Low-Noise Amplifiers (LNAs) and Power Amplifiers (PAs) for

ARL-TR-8871 ● JAN 2020

Low-Noise Amplifiers (LNAs) and Power Amplifiers (PAs) for Next-Generation S- and X-Band Radars by John E Penn

Approved for public release; distribution is unlimited.

NOTICES

Disclaimers

The findings in this report are not to be construed as an official Department of the Army position unless so designated by other authorized documents.

Citation of manufacturer’s or trade names does not constitute an official endorsement or approval of the use thereof.

Destroy this report when it is no longer needed. Do not return it to the originator.

ARL-TR-8871 ● JAN 2020

Low-Noise Amplifiers (LNAs) and Power Amplifiers (PAs) for Next-Generation S- and X-Band Radars John E Penn Sensors and Electron Devices Directorate, CCDC Army Research Laboratory


ii

REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing the burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS.

1. REPORT DATE (DD-MM-YYYY)

January 2020 2. REPORT TYPE

Technical Report 3. DATES COVERED (From - To)

March–September 2019 4. TITLE AND SUBTITLE

Low-Noise Amplifiers (LNAs) and Power Amplifiers (PAs) for Next-Generation S- and X-Band Radars

5a. CONTRACT NUMBER

5b. GRANT NUMBER

5c. PROGRAM ELEMENT NUMBER

6. AUTHOR(S)

John E Penn 5d. PROJECT NUMBER

5e. TASK NUMBER

5f. WORK UNIT NUMBER

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES)

CCDC Army Research Laboratory ATTN: FCDD-RLS-RE Adelphi, MD 20783-1138

8. PERFORMING ORGANIZATION REPORT NUMBER

ARL-TR-8871

9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES)

10. SPONSOR/MONITOR'S ACRONYM(S)

11. SPONSOR/MONITOR'S REPORT NUMBER(S)

12. DISTRIBUTION/AVAILABILITY STATEMENT


13. SUPPLEMENTARY NOTES ORCID ID: John Penn, 0000-0001-7535-0388 14. ABSTRACT

The US Army Combat Capabilities Development Command Army Research Laboratory (CCDC ARL) has been evaluating and designing efficient broadband high-power amplifiers for future adaptive multimode radar systems in addition to other circuits for use in communications, networking, and electronic warfare (EW). ARL submitted designs of broadband amplifiers, power amplifiers, high-dynamic range low-noise amplifiers, high-power switches, frequency multipliers, and other circuits for future radar, communications, EW, and sensor systems using Qorvo’s high-performance 0.15-µm gallium nitride (GaN) fabrication process. This technical note briefly summarizes several designs using Qorvo’s 0.15-µm high-power, efficient GaN on 4-mil silicon carbide process that were submitted to an ARL prototype wafer option fabrication.

15. SUBJECT TERMS

MMIC, power amplifier, low-noise amplifier, microwave, radar

16. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT

UU

18. NUMBER OF PAGES

38

19a. NAME OF RESPONSIBLE PERSON

John E Penn a. REPORT

Unclassified b. ABSTRACT

Unclassified

c. THIS PAGE

Unclassified

19b. TELEPHONE NUMBER (Include area code)

(301) 394-0423 Standard Form 298 (Rev. 8/98)

Prescribed by ANSI Std. Z39.18

https://orcid.org/0000-0001-7535-0388

iii

Contents

List of Figures iv

1. Introduction 1

2. S- and X-Band Low-Noise Amplifier 1

3. S- and X-Band Power Amplifier 5

4. S- and X-Band 10-W Power Amplifier 16

5. Broadband Feedback Amplifiers 21

6. Broadband Nonuniform Distributed Amplifier 24

7. Conclusions 27

8. References 29

List of Symbols, Abbreviations, and Acronyms 30

Distribution List 31

iv

List of Figures

Fig. 1 S- to X-band LNA ideal schematic (6- × 25-µm HEMT) .................... 2

Fig. 2 S- to X-band LNA simulation (ideal lossless match) ........................... 3

Fig. 3 S- to X-band LNA simulation (lossless input match, MMIC output match) ................................................................................................ 3

Fig. 4 S- to X-band LNA simulation (full MMIC layout) .............................. 4

