signal current.
Details of such a RF power supply design and the test results are presented in
this chapter.
+ ( U + V cosωt )
– ( U + V cosωt ) –
–
+ +
x
y z
Transformer/ Matching Circuit
Power Amplifier
RF Oscillator with AGC
Rectifier Feedback
feedback signal
Figure 7.1: Block diagram of the newly designed RF power supply for aquadrupole mass filter.
7.1.1 Component Selection
The design goal for this RF power supply is to provide enough voltage and current
to drive the mass filter rods at 1 MHz. As a result, the driving ability of each
component at 1 MHz is very important.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 138
The power amplifier second stage should be able to supply a 1-A current to
drive the transformer third stage. The output sine wave after the transformer
should have an amplitude of less than 500 V. It will be discussed in Section 7.2
that this power amplifier second stage is supplied by a 200-V DC voltage, and
is driven by a current signal from a centre-tapped transformer (details are illus-
trated by Fig. 7.4). Therefore, as the input and output of this power amplifier
second stage are both current signals, a power BJT becomes a better amplifying
transistor candidate over a MOSFET. Details of the power amplifier design will
be reported in Section 7.4.
As current signal is needed to drive the centre-tapped transformer mentioned
above, another BJT is chosen for this purpose. Such a BJT is the output transis-
tor of the RF oscillator first stage. Moreover, there is a crystal oscillator at this
first stage. The crystal oscillator provides voltage signal and prefers a FET load,
as suggested by the datesheet. As a result, a FET (instead of a BJT) is selected
to be driven by the crystal oscillator here. Details of this first stage circuit will
be discussed in Section 7.2.
7.1.2 Circuit Testing
For the circuit testing in this chapter, the DC power to the circuits is supplied by
the TTi DC power supplies mentioned in Chapter 4. The oscilloscope Tektronix
(Beaverton, Oregon) DPO2014 is used to measure the output of the oscillator,
and to perform the fast Fourier transform (FFT) on the output waveforms.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 139
7.2 RF Oscillator Design
A crystal oscillator is a good oscillating reference when designing a RF oscillator
demanding very stable frequency output. The goal is to have a single-frequency
sine wave at the output. Since most of the crystal oscillators output square waves,
a low-pass or a bandpass filter is essential for eliminating the high frequency
components in a square wave.
7.2.1 Crystal Oscillator
The schematic of the new RF oscillating source is shown in Fig. 7.2. A 1.000-MHz
high-stability crystal oscillator, HG-2150CA-SVH, from Epson Toyocom (Tokyo,
Japan),25 is introduced as a square-wave generator with a frequency tolerance of
±15×10−6. A 5-V DC voltage for this crystal oscillator is supplied by a regulator,
LM78L05, from Texas Instruments (Dallas, Texas, USA). The crystal drives a
N-channel enhancement-mode FET, 2N7002, from Fairchild Semiconductor (San
Jose, California). This FET has an input capacitance of around 20 pF at 1 MHz,
which was tested drivable by the chosen crystal oscillator. Its maximum drain-
source voltage rating of 60 V is also suitable for this application, in which less
than 12 V drain-source voltage will be supplied to this FET. By varying the
supply voltage to the FET 2N7002, the biasing drain current is changed, causing
a transconductance gm change. Therefore, the the output amplitude can be
altered. Here, the OE (output enable) pin of the crystal oscillator is connected
25See http://www.epsontoyocom.co.jp/english/product/OSC/set01/hg2150ca/index.html forinformation about high-stability oscillator HG-2150CA-SVH, accessed 15 August 2012.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 140
to the DC supply voltage (5 V in this case) to always enable this crystal oscillator.
Potentially, this OE pin can be used as a control terminal to disable the oscillator
output.
Gai
n C
on
tro
l
Cry
stal
O
sc.
Re
gula
tor
Byp
ass
Cap
. Em
itte
r Fo
llow
er
Ban
dp
ass
Filt
er
(a) Block diagram of the schematic shown in Fig. 7.2b.
Figure 7.2: Fixed frequency RF oscillating source with output amplitude con-trol.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 141
collector
3 41
5 2
U3 LT62
05R7 10
k
R10
56
C5 150p
C6 150p
12
OE1
GND
2OU
T3
Vcc
4
XTAL
1
OSC
MODU
LE
C2 0.1u
R8 11k
R11
1k
C9 0.1u
12
VCC
+12V
GND
C4 1n
R12
1k
C10
0.1u
12
R13
1.8k R14
2.2k
12
R3 3.6k
R5 270
12
R1 11k
R9 10k
C8 0.1u
Q1 2N70
02
R2 10k
Q2 DJT4
031N
VI8
VO1
GND 2/3/6/7
U2
78L0
5
VCC
12
C3 0.33u
12
R6 10
3 41
5 2
U4 LT62
05
VCC
R4 5k
C1 0.1u
3 21
8 4
U1:A LT62
06
5 67
8 4
U1:B
LT62
06FB
FIT P
OWER
RES
ISTO
R
C7 0.1u
R15
100
PA
BAND
-PAS
S FIL
TER
CV
(b) Schematic.
Figure 7.2: Fixed frequency RF oscillating source with output amplitude con-trol.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 142
7.2.2 Bandpass Filter
In the schematic shown in Fig. 7.2, two LT6205 operational amplifiers (op amps)
from Linear Technology (Milpitas, California, USA),26 construct a bandpass filter
and an output buffer. The op amp LT6205 has a gain-bandwidth product of
100 MHz, and can be supplied with DC voltage of 12 V. Therefore LT6205
(along with its dual version, LT6206, and quad version, LT6207) is widely used
in this oscillator circuit. This commonly known Deliyannis-Friend bandpass filter
(Deliyannis, 1968; Friend, 1970; Friend et al., 1975), consisting of the op amp
U3, resistors R7–R12, and capacitors C5–C6, is designed to have both gain and
Q around 10. A SPICE simulation is performed to test the frequency response of
this bandpass filter. The result in Fig. 7.3 shows a very narrow window for the
1-MHz signal, which is ideal for reshaping the 1-MHz square wave into a mostly
pure sine wave. The output of the bandpass filter fed a voltage follower op amp,
U4, to prevent the following stages from interfering with the filter.
7.2.3 Gain Control Scheme
An AGC scheme, which is similar to what was used in the Mathur and O’Connor’s
oscillator (Mathur and O’Connor, 2006), is utilized partially by two op amps,
LT6206.27 In the schematic shown in Fig. 7.2, both op amps (U1:B and U1:A)
are configured as comparators for the gain control signals, FB (feedback) and CV
26See Appendix A.3 for information about the op amp LT6205.27Op amp LT6206 is the dual version of LT6205. See Appendix A.3 for information about
the op amp LT6206.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 143
1 0 0 k 1 M 1 0 M
- 4 0
- 3 0
- 2 0
- 1 0
0
1 0
2 0Vo
ltage
Gain
(dB)
F r e q u e n c y ( H z )
L T 6 2 0 5 b a n d p a s s f i l t e r
Figure 7.3: The frequency response simulation of the Deliyannis-Friend band-pass filter in Fig. 7.2.
(control voltage), respectively. Note that here the voltage at pin CV can be also
generated by the resistor R3 and trimmer resistor R4. Pin CV is preserved for
a control input to assign the output amplitude. The existence of R3 and R4 is
just for testing convenience, and can be removed when a control voltage input is
provided to pin CV.
Pin FB is preserved for the feedback signal, which is designed to be propor-
tional to the amplitude of the RF signal applied to the quadrupole mass filter.
With no input, op amp U2B generates a 3 V DC voltage, which was found a
good value to prevent destabilization, for FET Q1. As the voltage on the FB
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 144
pin increases, the amplitude of the square wave output after Q1 will drop. In
general, the signal from terminal FB is to stabilize the output amplitude to an
assigned level, which is determined according to the voltage at terminal CV.
A power transistor, the NPN BJT, DJT4031N,28 from Diodes Inc. (Plano,
Texas), is able to be operated with a 3-A continuous collector current. The
transition frequency of 105 MHz reported on the datasheet indicates that this
BJT is suitable as a driving buffer for a 1 MHz signal here. This BJT is designed
as an output stage for driving the primary coil of a transformer with a centre-
tapped secondary winding to generate two symmetrical but 180-degree out of
phase sinusoidal waveforms. Such a transformer is illustrated in Fig. 7.4. The
’RF Source’ supplies a sine wave to the primary coil of this transformer. At the
output node ’V 0’ the output waveform will be a 180-degree out of phase of the
output at the node ’V 180.’ These waveforms can be used as the input signal for
the power amplifier second stage (formed of power BJTs), which will be reported
in Section 7.4.
7.2.4 RF Oscillator Printed Circuit Board
A single-layer PCB (sized roughly 72×25 mm), as shown in Fig. 7.5, is designed,
etched, and populated in-house. This single-layer board is for the purpose of
feedback signal testing. All components on the board are located inside a ground
ring/ground plate. The noise-sensitive crystal oscillator and its power supply,
28See http://www.diodes.com/products/catalog/detail.php?item-id=6044 for informationabout BJT, DJT4031N, accessed 15 August 2012.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 145
0
0
1
1
2
2
3
3
4
4
5
5
6
6
7
7
8
8
A A
B B
C C
D D
E E
F F
G G
RF Source
V_0
V_180
Figure 7.4: A transformer in which its secondary coil is centre-tapped. TheRF Source represents the circuit shown in Fig. 7.2, and drives the primary coilof this transformer. The output nodes ’V 0’ and ’V 180’ will be connected to thepower amplifier decried later in Section 7.4.
the 5-V DC regulator, (XTAL1 and U3 in the schematic shown in Fig. 7.2, re-
spectively) are located further away from other components to minimise noise
interference. The components of the Deliyannis-Friend bandpass filter are lo-
cated as close as possible to minimize the length of the traces, so that the stray
impedance from the traces is limited. For heat dissipation purpose, the collector
of the BJT, DJT4031N, is soldered on a large 11 × 11 mm metal plate located
around the top-right corner of the PCB inside the ground ring.
7.2.5 Testing Results
Figure 7.6 shows the output waveforms measured at (a) the OUT terminal of the
crystal oscillator XTAL1 and (b) the output of the op amp U4. The feedback
voltage at pin FB is set to around 2.0 V. At the OUT terminal of the crystal
oscillator, a 1.0-MHz square wave with a peak-to-peak amplitude of 5.0 V is
shown. The spectrum after FFT shows a 1.0-MHz peak with an amplitude of
7.2 dB, and a third harmonic at 3.0 MHz of -2.4 dB. After the FET Q1 and
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 146
(a) Printed circuit board layout file.
(b) Fully populated board, sized 72× 25 mm.
Figure 7.5: Printed circuit board of of the fixed frequency RF oscillating sourcewith output amplitude control.
the Deliyannis-Friend bandpass filter, the signal becomes a 1.0-MHz, 5.1-V sine
wave with a measured first harmonic of 5.3 dB. The third harmonic is hard to
be seen on the screen. However, the highest peak amplitude after 1.0 MHz is
-43 dB located at around 3.2 MHz. The odd harmonics from the square wave
generated by the crystal oscillator is effectively degraded. In particular, the third
harmonic at the op amp output is at least ∼40-dB lower than the original signal,
making the output of this oscillator an ideal fixed-frequency source for a mass
filter power supply.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 147
(a) The output of the crystal oscillator and the output waveformspectrum, measured at the OUT terminal of XTAL1 in Fig. 7.2.
(b) The output and its spectrum, measured at the output of op ampU4 when a ∼2.0 V voltage applied to the FB node.
Figure 7.6: The oscilloscope screen snapshots of the output waveforms (shown indark blue) from both the crystal oscillator and the bandpass filter. The measuredamplitudes shown on the screen are peak-to-peak values, and the peek intensityafter FFT (shown in red) is reported in dB.
Figure 7.7 shows the measured correlation between the feedback voltage at
pin FB and the output peak-to-peak amplitude at the output of op amp U4.
This is measured by using the oscilloscope DPO2014. Each point represents one
measurement. The power consumption is also monitored and plotted using the
y-axis on the right side. The power consumption of this RF oscillating source
depends on the output amplitude, and is in general less than 0.70 W. When
the feedback voltage is between 1.6 V and 2.4 V, an inversely linear relationship
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 148
is shown between the feedback voltage and the output amplitude. When the
feedback voltage is below 1.6 V, the output starts to saturate.
This FB pin is part of the AGC unit, and will intake the control signal gen-
erated by the precision rectifier (reported in Section 7.3). The precision rectifier
will be used to sense the output of the transformer stage (the third stage) of
a mass filter power supply. The generated feedback signal has to be adjusted
finely so that the feedback signal is limited between 1.6 V and 2.4 V and linearly
correlates to the output signal.
1 . 5 2 . 0 2 . 50
2
4
6
8
1 0
1 2
1 4
Outpu
t Amp
litude
(VPP
)
F e e d b a c k ( F B ) V o l t a g e ( V )
O u t p u t A m p l i t u d e P o w e r C o n s u m p t i o n
0 . 3 5
0 . 4 0
0 . 4 5
0 . 5 0
0 . 5 5
0 . 6 0
0 . 6 5
0 . 7 0
Powe
r Con
sump
tion (
W)
Figure 7.7: Measured RF oscillator output peak-to-peak amplitude (at the baseterminal of Q2 in Fig. 7.2) and the power consumption correlated to the appliedvoltage at pin FB (for a feedback signal).
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 149
7.3 Precision Rectifier Design
The precision rectifier is formed of a commonly used full-waveform rectifier with
op amps LT620729 and fast Schottky diodes HSMS-2800 from Avago Technologies
(San Jose, California, USA). This precision rectifier unit is designed to monitor
the output amplitude, and then convert the amplitude into a DC signal for the
feedback signal FB shown in Fig. 7.2. The schematic of the precision rectifier is
shown in Fig. 7.8.
The first part here is a commonly used precision rectifier design (Horowitz
and Hill, 1989) to ensure that the behaviour of this rectifier is close to ideal, so
that unwanted voltage drifting can be limited to achieve the output amplitude
stabilization goal of having a ∆V/V below 2.5× 10−4 (discussed in Section 3.3).
A unity-gain voltage follower formed by op amp U5:A is used to isolate the
input of the precision rectifier. After the buffer, the op amp U5:D generates the
first 1/2 wave, and the op amp U5:C finishes the full wave. Two more unity-gain
buffers (the op amp U6:A and U6:B) are utilized for isolation and for DC offset
removal. After the signal is smoothed by the capacitor C26 and the resistor R39,
a feedback DC signal is generated according to the amplitude of the sine wave
fed into the node ACFB shown in Fig. 7.8.
29Op amp LT6207 is the quad version of LT6205. See Appendix A.3 for information aboutthe op amp LT6207.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 150
R23
10k
AK
D1 HSMS
-2800
AK
D2 HSMS
-2800
R27
10k
R24
10k R2
510
k
12
C21
1n
R21
100k R2
210
0k
R28
10k R2
910
k
R26
10k
R30
10k
12
12
R35
10k
R36
10k
R37
10k
R38
10k
R31
1.5k
R32
1k
R33
270 R3
433
0
12
1212
AK
D3 HSMS
-2800
R39
1kC2
61n
FB
14 1516
4 13
U5:D
LT62
07
12 1110
4 13
U5:C
LT62
073 2
1
4 13
U5:A
LT62
07
12
3 21
8 4
U6:A
LT62
06
5 67
8 4
U6:B
LT62
06
BUFF
ER1/2
WAV
EFU
LL W
AVE
INV
BUFF
ERIN
V DC
OFF
SET R
EMOV
AL
12
C24
0.1u
C25
0.1u
C22
0.1u
C23
0.1u
12ACFB
Figure 7.8: Schematic of the precision rectifier.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 151
7.3.1 Precision Rectifier Printed Circuit Board
The PCB designed for the precision rectifier also houses the circuit reported in
Section 7.2. Namely, the circuits shown in both Fig. 7.2 and Fig. 7.8 are on this
board. This two-layer PCB (shown in Fig. 7.9) is sized about 75× 58 mm and is
designed, etched, and populated in-house, employing the same design concepts
mentioned in Section 7.2.4 (large ground plate, small area for the bandpass filter,
large heat dissipation plate for the output emitter follower Q2, and isolated area
for the crystal oscillator XTAL1, etc.)
Figure 7.9: Two-layer printed circuit board of the precision rectifier (only thetop layer is shown; the bottom layer is the ground/power plate.) This board alsohouses all of the components on the PCB shown in Fig. 7.5b.
When testing this board, the output of the emitter follower Q2 (pin PA in
Fig. 7.2) is connected to the input of the precision rectifier (pin ACFB in Fig. 7.8).
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 152
An amplitude-stabilized sine wave similar to the waveform reported in Fig. 7.6b
can be obtained. Note that such connection arrangement is for testing only.
When the mass filter power supply circuit is finalized, the output of the emitter
follower is feeding a power amplifier, and the input of the precision rectifier will be
coupled to a voltage divider, which divides the voltage applied to the quadrupole
rods to a reasonable level for the precision rectifier input.
7.4 Power Amplifier Design
The power amplifier stage employs the NPN bipolar power transistor, MJE18008,
from ON Semiconductor (Phoenix, Arizona, USA),30 because of its large maxi-
mum collector-emitter voltage rating of 450 V and its maximum collector current
rating of 8 A. A 200-V DC power supply will be used to supply this power ampli-
fier. This power amplifier is designed to be able to provide at least 1 A of current
to drive an output transformer. A maximum output amplitude of 500 V after
the transformer will be expected. A common-emitter configuration is set up, and
a NPN Darlington pair, BU323Z (with collector-emitter voltage rating of 350 V
and collector current rating of 10 A), from ON Semiconductor (Phoenix, Ari-
zona, USA),31 is utilized to supply the base current. Such characteristics enable
the potential to deliver large voltage and current into the quadrupole system.
This Darlington pair can provide a few amps of current, which may be necessary
30See Appendix A.4 for information about the bipolar power transistor MJE18008.31See http://www.onsemi.com/PowerSolutions/product.do?id=BU323Z for information about
the Darlington pair BU323Z, accessed 15 August 2012.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 153
to drive base terminals of both MJE18008 transistors when the current gain of
MJE18008 is set to around 10. The circuit schematic is shown in Fig. 7.10.
Figure 7.10: Schematic of the power amplifier stage.
The transformer mentioned in Section 7.2.3 (and illustrated in Fig. 7.4) is used
here to couple the oscillating sine wave (generated by the RF oscillator described
in Section 7.2) onto the base current supplied by the Darlington pair BU323Z
(Q3 provides current for the transformer T2 in Fig. 7.10). Such a transformer
has a ferrite core with three sets of 10-turn winding. Namely, the primary coil
and the both sides of the secondary coil have a winding of 10 turns.
Similar to the design reported by Mathur and O’Connor in 2006 (Mathur and
O’Connor, 2006), the load of the power transistor is an air-core transformer (T1)
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 154
in which both primary and secondary coils are centre-tapped. By increasing the
secondary-to-primary turns ratio of the transformer (assuming the primary coil
is the coil connected with the power BJT Q1 and Q1 in Fig. 7.10), the output
voltage can be increased to fit the need of a given quadrupole system. A DC offset
can be provided by the biasing voltage ’V bias’ to the centre-tap of the secondary
coil of the transformer T1. Here the output waveforms at nodes ’out1’ and ’out2’
become the signals described by Eq. (3.1) (also shown in Fig. 7.1), where the
U is the voltage supplied from the node ’V bias,’ and the V is the amplitude
of the sinusoidal waveform controlled by the gain control function described in
Section 7.2.3. In order to balance the outputs at node ’out1’ and ’out2’ (to have
the same output amplitude on both sides), the left side (components C4, Q1,
R3–R5 in Fig. 7.10) and right side (C7, Q2, R6–R8) of the oscillating circuitry
should be made identical, which makes the layout of the power amplifier PCB
crucial.
7.4.1 Power Amplifier Board
The PCB design follows the principle of making the layout of the oscillating pairs
Q1 and Q2 in Fig. 7.10 symmetrical. Therefore the component placement and
the routing have to be considered carefully. Figure 7.11 shows the fully populated
two-layer PCB. There are two electrically isolated heat sinks attached to each
other to ensure the working temperatures of the two power transistors remain
roughly the same. The components on the right side of the heat sink form a
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 155
bridge rectifier, which is not shown in the schematic of Fig. 7.10. Such a bridge
rectifier is for testing purpose and will be removed when the mass filter power
supply circuit is finalised.
7.5 Transformer
For transformers with centre-tapped coils such as the transformer T1 shown in
Fig. 7.10, it is easy to have a slightly mismatched turns ratio between the left and
right half of the coils. To gain a better balance, the bifilar winding is introduced.
A bifilar coil, as shown in Fig. 7.12, is wound with two closely spaced wires, which
is an excellent solution to have identical coils on both sides of a canter tap of the
transformer.
Figure 7.13 shows the photo of the hand-wound air-core bifilar transformer.
The primary coil is a 10-turn bifilar winding. By connecting the left end of
the first wire with the right end of the second wire, a centre tap is made, and
it becomes a centre-tapped coil with theoretically identical 10-turn winding on
both sides. Same method is employed for the 29-turn secondary coil, which is
wound on top of the primary coil.
Since the transformer is the load of the power amplifiers Q1 and Q2, the
impedance, determined by the the tubing dimension and the turns of winding, of
the transformer decides the gain of this power amplifier. Meanwhile, the resonant
frequency of the output system is a function of the impedance. The designed
transformer impedance has to be carefully considered. When the quadrupole
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 156
(a) Top view.
(b) Front view.
(c) Bottom layer.
Figure 7.11: Power amplifier board with components mounted.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 157
Figure 7.12: Winding of a bifilar coil with current flowing in the same direction(Ahn et al., 2011) [modified].
rods are connected, a resonant frequency of slightly less than 1 MHz is preferred
so that unwanted higher-order harmonics can be filtered out. If it is designed
that the resonant frequency of this output system is far away from 1 MHz, more
power will be needed to drive the quadrupole rods.
7.5.1 Testing Results
(Power Amplifier & Transformer)
Figure 7.14 shows the first test result of the power amplifier board and the air-
core transformer. To simplify the testing, only the differential voltage of the
two output nodes ’out1’ and ’out2’ in Fig. 7.10 is shown, and the centre tap
of the secondary coil is not connected to the biasing circuit for a DC offset.
As a result, for this test, the secondary-to-primary turns ratio of this output
transformer T1 becomes 58:10. The input signal is a 1-MHz sine wave with the
amplitude of ∼5.5 V (peak-to-peak value), generated by a function generator.
Under such conditions, an output waveform of around 1 MHz with the peak-to-
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 158
(a) Primary coil.
(b) Secondary coil.
Figure 7.13: Photo of the hand-wound air-core bifilar transformer.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 159
peak amplitude of 328 V is obtained.
Figure 7.14: The oscilloscope screen snapshot of the power amplifier output(shown in red), differential signal measured between the nodes ’out1’ and ’out2’after the output transformer (T1 in Fig. 7.10) with turns ratio of 58:10. TheFFT result is shown in light blue.
7.5.2 Future Transformer Modification
In the first attempt of making a transformer, the enameled wire (or the magnet
wire) with a size of American wire gauge (AWG) 16 was used. An AWG 16 wire
has a diameter of about 1.3 mm. However, due to the skin effect, in copper,
a 1-MHz signal is flowing within the skin depth of ∼66 µm. Therefore, most
of the signal current only flows within the ’66-µm skin’ of the wire used in the
home-made transformer mentioned above in Section 7.5. Alternatively, a Litz
wire, containing multiple stands which are electrically insulated from each other,
becomes a better choice when a large 1-MHz signal current is carried (more
cross-section area can be used to carry such a current). Figure 7.15 shows the
corss-section view of a Litz wire.32
32Figure 7.15 was obtained from the website http://www.rubadue.com/products/supplementary-2-layer-insulation-litz-wire-fep-insulation-double-insulated-wire,accessed 30 August 2012.
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 160
Figure 7.15: Cross-section view of a Litz wire.32
7.6 Conclusion & Future Work
As illustrated by Fig. 7.1, a completed quadrupole power supply consists of four
sections, a RF oscillator with an AGC unit, a power amplifier, a trasformer, and
a feedback signal generation system. Here, a RF oscillator with gain control unit
has been built and tested (reported in Section 7.2). A power amplifier and a
transformer has also been built and tested (reported in Section 7.4 and 7.5).
The rectifier feedback system shown in Fig. 7.1 has been partially built and
tested (reported in Section 7.3). Depending on the designed maximum output
amplitude being applied to the quadrupole rods, a voltage divider has to be
made to complete the feedback system. To assign such a maximum output am-
plitude, those parameters, such as the quadrupole size and ion mass, mentioned
in Section 3.2 have to be considered.
Recall that the impedance of the transformer and the quadrupole rods de-
termine the resonant frequency at the power amplifier output. The transformer
turns ratio also affects the output amplitude and transformer impedance. As a
7. Radio-Frequency Oscillator for a Quadrupole Mass Filter 161
result, to build a proper transformer that is suitable for a given mass filtering
solution calls for more careful system tunings.
A solution to reduce such a complexity is to design a transformer with resonant
frequency far away from the power supply frequency (1 MHz in this case). In
such a case, one no longer worry about the system tuning to have the resonant
frequency of the output system around 1 MHz. However, more current is needed
for driving such a system, and under such a condition the heat dissipation has to
be managed carefully. The dimension of the quadrupole to be used as a mass filter
can affect the output impedance and the design specifications of the RF power
supply. Since the dimension details have not been determined, this circuit (in
particular, the transformer and the quadrupole connection/mounting structure)
will be finalised after the design of the quadrupole.
CHAPTER 8
Conclusion
In this thesis, the history and important milestones of the mass spectrometry
(MS) development are reviewed. It is then followed by the studies of the ion
cyclotron resonance (ICR) technique, the ICR signal, and the quadrupole oper-
ation. Different electronic components for a high resolution/mass accuracy mass
spectrometer have been investigated in hope that the new designs and the learned
experiences contribute to the MS community. Two new preamplifiers for a 12-
T Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometer have
been designed, built and tested for the operation at room temperature. A new
quadrupole mass filter power supply has also been designed, built, and tested.
