TIPP 2011, Chicago 1
LED calibration systems for CALICE hadron calorimeter
Jiri Kvasnicka ([email protected])Institute of Physics, Prague
June 11, 2011
Under
HEAVY
CONSTRUCTION
TIPP 2011, Chicago 2
Outline• Calice prototype• SiPM Motivation (SiPM issues, temeperature drift..)• AHCAL 1m^2 solution
– Electronics solution– performance
• Embeded solution– Electronics solution– Performance
• Quasi-resonant LED driver– Electronics solution– Performance
• Light distribution
June 11, 2011
TIPP 2011, Chicago 3
AHCAL 1m2 Physics prototype• The AHCAL 1m2 - CALICE experiment
– built in 2005– Testbems 2006-2010 at CERN and FNAL. – Now in CERN as WHCAL– Tested together with ECAL (electromagnetic calorimeter) and
TCMT (Muon calorimeter)• 38 layers, 2cm Fe absorbers • 7608 photo detectors (SiPM) in total• One layer
– 216 scintillator tiles, 3x3, 6x6, 12 x 12 cm2,– Calibrating system (CMB) with 12 LEDs monitored by PIN-Photo
Diodes– Optical flash is distributed by fibre bundle individually to each
scintillator– 5 temperature sensors per layer - integrated circuits LM35
• Scintillating tile– 5mm thick Scintillator– WLS (wavelength shifting fiber), ~380nm~500nm)– SiPM photodetector attached to the WLS fiber + mirror
• SiPM (silicone photomultiplier)– 1156 pixels, each works in Geiger mode– Gain of SiPM ~105 to 106
June 11, 2011
ECAL
HCAL
TCMT
20mm Fe plates and
scintillators
3 cm1 mm
90 cm
AHCAL`
TIPP 2011, Chicago 4
Calibration Chain: ADC to MIP• AHCAL signal chain:
ParticleMIPsScintillating tilephotons (UV)Wavelength-shifting fibre photons (green)SiPMPhoto-electronsASIC readout
• Calibration task:Convert the detector signal to a number of MIP deposited by the particle
• The means on calibration: – LED light!– Charge injection– Beam: Muon tracks identification and calibration– Cosmic muons– Other means: Laser, Radioactive material
• Measured parameters:– SiPM gain (from Single Photon Spectrum)– Temperature (gain factor -2% per 1K)– Saturation function
June 11, 2011
TIPP 2011, Chicago 5
MUONS
Particle detection
LED
ligh
t
June 11, 2011
TIPP 2011, Chicago 6
CMB-ToDo
• El. Zapojeni, simulace
June 11, 2011
TIPP 2011, Chicago 7
CMB-results ToDo
• Jara’s calibration plots, SPS, etc.
June 11, 2011
TIPP 2011, Chicago 8
Integrated LED system
LEDs• Developed by DESY and Uni Wuppertal• Each Tile has its through-hole mounted LED• Each LED has its own driver circuitry.
– Operation: The current pulse though the LED is generated by discharging of the Capacitor by a fast transistor
– V-calib signal range: 3–10 V– System tuned for ~8 ns pulses
• Choice of the LED is critical for this driver– Several different LED types were tested (see next slide)– The technology of the LED is most important
• Only Single-quantum-well LEDs work well (usually UV-LED)• Usual (multi-quantum-well) LEDs have too big capacitance
and produce longer optical pulse. On the other hand, they are very bright
• Driver circuitry is now optimized and being manufactured on the new HBU for the technological prototype
June 11, 2011
5 ns
TIPP 2011, Chicago 9
Integrated LED system – Optimization
• Pulse of the Blue LED is much wider (~40 ns), than the UV LED (~5 ns)
• Light pulse width re-measured with a differential driver
– In this mode: LED is reverse biased, then for a short pulse forward biased and directly reverse biased again
– The reverse voltage helps to discharge the LED– Blue LED stops shining much faster in
differential mode
June 11, 2011
Blue LED
UV LED
Blue LED, differential
TIPP 2011, Chicago 10
Integrated LED system – SPS
• For longer (>30 ns) pulses, both UV and Blue LEDs produce equal optical pulses
• Observation: UV LED have much steeper rise time
• Driver circuitry is now optimized and being manufactured on the new HBU for the technological prototype
• Question: is short pulse necessary?– Answer: Yes, only 15 ns pulses and faster
produce decent Single Photon Spectra • Single Photon Spectrum (SPS)
– Short pulse -> improvement of the quality– Nice spectrum with UV-LED– Spectrum is more smeared with 30 ns
blue-LED
June 11, 2011
Blue LED, 30 ns
Blue LED, 15ns
UV LED, 7ns
TIPP 2011, Chicago 11
Integrated LED system – Light Yield• Measurements with key components
variation• Circuitry was finally tuned to deliver up
to 17K effective pixels – Light referenced to PMT signal
• Time behavior of the LED– Measured with PMT– Without tile: sharp pulse– With tile (and Wavelength shifting fibre)
long tail
June 11, 2011
Resistorvariation
Capacitorvariation
With Tile
25 ns
TIPP 2011, Chicago 12
QMB6-ToDo
June 11, 2011
TIPP 2011, Chicago 13
QMB6-ToDo
June 11, 2011
TIPP 2011, Chicago 14
Notched Fibre
June 11, 2011
• 24-notched fibre at the left figure. Illuminated by a green laser• Light is emitted from the notches• The notch is a special scratch to the fibre, which reflects the light to
the opposite direction• The size of the notch varies from the beginning to the end of the
fibre
First notch Middle notch End position notch
Emission from the fibre (side view)
TIPP 2011, Chicago 15
Optical fibre
• Measurements of the light yield – Through the 3mm hole on the
PCB (FR4 with filled inner layer)– 3 positions of the notch
according to the PCB thru-hole
June 11, 2011
“start” position “middle” position “end” position
TIPP 2011, Chicago 16
Notched fibers configuration
• 72 zarezova vlakna – vysledky linearity• LED vyzarovaci profil (smolda)• Konfigurace 3*24 zarezu
June 11, 2011
HBU6HBU5HBU4HBU3HBU2HBU1
TIPP 2011, Chicago 17
Development of new Quasi-resonant driver (QMB1)
• QMB1 (1-chanel LED driver):– Fixed
• Topology• Communicating bus (CAN)• CPU (Atmel AVR)• Trigger distribution (LVDS)• Trigger delay can be tuned by C trimmer (~10ns)
• Free to adjust: will be discussed at DESY in July calib meeting
– Mounting holes (fixation to support/HBU– Fibre(LED) position
• Set of notched fibers, semi-automat machine under development
– Set: 3*fiber with 24 notches, creating a line of 72 notches.
– 3 sets will be delivered
June 11, 2011
HBU6HBU5HBU4HBU3HBU2HBU1
TIPP 2011, Chicago 18
Conclusion
June 11, 2011