Fig. 5 S- to X-band LNA EM simulation (EM [solid] vs. MMIC [dash]) ...... 4

Fig. 6 S- to X-band LNA layout ................................................................... 5

Fig. 7 Power (green) and efficiency (magenta) contours at 6 GHz (8 × 125 µm) ...................................................................................... 7

Fig. 8 Initial ideal 3.5- and 7-W PA linear simulations (8 × 93 µm vs. two 8 × 93 µm) ......................................................................................... 8

Fig. 9 Power and PAE simulations at 6 GHz of ideal 3.5-W (solid) and 7-W PA (dotted)......................................................................................... 8

Fig. 10 Ideal PA no. 2 schematic (8- × 150-µm HEMT) ................................. 9

Fig. 11 Power and PAE simulations at 6 GHz of ideal PA no. 2 (8 × 150 µm) ...................................................................................... 9

Fig. 12 Revised S- to X-band 8- × 150-µm PA simulations: ideal (dash) vs. MMIC (solid) ................................................................................... 10

Fig. 13 Output match of S- to X-band 8- × 150-µm PA simulation: ideal (dash) vs. MMIC (solid) ................................................................... 11

Fig. 14 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, ideal lossless design (28 V) ................................................................................... 12

Fig. 15 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, with lossy MMIC elements (28 V) .................................................................... 13

Fig. 16 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, final layout EM (28 V) .............................................................................................. 14

Fig. 17 S- to X-band 8- × 150-µm PA simulations: MMIC (dash) vs. EM (solid)............................................................................................... 15

Fig. 18 Final compact layout of 8- × 150-µm S- to X-band PA ..................... 15

Fig. 19 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, ideal lossless design .................................................................................. 17

Fig. 20 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, lossy MMIC PA ........................................................................................ 18

Fig. 21 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, final layout EM (28 V) ............................................................................. 19

v

Fig. 22 S- to X-band 2X 8- × 150-µm 10-W PA linear simulations: 1X (dash) vs. 2X (solid).................................................................................... 20

Fig. 23 S- to X-band 2X 8- × 150-µm 10-W PA simulations: MMIC (dash) vs. EM (solid) ........................................................................................ 20

Fig. 24 Final compact layout of 2X 8- × 150-µm S- to X-band 10-W PA ..... 21

Fig. 25 Final compact layouts of broadband feedback amplifiers (6× 50 µm, 4× 50 µm, and 4× 65 µm) ................................................................. 22

Fig. 26 Simulation of 6- × 50-µm resistive feedback amplifier (MMIC [solid] vs. ideal [dot]) .................................................................................. 23

Fig. 27 Simulation of 4- × 50-µm source feedback amplifier (MMIC [solid] vs. ideal [dot]) .................................................................................. 23

Fig. 28 Simulation of 4- × 65-µm source feedback amplifier (MMIC [solid] vs. ideal [dot]) .................................................................................. 24

Fig. 29 Schematic of initial ideal uniform 4- × 20-µm DA ............................ 25

Fig. 30 Simulation of initial ideal uniform 4- × 20-µm DA ........................... 25

Fig. 31 Simulation of re-tuned ideal NUDA ................................................. 26

Fig. 32 EM simulation (solid) of nonuniform MMIC DA ............................. 26

Fig. 33 Layout of NUDA ............................................................................. 27

1

1. Introduction

The US Army Combat Capabilities Development Command Army Research Laboratory

(CCDC ARL) has been evaluating and designing efficient broadband high-power

amplifiers and robust low-noise amplifiers (LNAs) for future multimode radar systems

that could be used in other applications such as communications, networking, and

electronic warfare (EW). ARL submitted designs of broadband amplifiers, power

amplifiers (PAs), and high-dynamic range LNAs for future radar, communications, EW,

and sensor systems using Qorvo’s high-performance 0.15-µm gallium nitride (GaN)

fabrication process. The circuits most applicable to radar applications are documented in

this technical report. When the fabricated designs are returned and tested, future technical

reports will document results of circuit characterization.