162
8. Conclusion 163
8.1 FT-ICR Preamplifier
8.1.1 Preamplifier Using an Operational Amplifier
The first transimpedance preamplifier using an operational amplifier (op amp)
AD8099 shows a tested transimpedance of around 85 dBΩ between the frequency
of 3 kHz and 10 MHz (corresponding to the mass-to-charge ratio, m/z, of approx-
imately 18 to 61k for a 12-T FT-ICR system), when a single 18-kΩ (shunting a
0.8-pF feedback capacitor) feedback resistor is employed. It also has an tested
input current noise spectral density of around 1 pA/√
Hz. The total power con-
sumption of this circuit is around 310 mW when tested on the bench. The tran-
simpedance and the bandwidth can be adjusted by replacing passive components.
The feedback and bandwidth limitation of the circuit is discussed. With the max-
imum possible transimpedance of 5.3 MΩ when using an 0402 type surface mount
resistor, the preamplifier is estimated to be able to detect ∼110 charges.
8.1.2 Single-Transistor Preamplifier Using a T Feedback
Network
The second preamplifier employs a single-transistor design using a T-shaped feed-
back network. The T feedback network can bias the transistor gate and the ICR
detection plate. This single-transistor preamplification solution provides a tran-
simpedance of about 80 dBΩ between 1 kHz and 1 MHz (m/z, of around 180
to 180k for a 12-T FT-ICR system) and a low power consumption of ∼5.7 mW.
8. Conclusion 164
The T feedback system allows ∼100-fold less feedback resistance at a given tran-
simpedance, hence preserving bandwidth, which is important for the introduc-
tion of a more complicated modern ICR cell (namely, more input capacitance
is introduced by a modern ICR cell). In trading noise performance for higher
transimpedance, an alternative preamplifier design has also been studied with
a capability of a transimpedance of 120 dBΩ in the same bandwidth of about
1 MHz.
8.1.3 Cyclotron Frequency Correlation
The cyclotron frequency correlation has been discussed and estimated in Sec-
tion 5.2.2. It has been concluded that, for a transimpedance preamplifier, when
the magnitude of the feedback resistance is much greater than the magnitude of
the reactance of the effective input capacitance, R >>1
ωcyc(C/G), the magnitude
of the output voltage after such a preamplifier is independent of the cyclotron
frequency of the ion signal.
In a scenario when an input capacitance of 10 pF, and the cyclotron frequency
fcyc of signals between 10 kHz and 10 MHz (mass-to-charge ratio, m/z, of ∼18
to 18k for a 12-T FT-ICR system) is provided, a feedback resistance of much
larger than 16 GΩ is necessary for the elimination of the cyclotron frequency
dependency. However, with the introduction of protection circuitry at the input
node for performing excitation and detection on the same ICR electrode (Chen
et al., 2012), or the relocation of the preamplifier outside of the vacuum chamber
8. Conclusion 165
housing an ICR cell, more capacitance (from a longer cable and the feedthrough)
is coupled onto the input node of a preamplifier. For instance, when the input
capacitance is 200 pF, and the magnet is 21 T, (in which the cyclotron frequency
of a m/z 18k ion is about ∼18 kHz), the magnitude of1
ωcyc(C/G)becomes
around 440M. With the employment of a T feedback network in an op amp based
preamplifier, a transimpedance of much larger than 440 MΩ can be reachable.
As a result, the voltage output can be independent of the cyclotron frequency
when using a transimpedance preamplifier.
8.2 Power Supply for a Quadrupole Mass Filter
The oscillator reported in Chapter 7 is a simple solution to supply a quadrupole
mass filter. The most challenging part of the quadrupole mass filter power supply
design is the tuning after connecting each of the ’building blocks’ together. As
mentioned in Section 7.6, the design of the output transformer depends on many
factors, such as the quadrupole dimension, the designed gain and output ampli-
tude, and the resonant frequency, etc. More investigations are required before
finalising such a design.
Furthermore, if the output transformer is designed not to resonant at the
frequency of operation, the impedance tuning (to tune the resonant frequency of
the output stage to the operating frequency of the RF power supply) may not be
necessary. Instead, more current needs to be provided when operating the power
supply system under such a condition.
CHAPTER 9
Future Work
This chapter suggests the future works to extend this research in the area of
high performance MS related electronics. In the previous chapters, two tran-
simpedance preamplifiers for a 12-T Fourier-transform ion cyclotron resonance
(FT-ICR) mass spectrometer have been designed, built, and tested on the bench
for the operation at room temperature. It is planned to test both preamplifiers
using an electrically compensated ion cyclotron resonance (ICR) cell (Brustkern
et al., 2008). In this work, a new quadrupole mass filter power supply has also
been designed, built, and tested.
This thesis focus on reporting the bench testing results. It is hoped that in
the future those newly designed electronic components can be tested on a 12-T
FT-ICR mass spectrometer.
166
9. Future Work 167
9.1 Preamplifier
The bandwidth of both preamplifiers reported in this work is designed for a
12 T FT-ICR system. A 1-MHz bandwidth corresponds to the mass-to-charge
ratio, m/z, of approximately 180 at 12 T, which is suitable for most of the pro-
teomics applications using an FT-ICR mass spectrometer. The bandwidth of
both preamplifiers can be changed by replacing passive components, for the ap-
plications/tests in an FT-ICR system with different strength of magnetic field.
The following tests are suggested to extend this research.
9.1.1 T Feedback Network
The T feedback network tested in Chapter 6 is a resistive network, in which a
47-kΩ and a 18-kΩ resistor are connected in series and a third 180-Ω resistor
coupled between the junction node of the two series resistors and a reference
potential. For such specified resistive values, the T feedback network is proven
to have the ability to preserve bandwidth at a given transimpedance. For the
use of this T feedback network in a single-transistor preamplifier, one also takes
the advantage of the reference potential from the feedback network, which biases
the gate terminal of the transistor and the detection plate of the ion cyclotron
resonance (ICR) cell. One may claim that the lowered resistance value results
in more thermal noise current. However, each of the component in a T feedback
network can be replaced by reactive components to avoid thermal noise. It can be
capacitive, inductive, or both, for obtaining transimpedance characteristics which
9. Future Work 168
vary with frequency according to particular requirements in a given application.
Therefore, further research should be conducted to obtain the best combination
of the impedance values for each of the T feedback network components for
noise/bandwidth performance optimization.
9.1.2 Preamplifier Noise Performance
As of the noise performance of a preamplifier system, the noise power of any
given system can not be lower than the thermal noise power, which is a function
of the temperature and bandwidth (as described by Eq. (2.17)).
One may be able to take advantage of the cooling system of the FT-ICR su-
perconducting magnet to reduce the operating temperature of an FT-ICR pream-
plifier to a cryogenic level. The thermal noise power can be reduced by around
10 fold at the temperature of 4 K. With a much reduced noise level, a much
improved signal-to-noise performance may make it possible to detect a singly-
charged single ion in an ICR cell.
Reducing the bandwidth also limits the noise. In lieu of detecting the 1-MHz
bandwidth in one detection using a 1-MHz detection window, a narrower 100-kHz
window can be introduced to finish the same task in 10 detections.
Moreover, as reported in Chapter 5 and Chapter 6, the op amp based pream-
plifier provides more gain (transimpedance), whilst the single-transistor based
preamplifier reduces the noise and the power consumption. The permutation
between gain, bandwidth, noise performance, and power consumption (when a
9. Future Work 169
preamplifier is placed inside a vacuum chamber, the heat dissipation can affect the
vacuum condition) should be studied in the future for best system performance
for an FT-ICR mass spectrometer.
9.1.3 Cyclotron Frequency Correlation
The correlation of the cyclotron frequency and the preamplifier gain/transimpedance
is studied in Section 5.2.2 and in Section 8.1.3. The cyclotron frequency of the
ICR signal and the preamplifier gain/transimpedance can affect the output sig-
nal of the FT-ICR ion detector. Best calibration functions for processing the ion
signal should be studied when testing the transimpedance preamplifiers.
9.1.4 System Test
New compact PCBs using two layers should be built when testing the preampli-
fiers on an FT-ICR system. It is planned to first place the preamplifiers outside
of the vacuum chamber and to connect the detection plates via feedthroughs.
Then the performance variance when mounting the preamplifiers inside the vac-
uum chamber (here, the outgassing characteristics of the preamplifier components
should be evaluated) should be compared. The bandwidth tolerance, number of
ions being detected, dynamic range, and noise performance of the FT-ICR system
using new preamplifiers should be evaluated.
9. Future Work 170
9.2 Power Supply for a Quadrupole Mass Filter
As described in Chapter 7, a completed quadrupole power supply consists of four
sections, a RF oscillator with built-in automatic gain control (AGC), a power
amplifier, a transformer, and a feedback signal generation system. In this work,
a RF oscillator, a rectifier feedback system, and a power amplifier have been
built and tested. The dimension details of the quadrupole mass filter have not
been determined. As a result, the maximum output amplitude of this power
supply has not been defined. However, the output amplitude can be changed by
changing the turns ratio of the output transformer. The output impedance of this
system is defined by the transformer, quadrupole, and its connection/mounting
component. The resonant frequency at the output stage is affected by such an
output impedance. As a result, future investigations on those parameters are
required before finalising the power supply.
References
Ahn, M. C., Jang, J. Y., Ko, T. K., and Lee, H. (2011). Novel design of
the structure of a non-inductive superconducting coil. IEEE Transactions on
Applied Superconductivity, 21(3):1250–1253. (In this thesis: page 157.)
Aizikov, K., Mathur, R., and O’Connor, P. B. (2009). The spontaneous loss of
coherence catastrophe in Fourier transform ion cyclotron resonance mass spec-
trometry. Journal of the American Society for Mass Spectrometry, 20(2):247–
256. (In this thesis: page 33.)
Aizikov, K., Smith, D. F., Chargin, D. A., Ivanov, S., Lin, T.-Y., Heeren, R.
M. A., and O’Connor, P. B. (2011). Vacuum compatible sample positioning
device for matrix assisted laser desorption/ionization Fourier transform ion cy-
clotron resonance mass spectrometry imaging. Review of Scientific Instruments,
82(5):054102. (In this thesis: page 16.)
Amster, I. J. (1996). Fourier transform mass spectrometry. Journal of Mass
Spectrometry, 31(12):1325–1337. (In this thesis: page 16.)
171
REFERENCES 172
Analui, B. and Hajimiri, A. (2004). Bandwidth enhancement for transimpedance
amplifiers. IEEE Journal of Solid-State Circuits, 39(8):1263–1270. (In this
thesis: page 116.)
Appelt, S., Kuhn, H., Hasing, F. W., and Blumich, B. (2006). Chemical analysis
by ultrahigh-resolution nuclear magnetic resonance in the earth’s magnetic field.
Nature Physics, 2(2):105–109. (In this thesis: page 134.)
Austin, W. E., Holme, A. E., and Leck, J. H. (1976). The Mass Filter: De-
sign and Performance, in P. H. Dawson (Ed.), Quadrupole Mass Spectrometry
and Its Applications, chapter 6, pages 121–152. American Institute of Physics,
Woodbury, New York. Chapter VI. Reprinted as an ”American Vacuum Society
Classic” by American Institute of Physics. (In this thesis: page 56, 57, 65.)
Barros, M. A. M. (1982). Low-noise InSb photodetector preamp for the infrared.
IEEE Journal of Solid-State Circuits, 17(4):761–768. (In this thesis: page 114.)
Barrow, M. P., Witt, M., Headley, J. V., and Peru, K. M. (2010). Athabasca
oil sands process water: Characterization by atmospheric pressure photoioniza-
tion and electrospray ionization Fourier transform ion cyclotron resonance mass
spectrometry. Analytical Chemistry, 82(9):3727–3735. (In this thesis: page 16.)
Beynon, J. H. and Morgan, R. P. (1978). The development of mass spectroscopy:
An historical account. International Journal of Mass Spectrometry and Ion
Physics, 27(1):1–30. (In this thesis: page 9, 10.)
REFERENCES 173
Brustkern, A. M., Rempel, D. L., and Gross, M. L. (2008). An electrically com-
pensated trap designed to eighth order for FT-ICR mass spectrometry. Journal
of the American Society for Mass Spectrometry, 19(9):1281–1285. (In this thesis:
page 33, 166.)
Burlingame, A. L., Boyd, R. K., and Gaskell, S. J. (1998). Mass spectrometry.
Analytical Chemistry, 70(16):647–716. (In this thesis: page 13.)
Canterbury, J. D., Gladden, J., Buck, L., Olund, R., and MacCoss, M. J. (2010).
A high voltage asymmetric waveform generator for FAIMS. Journal of the
American Society for Mass Spectrometry, 21(7):1118–1121. (In this thesis: page
64.)
Caravatti, P. and Allemann, M. (1991). The ’infinity cell’: a new trapped-ion
cell with radiofrequency covered trapping electrodes for Fourier transform ion
cyclotron resonance mass spectrometry. Organic Mass Spectrometry, 26(5):514–
518. (In this thesis: page 33.)
Cermak, I. (2005). Compact radio-frequency power supply for ion and particle
guides and traps. Review of Scientific Instruments, 76(6):063302. (In this thesis:
page 59, 60, 65.)
Chang, B. T. and Mitchell, T. B. (2006). Frequency stabilized radio-frequency
generator for driving ion traps and other capacitive loads. Review of Scientific
Instruments, 77(6):063101. (In this thesis: page 60, 61, 62, 65.)
REFERENCES 174
Chen, R. Y., Hung, T.-S., and Hung, C.-Y. (2005). A CMOS infrared wireless
optical receiver front-end with a variable-gain fully-differential transimpedance
amplifier. IEEE Transactions on Consumer Electronics, 51(2):424–429. (In this
thesis: page 86.)
Chen, T., Kaiser, N. K., Beu, S. C., Hendrickson, C. L., and Marshall, A. G.
(2012). Excitation and detection with the same electrodes for improved FT-ICR
MS performance. In 60th ASMS Conference on Mass Spectrometry and Allied
Topics, Vancouver, BC, Canada. 20-24 May. (In this thesis: page 89, 164.)
Chien, F.-T. and Chan, Y.-J. (1999). Bandwidth enhancement of tran-
simpedance amplifier by a capacitive-peaking design. IEEE Journal of Solid-
State Circuits, 34(8):1167–1170. (In this thesis: page 116.)
Comisarow, M. B. (1978). Signal modeling for ion cyclotron resonance. Journal
of Chemical Physics, 69(9):4097–4104. (In this thesis: page 22, 23, 86, 107.)
Comisarow, M. B. (1993). Fundamental aspects of FT-ICR and applications to
chemistry. Hyperfine Interactions, 81(1-4):171–178. (In this thesis: page 17.)
Comisarow, M. B., Grassi, V., and Parisod, G. (1978). Fourier transform ion
cyclotron double resonance. Chemical Physics Letters, 57(3):413–416. (In this
thesis: page 31.)
Comisarow, M. B. and Marshall, A. G. (1974a). Fourier transform ion cyclotron
resonance spectroscopy. Chemical Physics Letters, 25(2):282–283. (In this thesis:
page 11, 17.)
REFERENCES 175
Comisarow, M. B. and Marshall, A. G. (1974b). Frequency-sweep Fourier trans-
form ion cyclotron resonance spectroscopy. Chemical Physics Letters, 26(4):489–
490. (In this thesis: page 17.)
Comisarow, M. B. and Marshall, A. G. (1976). Theory of Fourier transform
ion cyclotron resonance mass spectroscopy .I. Fundamental equations and low-
pressure line shape. Journal of Chemical Physics, 64(1):110–119. (In this thesis:
page 17.)
Cui, W., Rohrs, H. W., and Gross, M. L. (2011). Top-down mass spectrometry:
recent developments, applications and perspectives. Analyst, 136(19):3854–3864.
(In this thesis: page 16.)
Dawson, P. H. (1976). Quadrupole Mass Spectrometry and its Applications,
chapter 1, pages 1–7. American Institute of Physics, Woodbury, New York. (In
this thesis: page 48.)
de Hoffmann, E. and Stroobant, V. (2007). Mass Spectrometry: Principles and
Applications, pages 88–95. John Wiley & Sons. (In this thesis: page 53, 54.)
Deliyannis, T. (1968). High-Q factor circuit with reduced sensitivity. Electronics
Letters, 4(26):577–579. (In this thesis: page 142.)
Deshpande, N. P. (2008). Electron Devices & Circuits - Principles & Applica-
tions, page 221. McGraw-Hill Education (India) Pvt Limited. (In this thesis:
page 117.)
REFERENCES 176
Douglas, D. J. (2009). Linear quadrupoles in mass spectrometry. Mass Spec-
trometry Reviews, 28(6):937–960. (In this thesis: page 48, 49.)
Douglas, D. J., Frank, A. J., and Mao, D. (2005). Linear ion traps in mass
spectrometry. Mass Spectrometry Reviews, 24(1):1–29. (In this thesis: page 49.)
El-Diwany, M. H., Roulston, D. J., and Chamberlain, S. G. (1981). Design
of low-noise bipolar transimpedance preamplifiers for optical receivers. IEE
Proceedings G, Electronic Circuits and Systems, 128(6):299–306. (In this thesis:
page 86.)
Fabris, L. and Manfredi, P. (2002). Optimization of front-end design in imaging
and spectrometry applications with room temperature semiconductor detectors.
IEEE Transactions on Nuclear Science, 49(4):1978–1985. (In this thesis: page
92, 93.)
Fish, F. H. and Katz, R. S. (1977). Amplifier for fiber optics application. U.S.
Patent, 4,029,976. (In this thesis: page 114.)
Flora, J. W., Hannis, J. C., and Muddiman, D. C. (2001). High-mass accuracy of
product ions produced by SORI-CID using a dual electrospray ionization source
coupled with FTICR mass spectrometry. Analytical Chemistry, 73(6):1247–
1251. (In this thesis: page 31.)
Franceschi, P., Penasa, L., Ascenzi, D., Bassi, D., Scotoni, M., and Tosi, P.
(2007). A simple and cost-effective high voltage radio frequency driver for multi-
REFERENCES 177
polar ion guides. International Journal of Mass Spectrometry, 265(23):224–229.
(In this thesis: page 64.)
Friend, J. J. (1970). A single op amp biquadratic filter section. IEEE Inter-
national Symposium on Circuit Theory, pages 189–90. (In this thesis: page
142.)
Friend, J. J., Harris, C. A., and Hilberman, D. (1975). STAR: An active bi-
quadratic filter section. IEEE Transactions on Circuits and Systems, 22(2):115–
121. (In this thesis: page 142.)
Gauthier, J. W., Trautman, T. R., and Jacobson, D. B. (1991). Sustained off-
resonance irradiation for collision-activated dissociation involving Fourier trans-
form mass spectrometry. Collision-activated dissociation technique that emu-
lates infrared multiphoton dissociation. Analytica Chimica Acta, 246(1):211–
225. (In this thesis: page 31.)
Gerlich, D. (1992). Inhomogeneous RF fields: A versatile tool for the study
of processes with slow ions. Advances in Chemical Physics, 82:1–176. (In this
thesis: page 48.)
Green, R., Higgins, M., Joshi, H., and Leeson, M. (2008a). Bandwidth ex-
tension for optical wireless receiver-amplifiers. 10th Anniversary International
Conference on Transparent Optical Networks, 4:201–204. (In this thesis: page
86, 116.)
REFERENCES 178
Green, R. J. (1986). Experimental performance of a bandwidth enhancement
technique for photodetectors. Electronics Letters, 22(3):153–155. (In this thesis:
page 116.)
Green, R. J., Joshi, H., Higgins, M. D., and Leeson, M. S. (2008b). Recent
developments in indoor optical wireless systems. IET Communications, 2(1):3–
10. (In this thesis: page 134.)
Green, R. J. and McNeill, M. G. (1989). Bootstrap transimpedance ampli-
fier: a new configuration. IEE Proceedings G, Circuits, Devices and Systems,
136(2):57–61. (In this thesis: page 86.)
Headley, J. V., Barrow, M. P., Peru, K. M., Fahlman, B., Frank, R. A., Bicker-
ton, G., McMaster, M. E., Parrott, J., and Hewitt, L. M. (2011). Preliminary
fingerprinting of Athabasca oil sands polar organics in environmental samples
using electrospray ionization Fourier transform ion cyclotron resonance mass
spectrometry. Rapid Communications in Mass Spectrometry, 25(13):1899–1909.
(In this thesis: page 16.)
Hipple, J. A. and Shepherd, M. (1949). Mass spectrometry. Analytical Chem-
istry, 21(1):32–36. (In this thesis: page 13.)
Hiroki, S., Abe, T., and Murakami, Y. (1991). Development of a quadrupole
mass spectrometer using the second stable zone in Mathieu’s stability diagram.
Review of Scientific Instruments, 62(9):2121–2124. (In this thesis: page 54.)
REFERENCES 179
Horowitz, P. and Hill, W. (1989). The Art of Electronics. Cambridge University
Press, Cambridge, UK. (In this thesis: page 149.)
Hsu, C. S., Hendrickson, C. L., Rodgers, R. P., McKenna, A. M., and Marshall,
A. G. (2011). Petroleomics: advanced molecular probe for petroleum heavy
ends. Journal of Mass Spectrometry, 46(4):337–343. (In this thesis: page 16.)
Hullett, J. L. and Moustakas, S. (1981). Optimum transimpedance broadband
optical preamplifier design. Optical and Quantum Electronics, 13(1):65–69. (In
this thesis: page 86.)
Ivanov, B. I., Trgala, M., Grajcar, M., Il’ichev, E., and Meyer, H.-G. (2011).
Cryogenic ultra-low-noise SiGe transistor amplifier. Review of Scientific Instru-
ments, 82(10):104705. (In this thesis: page 43.)
Jau, Y.-Y., Benito, F. M., Partner, H., and Schwindt, P. D. D. (2011). Low
power high-performance radio frequency oscillator for driving ion traps. Review
of Scientific Instruments, 82(2):023118. (In this thesis: page 63, 64, 65.)
Jefferts, S. R. and Walls, F. L. (1989). A very low-noise FET input amplifier.
Review of Scientific Instruments, 60(6):1194–1196. (In this thesis: page 41.)
Johnson, J. B. (1925). The Schottky effect in low frequency circuits. Physical
Review, 26(1):71–85. (In this thesis: page 37.)
Johnson, J. B. (1928). Thermal agitation of electricity in conductors. Physical
Review, 32(1):97. (In this thesis: page 37.)
REFERENCES 180
Jones, R. M. and Anderson, S. L. (2000). Simplified radio-frequency generator
for driving ion guides, traps, and other capacitive loads. Review of Scientific
Instruments, 71(11):4335–4337. (In this thesis: page 57, 58, 65, 66.)
Jones, R. M., Gerlich, D., and Anderson, S. L. (1997). Simple radio-frequency
power source for ion guides and ion traps. Review of Scientific Instruments,
68(9):3357–3362. (In this thesis: page 57, 65, 66.)
Kaiser, N. K., Quinn, J. P., Blakney, G. T., Hendrickson, C. L., and Marshall,
A. G. (2011a). A novel 9.4 Tesla FTICR mass spectrometer with improved
sensitivity, mass resolution, and mass range. Journal of the American Society
for Mass Spectrometry, 22(8):1343–1351. (In this thesis: page 2, 46, 85.)
Kaiser, N. K., Savory, J. J., McKenna, A. M., Quinn, J. P., Hendrickson, C. L.,
and Marshall, A. G. (2011b). Electrically compensated Fourier transform ion
cyclotron resonance cell for complex mixture mass analysis. Analytical Chem-
istry, 83(17):6907–6910. (In this thesis: page 33.)
Kaur, P. and O’Connor, P. B. (2004). Use of statistical methods for estima-
tion of total number of charges in a mass spectrometry experiment. Analytical
Chemistry, 76(10):2756–2762. (In this thesis: page 41.)
Kim, S., Choi, M. C., Hur, M., Kim, H. S., Yoo, J. S., Hendrickson, C. L., and
Marshall, A. G. (2008). The ’hybrid cell’: A new compensated infinity cell for
larger radius ion excitation in Fourier transform ion cyclotron resonance mass
REFERENCES 181
spectrometry. Rapid Communications in Mass Spectrometry, 22(9):1423–1429.
(In this thesis: page 33.)
Kroto, H. (1997). Symmetry, space, stars and C60. Reviews of Modern Physics,
69(3):703–722. (In this thesis: page 13.)
Lawrence, E. O. (1934). Method and apparatus for the acceleration of ions.
U.S. Patent, 1,948,384. (In this thesis: page 11.)
Lawrence, E. O. and Livingston, M. S. (1932). The production of high speed
light ions without the use of high voltages. Physical Review, 40(1):19–35. (In
this thesis: page 11, 12, 17.)
Letzter, S. and Webster, N. (1970). Noise in amplifiers. IEEE Spectrum, 7(8):67–
75. (In this thesis: page 37, 38, 39, 40, 131.)
Li, H., Lin, T.-Y., Van Orden, S. L., Zhao, Y., Barrow, M. P., Pizarro, A. M.,
Qi, Y., Sadler, P. J., and O’Connor, P. B. (2011a). Use of top-down and bottom-
up Fourier transform ion cyclotron resonance mass spectrometry for mapping
calmodulin sites modified by platinum anticancer drugs. Analytical Chemistry,
83(24):9507–9515. (In this thesis: page 16.)
Li, H., Zhao, Y., Phillips, H. I. A., Qi, Y., Lin, T.-Y., Sadler, P. J., and OConnor,
P. B. (2011b). Mass spectrometry evidence for cisplatin as a protein cross-linking
reagent. Analytical Chemistry, 83(13):5369–5376. (In this thesis: page 16, 30,
31.)
REFERENCES 182
Li, Y., McIver, R. T., and Hunter, R. L. (1994). High-accuracy molecular mass
determination for peptides and proteins by Fourier transform mass spectrome-
try. Analytical Chemistry, 66(13):2077–2083. (In this thesis: page 25.)
Ligon, Woodfin V., J. (1979). Molecular analysis by mass spectrometry. Science,
205(4402):151–159. (In this thesis: page 14.)
Limbach, P. A., Grosshans, P. B., and Marshall, A. G. (1993). Experimental
determination of the number of trapped ions, detection limit, and dynamic range
in Fourier transform ion cyclotron resonance mass spectrometry. Analytical
Chemistry, 65(2):135–140. (In this thesis: page 41.)
Lin, C., Cournoyer, J. J., and O’Connor, P. B. (2006). Use of a double reso-
nance electron capture dissociation experiment to probe fragment intermediate
lifetimes. Journal of the American Society for Mass Spectrometry, 17(11):1605–
1615. (In this thesis: page 31.)
Lin, T.-Y., Green, R. J., and O’Connor, P. B. (2011). A gain and bandwidth
enhanced transimpedance preamplifier for Fourier-transform ion cyclotron reso-
nance mass spectrometry. Review of Scientific Instruments, 82(12):124101. (In
this thesis: page 33, 80, 83.)
Lin, T.-Y., Green, R. J., and O’Connor, P. B. (2012). A low noise single-
transistor transimpedance preamplifier for Fourier-transform mass spectrometry
using a T feedback network. Review of Scientific Instruments, 83(9):094102. (In
this thesis: page 33, 77, 109.)
REFERENCES 183
Little, D. P., Speir, J. P., Senko, M. W., O’Connor, P. B., and McLafferty,
F. W. (1994). Infrared multiphoton dissociation of large multiply charged ions
for biomolecule sequencing. Analytical Chemistry, 66(18):2809–2815. (In this
thesis: page 31.)
Lourette, N., Smallwood, H., Wu, S., Robinson, E. W., Squier, T. C., Smith,
R. D., and Pasa-Tolic, L. (2010). A top-down LC-FTICR MS-based strategy
for characterizing oxidized calmodulin in activated macrophages. Journal of the
American Society for Mass Spectrometry, 21(6):930–939. (In this thesis: page
16.)
Ly, T. and Julian, R. R. (2009). Ultraviolet photodissociation: Developments
towards applications for mass-spectrometry-based proteomics. Angewandte
Chemie International Edition, 48(39):7130–7137. (In this thesis: page 31.)
Makarov, A. (2000). Electrostatic axially harmonic orbital trapping: A high-
performance technique of mass analysis. Analytical Chemistry, 72(6):1156–1162.
(In this thesis: page 133.)
March, R. E. (1997). An introduction to quadrupole ion trap mass spectrometry.
Journal of Mass Spectrometry, 32(4):351–369. (In this thesis: page 49.)
March, R. E. and Todd, J. F. J. (2005). Quadrupole Ion Trap Mass Spectrometry.
John Wiley & Sons, Hoboken, NJ, 2nd edition. (In this thesis: page 49, 51.)
March, R. E. and Todd, J. F. J. (2009). An Appreciation and Historical Survey
of Mass Spectrometry, volume IV of Practical Aspects of Trapped Ion Mass
REFERENCES 184
Spectrometry: Theory and Instrumentation, chapter 1, pages 3–168. Taylor &
Francis, Boca Raton, FL. (In this thesis: page 49.)
Marshall, A. G., Hendrickson, C. L., and Jackson, G. S. (1998). Fourier trans-
form ion cyclotron resonance mass spectrometry: A primer. Mass Spectrometry
Reviews, 17(1):1–35. (In this thesis: page 21, 22, 27, 33, 34.)
Marshall, A. G. and Verdun, F. R. (1990). Fourier Transforms in NMR, Op-
tical, and Mass Spectrometry: A User’s Handbook, pages 144–147. Elsevier,
Amsterdam, The Netherlands. (In this thesis: page 24.)
Mathur, R., Knepper, R. W., and O’Connor, P. B. (2007). A low-noise, wide-
band preamplifier for a Fourier-transform ion cyclotron resonance mass spec-
trometer. Journal of the American Society for Mass Spectrometry, 18(12):2233–
2241. (In this thesis: page 41, 43, 44, 80.)
Mathur, R., Knepper, R. W., and O’Connor, P. B. (2008). A low-noise broad-
band cryogenic preamplifier operated in a high-field superconducting magnet.
IEEE Transactions on Applied Superconductivity, 18(4):1781–1789. (In this the-
sis: page 2, 43, 45, 46.)
Mathur, R. and O’Connor, P. B. (2006). Design and implementation of a high
power RF oscillator on a printed circuit board for multipole ion guides. Review
of Scientific Instruments, 77(11):114101. (In this thesis: page 4, 58, 59, 65, 66,
67, 137, 142, 153.)
REFERENCES 185
Mathur, R. and O’Connor, P. B. (2009). Artifacts in Fourier transform mass
spectrometry. Rapid Communications in Mass Spectrometry, 23(4):523–529. (In
this thesis: page 22, 23, 24, 99.)
McDonnell, L. A., Corthals, G. L., Willems, S. M., van Remoortere, A., van
Zeijl, R. J. M., and Deelder, A. M. (2010). Peptide and protein imaging mass
spectrometry in cancer research. Journal of Proteomics, 73(10):1921–1944. (In
this thesis: page 16.)
Misharin, A. S., Zubarev, R. A., and Doroshenko, V. M. (2010). Fourier trans-
form ion cyclotron resonance mass spectrometer with coaxial multi-electrode cell
(’O-trap’): first experimental demonstration. Rapid Communications in Mass
Spectrometry, 24(14):1931–1940. (In this thesis: page 33.)
Mohan, S. S., Hershenson, M. D. M., Boyd, S. P., and Lee, T. H. (2000).
Bandwidth extension in CMOS with optimized on-chip inductors. IEEE Journal
of Solid-State Circuits, 35(3):346–355. (In this thesis: page 116.)
Nikolaev, E. N., Boldin, I. A., Jertz, R., and Baykut, G. (2011). Initial ex-
perimental characterization of a new ultra-high resolution FTICR cell with dy-
namic harmonization. Journal of the American Society for Mass Spectrometry,
22(7):1125–1133. (In this thesis: page 33.)
Nyquist, H. (1928). Thermal agitation of electric charge in conductors. Physical
Review, 32(1):110. (In this thesis: page 37.)
REFERENCES 186
O’Connor, P. B. (2002). Considerations for design of a Fourier transform mass
spectrometer in the 4.2 K cold bore of a superconducting magnet. Rapid Com-
munications in Mass Spectrometry, 16(12):1160–1167. (In this thesis: page 43.)
O’Connor, P. B., Costello, C. E., and Earle, W. E. (2002). A high voltage RF
oscillator for driving multipole ion guides. Journal of the American Society for
Mass Spectrometry, 13(12):1370–1375. (In this thesis: page 50, 57, 65, 66, 67,
68.)
Painter, T. A., Markiewicz, W. D., Miller, J. R., Bole, S. T., Dixon, I. R.,
Cantrell, K. R., Kenney, S. J., Trowell, A. J., Dong Lak, K., Byoung Seob, L.,
Yeon Suk, C., Hyun Sik, K., Hendrickson, C. L., and Marshall, A. G. (2006).
Requirements and conceptual superconducting magnet design for a 21 T Fourier
transform ion cyclotron resonance mass spectrometer. IEEE Transactions on
Applied Superconductivity, 16(2):945–948. (In this thesis: page 89.)
Paul, W. (1990). Electromagnetic traps for charged and neutral particles. Re-
views of Modern Physics, 62(3):531–540. (In this thesis: page 48, 49.)
Perez Hurtado, P. and O’Connor, P. B. (2012). Deamidation of collagen. Ana-
lytical Chemistry, 84(6):3017–3025. (In this thesis: page 16.)
Robbins, M. D., Yoon, O. K., Zuleta, I., Barbula, G. K., and Zare, R. N. (2008).
Computer-controlled, variable-frequency power supply for driving multipole ion
guides. Review of Scientific Instruments, 79(3):034702. (In this thesis: page 63,
65.)
REFERENCES 187
Savory, J. J., Kaiser, N. K., McKenna, A. M., Xian, F., Blakney, G. T., Rodgers,
R. P., Hendrickson, C. L., and Marshall, A. G. (2011). Parts-per-billion Fourier
transform ion cyclotron resonance mass measurement accuracy with a walking
calibration equation. Analytical Chemistry, 83(5):1732–1736. (In this thesis:
page 30.)
Schmid, D. G., Grosche, P., Bandel, H., and Jung, G. (2000). FTICR-mass spec-
trometry for high-resolution analysis in combinatorial chemistry. Biotechnology
and Bioengineering, 71(2):149–161. (In this thesis: page 18, 20.)
Schottky, W. (1918). Uber spontane Stromschwankungen in verschiedenen Elek-
trizitatsleitern. Annalen Der Physik, 362(23):541–567. (In this thesis: page 37.)
Schottky, W. (1926). Small-shot effect and flicker effect. Physical Review,
28(1):74–103. (In this thesis: page 37.)
Smith, D., Kharchenko, A., Konijnenburg, M., Klinkert, I., Pasa-Tolic, L., and
Heeren, R. (2012). Advanced mass calibration and visualization for FT-ICR
mass spectrometry imaging. Journal of the American Society for Mass Spec-
trometry, 23(11):1865–1872. (In this thesis: page 30.)
Smith, D. F., Robinson, E. W., Tolmachev, A. V., Heeren, R. M. A., and Pasa-
Tolic, L. (2011). C60 secondary ion Fourier transform ion cyclotron resonance
mass spectrometry. Analytical Chemistry, 83(24):9552–9556. (In this thesis:
page 16.)
REFERENCES 188
Taban, I., Altelaar, A., van der Burgt, Y., McDonnell, L., Heeren, R., Fuchser,
J., and Baykut, G. (2007). Imaging of peptides in the rat brain using MALDI-
FTICR mass spectrometry. Journal of the American Society for Mass Spec-
trometry, 18(1):145–151. (In this thesis: page 16.)
Teloy, E. and Gerlich, D. (1974). Integral cross sections for ion-molecule reac-
tions. I. The guided beam technique. Chemical Physics, 4(3):417–427. (In this
thesis: page 48.)
Tolmachev, A. V., Robinson, E. W., Wu, S., Kang, H., Lourette, N. M., Pasa-
Tolic, L., and Smith, R. D. (2008). Trapped-ion cell with improved DC potential
harmonicity for FT-ICR MS. Journal of the American Society for Mass Spec-
trometry, 19(4):586–597. (In this thesis: page 33.)
Tsukakoshi, O., Hayashi, T., Yamamuro, K., and Nakamura, M. (2000). Devel-
opment of a high precision quadrupole mass filter using the zero-method control
circuit. Review of Scientific Instruments, 71(3):1332–1336. (In this thesis: page
64.)
Watson, J. T. and Sparkman, O. D. (2008). Introduction to Mass Spectrometry:
Instrumentation, Applications, and Strategies for Data Interpretation, pages 9–
22, 26. John Wiley & Sons, Chichester, UK, 4th edition. (In this thesis: page
9, 10, 11, 13, 27, 28.)
Weisbrod, C. R., Kaiser, N. K., Skulason, G. E., and Bruce, J. E. (2010). Excite-
coupled trapping ring electrode cell (eTREC): Radial trapping field control, lin-
REFERENCES 189
earized excitation, and improved detection. Analytical Chemistry, 82(14):6281–
6286. (In this thesis: page 33.)
Wu, S., Kam, K., Pommerenke, D., Cornelius, B., Shi, H., Herndon, M., and
Fan, J. (2010). Investigation of noise coupling from switching power supply to
signal nets. In IEEE International Symposium on Electromagnetic Compatibil-
ity, Fort Lauderdale, FL. 25-30 July. (In this thesis: page 80.)
Xian, F., Hendrickson, C. L., and Marshall, A. G. (2012). High resolution mass
spectrometry. Analytical Chemistry, 84(2):708–719. (In this thesis: page 89.)
Zhang, L. K., Rempel, D., Pramanik, B. N., and Gross, M. L. (2005). Accurate
mass measurements by Fourier transform mass spectrometry. Mass Spectrome-
try Reviews, 24(2):286–309. (In this thesis: page 25.)
Zubarev, R. A., Haselmann, K. F., Budnik, B., Kjeldsen, F., and Jensen, F.
(2002). Towards an understanding of the mechanism of electron-capture disso-
ciation: a historical perspective and modern ideas. European Journal of Mass
Spectrometry, 8(5):337–349. (In this thesis: page 31.)
Zubarev, R. A., Kelleher, N. L., and McLafferty, F. W. (1998). Electron capture
dissociation of multiply charged protein cations. A nonergodic process. Journal
of the American Chemical Society, 120(13):3265–3266. (In this thesis: page 31.)
APPENDIX A
Appendix
Datasheets of the key components used in this work are attached in the follow-
ing appendices. The datasheet of the JFET BF862, used in both preamplifiers
reported in Chapter 5 and 6, is attached in Appendix A.1.
The op amp AD8099 datasheet is attached in Appendix A.2. Such an op amp
is the main amplification stage in the preamplifier described in Chapter 5. In
Chapter 6, this op amp is also used to test the T feedback network.
Appendix A.3 includes the datasheet of the op amp LT6205, its dual version,
LT6206, and its quad version, LT6207. This LT6205/06/07 family of op amps is
used in the proposed power supply circuit for a quadrupole mass filter, reported
in Chapter 7.
The datasheet of the bipolar power transistor MJE18008 is included in Ap-
pendix A.4. Such a power transistor is the main component of the power amplifier
used in Chapter 7.
190
A. JFET BF862 191
A.1 Datasheet: JFET BF862
The datasheet of the JFET BF862 was downloaded from http://www.nxp.com/
pip/BF862.html, accessed 30 August 2012.
DATA SHEET
Product specificationSupersedes data of 1999 Jun 29
2000 Jan 05
DISCRETE SEMICONDUCTORS
BF862N-channel junction FET
ook, halfpage
M3D088
A. JFET BF862 192
2000 Jan 05 2
Philips Semiconductors Product specification
N-channel junction FET BF862
FEATURES
• High transition frequency for excellent sensitivity inAM car radios
• High transfer admittance.
APPLICATIONS
• Pre-amplifiers in AM car radios.
DESCRIPTION
Silicon N-channel symmetrical junction field-effecttransistor in a SOT23 package. Drain and source areinterchangeable.
PINNING SOT23
PIN DESCRIPTION
1 source
2 drain
3 gate
handbook, halfpage
s
dg
2 1
3MAM036Top view
Fig.1 Simplified outline and symbol.
Marking code : 2Ap.
QUICK REFERENCE DATA
SYMBOL PARAMETER CONDITIONS MIN. TYP. MAX. UNIT
VDS drain-source voltage − − 20 V
VGSoff gate-source cut-off voltage −0.3 −0.8 −1.2 V
IDSS drain-source current 10 − 25 mA
Ptot total power dissipation Ts ≤ 90 °C − − 300 mW
yfs transfer admittance 35 45 − mS
Tj junction temperature − − 150 °C
CAUTION
This product is supplied in anti-static packing to prevent damage caused by electrostatic discharge during transportand handling. For further information, refer to Philips specs.: SNW-EQ-608, SNW-FQ-302A and SNW-FQ-302B.
A. JFET BF862 193
2000 Jan 05 3
Philips Semiconductors Product specification
N-channel junction FET BF862
LIMITING VALUESIn accordance with the Absolute Maximum Rating System (IEC 134).
Note
1. Main heat transfer is via the gate lead.
THERMAL CHARACTERISTICS
Note
1. Soldering point of the gate lead.
SYMBOL PARAMETER CONDITIONS MIN. MAX. UNIT
VDS drain-source voltage − 20 V
VDG drain-gate voltage − 20 V
VGS gate-source voltage − −20 V
IDS drain-source current − 40 mA
IG forward gate current − 10 mA
Ptot total power dissipation Ts ≤ 90 °C; note 1 − 300 mW
Tstg storage temperature −65 +150 °CTj junction temperature − 150 °C
SYMBOL PARAMETER CONDITIONS VALUE UNIT
Rth j-s thermal resistance from junction to solderingpoint
note 1 200 K/W
handbook, halfpage
0 40Ts (°C)
Ptot(mW)
80 160
400
300
100
0
200
120
MCD808
Fig.2 Power derating curve.
A. JFET BF862 194
2000 Jan 05 4
Philips Semiconductors Product specification
N-channel junction FET BF862
STATIC CHARACTERISTICSTj = 25 °C; unless otherwise specified.
DYNAMIC CHARACTERISTICSCommon source; Tamb = 25 °C; VGS = 0; VDS = 8 V; unless otherwise specified.
SYMBOL PARAMETER CONDITIONS MIN. TYP. MAX. UNIT
V(BR)GSS gate-source breakdown voltage IGS = −1 µA; VDS = 0 −20 − − V
VGS gate-source forward voltage VDS = 0; IG = 1 mA − − 1 V
VGSoff gate-source cut-off voltage VDS = 8 V; ID = 1 µA −0.3 −0.8 −1.2 V
IGSS reverse gate current VGS = −15 V; VDS = 0 − − −1 nA
IDSS drain-source current VGS = 0; VDS = 8 V 10 − 25 mA
SYMBOL PARAMETER CONDITIONS MIN. TYP. MAX. UNIT
yfs common source forward transferadmittance
Tj = 25 °C 35 45 − mS
gos common source output conductance Tj = 25 °C − 180 400 µS
Ciss input capacitance f = 1 MHz − 10 − pF
Crss reverse transfer capacitance f = 1 MHz − 1.9 − pF
en equivalent noise input voltage f = 100 kHz − 0.8 − nV/√Hz
fT transition frequency − 715 − MHz
A. JFET BF862 195
2000 Jan 05 5
Philips Semiconductors Product specification
N-channel junction FET BF862
handbook, halfpage
0 −0.5 −1VGSoff (V)
IDSS(mA)
−1.5
40
30
10
0
20
MCD809
Fig.3 Drain saturation current as a function ofgate-source cut-off voltage; typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage
0
300
200
100
010 20 30
MCD810
IDSS (mA)
gos(µS)
Fig.4 Common-source output conductance as afunction of drain saturation current;typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage
0 10 20IDSS (mA)
yfs(mS)
30
60
50
30
20
40
MCD811
Fig.5 Forward transfer admittance as a functionof drain saturation current; typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage
0
60
40
20
010 20 30
MCD812
ID (mA)
yfs(mS)
Fig.6 Forward transfer admittance as a functionof drain current; typical values.
VDS = 8 V; Tj = 25 °C.
A. JFET BF862 196
2000 Jan 05 6
Philips Semiconductors Product specification
N-channel junction FET BF862
handbook, halfpage
−1
30
20
10
0−0.8 0
typ
min
−0.6 −0.4 −0.2
MCD813
VGS (V)
ID(mA)
max
Fig.7 Drain current as a function of gate-sourcevoltage; typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage20
0 4 120
8
4
8
12
16
MCD814
VDS (V)
ID(mA)
VGS = 0 V
−0.1 V
−0.2 V
−0.3 V
−0.4 V
−0.5 V
Fig.8 Drain current as a function of drain-sourcevoltage; typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage
250 5 10 15VDG (V)
IG(nA)
20
−104
−102
−1
−10−2
−10−4
MCD815
10 mA
1 mA
0.1 mA
IGSS
ID = 20 mA
Fig.9 Gate current as a function of drain-gatevoltage; typical values.
VDS = 8 V; Tj = 25 °C.
handbook, halfpage
−8 −6 −4 0
Cis
Crs
VGS (V)
C(pF)
12
0
4
8
−2
MCD816
Fig.10 Input and reverse transfer capacitance asfunctions of gate-source voltage; typicalvalues.
VDS = 8 V; f = 1 MHz; Tj = 25 °C.
A. JFET BF862 197
2000 Jan 05 7
Philips Semiconductors Product specification
N-channel junction FET BF862
handbook, halfpage102
10
1
10−1
10−2
MCD817
10−1 1 10f (MHz)
yis(mS)
102 103
gis
bis
Fig.11 Common-source input admittance as afunction of frequency; typical values.
VDS = 8 V; VGS = 0; Tamb = 25 °C.
handbook, halfpage10
1
10−1
MCD818
10−1
−102
−101 10
f (MHz)
yrs(mS)
ϕrs(deg)
102
−103
103
ϕrs
yrs
Fig.12 Common-source reverse admittance as afunction of frequency; typical values.
VDS = 5 V; VG2 = 4 V.
ID = 15 mA; Tamb = 25 °C.
handbook, halfpage102
10
1
MCD819
10−1
−10
−11 10
f (MHz)
yfs(mS)
ϕfs(deg)
102
−102
103
yfs
ϕfs
Fig.13 Common-source forward transferadmittance as a function of frequency;typical values.
VDS = 8 V; VGS = 0; Tamb = 25 °C.
handbook, halfpage102
10
1
10−1
MCD820
10−1 1 10f (MHz)
yos(mS)
102 103
gos
bos
Fig.14 Common-source output admittance as afunction of frequency; typical values.
VDS = 8 V; VGS = 0; Tamb = 25 °C.
A. JFET BF862 198
2000 Jan 05 8
Philips Semiconductors Product specification
N-channel junction FET BF862
PACKAGE OUTLINE
UNITA1
max.bp c D E e1 HE Lp Q wv
REFERENCESOUTLINEVERSION
EUROPEANPROJECTION ISSUE DATE
97-02-2899-09-13
IEC JEDEC EIAJ
mm 0.1 0.480.38
0.150.09
3.02.8
1.41.2
0.95
e
1.9 2.52.1
0.550.45 0.10.2
DIMENSIONS (mm are the original dimensions)
0.450.15
SOT23 TO-236AB
bp
D
e1
e
A
A1
Lp
Q
detail X
HE
E
w M
v M A
B
AB
0 1 2 mm
scale
A
1.10.9
c
X
1 2
3
Plastic surface mounted package; 3 leads SOT23
A. JFET BF862 199
2000 Jan 05 9
Philips Semiconductors Product specification
N-channel junction FET BF862
DEFINITIONS
LIFE SUPPORT APPLICATIONS
These products are not designed for use in life support appliances, devices, or systems where malfunction of theseproducts can reasonably be expected to result in personal injury. Philips customers using or selling these products foruse in such applications do so at their own risk and agree to fully indemnify Philips for any damages resulting from suchimproper use or sale.
Data sheet status
Objective specification This data sheet contains target or goal specifications for product development.
Preliminary specification This data sheet contains preliminary data; supplementary data may be published later.
Product specification This data sheet contains final product specifications.
Limiting values
Limiting values given are in accordance with the Absolute Maximum Rating System (IEC 134). Stress above one ormore of the limiting values may cause permanent damage to the device. These are stress ratings only and operationof the device at these or at any other conditions above those given in the Characteristics sections of the specificationis not implied. Exposure to limiting values for extended periods may affect device reliability.
Application information
Where application information is given, it is advisory and does not form part of the specification.
A. JFET BF862 200
A. Operational Amplifier AD8099 201
A.2 Datasheet: Operational Amplifier AD8099
The datasheet of the operational amplifier AD8099 was downloaded from http://
www.analog.com/en/high-speed-op-amps/low-noise-low-distortion-amplifiers/ad8099/
products/product.html, accessed 30 August 2012.
Ultralow Distortion, High Speed 0.95 nV/√Hz Voltage Noise Op Amp
AD8099
Rev. B Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners.
One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.326.8703 © 2004 Analog Devices, Inc. All rights reserved.
FEATURES Ultralow noise: 0.95 nV/√Hz, 2.6 pA/√Hz
Ultralow distortion 2nd harmonic RL = 1 kΩ , G = +2
−92 dB @ 10 MHz 3rd harmonic RL = 1 kΩ , G = +2
−105 dB @ 10 MHz High speed GBWP: 3.8 GHz
–3 dB bandwidth: 700 MHz (G = +2) 550 MHz (G = +10)
Slew rate: 475 V/µs (G = +2) 1350 V/µs (G = +10)
New pinout Custom external compensation, gain range –1, +2 to +10 Supply current: 15 mA Offset voltage: 0.5 mV max Wide supply voltage range: 5 V to 12 V
APPLICATIONS Pre-amplifiers Receivers Instrumentation Filters IF and baseband amplifiers A-to-D drivers DAC buffers Optical electronics
CONNECTION DIAGRAMS
DISABLE 1
FEEDBACK 2
–IN 3
+IN 4
+VS8
VOUT7
CC6
–VS5
0451
1-0-
001
0451
1-0-
002
1FEEDBACK
2–IN
3+IN
–VS 4
DISABLE8
+VS7
VOUT6
CC5
Figure 1. 8-Lead CSP (CP-8) Figure 2. 8-Lead SOIC-ED (RD-8)
GENERAL DESCRIPTION
The AD8099 is an ultralow noise (0.95 nV/√Hz) and distortion (–92 dBc @10 MHz) voltage feedback op amp, the combination of which make it ideal for 16- and 18-bit systems. The AD8099 features a new, highly linear, low noise input stage that increases the full power bandwidth (FPBW) at low gains with high slew rates. ADI’s proprietary next generation XFCB process enables such high performance amplifiers with relatively low power.
The AD8099 features external compensation, which lets the user set the gain bandwidth product. External compensation allows gains from +2 to +10 with minimal trade-off in band-width. The AD8099 also features an extremely high slew rate of 1350 V/µs, giving the designer flexibility to use the entire dynamic range without trading off bandwidth or distortion. The AD8099 settles to 0.1% in 18 ns and recovers from overdrive in 50 ns.
The AD8099 drives 100 Ω loads at breakthrough performance levels with only 15 mA of supply current. With the wide supply voltage range (5 V to 12 V), low offset voltage (0.1 mV typ), wide bandwidth (700 MHz for G = +2), and a GBWP up to 3.8 GHz, the AD8099 is designed to work in a wide variety of applications.
The AD8099 is available in a 3 mm × 3 mm lead frame chip scale package (LFCSP) with a new pinout that is specifically optimized for high performance, high speed amplifiers. The new LFCSP package and pinout enable the breakthrough performance that previously was not achievable with amplifiers. The AD8099 is rated to work over the extended industrial temperature range, −40°C to +125°C.