2. S- and X-Band Low-Noise Amplifier

Not all high-electron-mobility transistor (HEMT) models in the Qorvo 0.15-µm GaN

process design kit (PDK) had noise figure data, which are necessary for the design of an

LNA. Initially, gain, noise figure, quality factor (Q) of the optimal noise match, and

stability were simulated over 2–10 GHz for the recommended HEMT sizes with noise

data of 4 × 25 µm and 6 × 25 µm. Some source inductance was used to optimize the

performance tradeoffs, with the 6- × 25-µm HEMT chosen for the initial design. DC

biases of 5 and 10 V were recommended for low-noise applications, rather than the 25 to

28 V that would be typical for high-power applications. These designs should be able to

operate over a large range of DC, though the best noise figure is typically at a moderately

low-drain current.

Initial lossless matched LNA designs achieved good noise figure, very good gain, and

conditional stability over an octave bandwidth, 3–6 GHz, 4–8 GHz, or 5–10 GHz, but

achieving 3–10 GHz bandwidth with good stability, noise figure, gain, and return loss

was difficult. Large coupled lines, such as used on a prior Raytheon GaN fabrication1

were suitable for larger bandwidths but compact lumped element matching circuits

resulted in the best compromise of layout area, noise figure, stability, and gain. Figure 1

shows a schematic of the initial design using ideal lossless matching elements, while

Fig. 2 shows its corresponding linear simulation. The noise figure (green, right axis) is

well below 1 dB from 3 to 12 GHz, and gain is above the 10-dB goal from 3 GHz to

above 9 GHz. Note that with ideal lossless matching elements there is a stability problem

near 2 GHz (black, right axis). When lossy monolithic microwave integrated circuit

(MMIC) elements are substituted into the output matching network and the source

feedback inductors, the gain decreases but the stability improves. Figure 3 shows better

than 10 dB gain from 3 to 9 GHz, and the noise figure is still below 1 dB over the desired

2

range. Conditional stability is achieved for nominal loads (near 50 ohms), but there is

still a potential instability near 3 GHz for poor matches (Mu > 0.4). When the lossless

input match is converted to MMIC elements, unconditional stability is achieved, but the

gain drops about 1.5 dB, below the 10 dB goal, and the noise figure increases nearly 0.5

to a respectable 1.4 dB over a broad frequency range from 3.3 GHz up to 10 GHz. Linear

simulations of the full MMIC layout are shown in Fig. 4. The layout was then

electromagnetic (EM)-simulated and readjusted to achieve performance similar to the

original MMIC linear simulation (Fig. 5). Final layout of the S- to X-band LNA is shown

in Fig. 6.

Fig. 1 S- to X-band LNA ideal schematic (6- × 25-µm HEMT)

6x25

6x50

4x25

Rser=0

Lsrc=0.3Rser=0

Lsrc=0.8

Rshnt=2000

Rser=0Lsrc=0.65

Rshnt=2000

Wf=50

Wf=25

Wf=25

Rshnt=2000

PORTP=1Z=50 Ohm

PORTP=2Z=50 OhmRES

ID=TL3R=Rshnt Ohm

INDID=L8L=3.3 nH

G

D

S

QGAN15_ES_FET_V2ID=LN_1416_6x25_1Wu=Wf umN=6BIAS=5V 100 mA/mm

INDID=L9L=1.8 nH

CAPID=L5C=0.347 pF

CAPID=L6C=0.354 pF

CAPID=L7C=0.347 pF

INDID=L1L=Lsrc nH

INDID=L3L=1.8 nH

RESID=TL4R=Rser Ohm

CAPID=L4C=0.354 pF

G

D

S

QGAN15_ES_FET_V2ID=LN_1416_4x25_1Wu=Wf umN=4BIAS=5V 100 mA/mm

INDID=L2L=3.3 nH

QGAN15_ES_FETVIAID=FETVIA_1WIN=20 um

3

Fig. 2 S- to X-band LNA simulation (ideal lossless match)

Fig. 3 S- to X-band LNA simulation (lossless input match, MMIC output match)

4

Fig. 4 S- to X-band LNA simulation (full MMIC layout)

Fig. 5 S- to X-band LNA EM simulation (EM [solid] vs. MMIC [dash])

5

Fig. 6 S- to X-band LNA layout

3. S- and X-Band Power Amplifier

For the design of an S- to X-band PA, load pull simulations were performed on a

1-mm (8- × 125-µm) HEMT at 3, 6, and 9 GHz. The maximum allowed DC bias

for the 0.15-µm GaN process is 28 V and that was assumed for this PA design to

maximize power. Figure 7 shows an example of the power and power-added

efficiency (PAE) contours at 6 GHz for a 1-mm HEMT, yielding a maximum 60%

PAE for a load of 0.62 at 75° and a Q of 1.95. Maximum power of 37.1 dBm

(5.1 W) is shown for a load of 0.43 at 98° and a Q of 1.05, and a very good

compromise load of 0.47 at 94° and a Q of 1.2 yields 37 dBm (5 W) and 55% PAE.

Lower Q means larger achievable bandwidth, and the compromise target

impedance yields 5 W/mm with 55% PAE at 28 V. These load pull simulations

were used to create a simple scalable model of the target load impedance. A slightly

larger 8- × 150-µm (PDK limit) HEMT was chosen to achieve 5 W per device with

0.8-dB margin for losses. In order to achieve higher powers, multiple HEMTs

would have to be parallel combined. A 10-W PA design was later achieved by

6

paralleling two 8- × 150-µm HEMTs, starting with this initial 8- × 150-µm HEMT

5-W PA design. Another ideal lossless matched PA design based on 8- × 93-µm

HEMTs achieved very good gain but the bandwidth was less than desired. Figure

8 shows linear simulations of an initial 8- × 93-µm PA (solid lines) and a two-way

combined 8- × 93-µm PA at twice the output power. Very good maximum

efficiency (PAE) of 52% PAE was simulated with 3.5 W for the single 8- × 93-µm

PA and 7 W for the two-way combined 8- × 93-µm PA at 3-dB gain compression

and peak efficiency (Fig. 9). A redesign with a larger HEMT device was needed for

the 5-W and 10-W power goals, and additional compromises in performance would

be needed to achieve the bandwidth goals of S- to X-band (3–10 GHz). A larger

8- × 175-µm HEMT was used in a prior successful PA design in Qorvo’s 0.25-µm

GaN process, but the 0.15-µm GaN PDK would not allow scaling the nominal 8- ×

100-µm HEMT beyond 8- × 150-µm, so these second PA (no. 2) designs were

limited to 8 × 150 µm (1.2 mm). The ideal lossless matched single-stage 8- ×

150-µm PA no. 2, whose schematic is shown in Fig. 10, predicts 7 W (38.6 dBm)

at less than 3-dB gain compression with an excellent 54% PAE that is still rising

toward its peak (Fig. 11).

7

Fig. 7 Power (green) and efficiency (magenta) contours at 6 GHz (8 × 125 µm)

0 1.0

1.0

-1.0

10.0

10.0

-10.0

5.0

5.0

-5.0

2.0

2.0

-2.0

3.0

3.0

-3.0

4.0

4.0

-4.0

0.2

0.2

-0.2

0.4

0.4

-0.4

0.6

0.6

-0.6

0.8

0.8

-0.8

ContoursSwp Max

6e+009

Swp Min

1

m2

m5

m3

m1

PAE Max

PLoad Max

PAE Contours

PLoad Contours

Gamma Points

GPC_MAX(2,1,1,3)Load_Pull_8x125

Conj(S(2,2))Load_Pull_8x125

Conj(S(1,1))Load_Pull_8x125

m1: 60.046 %Mag 0.6186Ang 75.02 DegPAE = 60.046 %iPower = 1harm = 1

m3: 6 GHzMag 0.02044Ang 5.756 Deg

m5: 58 %Mag 0.7286Ang 81.64 DegPAE = 58 %iPower = 1harm = 1

m2: 37.105 dBmMag 0.4332Ang 97.81 DegPLoad = 37.105 dBmiPower = 1harm = 1

8

Fig. 8 Initial ideal 3.5- and 7-W PA linear simulations (8 × 93 µm vs. two 8 × 93 µm)

Fig. 9 Power and PAE simulations at 6 GHz of ideal 3.5-W (solid) and 7-W PA (dotted)

9

Fig. 10 Ideal PA no. 2 schematic (8- × 150-µm HEMT)

Fig. 11 Power and PAE simulations at 6 GHz of ideal PA no. 2 (8 × 150 µm)

In increase the gain bandwidth, the input match of the 8- × 150-µm PA was

increased in complexity to achieve broadband S- to X-band gain, 3–9 GHz. Stable

gain was achieved with nearly identical gain performance for the lossy MMIC PA

design versus the original ideal lossless PA as shown in Fig. 12. Comparing the

ideal output match design to the desired optimal parallel resistor-capacitor load

model shows excellent match from about 2.6 to 9 GHz (Fig. 13, dashed lines).