0451
1-A-
013
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–40
–110
–100
–90
–80
–70
–60
–50
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +2VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 3 . Harmonic Distortion vs. Frequency and Gain (SOIC)
A. Operational Amplifier AD8099 202
AD8099
Rev. B | Page 2 of 28
TABLE OF CONTENTS Specifications..................................................................................... 3
Specifications with ±5 V Supply................................................. 3
Specifications with +5 V Supply................................................. 4
Absolute Maximum Ratings............................................................ 5
Maximum Power Dissipation ..................................................... 5
ESD Caution.................................................................................. 5
Typical Performance Characteristics ............................................. 6
Theory of Operation ...................................................................... 15
Applications..................................................................................... 16
Using the AD8099 ...................................................................... 16
Circuit Components................................................................... 16
Recommended Values ............................................................... 17
Circuit Configurations .............................................................. 17
Performance vs. Component values ........................................ 19
Total Output Noise Calculations and Design......................... 20
Input Bias Current and DC Offset ........................................... 21
DISABLE Pin and Input Bias Cancellation............................. 21
16-Bit ADC Driver..................................................................... 22
Circuit Considerations .............................................................. 23
Design Tools and Technical Support ....................................... 23
Outline Dimensions ....................................................................... 25
Ordering Guide............................................................................... 26
REVISION HISTORY
6/04—Data Sheet changed from REV. A to REV. B
Change to General Description ...................................................... 1 Changes to Maximum Power Dissipation section ...................... 5 Changes to Applications section .................................................. 16 Changes to Table 7.......................................................................... 24 Changes to Ordering Guide .......................................................... 26
1/04—Data Sheet changed from REV. 0 to REV. A
Inserted new Figure 3................................................................... 1 Changes to Specifications ............................................................ 3 Inserted new Figures 22 to 34 ..................................................... 8 Inserted new Figures 51 to 55 ................................................... 14 Changes to Theory of Operation section ................................ 16 Changes to Circuit Components section................................. 17 Changes to Table 4...................................................................... 18 Changes to Figure 60.................................................................. 18 Changes to Total Output Noise Calculations and Design section........................................................................ 21 Changes to Figure 60.................................................................. 22 Changes to Figure 62.................................................................. 23 Changes to 16-Bit ADC Driver section ................................... 23 Changes to Table 6...................................................................... 23 Additions to PCB Layout section ............................................. 23
11/03—Revision 0: Initial Version
A. Operational Amplifier AD8099 203
AD8099
Rev. B | Page 3 of 28
SPECIFICATIONS SPECIFICATIONS WITH ±5 V SUPPLY TA = 25°C, G = +2, RL = 1 kΩ to ground, unless otherwise noted. Refer to Figure 60 through Figure 66 for component values and gain configurations . Table 1. Parameter Conditions Min Typ Max Unit DYNAMIC PERFORMANCE
–3 dB Bandwidth G = +5, VOUT = 0.2 V p-p 450 510 MHz G = +5, VOUT = 2 V p-p 205 235 MHz
Bandwidth for 0.1 dB Flatness (SOIC/CSP) G = +2, VOUT = 0.2 V p-p 34/25 MHz Slew Rate G = +10, VOUT = 6 V Step 1120 1350 V/µs G = +2, VOUT = 2 V Step 435 470 V/µs Settling Time to 0.1% G = +2, VOUT = 2 V Step 18 ns
NOISE/DISTORTION PERFORMANCE Harmonic Distortion (dBc) HD2/HD3 fC = 500 kHz, VOUT = 2 V p-p, G = +10 –102/–111 dBc
fC = 10 MHz, VOUT = 2 V p-p, G = +10 –84/–92 dBc Input Voltage Noise f = 100 kHz 0.95 nV/√Hz
Input Current Noise f = 100 kHz, DISABLE pin floating 2.6 pA/√Hz
f = 100 kHz, DISABLE pin = +VS 5.2 pA/√Hz
DC PERFORMANCE Input Offset Voltage 0.1 0.5 mV Input Offset Voltage Drift 2.3 µV/°C Input Bias Current DISABLE pin floating –6 –13 µA
DISABLE pin = +VS –0.1 –2 µA
Input Bias Current Drift 3 nA/°C Input Bias Offset Current 0.06 1 µA Open-Loop Gain 82 85 dB
INPUT CHARACTERISTICS Input Resistance Differential mode 4 kΩ Common mode 10 MΩ Input Capacitance 2 pF Input Common-Mode Voltage Range –3.7 to +3.7 V Common-Mode Rejection Ratio VCM = ±2.5 V 98 105 dB
DISABLE PIN
DISABLE Input Voltage Output disabled <2.4 V
Turn-Off Time 50% of DISABLE to < 10% of final VOUT, VIN = 0.5 V, G = +2
105 ns
Turn-On Time 50% of DISABLE to < 10% of final VOUT, VIN = 0.5 V, G = +2
39 ns
Enable Pin Leakage Current DISABLE =+5 V 17 21 µA
DISABLE Pin Leakage Current DISABLE = –5 V 35 44 µA
OUTPUT CHARACTERISTICS Output Overdrive Recovery Time (Rise/Fall) VIN = -2.5 V to 2.5 V, G =+2 30/50 ns Output Voltage Swing RL = 100 Ω –3.4 to +3.5 –3.6 to +3.7 V RL = 1 kΩ –3.7 to +3.7 –3.8 to +3.8 V Short-Circuit Current Sinking and sourcing 131/178 mA Off Isolation f = 1 MHz, DISABLE = low –61 dB
POWER SUPPLY Operating Range ±5 ±6 V Quiescent Current 15 16 mA Quiescent Current (Disabled) DISABLE = Low 1.7 2 mA
Positive Power Supply Rejection Ratio +VS = 4 V to 6 V, –VS = –5 V (input referred) 85 91 dB Negative Power Supply Rejection Ratio +VS = 5 V, –VS = –6 V to –4 V (input referred) 86 94 dB
A. Operational Amplifier AD8099 204
AD8099
Rev. B | Page 4 of 28
SPECIFICATIONS WITH +5 V SUPPLY VS = 5 V @ TA = 25°C, G = +2, RL = 1 kΩ to midsupply, unless otherwise noted. Refer to Figure 60 through Figure 66 for component values and gain configurations . Table 2. Parameter Conditions Min Typ Max Unit DYNAMIC PERFORMANCE
–3 dB Bandwidth G = +5, VOUT = 0.2 V p-p 415 440 MHz G = +5, VOUT = 2 V p-p 165 210 MHz
Bandwidth for 0.1 dB Flatness (SOIC/CSP) G = +2, VOUT = 0.2 V p-p 33/23 MHz Slew Rate G = +10, VOUT = 2 V Step 630 715 V/µs G = +2, VOUT = 2 V Step 340 365 V/µs Settling Time to 0.1% G = +2, VOUT = 2 V Step 18 ns
NOISE/DISTORTION PERFORMANCE Harmonic Distortion (dBc) HD2/HD3 fC = 500 kHz, VOUT = 1 V p-p, G = +10 –82/–94 dBc
fC = 10 MHz, VOUT = 1 V p-p, G = +10 –80/–75 dBc Input Voltage Noise f = 100 kHz 0.95 nV/√Hz
Input Current Noise f = 100 kHz, DISABLE pin floating 2.6 pA/√Hz
f = 100 kHz, DISABLE pin = +VS 5.2 pA/√Hz
DC PERFORMANCE Input Offset Voltage 0.1 0.5 mV Input Offset Voltage Drift 2.5 µV/°C Input Bias Current DISABLE pin floating –6.2 –13 µA
DISABLE pin = +VS –0.2 –2 µA
Input Bias Offset Current 0.05 1 µA Input Bias Offset Current Drift 2.4 nA/°C Open-Loop Gain VOUT = 1 V to 4 V 76 81 dB
INPUT CHARACTERISTICS Input Resistance Differential mode 4 kΩ Common mode 10 MΩ Input Capacitance 2 pF Input Common-Mode Voltage Range 1.3 to 3.7 V Common-Mode Rejection Ratio VCM = 2 V to 3 V 88 105 dB
DISABLE PIN
DISABLE Input Voltage Output disabled <2.4 V
Turn-Off Time 50% of DISABLE to <10% of Final VOUT, VIN = 0.5 V, G = +2
105 ns
Turn-On Time 50% of DISABLE to <10% of Final VOUT, VIN = 0.5 V, G = +2
61 ns
Enable Pin Leakage Current DISABLE = 5 V 16 21 µA
DISABLE Pin Leakage Current DISABLE = 0 V 33 44 µA
OUTPUT CHARACTERISTICS Overdrive Recovery Time (Rise/Fall) VIN = 0 to 2.5 V, G = +2 50/70 ns Output Voltage Swing RL = 100 Ω 1.5 to 3.5 1.2 to 3.8 V RL = 1 kΩ 1.2 to 3.8 1.2 to 3.8 V Short-Circuit Current Sinking and Sourcing 60/80 mA Off Isolation f = 1 MHz, DISABLE = Low –61 dB
POWER SUPPLY Operating Range ±5 ±6 V Quiescent Current 14.5 15.4 mA Quiescent Current (Disabled) DISABLE = Low 1.4 1.7 mA
Positive Power Supply Rejection Ratio +VS = 4.5 V to 5.5 V, –VS = 0 V (input referred) 84 89 dB Negative Power Supply Rejection Ratio +VS =5 V, -VS= –0.5 V to +0.5 V (input referred) 84 90 dB
A. Operational Amplifier AD8099 205
AD8099
Rev. B | Page 5 of 28
ABSOLUTE MAXIMUM RATINGS Table 3. Parameter Rating Supply Voltage 12.6 V Power Dissipation See Figure 4 Differential Input Voltage ±1.8 V Differential Input Current ±10mA Storage Temperature –65°C to +125°C Operating Temperature Range –40°C to +125°C Lead Temperature Range (Soldering 10 sec) 300°C
Junction Temperature 150°C
Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability.
MAXIMUM POWER DISSIPATION The maximum safe power dissipation in the AD8099 package is limited by the associated rise in junction temperature (TJ) on the die. The plastic encapsulating the die will locally reach the junction temperature. At approximately 150°C, which is the glass transition temperature, the plastic will change its properties. Even temporarily exceeding this temperature limit may change the stresses that the package exerts on the die, permanently shifting the parametric performance of the AD8099. Exceeding a junction temperature of 150°C for an extended period can result in changes in silicon devices, potentially causing failure.
The still-air thermal properties of the package and PCB (θJA), the ambient temperature (TA), and the total power dissipated in the package (PD) determine the junction temperature of the die. The junction temperature can be calculated as
( )JADAJ θPTT ×+=
The power dissipated in the package (PD) is the sum of the quiescent power dissipation and the power dissipated in the package due to the load drive for all outputs. The quiescent power is the voltage between the supply pins (VS) times the quiescent current (IS). Assuming the load (RL) is referenced to midsupply, the total drive power is VS/2 × IOUT, some of which is dissipated in the package and some in the load (VOUT × IOUT).
The difference between the total drive power and the load power is the drive power dissipated in the package.
PD = Quiescent Power + (Total Drive Power – Load Power)
( )L
2OUT
L
OUTSSSD R
V–
RV
2V
IVP ⎟⎟⎠
⎞⎜⎜⎝
⎛×+×=
RMS output voltages should be considered. If RL is referenced to VS–, as in single-supply operation, then the total drive power is VS × IOUT. If the rms signal levels are indeterminate, consider the worst case, when VOUT = VS/4 for RL to midsupply:
( ) ( )L
SSSD R
/VIVP
24+×=
In single-supply operation with RL referenced to VS–, worst case is VOUT = VS/2.
Airflow will increase heat dissipation, effectively reducing θJA. Also, more metal directly in contact with the package leads from metal traces, through holes, ground, and power planes will reduce the θJA. Soldering the exposed paddle to the ground plane significantly reduces the overall thermal resistance of the package. Care must be taken to minimize parasitic capaci-tances at the input leads of high speed op amps, as discussed in the PCB Layout section.
Figure 4 shows the maximum safe power dissipation in the package versus the ambient temperature for the exposed paddle (e-pad) SOIC-8 (70°C/W), and CSP (70°C/W), packages on a JEDEC standard 4-layer board. θJA values are approximations.
0451
1-0-
115
AMBIENT TEMPERATURE (°C)120–40 –20 0 20 40 60 80 100
MA
XIM
UM
PO
WER
DIS
SIPA
TIO
N (W
atts
)
0.0
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
LFCSP AND SOIC
Figure 4. Maximum Power Dissipation
ESD CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although this product features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality.
A. Operational Amplifier AD8099 206
AD8099
Rev. B | Page 6 of 28
TYPICAL PERFORMANCE CHARACTERISTICS Default Conditions: VS = ±5 V, TA = 25°C, RL = 1 kΩ tied to ground unless otherwise noted. Refer to Figure 63 through Figure 66 for component values and gain configurations.
FREQUENCY (MHz)
NO
RM
ALI
ZED
CLO
SED
-LO
OP
GA
IN (d
B)
1–10–9–8–7–6–5–4–3–2–1
01234
10 100 1000
0451
1-0-
074
G = +20
G = +5
G = +2
G = +10
G = –1
VOUT = 0.2V p-pVS = ±5VRLOAD = 1kΩ
Figure 5. Small Signal Frequency Response for Various Gains (SOIC)
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
17
10
9
8
14
13
12
11
16
15
17
10 100 1000
0451
1-0-
076
G = +5VS = ±5VVOUT = 0.2V p-p
RL = 100Ω, SOIC
RL = 1kΩ, SOIC
RL = 100Ω, CSPRL = 1kΩ, CSP
Figure 6. Small Signal Frequency Response for Various Load Resistors
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
6
5
4
3
2
1
10
9
8
7
11
10001 10 100
G = +2VS = ±5VRL = 1kΩ
+125°C
+85°C
–40°C
+25°C
0451
1-0-
098
VOUT = 0.2V p-p
Figure 7. Small Signal Frequency Response for Various Temperatures (SOIC)
FREQUENCY (MHz)
NO
RM
ALI
ZED
CLO
SED
-LO
OP
GA
IN (d
B)
1–10
–9–8–7–6–5–4–3–2–101234
10 100 1000
0451
1-0-
073
G = +20
G = –1
G = +2
G = +5
G = +10
VOUT = 0.2V p-pVS = ±5VRLOAD = 1kΩ
Figure 8. Small Signal Frequency Response for Various Gains (CSP)
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
17
10
9
8
14
13
12
11
16
15
17
10 100 1000
0451
1-0-
077
G = +5RL = 1kΩVOUT = 0.2V p-p
VS = ±2.5V, CSP
VS = ±2.5V, SOIC
VS = ±5V, SOIC
VS = ±5V, CSP
Figure 9. Small Signal Frequency Response for Various Supply Voltages
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
6
5
4
3
2
1
10
9
8
7
11
10001 10 100
G = +2VS = ±5VRL = 1kΩ
+125°C
+25°C
–40°C
+85°C
0451
1-0-
097
VOUT = 0.2V p-p
Figure 10. Small Signal Frequency Response for Various Temperatures (CSP)
A. Operational Amplifier AD8099 207
AD8099
Rev. B | Page 7 of 28
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
19
10
18
19
16
14
12
17
15
13
11
20
10 100 1000
5pF, CSP
5pF, SOIC
1pF, CSP
1pF, SOIC
0451
1-0-
104
G = +5VS = ±5V
Figure 11. Small Signal Frequency Response for Various Capacitive Loads
FREQUENCY (MHz)
NO
RM
ALI
ZED
CLO
SED
-LO
OP
GA
IN (d
B)
1 10 100 1000
0451
1-0-
011
1
–10
–9
–8
–7
–6
–5
–4
–3
–2
–1
0G = +2
G = +5
G = +10
G = +20
VS = ±5VVOUT = 2V p-pRLOAD = 1kΩ
Figure 12. Large Signal Frequency Response for Various Gains (SOIC)
0451
1-0-
009
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
15.5
10 100
6.5
6.4
6.3
6.2
6.1
6.0
5.9
5.8
5.7
5.6
VOUT = 200mV p-p
VOUT = 1.4V p-pVS = ±5VG = +2RL = 150Ω
Figure 13. 0.1 dB latness (SOIC) F
0451
1-0-
080
FREQUENCY (MHz)O
PEN
-LO
OP
GA
IN (d
B)
OPE
N-L
OO
P PH
ASE
(Deg
rees
)
0.001 0.01 0.1 1.0 10 100 1000–10
40
0
10
50
60
20
30
70
80
90
–180
–105
–165
–150
–90
–75
–135
–120
–60
–45
–30
VS = ±5VRL = 1kΩUNCOMPENSATED
PHASE
MAGNITUDE
Figu nse
re 14. Open Loop Frequency Respo
FREQUENCY (MHz)
NO
RM
ALI
ZED
CLO
SED
-LO
OP
GA
IN (d
B)
1 10 100 1000
0451
1-0-
012
2
–9
–8
–7
–6
–5
–4
–3
–2
–1
0
1
VS = ±5VVOUT = 2V p-pRLOAD = 1kΩ
G = +5
G = +10
G = +20
G = +2
Figure 15. Large Signal Frequency Response for Various Gains (CSP)
0451
1-0-
008
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
15.5
10 100
6.5
6.4
6.3
6.2
6.1
6.0
5.9
5.8
5.7
5.6
VOUT = 200mV p-p
VOUT = 1.4V p-pVS = ±5VG = +2RL = 150Ω
Figure 16. 0.1 d Flatness (CSP) B
A. Operational Amplifier AD8099 208
AD8099
Rev. B | Page 8 of 28
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
15
8
7
6
12
11
10
9
14
13
15
10 100 1000
0451
1-0-
078
G = +5VS = ±5VVOUT = 2V p-p
RL = 100Ω, SOIC
RL = 1kΩ, SOIC
RL = 100Ω, CSP
RL = 1kΩ, CSP
Figure 17. Large Sign us Load Resistances
al Frequency Response for Vario
FREQUENCY (MHz)
INPU
T IM
PED
AN
CE
(kΩ
)
10.001
10 100 1000
0.1
1.0
10.0
0.01
100.0
0451
1-0-
105
VS = ±5VG = +2
Figure 18. Input Impedance vs. Frequency
FREQUENCY (MHz)
OU
TPU
T IM
PED
AN
CE
(Ω)
0.1
0.01
1
10
100
1000.1 1 10 1000
G = +2
G = +5
G = +10
VS = ±5V
0451
1-0-
100
Figure 19. Output Impedance . Frequency for Various Gains vs
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
15
8
7
6
12
11
10
9
14
13
15
10 100 1000
0451
1-0-
079
G = +5RL = 1kΩVOUT = 2V p-p
VS = ±2.5V, CSP
VS = ±2.5V, SOIC
VS = ±5V, CSP
VS = ±5V, SOIC
Figu es
–80
–40
–60
–20
–50
–70
–30
–10
re 20. Large Signal Frequency Response for Various Supply Voltag
FREQUENCY (MHz)
OFF
ISO
LATI
ON
(dB
)
0.1–90
1 10 100 100004
511-
0-09
4
SOICCSP
G = +2RL = 1kΩVS = ±5VVDIS = 0V
Figure 21. Off Isolation vs. Frequency
0451
1-A-
008
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–120
–50
–100
–90
–80
–70
–60
–110SOLID LINES – SECOND HARMONICSDOTTED LINE – THIRD HARMONICSSOLID LINES – SECOND HARMONICS
RMONICSDOTTED LINES – THIRD HA
G = +5VOUT = 2V p-pVS = ±5VRL = 100Ω
SOIC
CSP
Figure 22. Harmonic D tion vs. Frequency istor
A. Operational Amplifier AD8099 209
AD8099
Rev. B | Page 9 of 28
0451
1-A-
009
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–50
–110
–100
–90
–80
–70
–60
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VOUT = 2V p-pVS = ±5VRL = 1kΩ
Fig C)
ure 23. Harmonic Distortion vs. Frequency (SOI
0451
1-A-
010
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–40
–110
–100
–90
–80
–70
–60
–50
–120SOLID LINES – SECOND HARMONICSDOTTED LINE – THIRD HARMONICS
SOLID LINE – SECOND HARMONICCDOTTED LINE – THIRD HARMONI
G = +2VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 24. Harmonic Disto tion vs. Frequency (SOIC)
r
0451
1-A-
011
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–40
–110
–100
–90
–80
–70
–60
–50
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = –1VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 25. Harmonic Disto tion vs. Frequency (SOIC) r
0451
1-A-
012
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–50
–110
–100
–90
–80
–70
–60
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 26. Harmonic Distortion vs. Frequency (CSP)
0451
1-A-
013
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–40
–110
–100
–90
–80
–70
–60
–50
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +2VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 27. Harmonic Distortion vs. Frequency (CSP)
0451
1-A-
014
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–130
–40
–110
–100
–90
–80
–70
–60
–50
–120SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = –1VOUT = 2V p-pVS = ±5VRL = 1kΩ
Figure 28. Harmonic Distortion vs. Frequency (CSP)
A. Operational Amplifier AD8099 210
AD8099
Rev. B | Page 10 of 28
0451
1-A-
015
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–120
–50
–100
–90
–80
–70
–60
–110SOLID LINES – SECOND HARMONICSDOTTED LINES – THIRD HARMONICS
G = +10RL = 1kΩ
VS = ±2.5VVOUT = 1V p-p
VS = ±5VVOUT = 2V p-p
Figure 29. ge (SOIC)
Harmonic Distortion vs. Frequency and Supply Volta
0451
1-A-
016
OUTPUT AMPLITUDE (V p-p)71 2 3 4 5 6
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–110
–40
–90
–80
–70
–60
–50
–100SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VS = ±5Vf = 10MHzRL = 100Ω
Figure 30. Harmonic Distortion vs. Output Amplitude (SOIC)
0451
1-A-
017
OUTPUT AMPLITUDE (V p-p)
0451
1-A-
018
FREQUENCY (MHz)0.1 1.0 10.0
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)–120
–110
–50
–90
–80
–70
–60
–100
SOLID LINES – SECOND HARMONICSDOTTED LINE – THIRD HARMONICSSOLID LINES – SECOND HARMONICSDOTTED LINES – THIRD HARMONICS
G = +10RL = 1kΩ
VS = ±2.5VVOUT = 1V p-p
VS = ±5VVOUT = 2V p-p
Figur CSP) e 32. Harmonic Distortion vs. Frequency for Various Supplies (
0451
1-A-
019
OUTPUT AMPLITUDE (V p-p)71 2 3 4 5 6
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–110
–40
–90
–80
–70
–60
–50
–100SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VS = ±5Vf = 10MHzRL = 100Ω
Figure 33. Harmonic Distortion vs. Output Amplitude (CSP)
0451
1-A-
021
OUTPUT AMPLITUDE (V p-p)71 2 3 4 5 6
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–120
–110
–40
–90
–80
–70
–60
–50
–100
SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VS = ±5Vf = 10MHzRL = 1kΩ
Figure 34. Harmonic Distortion vs. Output Amplitude (CSP)
71 2 3 4 5 6
HA
RM
ON
IC D
ISTO
RTI
ON
(dB
c)
–120
–110
–40
–90
–80
–70
–60
–50
–100
SOLID LINE – SECOND HARMONICDOTTED LINE – THIRD HARMONIC
G = +5VS = ±5Vf = 10MHzRL = 1kΩ
Figure 31. Harmonic Distortion vs. Output Amplitude (SOIC)
A. Operational Amplifier AD8099 211
AD8099
Rev. B | Page 11 of 28
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 5 10–0.20
–0.05
–0.10
–0.15
0.15
0.10
0.05
0
0.20
15 20 25 30 35 40 45 50
1pF
10pF, 20Ω RSNUB
0451
1-0-
095
RSNUB
CL RLG = +5VS = ±5VRL = 1kΩ
Figur oads (SOIC)
e 35. Small Signal Transient Response for Various Capacitive L
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 10–0.15
–0.05
–0.10
0.10
0.05
0
0.15
20 30 40 50
G = +10RL = 1kΩ
VS = ±5.0VAND ±2.5V, SOIC
VS = ±5.0VAND ±2.5V, CSP
0451
1-0-
107
t ResponFigure 36. Small Signal Transien se for Various Supply Voltages
0451
1-A-
017
TIME (ns)10000 100 200 300 400 500 600 700 800 900
OU
TPU
T VO
LTA
GE
(V)
–5
5
4
3
2
1
0
–1
–2
–3
–4
RL = 1kΩ
RL = 100Ω
INPUT × 2
Figure 37. Output Overdrive Rec very for Various Resistive Loads
o
0451
1-0-
096
TIME (ns)O
UTP
UT
VOLT
AG
E (V
)
0 5 10–0.20
–0.05
–0.10
–0.15
0.15
0.10
0.05
0
0.20
15 20 25 30 35 40 45 50
10pF, 20Ω RSNUB
1pF
RSNUB
CL RLG = +5VS = ±5VRL = 1kΩ
Figure 38. Small Signal Transient Response for Various Capacitive Loads (CSP)
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 10–0.20
–0.05
–0.10
–0.15
0.15
0.10
0.05
0
0.20
20 30 40 50
RL = 1kΩ, 100ΩVOUT = 200mV p-pG = +5
VS = ±2.5VCSP VS = ±5.0V
CSP
VS = ±5.0VSOIC
VS = ±2.5VSOIC
0451
1-0-
102
Figure 39. Small Signal Transient Response for Various Supply Voltages
0451
1-0-
010
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
–0.52000 50 100 150
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0TURN ON
TURN ONINPUT
TURN OFF
TURN OFFINPUT
VS = ±5VG = 2
Figure 40. Disable/Enable Switching Speed
A. Operational Amplifier AD8099 212
AD8099
Rev. B | Page 12 of 28
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 10–1.5
–0.5
–1.0
1.0
0.5
0
1.5
20 30 40 50
G = +10RL = 1kΩ
VS = ±2.5V
VS = ±5.0V
0451
1-0-
106
Figure 41. L ltage (CSP)
arge Signal Transient Response vs. Supply Vo
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 10–1.5
–0.5
–1.0
1.0
0.5
0
1.5
20 30 40 50
VS = ±5.0V
G = +10RL = 1kΩ
VS = ±2.5V
0451
1-0-
118
F ) igure 42. Large Signal Frequency Response vs. Supply Voltage (SOIC
TIME (ns)
OU
TPU
T VO
LTA
GE
(V)
0 10–1.5
–0.5
–1.0
1.0
0.5
0
1.5
20 30 40 50
RL = 1kΩ, 100ΩG = +5
VS = ±5V
VS = ±2.5V
0451
1-0-
101
Fig d Load Resistance (SOIC and CSP)
ure 43. Large Signal Transient Response for Various Supply Voltages ans
0451
1-0-
052
TIME (ns)
OU
TPU
T/IN
PUT
VOLT
AG
E (V
)0 5 10 15 20 25 30 35 40
–1.5
–0.5
–1.0
0
0.5
1.0
1.5
–0.3%
–0.1%
–0.2%
0%
0.1%
0.2%
0.3%
45
INPUT
ERROR
OUTPUT
G = +2RLOAD = 1kΩVs = ±5V
Figu )
re 44. Short Term Settling Time (CSP
0451
1-0-
051
TIME (ns)
OU
TPU
T/IN
PUT
VOLT
AG
E (V
)
0 5 10 15 20 25 30 35 40–1.5
–0.5
–1.0
0
0.5
1.0
1.5
–0.3%
–0.1%
–0.2%
0%
0.1%
0.2%
0.3%
45
G = +2RLOAD = 1kΩVs = ±5V
INPUT
ERROR
OUTPUT
Figure 45. Short Term Settling Time (SOIC)
0451
1-0-
050
TIME (µs)
OU
TPU
T/IN
PUT
VOLT
AG
E (V
)
0 50 100 150 200 250 300 350 400 450–1.5
–0.5
–1.0
0
0.5
1.0
1.5
–0.30%
–0.10%
–0.20%
0%
0.10%
0.20%
0.30%
500
INPUT
ERROR
OUTPUTG = +2VS = ±5V
Figure 46. Long Term Settling Time
A. Operational Amplifier AD8099 213
AD8099
Rev. B | Page 13 of 28
0451
1-0-
113
FREQUENCY (MHz)10000.1 1.0 10 100
CO
MM
ON
-MO
DE
REJ
ECTI
ON
(dB
)
–110
–20
–30
–50
–70
–40
–60
–80
–90
–100
G = +2RL = 1kΩ
Figure 4 uency
7. Common-Mode Rejection vs. Freq
0451
1-0-
004
1 10 100 1k 10k 100k 1M 10M 100M 1G1
10
100
1000
FREQUENCY (Hz)
INPU
T C
UR
REN
T N
OIS
E (p
A H
z)
Figure 48. Input Current Noise v . Frequency (DISABLE = Open) s
0451
1-0-
005
1 10 100 1k 10k 100k 1M 10M 100M 1G0.1
1
10
1000
FREQUENCY (Hz)
INPU
T VO
LTA
GE
NO
ISE
(nV
Hz)
Figure 49. Input Voltage Noise vs. Frequency
0451
1-0-
114
FREQUENCY (MHz)10000.01 0.10 1.0 10 100
POW
ER S
UPP
LY R
EJEC
TIO
N (d
B)
–100
0
–20
–10
–40
–60
–30
–50
–70
–80
–90
G = +5RL = 1kΩ
POSITIVE
NEGATIVE
Fi
gure 50. Power Supply Rejection vs. Frequency
0451
1-0-
003
FREQUENCY (Hz)
INPU
T C
UR
REN
T N
OIS
E (p
A H
z)
1 10 100 1k 10k 100k 1M 10M 100M 1G1
10
100
1000
Figure 51. Input Current Noise vs. Frequency (DISABLE = +VS)
VOFFSET (µV)
CO
UN
T
–3000
60
40
20
100
80
120
–200 0–100 100 200
0451
1-0-
075
VS = ±5VN = 1,200X = –70µVσ = 80µVX
Figure 52. Input Offset Voltage Distribution
A. Operational Amplifier AD8099 214
AD8099
Rev. B | Page 14 of 28
0451
1-A-
003
TEMPERATURE (C)125–40 –25 –10 5 20 35 50 65 80 95 110
OFF
SET
VOLT
AG
E (µ
V)
–200
400
200
100
300
0
–100
VS = 5V
VS = ±5V
F
igure 53. Input Offset Voltage vs. Temperature
0451
1-A-
004
TEMPERATURE (C)125–40 –25 –10 5 20 35 50 65 80 95 110
BIA
S C
UR
REN
T (µ
A)
–6.6
–5.4
–5.8
–6.0
–5.6
–6.2
–6.4
IB+, VS = ±5V
IB+, VS = 5V
IB–, VS = ±5V
IB–, VS = 5V
Figure 54. Input Bias Current vs. Temperature (DISABLE Pin Floating)
0451
1-A-
005
TEMPERATURE (C)125–40 –25 –10 –5 20 35 50 65 80 95 110
OU
TPU
T SA
TUR
ATI
ON
VO
LTA
GE
(V)
1.12
1.24
1.20
1.18
1.22
1.16
1.14+VS – VOUT
VS = 5V
VS = ±5V
+VS – VOUT
–VS + VOUT
–VS + VOUT
Figure 55. Output Saturation Voltage vs. Temperature
0451
1-A-
006
TEMPERATURE (C)125–40 –25 –10 5 20 35 50 65 80 95 110
SUPP
LY C
UR
REN
T (m
A)
8
20
16
14
18
12
10
VS = 5V
VS = ±5V
Figure 56. Supply Current vs. Temperature
0451
1-A-
007
TEMPERATURE (C)125–40 –25 –10 5 20 35 50 65 80 95 110
BIA
S C
UR
REN
T (µ
A)
–1.0
1.0
–0.2
–0.4
0
0.2
0.4
0.6
0.8
–0.6
–0.8
IB+, VS = 5V
IB+, VS = ±5V
IB–, VS = 5V
IB–, VS = ±5V
Figure 57. Input Bias Current vs. Temperature (DISABLE Pin = +VS)
A. Operational Amplifier AD8099 215
AD8099
Rev. B | Page 15 of 28
THEORY OF OPERATION The AD8099 is a voltage feedback op amp that employs a new highly linear low noise input stage. With this input stage, the AD8099 can achieve better than 90 dB distortion for a 2 V p-p, 10 MHz output signal with an input referred voltage noise of less than 1 nV/√Hz. This noise level and distortion performance has been previously achievable only with fully uncompensated amplifiers. The AD8099 achieves this level of performance for gains as low as +2. This new input stage also triples the achievable slew rate for comparably compensated 1 nV/√Hz amplifiers.