Converting the broadband output match to actual MMIC elements reduced the

bandwidth slightly with minimal losses of 0.5–0.6 dB across most of the band,

rolling off below 3 GHz and above 9 GHz (Fig. 13, solid lines). A 0.5-dB loss

corresponds to about 12% power and efficiency loss in the output match compared

to the ideal case. Output power and PAE performance simulations for the ideal

INDID=L7L=2.37 nH

INDID=L16L=1 nH

CAPID=L17C=3 pF

CAPID=L15C=1.8 pF

INDID=L8L=2.07 nH

CAPID=L4C=0.435 pF

INDID=L6L=2.37 nH

INDID=L14L=0.40 nH RES

ID=R2R=Res2 Ohm

CAPID=L5C=0.49 pF

CAPID=L3C=100 pF

G

D

QGAN15_ES_FET2_V2ID=LIN_4852_8x150_1Wu=150 umN=8VIA_count_per_pad=2VIA=No ShareBIAS=28V 100 mA/mm

RESID=R1R=Res1 Ohm

PORTP=1Z=50 Ohm

PORTP=2Z=50 Ohm

TLINID=TL1Z0=29 OhmEL=90 DegF0=5 GHz

CAPID=L18C=0.536 pF IND

ID=L13L=1.25 nH

Res2=50WF=150

Res1=2w50=88

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Power (dBm)

Pcomp_8x150_Amp_5G

0

5

10

15

20

25

30

35

40

45

50

55

60

p2

p1p3

25 dBm13.55 dB

0 dBm16.39 dB

25 dBm38.55 dBm

25 dBm53.99

DB(|Pcomp(PORT_2,1)|)[*,X] (dBm)AMP_8x150_5G_NL.AP_HB

PAE(PORT_1,PORT_2)[*,X]AMP_8x150_5G_NL.AP_HB

DB(PGain(PORT_1,PORT_2))[*,X]AMP_8x150_5G_NL.AP_HB

p3: Freq = 5 GHz p1: Freq = 5 GHz

p2: Freq = 5 GHz

10

lossless S- to X-band 8- × 150-µm PA from 3 to 9 GHz are shown in Fig. 14.

Efficiency (PAE) ranges from a peak of 60% at 3 GHz to 52% at 9 GHz, while

output power is approximately 6.75 W (38.3 dBm) from 3 to 7 GHz, falling off

some at 8 and 9 GHz, with an input power of 0.5 W (27 dBm). Once the 0.5-dB

losses of the actual MMIC matching elements are factored in, the output power

drops to 6 W (37.8 dBm), falling even more at 8 and 9 GHz, with an input power

of 0.5 W (27 dBm), while PAE drops to a peak of 51% at 3 GHz, dropping farther

to a peak PAE of 44% at 9 GHz. Figure 15 shows the simulated output power and

PAE performance for the S- to X-band 8- × 150-µm PA with lossy MMIC matching

circuits from 3 to 9 GHz. When the MMIC layout is EM-simulated and the layouts

readjusted to get back to the prior performance, the output power is similar, as are

the peak efficiencies which range from about 44% to 50% as shown in Fig. 16. The

small signal linear performance of the final layout (EM) is shown in Fig. 17 and is

very similar to the original MMIC linear simulation. Final layout plot of the

compact 5- to 6-W broadband 3- to 9-GHz PA is shown in Fig. 18.