The simplified AD8099 topology is shown in Figure 58. The amplifier is a single gain stage with a unity gain output buffer fabricated in Analog Devices’ extra fast complimentary bipolar process (XFCB). The AD8099 has 85 dB of open-loop gain and maintains precision specifications such as CMRR, PSRR, VOS, and ∆VOS/∆T to levels that are normally associated with topologies having two or more gain stages.
BUFFERgmCCR1 RL
VOUT
0451
1-0-
060
Figure 58. AD8099 Topology
The AD8099 can be externally compensated down to a gain of 2 through the use of an RC network. Above gains of 15, no exter-nal compensation network is required. To realize the full gain bandwidth product of the AD8099, no PCB trace should be connected to or within close proximity of the external compen-sation pin for the lowest possible capacitance.
External compensation allows the user to optimize the closed-loop response for minimal peaking while increasing the gain bandwidth product in higher gains, lowering distortion errors that are normally more prominent with internally compensated parts in higher gains. For a fixed gain bandwidth, wideband distortion products would normally increase by 6 dB going from a closed-loop gain of 2 to 4. Increasing the gain bandwidth product of the AD8099 eliminates this effect with increasing closed-loop gain.
The AD8099 is available in both a SOIC and an LFCSP, each of which has a thermal pad for lower operating temperature. To help avoid this pad in board layout, both packages have an extra output pin on the opposite side of the package for ease in con-necting a feedback network to the inputs. The secondary output pin also isolates the interaction of any capacitive load on the
output and self-inductance of the package and bond wire from the feedback loop. While using the secondary output for feed-back, inductance in the primary output will now help to isolate capacitive loads from the output impedance of the amplifier. Since the SOIC has greater inductance in its output, the SOIC will drive capacitive loads better than the LFCSP. Using the primary output for feedback with both packages will result in the LFCSP driving capacitive load better than the SOIC.
The LFCSP and SOIC pinouts are identical, except for the rotation of all pins counterclockwise by one pin on the LFCSP. This isolates the inputs from the negative power supply pin, removing a mutually inductive coupling that is most prominent while driving heavy loads. For this reason, the LFCSP second harmonic, while driving a heavy load, is significantly better than that of the SOIC.
A three-state input pin is provided on the AD8099 for a high impedance power-down and an optional input bias current cancellation circuit. The high impedance output allows several AD8099s to drive the same ADC or output line time inter-leaved. Pulling the DISABLE pin low activates the high impedance state. See Table 5 for threshold levels. When the DISABLE pin is left floating, the AD8099 operates normally. With the DISABLE pin pulled within 0.7 V of the positive supply, an optional input bias current cancellation circuit is turned on, which lowers the input bias current to less than 200 nA. In this mode, the user can drive the AD8099 with a high dc source impedance and still maintain minimal output referred offset without having to use impedance matching techniques. In addition, the AD8099 can be ac-coupled while setting the bias point on the input with a high dc impedance network. The input bias current cancellation circuit will double the input referred current noise, but this effect is minimal as long as wideband impedance is kept low (see Figure 48 and Figure 51).
A pair of internally connected diodes limits the differential voltage between the noninverting input and the inverting input of the AD8099. Each set of diodes has two series diodes, which are connected in anti-parallel. This limits the differential voltage between the inputs to approximately ±1.8 V. All of the AD8099 pins are ESD protected with voltage limiting diodes connected between both rails. The protection diodes can handle 5 mA of steady state current. Currents should be limited to 5 mA or less through the use of a series limiting resistor.
A. Operational Amplifier AD8099 216
AD8099
Rev. B | Page 16 of 28
APPLICATIONS USING THE AD8099 The AD8099 offers unrivaled noise and distortion performance in low signal gain configurations. In low gain configurations (less than15), the AD8099 requires external compensation. The amount of gain and performance needed will determine the compensation network.
Understanding the subtleties of the AD8099 gives the user insight on how to exact its peak performance. Use the component values and circuit configurations shown in the Applications section as starting points for designs. Specific circuit applications will dictate the final configuration and value of your components.
CIRCUIT COMPONENTS The circuit components are referenced in Figure 59, the recommended noninverting circuit schematic for the AD8099. See Table 4 for typical component values and performance data.
1
84
7
53
6
2
AD8099
C50.1µF
CC
CF
C1
C410µF
C210µF
C30.1µF
RC
RF
RG
RS
R1
+VS
–VS
VOUT
DISABLE
0451
1-0-
061
VIN
Figure 59. Wideband Noninverting Gain Configuration (SOIC)
RF and RG—The feedback resistor and the gain set resistor determine the noise gain of the amplifier; typical RF values range from 250 Ω to 499 Ω.
CF—Creates a zero in the loop response to compensate the pole created by the input capacitance (including stray capacitance) and the feedback resistor RF. CF helps reduce high frequency peaking and ringing in the closed-loop response. Typical range is 0.5 pF to 1.5 pF for evaluation circuits used here.
R1—This resistor terminates the input of the amplifier to the source resistance of the signal source, typically 50 Ω. (This is application specific and not always required.)
RS—Many high speed amplifiers in low gain configurations require that the input stage be terminated into a nominal impedance to maintain stability. The value of RS should be kept to 50 Ω or lower to maintain low noise performance. At higher gains, RS may be reduced or even eliminated. Typical range is 0 Ω to 50 Ω.
CC—The compensation capacitor decreases the open-loop gain at higher frequencies where the phase is degrading. By decreas-ing the open-loop gain here, the phase margin is increased and the amplifier is stabilized. Typical range is 0 pF to 5 pF. The value of CC is gain dependent.
RC—The series lead inductance of the package and the com-pensation capacitance (CC) forms a series resonant circuit. RC dampens this resonance and prevents oscillations. The recommended value of RC is 50 Ω for a closed-loop gain of 2. This resistor introduces a zero in the open-loop response and must be kept low so that this zero occurs at a higher frequency. The purpose of the compensation network is to decrease the open-loop gain. If the resistance becomes too large, the gain will be reduced to the resistor value, and not necessarily to 0 Ω, which is what a single capacitor would do over frequency. Typical value range is 0 Ω to 50 Ω.
C1—To lower the impedance of RC , C1 is placed in parallel with RC. C1 is not required, but greatly reduces peaking at low closed-loop gains. The typical value range is 0 pF to 2 pF.
C2 and C3—Bypass capacitors are connected between both supplies for optimum distortion and PSRR performance. These capacitors should be placed as close as possible to the supply pins of the amplifier. For C3, C5, a 0508 case size should be used. The 0508 case size offers reduced inductance and better frequency response.
C4 and C2—Electrolytic bypass capacitors.
A. Operational Amplifier AD8099 217
AD8099
Rev. B | Page 17 of 28
RECOMMENDED VALUES Table 4. Recommended Values and AD8099 Performance
Feedback Network Values
Compensation Network Values
Gain Package RF RG RS CF RC CC C1
−3 dB SS Bandwidth
(MHz) Slew Rate
(V/µs) Peaking
(dB)
Output Noise (AD8099 Only)
(nV/√Hz)
Total Output NoiseIncluding Resistors
(nV/√Hz)
−1, 2 SOIC 250 250 50 1.5 50 4 1.5 440/700 515 0.3/3.1 2.1 4
2 CSP 250 250 50 0.5 50 5 2 700 475 3.2 2.1 4
−1 CSP 250 250 50 1.0 50 5 2 420 475 0.8 2.1 4
5 CSP/SOIC 499 124 20 0.5 50 1 0 510 735 1.4 4.9 8.6
10 CSP/SOIC 499 54 0 0 0 0.5 0 550 1350 0.8 9.6 13.3
20 CSP/SOIC 499 26 0 0 0 0 0 160 1450 0 19 23.3
CIRCUIT CONFIGURATIONS Figure 60 through Figure 66 show typical schematics for the AD8099 in various gain configurations. Table 4 data was collected using the schematics shown in Figure 60 through Figure 66. Resistor R1, as shown in Figure 60 through Figure 66,
is the test equipment termination resistor. R1 is not required for normal operation, but is shown in the schematics for completeness.
1
84
7
53
6
2
AD8099
C50.1µF
CC4pF
CF1.5pF
C11.5pF
C410µF
C210µF
C30.1µF
RC50Ω
RF250Ω
RG250Ω
RS50Ω
R150Ω
+VS
–VS
VOUT
0451
1-0-
116
VIN
RL1kΩ
DISABLE
Figure 60. Amplifier Configuration for SOIC Package, Gain = –1
1
84
7
53
6
2
AD8099
C50.1µF
CC4pF
CF1.5pF
C11.5pF
C410µF
C210µF
C30.1µF
RC50Ω
RF250Ω
RG250Ω
RS50Ω
R150Ω
+VS
–VS
VOUT
0451
1-0-
054
VINRL1kΩ
DISABLE
Figure 61. Amplifier Configuration for SOIC Package, Gain = +2
2
15
8
64
7
3
AD8099
C50.1µF
CC5pF
CF1pF
C12pF
C410µF
C210µF
C30.1µF
RC50Ω
RF250Ω
RS50Ω
+VS
–VS
VOUT
0451
1-0-
108
RG250Ω
R150Ω
VIN
RL1kΩ
DISABLE
Figure 62. Amplifier Configuration for CSP Package, Gain =–1
2
15
8
64
7
3
AD8099
C50.1µF
CC5pF
CF0.5pF
C12pF
C410µF
C210µF
C30.1µF
RC50Ω
RF250Ω
RG250Ω
RS50Ω
R150Ω
+VS
–VS
VOUT
0451
1-0-
053
VINRL1kΩ
DISABLE
Figure 63. Amplifier Configuration for CSP Package, Gain = +2
A. Operational Amplifier AD8099 218
AD8099
Rev. B | Page 18 of 28
FB
AD8099
C50.1µF
CC1pF
CF0.5pF
C410µF
C210µF
C30.1µF
RC50Ω
RF499Ω
RG124Ω
RS20Ω
R150Ω
+VS
–VS
VOUT04
511-
0-05
5
VINRL1kΩ
–
+
D–V
CC
VO
+V
DISABLE
Figure 64. Amplifier Configuration for CSP and SOIC Package, Gain = +5
AD8099
C50.1µF
CC0.5pFC4
10µF
C210µF
C30.1µF
RF499Ω
RG54Ω
R150Ω
+VS
–VS
VOUT
0451
1-0-
056
VINRL1kΩ
FB–
+
D–V
CC
VO
+V
DISABLE
Figure 65. Amplifier Configuration for CSP and SOIC Packages, Gain = +10
AD8099
C50.1µF
C410µF
C210µF
C30.1µF
RF499Ω
RG26Ω
R150Ω
+VS
–VS
VOUT
0451
1-0-
057
VINRL1kΩ
FB–
+
D–V
CC
VO
+V
DISABLE
Figure 66. Amplifier Configuration for CSP and SOIC Packages, Gain = +20
A. Operational Amplifier AD8099 219
AD8099
Rev. B | Page 19 of 28
PERFORMANCE VS. COMPONENT VALUES The influence that each component has on the AD8099 frequency response can be seen in Figure 67 and Figure 68. In Figure 67 and Figure 68, all component values are held constant, except for the individual component shown, which is varied. For example, in the RS performance plot of Figure 68, all components are held constant except RS, which is varied from 0 Ω to 50 Ω.; and clearly indicates that RS has a major influence on peaking and bandwidth of the AD8099.
1
84
7
53
6
2
AD8099
C50.1µF
CC
CF
C1
C410µF
C210µF
C30.1µF
RC
RF
RG
RS
R1
+VS
–VS
VOUT
SOIC PINOUT SHOWN
VIN
DISABLE
0451
1-0-
117
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
–23000
0451
1-0-
020
1 10 100 1000
9
8
7
6
5
4
3
2
1
0
–1
VS = ±5VG = +2RLOAD = 1kΩSOIC PACKAGE
C1 = 0pF
C1 = 2pF
C1 = 1.5pF
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
–13000
0451
1-0-
024
1 10 100 1000
10
9
8
7
6
5
4
3
2
1
0
CC = 3pF
CC = 4pF
CC = 5pF
VS = ±5VG = +2RLOAD = 1kΩSOIC PACKAGE
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
–13000
0451
1-0-
030
1
10
8
9
7
6
5
4
3
2
1
0
VS = ±5VG = +2RLOAD = 1kΩSOIC PACKAGE
10 100 1000
RC = 50Ω
RC = 35Ω
RC = 20Ω
Figure 67. Frequency Response for Various Values of C1, CC, RC
A. Operational Amplifier AD8099 220
AD8099
Rev. B | Page 20 of 28
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
–13000
0451
1-0-
032
1
10
8
9
7
6
5
4
3
2
1
0
VS = ±5VG = +2RLOAD = 1kΩSOIC PACKAGE
10 100 1000
RF = RG = 200
RF = RG = 250
RF = RG = 300
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
10
9
8
7
6
5
4
3
2
1
100 1000 3000
0451
1-0-
058
1–1
0
10
CF = 1pF
CF = 0.5pF
CF = 1.5pF
V = ±5VSG = +2RLOAD = 1kΩSOIC PACKAGE
FREQUENCY (MHz)
CLO
SED
-LO
OP
GA
IN (d
B)
010000
0451
1-0-
034
1
12
10
11
9
8
7
6
5
4
3
2
1
VS = ±5VG = +2RLOAD = 1kΩSOIC PACKAGE
10 100 1000
RS = 0
RS = 50
RS = 20
1
84
7
53
6
2
AD8099
C50.1µF
CC
CF
C1
C410µF
C210µF
C30.1µF
RC
RF
RG
RS
R1
+VS
–VS
VOUT
SOIC PINOUT SHOWN
VIN
DISABLE
0451
1-0-
117
Figure 68. Frequency Response for Various Values of RF, CF, RS
TOTAL OUTPUT NOISE CALCULATIONS AND DESIGN To analyze the noise performance of an amplifier circuit, the individual noise sources must be identified. Then determine if the source has a significant contribution to overall noise perfor-mance of the amplifier. To simplify the noise calculations, we will work with noise spectral densities, rather than actual voltages to leave bandwidth out of the expressions (noise spectral density, which is generally expressed in nV/√Hz, is equivalent to the noise in a 1 Hz bandwidth).
The noise model shown in Figure 69 has six individual noise sources: the Johnson noise of the three resistors, the op amp voltage noise, and the current noise in each input of the amplifier. Each noise source has its own contribution to the
noise at the output. Noise is generally specified RTI (referred to input), but it is often simpler to calculate the noise referred to the output (RTO) and then divide by the noise gain to obtain the RTI noise.
All resistors have a Johnson noise of √(4kBTR), where k is Boltzmann’s Constant (1.38 × 10–23 J/K), T is the absolute temperature in Kelvin, B is the bandwidth in Hz, and R is the resistance in ohms. A simple relationship, which is easy to remember, is that a 50 Ω resistor generates a Johnson noise of 1 nV√Hz at 25°C. The AD8099 amplifier has roughly the same equivalent noise as a 50 Ω resistor.
A. Operational Amplifier AD8099 221
AD8099
Rev. B | Page 21 of 28
0451
1-0-
070
GAIN FROM"B" TO OUTPUT = – R2
R1
GAIN FROM"A" TO OUTPUT =
NOISE GAIN =
NG = 1 + R2R1
IN–
VN
VN, R1
VN, R3
R1
R2
IN+R3
4kTR2
4kTR1
4kTR3
VN, R2
B
A
VN2 + 4kTR3 + 4kTR1 R2 2
R1 + R2
IN+2R32 + IN–2 R1 × R2 2 + 4kTR2 R1 2
R1 + R2 R1 + R2RTI NOISE =
RTO NOISE = NG × RTI NOISE
VOUT
+
Figure 69. Op Amp Noise Analysis Model
In applications where noise sensitivity is critical, care must be taken not to introduce other significant noise sources to the amplifier. Each resistor is a noise source. Attention to the following areas is critical to maintain low noise performance: design, layout, and component selection. A summary of noise performance for the amplifier and associated resistors can be seen in Table 4.
INPUT BIAS CURRENT AND DC OFFSET In high noise gain configurations, the effects of output offset voltage can be significant, even with low input bias currents and input offset voltages. Figure 70 shows a comprehensive offset voltage model, which can be used to determine the referred to output (RTO) offset voltage of the amplifier or referred to input (RTI) offset voltage.
0451
1-0-
071
GAIN FROM"B" TO OUTPUT = – R2
R1
GAIN FROM"A" TO OUTPUT =
NOISE GAIN =
NG = 1 + R2R1
IB–
VOS
R1
R2
IB+R3
B
A
OFFSET (RTO) = VOS 1 + R2 + IB+ × R3 1 + R2 – IB– × R2R1 R1
OFFSET (RTI) = VOS + IB+ × R3 – IB–R1 × R2R1 + R2
OFFSET (RTI) = VOS IF IB+ = IB– AND R3 = R1 × R2 R1 + R2
VOUT
FOR BIAS CURRENT CANCELLATION:
Figure 70. Op Amp Total Offset Voltage Model
For RTO calculations, the input offset voltage and the voltage generated by the bias current flowing through R3 are multiplied by the noise gain of the amplifier. The voltage generated by IB– through R2 is summed together with the previous offset voltages to arrive at a final output offset voltage. The offset voltage can also be referred to the input (RTI) by dividing the calculated output offset voltage by the noise gain.
As seen in Figure 70 if IB+ and IB– are the same and R3 equals the parallel combination of R1 and R2, then the RTI offset voltage can be reduced to only VOS. This is a common method used to reduce output offset voltage. Keeping resistances low helps to minimize offset error voltage and keeps the voltage noise low.
DISABLE PIN AND INPUT BIAS CANCELLATION
The AD8099 DISABLE pin performs three functions; enable, disable, and reduction of the input bias current. When the DISABLE pin is brought to within 0.7 V of the positive supply, the input bias current is reduced by an approximate factor of 60. However, the input current noise doubles to 5.2 pA/√Hz. Table 5 outlines the DISABLE pin functionality.
Table 5. DISABLE Pin Truth Table Supply Voltage ±5 V +5 V
Disable –5 to +2.4 0 to 2.4 Enable Open Open Low Input Bias Current 4.3 to 5 4.3 to 5
A. Operational Amplifier AD8099 222
AD8099
Rev. B | Page 22 of 28
8
1
4
7
53
6
2
AD8099
AD7667
C410µF
CC9pF
C12pF
C50.1µF
C110µF
C20.1µF
RC50Ω
RF150Ω
RG150Ω
RS50Ω
R1590Ω
R2590Ω
+VS
–VS
DISABLE
VIN
0451
1-0-
072
IN
REFGND
REF
INGND
AGND AVDD DGND DVDD
DVDD
C62.7nF
R715Ω
AVDD
1µF 47µF
REF
0.1µF 0.1µF
+2.5V
Figure 71. ADC Driver
16-BIT ADC DRIVER Ultralow noise and distortion performance make the AD8099 an ideal ADC driver. Even though the AD8099 is not unity gain stable, it can be configured to produce a net gain of +1 amplifier, as shown in Figure 71. This is achieved by combining a gain of +2 and a gain of –1 for a net gain of +1. The input range of the ADC is 0 V to 2.5 V.
Table 6 shows the performance data of the AD8099 and the Analog Devices AD7667 a 1 MSPS 16-bit ADC.
Table 6. ADC Driver Performance, fC = 20 kHz, VOUT = 2.24 V p-p Parameter Measurement (dB) Second Harmonic Distortion –111.4 Third Harmonic Distortion –103.2 THD –101.4 SFDR 102.2 SNR 88.1
A. Operational Amplifier AD8099 223
AD8099
Rev. B | Page 23 of 28
CIRCUIT CONSIDERATIONS Optimizing the performance of the AD8099 requires attention to detail in layout and signal routing of the board. Power supply bypassing, parasitic capacitance, and component selection all contribute to the overall performance of the amplifier. The AD8099 features an exposed paddle on the backs of both the CSP and SOIC packages. The exposed paddle provides a low thermal resistive path to the ground plane. For best performance, solder the exposed paddle to the ground plane.
PCB Layout
The compensation network is determined by the amplifier gain requirements. For lower gains, the layout and component placement are more critical. For higher gains, there are fewer compensation components, which results in a less complex layout. With diligent consideration to layout, grounding, and component placement, the AD8099 evaluation boards have been optimized for peak performance. These are the same evaluation boards that are available to customers; see Table 7 for ordering information. The noninverting evaluation board art-work for SOIC and CSP layouts are shown in Figure 72 and Figure 73. Incorporating the layout information shown in Figure 72 and Figure 73 into new designs is highly recom-mended and helps to ensure optimal circuit performance. The concepts of layout, grounding, and component placement, llustrated in Figure 72 and Figure 73,also apply to inverting configurations. For scale, the boards are 2” × 2”.
Parasitics
The area surrounding the compensation pin is very sensitive to parasitic capacitance. To realize the full gain bandwidth product of the AD8099, there should be no trace connected to or within close proximity of the external compensation pin for the lowest possible capacitance. When compensation is required, the traces to the compensation pin, the negative supply, and the interconnect between components (i.e. CC, C1, and RC in Figure 59) should be made as wide as possible to minimize inductance.