Fig. 12 Revised S- to X-band 8- × 150-µm PA simulations: ideal (dash) vs. MMIC (solid)

11

Fig. 13 Output match of S- to X-band 8- × 150-µm PA simulation: ideal (dash) vs. MMIC

(solid)

12

Fig. 14 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, ideal lossless design (28 V)

0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930

Power (dBm)

Pcomp_8x150_Amp_5G

10

15

20

25

30

35

40

0

10

20

30

40

50

60

p21p20p19p18p17p16

p15

p11p10p9

p6

p5p2

p1

p8

p4

p3

27 dBm55.92Freq = 5 GHz

27 dBm38.24 dBmFreq = 5 GHz

0 dBm18.29 dBFreq = 3 GHz



DB(|Pcomp(PORT_2,1)|)[*,X] (L, dBm)AMP_8x150_5G_NL.AP_HB

PAE(PORT_1,PORT_2)[*,X] (R)AMP_8x150_5G_NL.AP_HB

DB(PGain(PORT_1,PORT_2))[*,X] (L)AMP_8x150_5G_NL.AP_HB

p3: Freq = 3 GH z p4: Freq = 4 GH z p7: Freq = 5 GH z p8: Freq = 6 GH z p12: Freq = 7 G H z

p13: Freq = 8 G H zp14: Freq = 9 G H zp1: Freq = 3 GH z p2: Freq = 4 GH z p5: Freq = 5 GH z

p6: Freq = 6 GH z p9: Freq = 7 GH z p10: Freq = 8 G H zp11: Freq = 9 G H zp15: Freq = 3 G H z

p16: Freq = 4 G H zp17: Freq = 5 G H zp18: Freq = 6 G H zp19: Freq = 7 G H zp20: Freq = 8 G H z

p21: Freq = 9 G H z

13

Fig. 15 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, with lossy MMIC elements (28 V)

0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930

Power (dBm)

Pcomp_8x150_Amp_5G

10

15

20

25

30

35

40

0

10

20

30

40

50

60

p21

p20p19p18

p17p16p15

p12

p11

p9

p6

p5p2

p1

p14p10

p8

p3









p3: Freq = 3 GH z p4: Freq = 4 GH z p7: Freq = 5 GH z p8: Freq = 6 GH z p10: Freq = 7 G H z

p13: Freq = 8 G H zp14: Freq = 9 G H zp1: Freq = 3 GH z p2: Freq = 4 GH z p5: Freq = 5 GH z

p6: Freq = 6 GH z p9: Freq = 7 GH z p11: Freq = 8 G H zp12: Freq = 9 G H zp15: Freq = 3 G H z

p16: Freq = 4 G H zp17: Freq = 5 G H zp18: Freq = 6 G H zp19: Freq = 7 G H zp20: Freq = 8 G H z

p21: Freq = 9 G H z

14

Fig. 16 Power and PAE of 8- × 150-µm PA from 3 to 9 GHz, final layout EM (28 V)

0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930

Power (dBm)

Pcomp_8x150_Amp_5G

10

15

20

25

30

35

40

0

10

20

30

40

50

60

p18p17

p16p15p14

p13p12

p11p10p9

p5

p4p2

p1

p8

p3









p3: Freq = 3 GH z p6: Freq = 4 GH z p7: Freq = 5 GH z p8: Freq = 6 GH z p19: Freq = 7 GH z

p20: Freq = 8 GH zp21: Freq = 9 GH zp1: Freq = 3 GH z p2: Freq = 4 GH z p4: Freq = 5 GH z

p5: Freq = 6 GH z p9: Freq = 7 GH z p10: Freq = 8 GH zp11: Freq = 9 GH zp12: Freq = 3 GH z

p13: Freq = 4 GH zp14: Freq = 5 GH zp15: Freq = 6 GH zp16: Freq = 7 GH zp17: Freq = 8 GH z

p18: Freq = 9 GH z

15

Fig. 17 S- to X-band 8- × 150-µm PA simulations: MMIC (dash) vs. EM (solid)

Fig. 18 Final compact layout of 8- × 150-µm S- to X-band PA

16

4. S- and X-Band 10-W Power Amplifier

The previous 5- to 6-W S- to X-band PA design was used as the basis for parallel

combining two 8- × 150-µm HEMTs to achieve a 10-W output power goal. First,

an ideal double-tuned lossless output match for an 8- × 150-µm HEMT was

designed based on the prior ideal match to 50 ohms but translated to 100 ohms.