All ground and power planes under the pins of the AD8099 should be cleared of copper to prevent parasitic capacitance between the input and output pins to ground. A single mount-ing pad on a SOIC footprint can add as much as 0.2 pF of capacitance to ground as a result of not clearing the ground or power plane under the AD8099 pins. Parasitic capacitance can cause peaking and instability, and should be minimized to ensure proper operation.
The new pinout of the AD8099 reduces the distance between the output and the inverting input of the amplifier. This helps to minimize the parasitic inductance and capacitance of the feedback path, which, in turn, reduces ringing and second harmonic distortion.
Grounding
When possible, ground and power planes should be used. Ground and power planes reduce the resistance and inductance of the power supply feeds and ground returns. If multiple planes are used, they should be “stitched” together with multiple vias. The returns for the input, output terminations, bypass capacitors, and RG should all be kept as close to the AD8099 as possible. Ground vias should be placed at the very end of the component mounting pad to provide a solid ground return. The output load ground and the bypass capacitor grounds should be returned to a common point on the ground plane to minimize parasitic inductance and improve distortion performance. The AD8099 packages feature an exposed paddle. For optimum performance, solder this paddle to ground. For more information on PCB layout and design considerations, refer to section 7-2 of the 2002 Analog Devices Op Amp Applications book.
Power Supply Bypassing
The AD8099 power supply bypassing has been optimized for each gain configuration as shown in Figure 60 through Figure 66 in the Circuit Configurations section. The values shown should be used when possible. Bypassing is critical for stability, frequency response, distortion, and PSRR performance. The 0.1 µF capacitors shown in Figure 60 through Figure 66 should be as close to the supply pins of the AD8099 as possible and the electrolytic capacitors beside them.
Component Selection
Smaller components less than 1206 SMT case size, offer smaller mounting pads, which have less parasitics and allow for a more compact layout. It is critical for optimum performance that high quality, tight tolerance (where critical), and low drift compo-nents be used. For example, tight tolerance and low drift is critical in the selection of the feedback capacitor used in Figure 60. The feedback compensation capacitor in Figure 60 is 1.5pF. This capacitor should be specified with NPO material. NPO material typically has a ±30 ppm/°C change over –55°C to +125°C temperature range. For a 100°C change, this would result in a 4.5 fF change in capacitance, compared to an X7R material, which would result in a 0.23 pF change, a 15% change from the nominal value. This could introduce excessive peaking, as shown in Figure 68, CF vs. Frequency Response.
DESIGN TOOLS AND TECHNICAL SUPPORT Analog Devices is committed to the design process by providing technical support and online design tools. ADI offers technical support via free evaluation boards, sample ICs, SPICE models, interactive evaluation tools, application notes, phone and email support—all available at www.analog.com.
A. Operational Amplifier AD8099 224
A. Operational Amplifier LT6205/06/07 225
A.3 Datasheet: Operational Amplifier
LT6205/LT6206/LT6207
The datasheet of the operational amplifier LT6205/LT6206/LT6207 was down-
loaded from http://www.linear.com/product/LT6205, accessed 30 August 2012.
LT6205/LT6206/LT6207
1620567fc
TYPICAL APPLICATION
FEATURES
APPLICATIONS
DESCRIPTION
Single/Dual/Quad Single Supply 3V,
100MHz Video Op Amps
The LT®6205/LT6206/LT6207 are low cost single/dual/quad voltage feedback amplifi ers that feature 100MHz gain-bandwidth product, 450V/μs slew rate and 50mA output current. These amplifi ers have an input range that includes ground and an output that swings within 60mV of either supply rail, making them well suited for single supply operation.
These amplifi ers maintain their performance for supplies from 2.7V to 12.6V and are specifi ed at 3V, 5V and ±5V. The inputs can be driven beyond the supplies without damage or phase reversal of the output. Isolation between channels is high, over 90dB at 10MHz.
The LT6205 is available in the 5-pin SOT-23, and the LT6206 is available in an 8-lead MSOP package with standard op amp pinouts. For compact layouts the quad LT6207 is available in the 16-pin SSOP package. These devices are specifi ed over the commercial, industrial and automotive temperature ranges.
Baseband Video Splitter/Cable Driver
n 450V/μs Slew Raten 100MHz Gain Bandwidth Productn Wide Supply Range 2.7V to 12.6Vn Output Swings Rail-to-Railn Input Common Mode Range Includes Groundn High Output Drive: 50mAn Channel Separation: 90dB at 10MHzn Specifi ed on 3V, 5V and ±5V Suppliesn Input Offset Voltage: 1mVn Low Power Dissipation: 20mW per Amplifi er on
Single 5Vn Operating Temperature Range: –40°C to 125°Cn Low Profi le (1mm) SOT-23 (ThinSOT™) Package
n Video Line Drivern Automotive Displaysn RGB Amplifi ersn Coaxial Cable Driversn Low Voltage High Speed Signal Processing
Output Step Response
–
+
–
+
LT6206
VIN
1μF
75Ω
75Ω
75Ω
75Ω
75Ω
VOUT1
VOUT2
3.3V
499Ω 499Ω
499Ω 499Ω
1
7
4
8
6
5
3
2
F3dB 50MHz
IS 25mA
620567 TA01a
VIN
VOUT
20ns/DIVVS = 3.3VVIN = 0.1V TO 1.1Vf = 10MHz
0V
0V
620567 TA01b
L, LT, LTC, LTM, Linear Technology and the Linear logo are registered trademarks and ThinSOT is a trademark of Linear Technology Corporation. All other trademarks are the property of their respective owners.
A. Operational Amplifier LT6205/06/07 226
LT6205/LT6206/LT6207
2620567fc
ABSOLUTE MAXIMUM RATINGSTotal Supply Voltage (V+ to V–) ..............................12.6VInput Current .......................................................±10mAInput Voltage Range (Note 2) ....................................±VS Output Short-Circuit Duration (Note 3) ........... Indefi nitePin Current While Exceeding Supplies (Note 9) ...±25mAOperating Temperature Range (Note 4) LT6205C/LT6206C/LT6207C,
LT6205I/LT6206I/LT6207I .................... –40°C to 85°C LT6205H ........................................... –40°C to 125°C
(Note 1)
5 V+
4 –IN
OUT 1
TOP VIEW
S5 PACKAGE5-LEAD PLASTIC TSOT-23
V– 2
+IN 3+ –
TJMAX = 150°C, θJA = 250°C/W
1
2
3
4
OUT A
–IN A
+IN A
V–
8
7
6
5
V+
OUT B
–IN B
+IN B
TOP VIEW
MS8 PACKAGE8-LEAD PLASTIC MSOP
+
–+
–
TJMAX = 150°C, θJA = 250°C/W
TOP VIEW
GN PACKAGE16-LEAD NARROW PLASTIC SSOP
1
2
3
4
5
6
7
8
16
15
14
13
12
11
10
9
OUT A
–IN A
+IN A
V+
+IN B
–IN B
OUT B
NC
OUT D
–IN D
+IN D
V–
+IN C
–IN C
OUT C
NC
C–
+
D+
–
B–
+
A+
–
TJMAX = 150°C, θJA = 135°C/W
PIN CONFIGURATION
ORDER INFORMATIONLEAD FREE FINISH TAPE AND REEL PART MARKING* PACKAGE DESCRIPTION SPECIFIED TEMPERATURE RANGE
LT6205CS5#PBF LT6205CS5#TRPBF LTAEM 5-Lead Plastic TSOT-23 0°C to 70°C
LT6205IS5#PBF LT6205IS5#TRPBF LTAEM 5-Lead Plastic TSOT-23 –40°C to 85°C
LT6205HS5#PBF LT6205HS5#TRPBF LTAEM 5-Lead Plastic TSOT-23 –40°C to 125°C
LT6206CMS8#PBF LT6206CMS8#TRPBF LTH3 8-Lead Plastic MSOP 0°C to 70°C
LT6206IMS8#PBF LT6206IMS8#TRPBF LTH4 8-Lead Plastic MSOP –40°C to 85°C
LT6207CGN#PBF LT6207CGN#TRPBF 6207 16-Lead Narrow Plastic SSOP 0°C to 70°C
LT6207IGN#PBF LT6207IGN#TRPBF 6207I 16-Lead Narrow Plastic SSOP –40°C to 85°C
Consult LTC Marketing for parts specifi ed with wider operating temperature ranges. *The temperature grade is identifi ed by a label on the shipping container.
Consult LTC Marketing for information on non-standard lead based fi nish parts.
For more information on lead free part marking, go to: http://www.linear.com/leadfree/ For more information on tape and reel specifi cations, go to: http://www.linear.com/tapeandreel/
Specifi ed Temperature Range (Note 4) LT6205C/LT6206C/LT6207C ..................... 0°C to 70°C LT6205I/LT6206I/LT6207I .................... –40°C to 85°C LT6205H ........................................... –40°C to 125°CStorage Temperature Range .................. –65°C to 150°CMaximum Junction Temperature .......................... 150°CLead Temperature (Soldering, 10 sec) ................. 300°C
A. Operational Amplifier LT6205/06/07 227
LT6205/LT6206/LT6207
3620567fc
ELECTRICAL CHARACTERISTICS
SYMBOL PARAMETER CONDITIONS
LT6205C/LT6206C/LT6207CLT6205I/LT6206I/LT6207I
UNITSMIN TYP MAX
VOS Input Offset Voltagel
1 3.55
mVmV
Input Offset Voltage Match(Channel-to-Channel) (Note 5) l
1 34
mVmV
Input Offset Voltage Drift (Note 6) l 7 15 μV/°C
IB Input Bias Current l 10 30 μA
IOS Input Offset Current l 0.6 3 μA
Input Noise Voltage 0.1Hz to 10Hz 2 μVP-P
en Input Noise Voltage Density f = 10kHz 9 nV/√Hz
in Input Noise Current Density f = 10kHz 4 pA/√Hz
Input Resistance VCM = 0V to V+ – 2V 1 MΩ
Input Capacitance 2 pF
CMRR Common Mode Rejection Ratio VCM = 0V to V+ – 2V l 78 90 dB
Input Voltage Range l 0 V+ – 2 V
PSRR Power Supply Rejection Ratio VS = 3V to 12VVCM = VOUT = 0.5V
l 67 75 dB
Minimum Supply Voltage VCM = 0.5V l 2.7 V
AVOL Large-Signal Voltage Gain VS = 5V, VO = 0.5V to 4.5V, RL = 1kVS = 5V, VO = 1V to 3V, RL = 150ΩVS = 3V, VO = 0.5V to 2.5V, RL = 1k
l
l
l
305
20
1002060
V/mVV/mVV/mV
VOL Output Voltage Swing Low (Note 7) No Load, Input Overdrive = 30mVISINK = 5mAVS = 5V, ISINK = 25mAVS = 3V, ISINK = 15mA
l
l
l
l
1075300200
25150500350
mVmVmVmV
VOH Output Voltage Swing High (Note 7) No Load, Input Overdrive = 30mVISOURCE = 5mAVS = 5V, ISOURCE = 25mAVS = 3V, ISOURCE = 15mA
l
l
l
l
60150650300
1002501200500
mVmVmVmV
ISC Short-Circuit Current VS = 5V, Output Shorted to GNDl
3520
60 mAmA
VS = 3V, Output Shorted to GNDl
3020
50 mAmA
IS Supply Current per Amplifi erl
3.75 55.75
mAmA
GBW Gain Bandwidth Product f = 2MHz l 65 100 MHz
SR Slew Rate VS = 5V, AV = 2, RF = RG = 1kVO = 1V to 4V, Measured from 1.5V to 3.5V
450 V/μs
Channel Separation f = 10MHz 90 dB
FPBW Full Power Bandwidth VOUT = 2VP-P (Note 8) 71 MHz
ts Settling Time to 3%Settling Time to 1%
VS = 5V, ΔVOUT = 2V, AV = –1, RL = 150Ω 1525
nsns
Differential GainDifferential Phase
VS = 5V, AV = 2, RL = 150Ω, Output Black Level = 1VVS = 5V, AV = 2, RL = 150Ω, Output Black Level = 1V
0.050.08
%Deg
The l denotes specifi cations which apply over the specifi ed temperature range, otherwise specifi cations are at TA = 25°C. VS = 3V, 0V; VS = 5V, 0V; VCM = VOUT = 1V, unless otherwise noted.
A. Operational Amplifier LT6205/06/07 228
LT6205/LT6206/LT6207
4620567fc
ELECTRICAL CHARACTERISTICS
SYMBOL PARAMETER CONDITIONS
LT6205C/LT6206C/LT6207CLT6205I/LT6206I/LT6207I
UNITSMIN TYP MAX
VOS Input Offset Voltagel
1 4.56
mVmV
Input Offset Voltage Match(Channel-to-Channel) (Note 5) l
1 34
mVmV
Input Offset Voltage Drift (Note 6) l 10 18 μV/°C
IB Input Bias Current l 18 30 μA
IOS Input Offset Current l 0.6 3 μA
Input Noise Voltage 0.1Hz to 10Hz 2 μVP-P
en Input Noise Voltage Density f = 10kHz 9 nV/√Hz
in Input Noise Current Density f = 10kHz 4 pA/√Hz
Input Resistance VCM = –5V to 3V 1 MΩ
Input Capacitance 2 pF
CMRR Common Mode Rejection Ratio VCM = –5V to 3V l 78 90 dB
Input Voltage Range l –5 3 V
PSRR Power Supply Rejection Ratio VS = ±2V to ±6V l 67 75 dB
AVOL Large-Signal Voltage Gain VO = –4V to 4V, RL = 1k l 50 133 V/mV
VO = –3V to 3V, RL = 150Ω l 7.5 20 V/mV
Output Voltage Swing No Load, Input Overdrive = 30mVIOUT = ±5mAIOUT = ±25mA
l
l
l
±4.88±4.75±3.8
±4.92±4.85±4.35
VVV
ISC Short-Circuit Current Short to Groundl
±40±30
±60 mAmA
IS Supply Current per Amplifi erl
4 5.66.5
mAmA
GBW Gain Bandwidth Product f = 2MHz l 65 100 MHz
SR Slew Rate AV = –1, RL = 1kVO = –4V to 4V, Measured from –3V to 3V
350 600 V/μs
Channel Separation f = 10MHz 90 dB
FPBW Full Power Bandwidth VOUT = 8VP-P (Note 8) 14 24 MHz
ts Settling Time to 3%Settling Time to 1%
ΔVOUT = 2V, AV = –1, RL = 150Ω 1525
nsns
Differential GainDifferential Phase
AV = 2, RL = 150Ω, Output Black Level = 1VAV = 2, RL = 150Ω, Output Black Level = 1V
0.050.08
%Deg
The l denotes specifi cations which apply over the specifi ed temperature range, otherwise specifi cations are at TA = 25°C. VS = ±5V; VCM = VOUT = 0V, unless otherwise noted.
The l denotes specifi cations which apply over the full specifi ed temperature range, –40°C ≤ TA ≤ 125°C, otherwise specifi cations are at TA = 25°C. VS = 3V, 0V; VS = 5V, 0V; VCM = VOUT = 1V, unless otherwise noted.
SYMBOL PARAMETER CONDITIONS
LT6205H
UNITSMIN TYP MAX
VOS Input Offset Voltagel
1 3.56
mVmV
Input Offset Voltage Drift (Note 6) l 20 μV/°C
IB Input Bias Current l 45 μA
A. Operational Amplifier LT6205/06/07 229
LT6205/LT6206/LT6207
5620567fc
ELECTRICAL CHARACTERISTICS
SYMBOL PARAMETER CONDITIONS
LT6205H
UNITSMIN TYP MAX
IOS Input Offset Current l 5 μA
Input Noise Voltage 0.1Hz to 10Hz 2 μVP-P
en Input Noise Voltage Density f = 10kHz 9 nV/√Hz
in Input Noise Current Density f = 10kHz 4 pA/√Hz
Input Resistance VCM = 0V to V+ – 2V 1 MΩ
Input Capacitance 2 pF
CMRR Common Mode Rejection Ratio VCM = 0V to V+ – 2V l 72 dB
Input Voltage Range l 0 V+ – 2 V
PSRR Power Supply Rejection Ratio VS = 3V to 12VVCM = VOUT = 0.5V
l 62 dB
Minimum Supply Voltage VCM = 0.5V l 2.7 V
AVOL Large-Signal Voltage Gain VS = 5V, V0 = 0.5V to 4.5V, RL = 1kVS = 5V, V0 = 1V to 3V, RL = 150ΩVS = 3V, V0 = 0.5V to 2.5V, RL = 1k
l
l
l
253.515
V/mVV/mVV/mV
VOL Output Voltage Swing Low (Note 7) No Load, Input Overdrive = 30mVISINK = 5mAVS = 5V, ISINK = 25mAVS = 3V, ISINK = 15mA
l
l
l
l
40200600400
mVmVmVmV
VOH Output Voltage Swing High (Note 7) No Load, Input Overdrive = 30mVISOURCE = 5mAVS = 5V, ISOURCE = 25mAVS = 3V, ISOURCE = 15mA
l
l
l
l
1253001400600
mVmVmVmV
ISC Short-Circuit Current VS = 5V, Output Shorted to GNDl
3520
60 mAmA
VS = 3V, Output Shorted to GNDl
3015
50 mAmA
IS Supply Current per Amplifi erl
3.75 56.5
mAmA
GBW Gain Bandwidth Product f = 2MHz l 50 MHz
SR Slew Rate VS = 5V, AV = 2, RF = RG = 1kVO = 1V to 4V, Measured from 1.5V to 3.5V
450 V/μs
Channel Separation f = 10MHz 90 dB
FPBW Full Power Bandwidth VOUT = 2VP-P (Note 8) 71 MHz
ts Settling Time to 3%Settling Time to 1%
VS = 5V, ΔVOUT = 2V, AV = –1, RL = 150Ω 1525
nsns
Differential GainDifferential Phase
VS = 5V, AV = 2, RL = 150Ω, Output Black Level = 1VVS = 5V, AV = 2, RL = 150Ω, Output Black Level = 1V
0.050.08
%Deg
The l denotes specifi cations which apply over the full specifi ed temperature range, –40°C ≤ TA ≤ 125°C, otherwise specifi cations are at TA = 25°C. VS = 3V, 0V; VS = 5V, 0V; VCM = VOUT = 1V, unless otherwise noted.
The l denotes specifi cations which apply over the full specifi ed temperature range, –40°C ≤ TA ≤ 125°C, otherwise specifi cations are at TA = 25°C. VS = ±5V; VCM = VOUT = 0V, unless otherwise noted.
SYMBOL PARAMETER CONDITIONS
LT6205H
UNITSMIN TYP MAX
VOS Input Offset Voltagel
1.3 4.57
mVmV
Input Offset Voltage Drift (Note 6) l 25 μV/°C
A. Operational Amplifier LT6205/06/07 230
LT6205/LT6206/LT6207
6620567fc
ELECTRICAL CHARACTERISTICS
Note 1: Stresses beyond those listed under Absolute Maximum Ratings
may cause permanent damage to the device. Exposure to any Absolute
Maximum Rating condition for extended periods may affect device
reliability and lifetime.
Note 2: The inputs are protected by back-to-back diodes. If the differential
input voltage exceeds 1.4V, the input current should be limited to less than
10mA.
Note 3: A heat sink may be required to keep the junction temperature
below absolute maximum. This depends on the power supply voltage and
how many amplifi ers are shorted.
Note 4: The LT6205C/LT6206C/LT6207C are guaranteed to meet specifi ed
performance from 0°C to 70°C and are designed, characterized and
expected to meet specifi ed performance from –40°C to 85°C but are not
tested or QA sampled at these temperatures. The LT6205I/LT6206I/LT6207I
are guaranteed to meet specifi ed performance from –40°C to 85°C. The
LT6205H is guaranteed to meet specifi ed performance from –40°C to 125°C.
Note 5: Matching parameters are the difference between the two amplifi ers
A and D and between B and C of the LT6207; between the two amplifi ers of
the LT6206.
Note 6: This parameter is not 100% tested.
Note 7: Output voltage swings are measured between the output and
power supply rails.
Note 8: Full power bandwidth is calculated from the slew rate
measurement: FPBW = SR/2πVPEAK.
Note 9: There are reverse biased ESD diodes on all inputs and outputs.
If these pins are forced beyond either supply, unlimited current will fl ow
through these diodes. If the current is transient in nature and limited to
less than 25mA, no damage to the device will occur.
SYMBOL PARAMETER CONDITIONS
LT6205H
UNITSMIN TYP MAX
IB Input Bias Current l 50 μA
IOS Input Offset Current l 5 μA
Input Noise Voltage 0.1Hz to 10Hz 2 μVP-P
en Input Noise Voltage Density f = 10kHz 9 nV/√Hz
in Input Noise Current Density f = 10kHz 4 pA/√Hz
Input Resistance VCM = –5V to 3V 1 MΩ
Input Capacitance 2 pF
CMRR Common Mode Rejection Ratio VCM = –5V to 3V l 72 dB
Input Voltage Range l –5 3 V
PSRR Power Supply Rejection Ratio VS = ±2V to ±6V l 62 dB
AVOL Large-Signal Voltage Gain VO = –4V to 4V, RL = 1k l 40 V/mV
VO = –3V to 3V, RL = 150Ω l 5 V/mV
Output Voltage Swing No Load, Input Overdrive = 30mVIOUT = ±5mAIOUT = ±25mA
l
l
l
±4.85±4.65±3.5
VVV
ISC Short-Circuit Current Short to Groundl
±40±20
±60 mAmA
IS Supply Current per Amplifi erl
4 5.67.5
mAmA
GBW Gain Bandwidth Product f = 2MHz l 50 MHz
SR Slew Rate AV = –1, RL = 1kVO = –4V to 4V, Measured from –3V to 3V
350 600 V/μs
Channel Separation f = 10MHz 90 dB
FPBW Full Power Bandwidth VOUT = 8VP-P (Note 8) 14 24 MHz
ts Settling Time to 3%Settling Time to 1%
ΔVOUT = 2V, AV = –1, RL = 150Ω 1525
nsns
Differential GainDifferential Phase
AV = 2, RL = 150Ω, Output Black Level = 1VAV = 2, RL = 150Ω, Output Black Level = 1V
0.050.08
%Deg
The l denotes specifi cations which apply over the full specifi ed temperature range, –40°C ≤ TA ≤ 125°C, otherwise specifi cations are at TA = 25°C. VS = ±5V; VCM = VOUT = 0V, unless otherwise noted.
A. Operational Amplifier LT6205/06/07 231
LT6205/LT6206/LT6207
7620567fc
TYPICAL PERFORMANCE CHARACTERISTICS
VOS DistributionSupply Current per Amplifi ervs Supply Voltage Minimum Supply Voltage
Change in Offset Voltage vs Input Common Mode Voltage
Input Bias Current vs Input Common Mode Voltage Input Bias Current vs Temperature
Output Saturation Voltage vs Load Current (Output Low)
Output Saturation Voltage vs Load Current (Output High)
Short-Circuit Current vs Temperature
INPUT OFFSET VOLTAGE (mV)
–3 –2 –1 0 1 2 3
PE
RC
EN
T O
F U
NIT
S (
%)
620567 G01
40
30
35
25
20
15
10
5
0
VS = 5V, 0VVCM = 1V
TOTAL SUPPLY VOLTAGE (V)
0 1 2 3 4 5 6 7 8 9 10 11 12
SU
PP
LY C
UR
RE
NT
PE
R A
MP
LIF
IER
(m
A)
620567 G02
5
4
3
2
1
0
TA = 25°C
TA = –55°C
TA = 125°C
TOTAL SUPPLY VOLTAGE (V)
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
CH
AN
GE
IN
IN
PU
T O
FFS
ET
VO
LTA
GE
(μV
)620567 G03
100
0
–100
–200
–300
–400
–500
–600
TA = 25°C
TA = –55°C
TA =125°C
INPUT COMMON MODE VOLTAGE (V)
0 1 2 3 4 5
OFF
SE
T V
OLT
AG
E C
HA
NG
E (
μV
)
620567 G04
1000
800
600
400
200
0
TA = 25°C
TA = –55°C
TA =125°C
VS = 5V, 0V
INPUT COMMON MODE VOLTAGE (V)
0 1 2 3 4 5
INP
UT
BIA
S C
UR
RE
NT
(μ
A)
–2
–3
–4
–5
–6
–7
–9
–12
–8
–11
–10
VS = 5V, 0V
TA = 25°C
TA = –55°C
TA = 125°C
620567 G05
TEMPERATURE (°C)
–50 –25 0 25 50 75 100 125
IN
PU
T B
IAS
CU
RR
EN
T (
μA
)
620567 G06
–4
–6
–5
–7
–8
–9
–10
–11
–12
VS = 5V, 0VVCM = 1V
LOAD CURRENT (mA)
0.1
OU
TP
UT
SA
TU
RA
TIO
N V
OLT
AG
E (
V)
1
0.01 1 10 100
620567 G07
0.010.1
10VS = 5V, 0VVOD = 30mV
TA = 25°C
TA = –55°C
TA = 125°C
LOAD CURRENT (mA)
0.1
OU
TP
UT
SA
TU
RA
TIO
N V
OLT
AG
E (
V)
1
0.01 1 10 100
620567 G08
0.010.1
10
TA = 25°C
TA = –55°C
TA = 125°C
VS = 5V, 0VVOD = 30mV
TEMPERATURE (°C)
–50 –25 0 25 50 75 100 125
OU
TP
UT
SH
OR
T-C
IRC
UIT
CU
RR
EN
T (
mA
)
620567 G09
75
65
70
60
55
50
45
40
35
VS = 5V, 0VVCM = 1V
VS = 3V, 0VVCM = 1V
SINKING
SOURCING
SINKING
SOURCING
A. Operational Amplifier LT6205/06/07 232
LT6205/LT6206/LT6207
8620567fc
TYPICAL PERFORMANCE CHARACTERISTICS
Short-Circuit Current vs Temperature Open-Loop Gain Open-Loop Gain
Warm Up Drift vs Time (LT6206)Input Noise Voltage Density vs Frequency
Input Noise Current Density vs Frequency
0.1Hz to 10Hz Noise Voltage Gain and Phase vs FrequencyGain Bandwidth and Phase Margin vs Supply Voltage
TEMPERATURE (°C)
–50 –25 0 25 50 75 100 125
OU
TP
UT
SH
OR
T-C
IRC
UIT
CU
RR
EN
T (
mA
)
620567 G10
90
80
70
60
50
40
3O
VS = 5V
SINKING
SOURCING
OUTPUT VOLTAGE (V)
0 1.0 2.0 3.0 4.00.5 1.5 2.5 3.5 4.5 5.0
INP
UT
VO
LTA
GE
(μV
)
500
400
300
200
100
0
–200
–500
–100
–400
–300
RL = 150Ω
RL = 1k
620567 G11
VS = 5V, 0VVCM = 1VTA = 25°C
OUTPUT VOLTAGE (V)
–5 –3 –1 1 3–4 –2 0 2 4 5
INP
UT
VO
LTA
GE
( μ
V)
500
400
300
200
100
0
–200
–500
–100
–400
–300
RL = 150Ω
RL = 1k
620567 G12
VS = 5VTA = 25°C
TIME AFTER POWER-UP (s)
0 20 40 60 8010 30 50 70 90 100
CH
AN
GE
IN
OFF
SE
T V
OLT
AG
E (
μV
)
120
100
80
40
0
60
20
620567 G13
VS = ±5V
VS = 5V, 0V
TA = 25°C
FREQUENCY (Hz)
100 1k 10k 100k
INP
UT
NO
ISE
VO
LTA
GE
DE
NS
ITY
(n
V/
Hz)
30
25
20
15
10
5
0
620567 G14
VS = 5V, 0VVCM = 1VTA = 25°C
FREQUENCY (Hz)
100 1k 10k 100k
16
14
12
10
8
6
4
2
0
620567 G15
VS = 5V, 0VVCM = 1VTA = 25°C
INP
UT
NO
ISE
CU
RR
EN
T D
EN
SIT
Y (
pA
/H
z)
TIME (2 SEC/DIV)
NO
ISE
VO
LTA
GE
(1μ
V/D
IV)
620567 G16
VS = 5V, 0VVCM = 1VTA = 25°C
FREQUENCY (Hz)
GA
IN (
dB
)
70
60
50
40
30
20
10
0
–10
–20
PH
AS
E (D
EG
)
140
120
100
80
60
40
20
0
-20
-40100k 10M 100M 500M
620567 G17
1M
PHASE
GAIN
VS = 3V, 0V
TA = 25°CRL = 1kCL = 5pF
VS = 5V
VS = 5V
VS = 3V, 0V
TOTAL SUPPLY VOLTAGE (V)
0 2 4 6 8 10 12
GA
IN B
AN
DW
IDT
H (
MH
z)
620567 G18
110
105
100
95
TA = 25°CRF = RG = 1kCL = 5pF
PHASE MARGIN
GAIN BANDWIDTH
PH
AS
E M
AR
GIN
(DE
G)
50
45
40
35
A. Operational Amplifier LT6205/06/07 233
LT6205/LT6206/LT6207
9620567fc
TYPICAL PERFORMANCE CHARACTERISTICS
Gain Bandwidth and Phase Margin vs Temperature Slew Rate vs Temperature Slew Rate vs Closed-Loop Gain
Closed-Loop Gain vs Frequency Output Impedance vs FrequencyPower Supply Rejection Ratio vs Frequency
Common Mode Rejection Ratio vs Frequency Channel Separation vs Frequency
Series Output Resistor vs Capacitive Load
TEMPERATURE (°C)
–50 –25 0 25 50 75 125100
GA
IN B
AN
DW
IDT
H (
MH
z)
620567 G19
120
110
100
90
80
PH
AS
E M
AR
GIN
(DE
G)
55
50
45
40
35
GAIN BANDWIDTH
PHASE MARGIN
VS = 5V
VS = 5V
VS = 3V, 0V
VS = 3V, 0V
RL = 1kCL = 5pF
TEMPERATURE (°C)
–50 –25 0 25 50 75 125100
SL
EW
RA
TE
(V
/μs)
620567 G20
750
650
600
700
550
500
450
400
350
RISING VS = 5V
AV = –1RG = RF = 1kRL = 1k
RISING VS = 5V, 0V
FALLING VS = 5V, 0V
FALLING VS = 5V
GAIN (AV)
2 3 4 5
SL
EW
RA
TE
(V
/μs)
620567 G21
750
650
600
700
550
500
400
450
RISING
VS = 5VVO = –4V to 4VRL = 1kTA = 25°C
FALLING
FREQUENCY (Hz)
GA
IN (
dB
)
15
12
9
6
3
0
–3
–6
–9
–15
–12
100k 10M 100M 500M
620567 G22
1M
VS = 5VVCM = 0V
VS = 3VVCM = 1V
TA = 25°CCL = 5pFAV = +1
FREQUENCY (Hz)
100k0.1
OU
TP
UT
IM
PE
DA
NC
E (
Ω)
10
1000
10M 100M1M 500M
620567 G23
1
100
VS = 5V, 0VTA = 25°C
AV = 10
AV = 2AV = 1
FREQUENCY (Hz)
PO
WE
R S
UP
PLY
RE
JEC
TIO
N R
AT
IO (
dB
)
90
80
70
60
50
40
30
20
10
010k 1M 10M 100M
620567 G24
100k
VS = 5V, 0VTA = 25°C
–PSRR+PSRR
FREQUENCY (Hz)
CO
MM
ON
MO
DE
RE
JEC
TIO
N R
AT
IO (
dB
)
10k 1M 10M 1G
620567 G25
100k
100
90
80
70
60
50
40
30
20
10
0100M
VS = 5VTA = 25°C
FREQUENCY (Hz)
1M
120
110
100
90
80
70
60
50
40
VO
LTA
GE
GA
IN (
dB
)
10M 100M
620567 G26
VS = 5VLT6206 CH A-BLT6207 CH A-D, CH B-CTA = 25°C
CAPACITIVE LOAD (pF)
10
40
35
30
25
20
15
10
5
0
OV
ER
SH
OO
T (
%)
100 1000
620567 G27
VS = 5V, 0VAV = 1TA = 25°C RS = 10Ω, RL =
RS = 20Ω, RL =
RL = RS = 50Ω
A. Operational Amplifier LT6205/06/07 234
LT6205/LT6206/LT6207
10620567fc
TYPICAL PERFORMANCE CHARACTERISTICS
Series Output Resistor vs Capacitive Load
Maximum Undistorted Output Signal vs Frequency Distortion vs Frequency
Distortion vs Frequency Distortion vs Frequency Distortion vs Frequency
Large Signal Response VS = 5V, 0V
Small Signal Response VS = 5V, 0V
CAPACITIVE LOAD (pF)
10
40
35
30
25
20
15
10
5
0
OV
ER
SH
OO
T (
%)
100 1000
620567 G28
VS = 5V, 0VAV = 2TA = 25°C RS = 10Ω, RL =
RS = 20Ω, RL =
RL = RS = 50Ω
FREQUENCY (MHz)
0.1 1 10 100
OU
TP
UT
VO
LTA
GE
SW
ING
(V
P-P
)
10
9
8
7
6
5
4
0
3
2
1
620567 G30
VS = 5V
TA = 25°CHD2, HD3 < –30dBc
AV = 2
AV = –1
FREQUENCY (MHz)
0.01 0.1 1 10
DIS
TO
RT
ION
(d
B)
–30
–40
–50
–60
–70
–80
–90
–100
620567 G31
AV = +1VO = 2VP-P
VS = 5V, 0V
RL = 1k, 2ND
RL = 1k, 3RD
RL = 150Ω, 3RD
RL = 150Ω, 2ND
FREQUENCY (MHz)
0.01 0.1 1 10
DIS
TO
RT
ION
(d
B)
–30
–40
–50
–60
–70
–80
–90
–100
620567 G32
AV = +2VO = 2VP-P
VS = 5V, 0V
RL = 1k, 2ND
RL = 1k, 3RD
RL = 150Ω, 3RD
RL = 150Ω, 2ND
FREQUENCY (MHz)
0.01 0.1 1 10
DIS
TO
RT
ION
(d
B)
–30
–40
–50
–60
–70
–80
–90
–100
620567 G33
AV = +1VO = 2VP-P
VS = 5V
RL = 1k, 2ND RL = 1k, 3RD
RL = 150Ω, 3RD
RL = 150Ω, 2ND
FREQUENCY (MHz)
0.01 0.1 1 10
DIS
TO
RT
ION
(d
B)
–30
–40
–50
–60
–70
–80
–90
–100
620567 G34
AV = +2VO = 2VP-P
VS = 5V
RL = 1k, 3RD
RL = 150Ω, 3RD
RL = 150Ω, 2ND
RL = 1k, 2ND
50
0m
V/D
IV
50ns/DIV 620567 G35
0V
VS = 5V, 0VAV = 1RL = 150Ω
50
0m
V/D
IV
50ns/DIV 620567 G36
2.5V
VS = 5V, 0VAV = 1RL = 150Ω
A. Operational Amplifier LT6205/06/07 235
LT6205/LT6206/LT6207
11620567fc
TYPICAL PERFORMANCE CHARACTERISTICS
Large Signal Response VS = ±5V Small Signal Response VS = ±5V Output-Overdrive Recovery
1V
/DIV
50ns/DIV 620567 G37
0V
VS = ±5VAV = 1RL = 150Ω
50
mV
/DIV
50ns/DIV 620567 G38
0V
VS = ±5VAV = 1RL = 150Ω
VIN
(1V
/DIV
VO
UT
(2V
/DIV
100ns/DIV 620567 G39
0V
0V
VS = 5V, 0VAV = 2
APPLICATIONS INFORMATION
Figure 1. Simplifi ed Schematic
Q7
Q9 Q10
Q11 Q12
Q8
Q5
Q6
Q3
Q13
Q14
Q4
Q2
Q1
D3
D4
D1
D2
RIN150Ω
RIN150Ω
DESD1
DESD2
DESD5
DESD6
DESD3
DESD4
+IN
–IN
I1 I2 I3 R2
R1
R3
I4 R4 R5
COMPLEMENTARYDRIVE
GENERATOR
CM
V+
V–
V+
V–
V+
V–
V+
V–
OUT
620567 F01
A. Operational Amplifier LT6205/06/07 236
LT6205/LT6206/LT6207
12620567fc
APPLICATIONS INFORMATIONAmplifi er Characteristics
Figure 1 shows a simplifi ed schematic of the LT6205/LT6206/LT6207. The input stage consists of transistors Q1 to Q8 and resistor R1. This topology allows for high slew rates at low supply voltages. The input common mode range extends from ground to typically 1.75V from VCC, and is limited by 2 VBEs plus a saturation voltage of a current source. There are back-to-back series diodes, D1 to D4, across the + and – inputs of each amplifi er to limit the differential voltage to ±1.4V. RIN limits the current through these diodes if the input differential voltage exceeds ±1.4V. The input stage drives the degeneration resistors of PNP and NPN current mirrors, Q9 to Q12, which convert the differential signals into a single-ended output. The complementary drive generator supplies current to the output transistors that swing from rail-to-rail.
The current generated through R1, divided by the capacitor CM, determines the slew rate. Note that this current, and hence the slew rate, are proportional to the magnitude of the input step. The input step equals the output step divided by the closed loop gain. The highest slew rates are therefore obtained in the lowest gain confi gurations. The Typical Performance Characteristics curve of Slew Rate vs Closed-Loop Gain shows the details.
ESD
The LT6205/LT6206/LT6207 have reverse-biased ESD protection diodes on all inputs and outputs as shown in Figure 1. If these pins are forced beyond either supply unlimited current will fl ow through these diodes. If the current is transient, and limited to 25mA or less, no dam-age to the device will occur.
Layout and Passive Components
With a gain bandwidth product of 100MHz and a slew rate of 450V/μs the LT6205/LT6206/LT6207 require special attention to board layout and supply bypassing. Use a ground plane, short lead lengths and RF quality low ESR supply bypass capacitors. The positive supply pin should be bypassed with a small capacitor (typically 0.01μF to 0.1μF) within 0.25 inches of the pin. When driving heavy loads, an additional 4.7μF electrolytic capacitor should be used. When using split supplies, the same is true for the
negative supply pin. For optimum performance all feedback components and bypass capacitors should be contained in a 0.5 inch by 0.5 inch area. This helps ensure minimal stray capacitances.
The parallel combination of the feedback resistor and gain setting resistor on the inverting input can combine with the input capacitance to form a pole which can degrade stability. In general, use feedback resistors of 1k or less.
Capacitive Load
The LT6205/LT6206/LT6207 are optimized for wide band-width video applications. They can drive a capacitive load of 20pF in a unity-gain confi guration. When driving a larger capacitive load, a resistor of 10Ω to 50Ω should be connected between the output and the capacitive load to avoid ringing or oscillation. The feedback should still be taken from the output pin so that the resistor will isolate the capacitive load and ensure stability. The Typi-cal Performance Characteristics curves show the output overshoot when driving a capacitive load with different series resistors.
Video Signal Characteristics
Composite video is the most commonly used signal in broadcast grade products and includes luma (or lumi-nance, the intensity information), chroma (the colorimetry information) and sync (vertical and horizontal raster tim-ing) elements combined into a single signal, NTSC and PAL being the common formats. Component video for entertainment systems include separate signal(s) for the luma and chroma (i.e., Y/C or YPbPr) with sync generally applied to the luma channel (Y signal). In some instances, native RGB signals (separate intensity information for each primary color: red, green, blue) will have sync included as well. All the signal types that include sync are electrically similar from a voltage-swing standpoint, though various timing and bandwidth relationships exist depending on the applicable standard.
The typical video waveforms that include sync (includ-ing full composite) are specifi ed to have nominal 1VP-P amplitude. The lower 0.3V is reserved for sync tips that carry timing information, and by being at a lower potential than all the other information, represents blacker-than-
A. Operational Amplifier LT6205/06/07 237
LT6205/LT6206/LT6207
13620567fc
APPLICATIONS INFORMATIONblack intensity, thereby causing scan retrace activity to be invisible on a CRT. The black level of the waveform is at (or set up very slightly above) the upper limit of the sync information. Waveform content above the black level is intensity information, with peak brightness represented at the maximum signal level. In the case of composite video, the modulated color subcarrier is superimposed on the waveform, but the dynamics remain inside the 1VP-P limit (a notable exception is the chroma ramp used for differential-gain and differential-phase measurements, which can reach 1.15VP-P).
DC-Coupled Video Amplifi er Considerations
Typically video amplifi ers drive cables that are series terminated (back-terminated) at the source and load-ter-minated at the destination with resistances equal to the cable characteristic impedance, Z0 (usually 75Ω). This confi guration forms a 2:1 resistor divider in the cabling that must be accounted for in the driver amplifi er by delivering 2VP-P output into an effective 2 • Z0 load (e.g., 150Ω). Driving the cable can require more than 13mA while the output is approaching the saturation limits of the amplifi er output. The absolute minimum supply is: VMIN = 2 + VOH +VOL. For example, the LT6206 dual operating on 3.3V as shown on the front page of this data sheet, with exceptionally low VOH ≤ 0.5V and VOL ≤ 0.35V, provides a design margin of 0.45V. The design margin must be large enough to include supply variations and DC bias accuracy for the DC-coupled video input.
Handling AC-Coupled Video Signals
AC-coupled video inputs are intrinsically more diffi cult to handle than those with DC-coupling because the average signal voltage of the video waveform is effected by the picture content, meaning that the black level at the amplifi er wanders with scene brightness. The wander is measured as 0.56V for a 1VP-P NTSC waveform changing from black fi eld to white fi eld and vice-versa, so an additional 1.12V allowance must be made in the amplifi er supply (assum-ing gain of 2, so VMIN = 3.12 + VOH +VOL). For example, an LT6205 operating on 5V has a conservative design margin of 1.03V. The amplifi er output (for gain of 2) must swing +1.47V to –1.65V around the DC-operating point, so the biasing circuitry needs to be designed accordingly for optimal fi delity.
Clamped AC-Input Cable Driver
A popular method of further minimizing supply require-ments with AC-coupling is to employ a simple clamping scheme, as shown in Figure 2. In this circuit, the LT6205 operates from 3.3V by having the sync tips control the charge on the coupling capacitor C1, thereby reducing the black level input wander to ≈ 0.07V. The only minor drawback to this circuit is the slight sync tip compression (≈ 0.025V at input) due to the diode conduction current, though the picture content remains full fi delity. This circuit has nearly the design margin of its DC-coupled counter-part, at 0.31V (for this circuit, VMIN = 2.14 + VOH +VOL). The clamp diode anode bias is selected to set the sync tip output voltage at or slightly above VOL.
YPbPr to RGB Component Video Converter
The back page application uses the LT6207 quad to imple-ment a minimum amplifi er count topology to transcode consumer component video into RGB. In this circuit, signals only pass through one active stage from any input to any output, with passive additions being performed by the cable back-termination resistors. The compromise in using passive output addition is that the amplifi er outputs must be twice as large as that of a conventional cable driver. The Y-channel section also has the demanding requirement that it single-handedly drives all three outputs to full brightness during times of white content, so a helper current source is used to assure unclipped video when operating from ±5V supplies. This circuit maps sync-on-Y to sync on all the RGB channels, and for best results should have input black levels at 0V nominal to prevent clipping.
A. Operational Amplifier LT6205/06/07 238
LT6205/LT6206/LT6207
14620567fc
TYPICAL APPLICATION
Figure 2. Clamped AC-Input Video Cable Driver
–
+
LT6205COMPOSITE
VIDEO IN 1VP-P
0.1μF
C1
4.7μF
C2
4.7μF
BAT54
75Ω
10k
75Ω
VIDEO OUT
3.3V
1k 1k
2.4k
1
2
5
3
4
IS 19mA
620567 TA02
470Ω
A. Operational Amplifier LT6205/06/07 239
LT6205/LT6206/LT6207
15620567fc
PACKAGE DESCRIPTION
1.50 – 1.75(NOTE 4)
2.80 BSC
0.30 – 0.45 TYP 5 PLCS (NOTE 3)
DATUM ‘A’
0.09 – 0.20(NOTE 3) S5 TSOT-23 0302
PIN ONE
2.90 BSC(NOTE 4)
0.95 BSC
1.90 BSC
0.80 – 0.90
1.00 MAX0.01 – 0.10
0.20 BSC
0.30 – 0.50 REF
NOTE:1. DIMENSIONS ARE IN MILLIMETERS2. DRAWING NOT TO SCALE3. DIMENSIONS ARE INCLUSIVE OF PLATING4. DIMENSIONS ARE EXCLUSIVE OF MOLD FLASH AND METAL BURR5. MOLD FLASH SHALL NOT EXCEED 0.254mm6. JEDEC PACKAGE REFERENCE IS MO-193
3.85 MAX
0.62MAX
0.95REF
RECOMMENDED SOLDER PAD LAYOUTPER IPC CALCULATOR
1.4 MIN2.62 REF
1.22 REF
S5 Package5-Lead Plastic TSOT-23
(Reference LTC DWG # 05-08-1635)
A. Operational Amplifier LT6205/06/07 240
LT6205/LT6206/LT6207
16620567fc
PACKAGE DESCRIPTIONMS8 Package
8-Lead Plastic MSOP(Reference LTC DWG # 05-08-1660)
MSOP (MS8) 0307 REV F
0.53 ± 0.152(.021 ± .006)
SEATINGPLANE
NOTE:1. DIMENSIONS IN MILLIMETER/(INCH)2. DRAWING NOT TO SCALE3. DIMENSION DOES NOT INCLUDE MOLD FLASH, PROTRUSIONS OR GATE BURRS. MOLD FLASH, PROTRUSIONS OR GATE BURRS SHALL NOT EXCEED 0.152mm (.006") PER SIDE4. DIMENSION DOES NOT INCLUDE INTERLEAD FLASH OR PROTRUSIONS. INTERLEAD FLASH OR PROTRUSIONS SHALL NOT EXCEED 0.152mm (.006") PER SIDE5. LEAD COPLANARITY (BOTTOM OF LEADS AFTER FORMING) SHALL BE 0.102mm (.004") MAX
0.18(.007)
0.254(.010)
1.10(.043)MAX
0.22 – 0.38(.009 – .015)
TYP
0.1016 ± 0.0508(.004 ± .002)
0.86(.034)REF
0.65(.0256)
BSC
0° – 6° TYP
DETAIL “A”
DETAIL “A”
GAUGE PLANE
1 2 3 4
4.90 ± 0.152(.193 ± .006)
8 7 6 5
3.00 ± 0.102(.118 ± .004)
(NOTE 3)
3.00 ± 0.102(.118 ± .004)
(NOTE 4)
0.52(.0205)
REF
5.23(.206)MIN
3.20 – 3.45(.126 – .136)
0.889 ± 0.127(.035 ± .005)
RECOMMENDED SOLDER PAD LAYOUT
0.42 ± 0.038(.0165 ± .0015)
TYP
0.65(.0256)
BSC
GN Package16-Lead Plastic SSOP (Narrow .150 Inch)
(Reference LTC DWG # 05-08-1641)
GN16 (SSOP) 0204
1 2 3 4 5 6 7 8
.229 – .244(5.817 – 6.198)
.150 – .157**(3.810 – 3.988)
16 15 14 13
.189 – .196*(4.801 – 4.978)
12 11 10 9
.016 – .050(0.406 – 1.270)
.015 ± .004(0.38 ± 0.10)
× 45°
0° – 8° TYP.007 – .0098(0.178 – 0.249)
.0532 – .0688(1.35 – 1.75)
.008 – .012(0.203 – 0.305)
TYP
.004 – .0098(0.102 – 0.249)
.0250(0.635)
BSC
.009(0.229)
REF
.254 MIN
RECOMMENDED SOLDER PAD LAYOUT
.150 – .165
.0250 BSC.0165 ± .0015
.045 ±.005
*DIMENSION DOES NOT INCLUDE MOLD FLASH. MOLD FLASH SHALL NOT EXCEED 0.006" (0.152mm) PER SIDE**DIMENSION DOES NOT INCLUDE INTERLEAD FLASH. INTERLEAD FLASH SHALL NOT EXCEED 0.010" (0.254mm) PER SIDE
INCHES(MILLIMETERS)
NOTE:1. CONTROLLING DIMENSION: INCHES
2. DIMENSIONS ARE IN
3. DRAWING NOT TO SCALE
A. Operational Amplifier LT6205/06/07 241
LT6205/LT6206/LT6207
17620567fc
Information furnished by Linear Technology Corporation is believed to be accurate and reliable. However, no responsibility is assumed for its use. Linear Technology Corporation makes no representa-tion that the interconnection of its circuits as described herein will not infringe on existing patent rights.
(Revision history begins at Rev C)
REV DATE DESCRIPTION PAGE NUMBER
C 3/10 C Grade Specifi ed Temperature Range Changed in the Order Information Section 2
REVISION HISTORY
A. Operational Amplifier LT6205/06/07 242
LT6205/LT6206/LT6207
18620567fc
Linear Technology Corporation1630 McCarthy Blvd., Milpitas, CA 95035-7417 (408) 432-1900 FAX: (408) 434-0507 www.linear.com © LINEAR TECHNOLOGY CORPORATION 2003
LT 0310 REV C • PRINTED IN USA
RELATED PARTS
TYPICAL APPLICATION
PART NUMBER DESCRIPTION COMMENTS
LT1253/LT1254 Low Cost Dual and Quad Video Amplifi ers –3dB Bandwidth = 90MHz, Current Feedback
LT1395/LT1396/LT1397 Single Dual Quad 400MHz Current Feedback Amplifi ers 0.1dB Flatness to 100MHz, 80mA Output Drive
LT1675 RGB Multiplexer with Current Feedback Amplifi ers –3dB Bandwidth = 250MHz, 100MHz Pixel Switching
LT1809/LT1810 Single/Dual, 180MHz, Rail-to-Rail Input and Output Amplifi ers 350V/μs Slew Rate, Shutdown, Low Distortion –90dBc at 5MHz
LT6550/LT6551 3.3V Triple and Quad Video Amplifi ers Internal Gain of 2, 110MHz –3dB Bandwidth, Input Common Modes to Ground
LT6552 3.3V Single Supply Video Difference Amplifi er Differential or Single-Ended Gain Block, 600V/μs Slew Rate, Input Common Modes to Ground
YPBPR to RGB Converter
150Ω
499Ω
150Ω
150Ω
150Ω
107Ω
80.6Ω
499Ω
75Ω
75Ω
R
B
150Ω
150Ω
75Ω
G
F3dB 40MHz
IS 60mA
BLACK LEVELS 0V
R = Y + 1.4 • PR
B = Y + 1.8 • PB
G = Y – 0.34 • PB – 0.71 • PR 620567 TA03
LT6207
14
16
15
12
11
10
13
4
6
7
5
3
2
1
1μF
133Ω
75Ω
36Ω
FMMT3906
CMPD6001S
4.7k
95.3Ω174Ω
5V
1μF
–5V
499Ω365Ω
499Ω165Ω
PR
PB
Y
–
+
–
+
–
+
–
+
A. Operational Amplifier LT6205/06/07 243
A. Bipolar Power Transistor MJE18008 244
A.4 Datasheet: Bipolar Power Transistor MJE18008
The datasheet of the bipolar power transistor MJE18008 was downloaded from
http://www.onsemi.com/PowerSolutions/product.do?id=MJE18008, accessed 30 Au-
gust 2012.
© Semiconductor Components Industries, LLC, 2006
April, 2006 − Rev. 61 Publication Order Number:
MJE18008/D
MJE18008, MJF18008Preferred Device
SWITCHMODE
NPN Bipolar Power TransistorFor Switching Power Supply Applications
The MJE/MJF18008 have an applications specific state−of−the−artdie designed for use in 220 V line−operated SWITCHMODE Powersupplies and electronic light ballasts.