Then, two of these circuits were combined in parallel to create the output match to

50 ohms. Output power and PAE performance simulations for the ideal lossless S-

to X-band 2X 8- × 150-µm PA from 3 to 9 GHz are shown in Fig. 19. Efficiency

(PAE) ranges from a peak of 58% at 3 GHz, to 49% at 9 GHz, while output power

is up to 16 W (42 dBm) with an input power of 1 W (30 dBm). Once the 0.75-dB

losses of the actual MMIC matching elements are factored in, the output power is

as high as 13.2 W (41.1 dBm), falling off at 7 to 9 GHz, with an input power of 1

W (30 dBm), while PAE is a peak of 50% at 4 GHz, dropping to a peak PAE of

31% at 9 GHz. Figure 20 shows the simulated output power and PAE performance

for the S- to X-band 2X 8- × 150-µm 10-W PA with lossy MMIC matching circuits

from 3 to 9 GHz. When the MMIC layout is EM-simulated and the layouts

readjusted to get back to the prior performance, the output power ranges from 10 to

12 W from 3 to 9 GHz, while PAE ranges from 39% to 48% with an input power

of 1 W (30 dBm), as shown in Fig. 21. The small signal linear performance of the

final layout of the parallel 10-W PA versus the prior single 8- × 150-µm PA is

shown in Fig. 22. It is similar in shape but the gain is 2 to 3 dB lower than the prior

5- to 6-W PA. The small signal linear performance of the final layout (EM) is shown

in Fig. 23 and is very similar to the original MMIC linear simulation. Final layout

plot of the compact 10- to 12-W broadband 3- to 9-GHz PA is shown in Fig. 24.

Note the odd-mode stabilizing resistors between the gates and drains of the two

parallel combined 8- × 150-µm HEMTs shown in the layout of Fig. 24. An analysis

of the odd-mode stability showed a potential problem around 5.2–6.6 GHz that

could be removed with either an odd-mode resistor on the gates of less than

150 ohms or an odd-mode resistor on the drains of less than 250 ohms. To provide

margin in avoiding an odd-mode oscillation, both a 100-ohm resistor placed on the

gates and a 200-ohm resistor placed on the drains were added to the two parallel

8- × 150-µm HEMTs.

17

Fig. 19 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, ideal lossless design

0 1 2 3 4 5 6 7 8 9 101112131415161718192021222324252627282930313233

Power (dBm)

Pcomp_8x150_Amp_5G_2X

10

15

20

25

30

35

40

45

0

10

20

30

40

50

60

70

p19p18p17p16p15p14

p13

p12p11

p7

p6p5

p2

p1

p10

p4p3






DB(|Pcomp(PORT_2,1)|)[*,X] (L, dBm)

AMP_8x150_5G_NL_2X.AP_HB

PAE(PORT_1,PORT_2)[*,X] (R)

AMP_8x150_5G_NL_2X.AP_HB

DB(PGain(PORT_1,PORT_2))[*,X] (L)

AMP_8x150_5G_NL_2X.AP_HBp3: Freq = 3 GHz p4: Freq = 4 GHz p8: Freq = 5 GHz p9: Freq = 6 GHz p10: Freq = 7 GHz

p20: Freq = 8 GHzp21: Freq = 9 GHzp1: Freq = 3 GHz p2: Freq = 4 GHz p5: Freq = 5 GHz

p6: Freq = 6 GHz p7: Freq = 7 GHz p11: Freq = 8 GHzp12: Freq = 9 GHzp13: Freq = 3 GHz

p14: Freq = 4 GHzp15: Freq = 5 GHzp16: Freq = 6 GHzp17: Freq = 7 GHzp18: Freq = 8 GHz

p19: Freq = 9 GHz

18

Fig. 20 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, lossy MMIC PA

19

Fig. 21 Power and PAE of 2X 8- × 150-µm 10-W PA from 3 to 9 GHz, final layout EM (28 V)

20

Fig. 22 S- to X-band 2X 8- × 150-µm 10-W PA linear simulations: 1X (dash) vs. 2X (solid)

Fig. 23 S- to X-band 2X 8- × 150-µm 10-W PA simulations: MMIC (dash) vs. EM (solid)

21

Fig. 24 Final compact layout of 2X 8- × 150-µm S- to X-band 10-W PA

5. Broadband Feedback Amplifiers

A simple approach to a broadband amplifier uses feedback, typically resistive

feedback from drain to gate. Several broadband feedback amplifiers were designed

in a prior Raytheon GaN process, demonstrating excellent noise figure as well as

broadband gain. Using the two provided noise figure HEMT models for the Qorvo

0.15-µm GaN process, variations of a broadband feedback amplifier were designed.