Features• Improved Efficiency Due to Low Base Drive Requirements:
♦ High and Flat DC Current Gain hFE♦ Fast Switching♦ No Coil Required in Base Circuit for Turn−Off (No Current Tail)
• Tight Parametric Distributions are Consistent Lot−to−Lot
• Two Package Choices: Standard TO−220 or Isolated TO−220
• MJF18008, Case 221D, is UL Recognized at 3500 VRMS: File#E69369
• Pb−Free Packages are Available*
MAXIMUM RATINGS
Rating Symbol Value Unit
Collector−Emitter Sustaining Voltage VCEO 450 Vdc
Collector−Base Breakdown Voltage VCES 1000 Vdc
Emitter−Base Voltage VEBO 9.0 Vdc
Collector Current − Continuous− Peak (Note 1)
ICICM
8.016
Adc
Base Current − Continuous− Peak (Note 1)
IBIBM
4.08.0
Adc
RMS Isolation Voltage (Note 2)Test No. 1 Per Figure 22aTest No. 1 Per Figure 22bTest No. 1 Per Figure 22c
(for 1 sec, R.H. < 30%, TA = 25C)
VISOL MJF18008450035001500
V
Total Device Dissipation @ TC = 25CMJE18008MJF18008
Derate above 25°C MJE18008MJF18008
PD125451.00.36
WW/C
Operating and Storage Temperature TJ, Tstg −65 to 150 C
THERMAL CHARACTERISTICS
Characteristics Symbol Max Unit
Thermal Resistance, Junction−to−CaseMJE18008MJF18008
RJC1.02.78
C/W
Thermal Resistance, Junction−to−Ambient RJA 62.5 C/W
Maximum Lead Temperature for SolderingPurposes 1/8″ from Case for 5 Seconds
TL 260 C
Maximum ratings are those values beyond which device damage can occur.Maximum ratings applied to the device are individual stress limit values (notnormal operating conditions) and are not valid simultaneously. If these limits areexceeded, device functional operation is not implied, damage may occur andreliability may be affected.1. Pulse Test: Pulse Width = 5 ms, Duty Cycle ≤ 10%.2. Proper strike and creepage distance must be provided.
POWER TRANSISTOR8.0 AMPERES1000 VOLTS
45 and 125 WATTS
TO−220ABCASE 221A−09
STYLE 11
http://onsemi.com
MARKINGDIAGRAMS
23
Preferred devices are recommended choices for future useand best overall value.
G = Pb−Free PackageA = Assembly LocationY = YearWW = Work Week
MJE18008GAYWW
See detailed ordering and shipping information in the packagedimensions section on page 7 of this data sheet.
ORDERING INFORMATION
TO−220 FULLPACKCASE 221D
STYLE 2UL RECOGNIZED
31
2
MJF18008GAYWW
*For additional information on our Pb−Free strategy and soldering details, please download theON Semiconductor Soldering and MountingTechniques Reference Manual, SOLDERRM/D.
A. Bipolar Power Transistor MJE18008 245
MJE18008, MJF18008
http://onsemi.com2
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ELECTRICAL CHARACTERISTICS (TC = 25C unless otherwise specified)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Characteristic ÎÎÎÎÎÎÎÎ
SymbolÎÎÎÎÎÎÎÎ
Min ÎÎÎÎÎÎ
TypÎÎÎÎÎÎÎÎ
Max ÎÎÎÎÎÎ
Unit
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
OFF CHARACTERISTICS
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎCollector−Emitter Sustaining Voltage (IC = 100 mA, L = 25 mH) ÎÎÎÎVCEO(sus)ÎÎÎÎ450 ÎÎÎ− ÎÎÎÎ− ÎÎÎVdcÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎCollector Cutoff Current (VCE = Rated VCEO, IB = 0)
ÎÎÎÎÎÎÎÎICEO
ÎÎÎÎÎÎÎÎ−
ÎÎÎÎÎÎ−
ÎÎÎÎÎÎÎÎ100
ÎÎÎÎÎÎAdcÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Collector Cutoff Current (VCE = Rated VCES, VEB = 0)(TC = 125C)
Collector Cutoff Current (VCE = 800 V, VEB = 0) (TC = 125C)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ICESÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
−−−
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−−
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
100500100
ÎÎÎÎÎÎÎÎÎÎÎÎ
Adc
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎEmitter Cutoff Current (VEB = 9.0 Vdc, IC = 0) ÎÎÎÎIEBO ÎÎÎÎ− ÎÎÎ− ÎÎÎÎ100 ÎÎÎAdcÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎON CHARACTERISTICSÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Base−Emitter Saturation Voltage (IC = 2.0 Adc, IB = 0.2 Adc)Base−Emitter Saturation Voltage (IC = 4.5 Adc, IB = 0.9 Adc)
ÎÎÎÎÎÎÎÎÎÎÎÎ
VBE(sat)ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
0.820.92
ÎÎÎÎÎÎÎÎÎÎÎÎ
1.11.25
ÎÎÎÎÎÎÎÎÎ
Vdc
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Collector−Emitter Saturation Voltage(IC = 2.0 Adc, IB = 0.2 Adc)
(TC = 125C)(IC = 4.5 Adc, IB = 0.9 Adc)
(TC = 125C)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
VCE(sat)ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
−−−−
ÎÎÎÎÎÎÎÎÎÎÎÎ
0.30.30.350.4
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
0.60.650.70.8
ÎÎÎÎÎÎÎÎÎÎÎÎ
Vdc
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
DC Current Gain (IC = 1.0 Adc, VCE = 5.0 Vdc)(TC = 125C)
DC Current Gain (IC = 4.5 Adc, VCE = 1.0 Vdc)(TC = 125C)
DC Current Gain (IC = 2.0 Adc, VCE = 1.0 Vdc)(TC = 125C)
DC Current Gain (IC = 10 mAdc, VCE = 5.0 Vdc)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
hFEÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
14−
6.05.0111110
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
−289.08.0151620
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
34−−−−−−
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
−
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎDYNAMIC CHARACTERISTICSÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Current Gain Bandwidth (IC = 0.5 Adc, VCE = 10 Vdc, f = 1.0 MHz)ÎÎÎÎÎÎÎÎ
fTÎÎÎÎÎÎÎÎ
−ÎÎÎÎÎÎ
13ÎÎÎÎÎÎÎÎ
−ÎÎÎÎÎÎ
MHzÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Output Capacitance (VCB = 10 Vdc, IE = 0, f = 1.0 MHz) ÎÎÎÎÎÎÎÎ
CobÎÎÎÎÎÎÎÎ
− ÎÎÎÎÎÎ
100ÎÎÎÎÎÎÎÎ
150 ÎÎÎÎÎÎ
pFÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Input Capacitance (VEB = 8.0 V) ÎÎÎÎÎÎÎÎ
Cib ÎÎÎÎÎÎÎÎ
− ÎÎÎÎÎÎ
1750ÎÎÎÎÎÎÎÎ
2500 ÎÎÎÎÎÎ
pF
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Dynamic Saturation Voltage:
D t i d 1 0 d
ÎÎÎÎÎÎÎÎÎÎ
(IC = 2.0 AdcIB = 200 mAdc
ÎÎÎÎÎÎÎÎ
1.0 sÎÎÎÎÎÎÎÎÎÎ(TC = 125°C)
ÎÎÎÎÎÎÎÎ
VCE(dsat)ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
5.511.5ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
Vdc
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Determined 1.0 s and3.0 s respectively afterrising IB1 reaches 90% of
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
IB1 = 200 mAdcVCC = 300 V)
ÎÎÎÎÎÎÎÎÎÎÎÎ
3.0 sÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
3.56.5
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎ
rising IB1 reaches 90% offinal IB1(see Figure 18)
ÎÎÎÎÎÎÎÎÎÎ
(IC = 5.0 AdcIB = 1 0 Adc
ÎÎÎÎÎÎÎÎ
1.0 sÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)ÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
11.514.5ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
IB1 = 1.0 AdcVCC = 300 V)
ÎÎÎÎÎÎÎÎÎÎÎÎ
3.0 sÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
2.49.0
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎSWITCHING CHARACTERISTICS: Resistive Load (D.C. 10%, Pulse Width = 20 s)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Turn−On Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(IC = 2.0 Adc, IB1 = 0.2 Adc,IB2 = 1.0 Adc, VCC = 300 V) ÎÎÎÎÎ
ÎÎÎÎÎ(TC = 125°C)ÎÎÎÎ
ÎÎÎÎ
ton ÎÎÎÎÎÎÎÎ
−− ÎÎÎÎÎÎ
200190ÎÎÎÎÎÎÎÎ
300− ÎÎÎÎÎÎ
ns
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Turn−Off Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
B2 , CC )
ÎÎÎÎÎÎÎÎÎÎ(TC = 125°C)
ÎÎÎÎÎÎÎÎ
toff ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
1.21.5ÎÎÎÎÎÎÎÎ
2.5−ÎÎÎÎÎÎ
s
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Turn−On TimeÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(IC = 4.5 Adc, IB1 = 0.9 Adc,IB2 = 2.25 Adc, VCC = 300 V)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
ton
ÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
100250
ÎÎÎÎÎÎÎÎÎÎÎÎ
180−
ÎÎÎÎÎÎÎÎÎ
ns
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Turn−Off Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
B2 , CC )
ÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)ÎÎÎÎÎÎÎÎ
toff ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
1.62.0ÎÎÎÎÎÎÎÎ
2.5−ÎÎÎÎÎÎ
s
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
SWITCHING CHARACTERISTICS: Inductive Load (Vclamp = 300 V, VCC = 15 V, L = 200 H)ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Fall Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(IC = 2.0 Adc, IB1 = 0.2 Adc,IB2 = 1.0 Adc)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
tfiÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
100120
ÎÎÎÎÎÎÎÎÎÎÎÎ
180−
ÎÎÎÎÎÎÎÎÎ
ns
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Storage Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
B2 )
ÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)ÎÎÎÎÎÎÎÎ
tsi ÎÎÎÎÎÎÎÎ
−− ÎÎÎÎÎÎ
1.51.9ÎÎÎÎÎÎÎÎ
2.75− ÎÎÎÎÎÎ
s
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Crossover Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎ(TC = 125°C)
ÎÎÎÎÎÎÎÎ
tc ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
250230ÎÎÎÎÎÎÎÎ
350−ÎÎÎÎÎÎ
ns
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Fall TimeÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(IC = 4.5 Adc, IB1 = 0.9 Adc,IB2 = 2.25 Adc)
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
tfiÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
85135
ÎÎÎÎÎÎÎÎÎÎÎÎ
150−
ÎÎÎÎÎÎÎÎÎ
ns
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Storage Time ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
B2 )
ÎÎÎÎÎÎÎÎÎÎ(TC = 125°C)
ÎÎÎÎÎÎÎÎ
tsi ÎÎÎÎÎÎÎÎ
−−ÎÎÎÎÎÎ
2.02.6ÎÎÎÎÎÎÎÎ
3.2−ÎÎÎÎÎÎ
s
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Crossover TimeÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
(TC = 125°C)
ÎÎÎÎÎÎÎÎÎÎÎÎ
tcÎÎÎÎÎÎÎÎÎÎÎÎ
−−
ÎÎÎÎÎÎÎÎÎ
210250
ÎÎÎÎÎÎÎÎÎÎÎÎ
300−
ÎÎÎÎÎÎÎÎÎ
ns
3. Pulse Test: Pulse Width = 5.0 ms, Duty Cycle 10%.4. Proper strike and creepage distance must be provided.
A. Bipolar Power Transistor MJE18008 246
MJE18008, MJF18008
http://onsemi.com3
hF
E, D
C C
UR
RE
NT
GA
IN
IC, COLLECTOR CURRENT (AMPS)
TJ = 125°C
C, C
APA
CIT
AN
CE
(pF
)
0.01
100
IC, COLLECTOR CURRENT (AMPS)
Figure 1. DC Current Gain @ 1 Volt
hF
E, D
C C
UR
RE
NT
GA
IN
Figure 2. DC Current Gain @ 5 Volts
VC
E, V
OLT
AG
E (
VO
LTS
)
Figure 3. Collector Saturation Region Figure 4. Collector−Emitter Saturation Voltage
Figure 5. Base−Emitter Saturation Region Figure 6. Capacitance
10
11 10
100
10
10.01 0.1 1 10
2
0.01
IB, BASE CURRENT (AMPS)
10
1
0.010.01
IC COLLECTOR CURRENT (AMPS)
0.1
1.3
1
0.8
0.40.01
IC, COLLECTOR CURRENT (AMPS)
0.1 1 10
1000
100
1
VCE, COLLECTOR−EMITTER VOLTAGE (VOLTS)
1 1000
1
0
0.1
1 10
10000
10
0.1
0.1 1 10
10
TJ = 25°C
TJ = −20°C
TJ = 125°C
TJ = 25°C
VC
E, V
OLT
AG
E (
VO
LTS
)
IC/IB = 10
IC/IB = 5
VB
E, V
OLT
AG
E (
VO
LTS
) 1.1
0.9
0.6
0.5
0.5
1.5
1.2
TJ = 25°C
3 A 5 A 8 A 10 A
TJ = 25°C
TJ = 125°C
TJ = 25°C
TJ = 125°C IC/IB = 5
IC/IB = 10
TJ = −20°C
IC = 1 A
0.7
Cob
100
Cib
TYPICAL STATIC CHARACTERISTICS
VCE = 1 V VCE = 5 V
TJ = 25°C
f = 1 MHz
A. Bipolar Power Transistor MJE18008 247
MJE18008, MJF18008
http://onsemi.com4
Figure 7. Resistive Switching, ton Figure 8. Resistive Switching, toff
IC, COLLECTOR CURRENT (AMPS)
IC COLLECTOR CURRENT (AMPS)
IC, COLLECTOR CURRENT (AMPS)
0
1500
IC, COLLECTOR CURRENT (AMPS)
t, T
IME
(ns
)
Figure 9. Inductive Storage Time, tsi Figure 10. Inductive Storage Time, tsi(hFE)
Figure 11. Inductive Switching, tc and tfiIC/IB = 5
Figure 12. Inductive Switching, tc and tfiIC/IB = 10
1000
04 8
2000
0
3500
3
hFE, FORCED GAIN
6
400
50
0
IC, COLLECTOR CURRENT (AMPS)
4 8
250
200
50
2000
012 15
300
150
2
2 5 8
IC/IB = 5
t si, S
TO
RA
GE
TIM
E (
ns)
200
150
100
6
500
IC/IB = 10
4 82 6
500
1000
1500
2500
3000
3500
t, T
IME
(ns
)
t, T
IME
(ns
)
1 3 4 6 7
1000
1500
2500
9
5000
2000
0
500
1000
1500
2500
3000
3500
1 2 3 5
t, T
IME
(ns
)
4 7 81 2 3 5 6
t, T
IME
(ns
)
1 3 5 7 1 3 5 7
500
3000
4 5 7 8 10 11 13 14
250
100
IC/IB = 10
IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
4000
4500
300
350
IB(off) = IC/2
VCC = 300 V
PW = 20 s
IC/IB = 5
IC/IB = 10
TJ = 125°C
TJ = 25°C
4500
4000
IC/IB = 5
IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
IC = 2 A
IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
6 7
tfi
tc
tfi
tc
TJ = 25°C
TJ = 125°C
TYPICAL SWITCHING CHARACTERISTICS(IB2 = IC/2 for all switching)
IC = 4.5 A
IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
IB(off) = IC/2
VCC = 300 V
PW = 20 s
TJ = 25°C
TJ = 125°C
TJ = 25°C
TJ = 125°C
TJ = 25°C
TJ = 125°C
TJ = 25°C
TJ = 125°C
A. Bipolar Power Transistor MJE18008 248
MJE18008, MJF18008
http://onsemi.com5
hFE, FORCED GAIN
TC
, CR
OS
SO
VE
R T
IME
(ns
)
3
160
hFE, FORCED GAIN
Figure 13. Inductive Fall Time
t fi, F
ALL
TIM
E (
ns)
Figure 14. Inductive Crossover Time
605 15
400
200
504 6 7 8 9 10 11 12 13 14
70
80
140
3 5 154 6 7 8 9 10 11 12 13 14
300
100
IC = 2 A
IC = 4.5 A
TJ = 25°C
TJ = 125°C
100
120
350IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
150
130
110
90 150
250
IB(off) = IC/2
VCC = 15 V
VZ = 300 V
LC = 200 H
IC = 2 A
IC = 4.5 A
TYPICAL SWITCHING CHARACTERISTICS(IB2 = IC/2 for all switching)
TJ = 25°C
TJ = 125°CI C
, CO
LLE
CTO
R C
UR
RE
NT
(AM
PS
)
VCE, COLLECTOR−EMITTER VOLTAGE (VOLTS)
I C, C
OLL
EC
TOR
CU
RR
EN
T (A
MP
S)
Figure 15. Forward Bias Safe Operating Area Figure 16. Reverse Bias Switching SafeOperating Area
Figure 17. Forward Bias Power Derating
100
10
VCE, COLLECTOR−EMITTER VOLTAGE (VOLTS)
9
6
00 200
1,0
0,8
0,2
0,020
TC, CASE TEMPERATURE (°C)
80 140 160
1
0.01
3
600 1000100 1000
DC (MJE18008)
5 ms
PO
WE
R D
ER
ATIN
G F
AC
TO
R
0,6
0,4
10
0.1
EXTENDED
SOA
1 ms 10 s 1 s
400
2
1
4
5
40 60 100 120
SECOND BREAKDOWN
DERATING
DC (MJF18008)
7
8
limits of the transistor that must be observed for reliableoperation; i.e., the transistor must not be subjected to greaterdissipation than the curves indicate. The data of Figure 15 isbased on TC = 25°C; TJ(pk) is variable depending on powerlevel. Second breakdown pulse limits are valid for dutycycles to 10% but must be derated when TC > 25°C. Secondbreakdown limitations do not derate the same as thermallimitations. Allowable current at the voltages shown inFigure 15 may be found at any case temperature by using theappropriate curve on Figure 17. TJ(pk) may be calculatedfrom the data in Figure 20 and 21. At any case temperatures,thermal limitations will reduce the power that can be handledto values less than the limitations imposed by secondbreakdown. For inductive loads, high voltage and currentmust be sustained simultaneously during turn−off with thebase−to−emitter junction reverse−biased. The safe level isspecified as a reverse−biased safe operating area (Figure 16).This rating is verified under clamped conditions so that thedevice is never subjected to an avalanche mode.
800
−1, 5 V
−5 V
TC ≤ 125°C
IC/IB ≥ 4
LC = 500 H
GUARANTEED SAFE OPERATING AREA INFORMATION
THERMAL DERATING
VBE(off) = 0 V
There are two limitations on the power handling abilityof a transistor: average junction temperature and secondbreakdown. Safe operating area curves indicate IC −VCE
A. Bipolar Power Transistor MJE18008 249
MJE18008, MJF18008
http://onsemi.com6
−5
−4
−3
−2
−1
0
1
2
3
4
5
0 1 2 3 4 5 6 7 8
Figure 18. Dynamic Saturation Voltage Measurements
TIME
VCE
VO
LTS
IB
Figure 19. Inductive Switching Measurements
1 s
3 s
90% IB
dyn 1 s
dyn 3 s
10
9
8
7
6
5
4
3
2
1
00 1 2 3 4 5 6 7 8
TIME
IB
IC
tsi
VCLAMP 10% VCLAMP
90% IB1
10% ICtc
90% ICtfi
Table 1. Inductive Load Switching Drive Circuit
+15 V
1 F150
3 W
100
3 W
MPF930
+10 V
50
COMMON
−Voff
500 F
MPF930
MTP8P10
MUR105
MJE210
MTP12N10
MTP8P10
150
3 W
100 F
Iout
A
1 F
IC PEAK
VCE PEAK
VCE
IB
IB1
IB2
V(BR)CEO(sus)
L = 10 mH
RB2 = ∞VCC = 20 VOLTS
IC(pk) = 100 mA
INDUCTIVE SWITCHING
L = 200 H
RB2 = 0
VCC = 15 VOLTS
RB1 SELECTED FOR
DESIRED IB1
RBSOA
L = 500 H
RB2 = 0
VCC = 15 VOLTS
RB1 SELECTED
FOR DESIRED IB1
RB2
RB1
A. Bipolar Power Transistor MJE18008 250
MJE18008, MJF18008
http://onsemi.com7
0.01
t, TIME (ms)
Figure 20. Typical Thermal Response (ZJC(t)) for MJE18008
r(t)
, TR
AN
SIE
NT
TH
ER
MA
L R
ES
ISTA
NC
E
(NO
RM
ALI
ZE
D)
RJC(t) = r(t) RJC
RJC = 1.0°C/W MAX
D CURVES APPLY FOR POWER
PULSE TRAIN SHOWN
READ TIME AT t1TJ(pk) − TC = P(pk) RJC(t)
P(pk)
t1t2
DUTY CYCLE, D = t1/t2
0.2
0.02
0.1
D = 0.5
SINGLE PULSE
0.01 0.1 1 10 100 1000
0.1
1
0.05
TYPICAL THERMAL RESPONSE
0.01
Figure 21. Typical Thermal Response (ZJC(t)) for MJF18008
r(t)
, TR
AN
SIE
NT
TH
ER
MA
L R
ES
ISTA
NC
E
(NO
RM
ALI
ZE
D)
RJC(t) = r(t) RJC
RJC = 2.78°C/W MAX
D CURVES APPLY FOR POWER
PULSE TRAIN SHOWN
READ TIME AT t1TJ(pk) − TC = P(pk) RJC(t)
P(pk)
t1t2
DUTY CYCLE, D = t1/t2
0.2
0.02
0.1
SINGLE PULSE
0.01 0.1 1 10 100 100000
0.1
1
1000 10000
0.05
D = 0.5
t, TIME (ms)
ORDERING INFORMATION
Device Package Shipping
MJE18008 TO−220AB 50 Units / Rail
MJE18008G TO−220AB(Pb−Free)
50 Units / Rail
MJF18008 TO−220 (Fullpack) 50 Units / Rail
MJF18008G TO−220 (Fullpack)(Pb−Free)
50 Units / Rail
A. Bipolar Power Transistor MJE18008 251
MJE18008, MJF18008
http://onsemi.com8
MOUNTED
FULLY ISOLATED
PACKAGE
LEADS
HEATSINK
0.110″ MIN
Figure 22a. Screw or Clip Mounting Positionfor Isolation Test Number 1
*Measurement made between leads and heatsink with all leads shorted together
CLIP
MOUNTED
FULLY ISOLATED
PACKAGE
LEADS
HEATSINK
CLIP 0.099″ MIN
MOUNTED
FULLY ISOLATED
PACKAGE
LEADS
HEATSINK
0.099″ MIN
Figure 22b. Clip Mounting Positionfor Isolation Test Number 2
Figure 22c. Screw Mounting Positionfor Isolation Test Number 3
TEST CONDITIONS FOR ISOLATION TESTS*
4−40 SCREW
PLAIN WASHER
HEATSINK
COMPRESSION WASHER
NUT
CLIP
HEATSINK
Laboratory tests on a limited number of samples indicate, when using the screw and compression washer mountingtechnique, a screw torque of 6 to 8 in . lbs is sufficient to provide maximum power dissipation capability. The compres-sion washer helps to maintain a constant pressure on the package over time and during large temperature excursions. Destructive laboratory tests show that using a hex head 4−40 screw, without washers, and applying a torque in excessof 20 in . lbs will cause the plastic to crack around the mounting hole, resulting in a loss of isolation capability. Additional tests on slotted 4−40 screws indicate that the screw slot fails between 15 to 20 in . lbs without adverselyaffecting the package. However, in order to positively ensure the package integrity of the fully isolated device, ON Semi-conductor does not recommend exceeding 10 in . lbs of mounting torque under any mounting conditions.
Figure 23. Typical Mounting Techniquesfor Isolated Package
Figure 23a. Screw−Mounted Figure 23b. Clip−Mounted
MOUNTING INFORMATION**
** For more information about mounting power semiconductors see Application Note AN1040.
A. Bipolar Power Transistor MJE18008 252
MJE18008, MJF18008
http://onsemi.com9
PACKAGE DIMENSIONS
NOTES:1. DIMENSIONING AND TOLERANCING PER ANSI
Y14.5M, 1982.2. CONTROLLING DIMENSION: INCH.3. DIMENSION Z DEFINES A ZONE WHERE ALL
BODY AND LEAD IRREGULARITIES AREALLOWED.
DIM MIN MAX MIN MAX
MILLIMETERSINCHES
A 0.570 0.620 14.48 15.75
B 0.380 0.405 9.66 10.28
C 0.160 0.190 4.07 4.82
D 0.025 0.035 0.64 0.88
F 0.142 0.147 3.61 3.73
G 0.095 0.105 2.42 2.66
H 0.110 0.155 2.80 3.93
J 0.018 0.025 0.46 0.64
K 0.500 0.562 12.70 14.27
L 0.045 0.060 1.15 1.52
N 0.190 0.210 4.83 5.33
Q 0.100 0.120 2.54 3.04
R 0.080 0.110 2.04 2.79
S 0.045 0.055 1.15 1.39
T 0.235 0.255 5.97 6.47
U 0.000 0.050 0.00 1.27
V 0.045 −−− 1.15 −−−
Z −−− 0.080 −−− 2.04
B
Q
H
Z
L
V
G
N
A
K
F
1 2 3
4
D
SEATINGPLANE−T−
CST
U
R
J
TO−220ABCASE 221A−09
ISSUE AA
STYLE 1:PIN 1. BASE
2. COLLECTOR3. EMITTER4. COLLECTOR
TO−220 FULLPAKCASE 221D−03
ISSUE G
DIM
A
MIN MAX MIN MAX
MILLIMETERS
0.625 0.635 15.88 16.12
INCHES
B 0.408 0.418 10.37 10.63
C 0.180 0.190 4.57 4.83
D 0.026 0.031 0.65 0.78
F 0.116 0.119 2.95 3.02
G 0.100 BSC 2.54 BSC
H 0.125 0.135 3.18 3.43
J 0.018 0.025 0.45 0.63
K 0.530 0.540 13.47 13.73
L 0.048 0.053 1.23 1.36
N 0.200 BSC 5.08 BSC
Q 0.124 0.128 3.15 3.25
R 0.099 0.103 2.51 2.62
S 0.101 0.113 2.57 2.87
U 0.238 0.258 6.06 6.56
−B−
−Y−
GN
DL
K
H
A
F
Q
3 PL
1 2 3
MBM0.25 (0.010) Y
SEATINGPLANE
−T−
U
C
S
JR
NOTES:1. DIMENSIONING AND TOLERANCING PER ANSI
Y14.5M, 1982.2. CONTROLLING DIMENSION: INCH3. 221D−01 THRU 221D−02 OBSOLETE, NEW
STANDARD 221D−03.
STYLE 2:PIN 1. BASE
2. COLLECTOR3. EMITTER
A. Bipolar Power Transistor MJE18008 253