One amplifier uses a 6- × 50-µm HEMT with resistive feedback, and the other two

designs use source inductance to create broadband LNAs. These small compact

designs are intended to be biased through an external bias tee for the drain supply.

Figure 25 shows the layouts of a 6- × 50-µm feedback amplifier using a series 450-

ohm resistor and capacitor for the feedback path, plus a 4- × 50-µm and a 4- × 65-

µm feedback amplifier that use source inductance. The 6- × 50-µm resistive

feedback amplifier simulation shows 13.5-dB gain at 0.5 GHz rolling off to 10.1-

dB gain at 7 GHz with a minimum noise figure of 1.5 dB increasing gradually to 2

dB at 7 GHz. Simulations of the 6- × 50-µm MMIC layout (solid) versus an ideal

lossless amplifier are shown in Fig. 26. Next, the simulations of the 4- × 50-µm

22

feedback amplifier show the MMIC element performance (solid) versus an ideal

lossless design in Fig. 27. The 4- × 50-µm source feedback amplifier simulation

shows 15.5-dB gain at 0.5 GHz rolling off to 10.5-dB gain at 7 GHz with a

minimum noise figure of 1.1 dB increasing gradually to 1.25 dB at 7 GHz. The

second source feedback amplifier using a slightly larger 4- × 65-µm HEMT

provides higher low-frequency gain and slightly better noise figure. Simulations of

the 4- × 65-µm feedback amplifier show the MMIC element performance (solid)

versus an ideal lossless design in Fig. 28. The 4- × 65-µm source feedback amplifier

simulation shows 17.3-dB gain at 0.5 GHz rolling off to 10.3-dB gain at 7 GHz

with a minimum noise figure of 1.0 dB increasing gradually to 1.2 dB at 7 GHz.

Fig. 25 Final compact layouts of broadband feedback amplifiers (6× 50 µm, 4× 50 µm, and

4× 65 µm)

23

Fig. 26 Simulation of 6- × 50-µm resistive feedback amplifier (MMIC [solid] vs. ideal [dot])

Fig. 27 Simulation of 4- × 50-µm source feedback amplifier (MMIC [solid] vs. ideal [dot])

24

Fig. 28 Simulation of 4- × 65-µm source feedback amplifier (MMIC [solid] vs. ideal [dot])

6. Broadband Nonuniform Distributed Amplifier

Distributed amplifiers (DAs) provide very broadband gain and decades of

bandwidth but are typically a compromise on performance when power efficiency

or low noise figure is desired. Nonuniform DAs have been explored to achieve large

bandwidth gain with better efficiencies for PAs. A nonuniform approach was

explored, but with lower noise figure as the main objective. Using the 4-finger

HEMT model with noise data, a simple uniform ideal DA was designed using 4- ×

20-µm HEMT devices. Figure 29 shows the simple ideal schematic and Fig. 30

shows the preliminary simulation with good flat 10-dB gain to nearly 30 GHz. Note

the stability problem near 40 GHz that needs to be addressed. The noise figure of

the ideal DA is almost 2 dB at its minimum and is below 2.5 dB from about 8 to

26 GHz. This initial design was retuned, allowing varying HEMT sizes resulting in

a better noise figure, below 2.5 dB from below 5 GHz up to 26 GHz with a

minimum noise figure of 1.8 dB at 8 GHz (Fig. 31). Gain of this retuned DA is

higher, and flat to 30 GHz, while stability is vastly improved. Once the ideal DA

was converted to MMIC elements, EM-simulated, and re-tuned, the final EM

25

simulations show a higher noise figure closer to 2.5-dB average over 6 to 26 GHz,

with a minimum of 2.2 dB at 8 GHz. Gain is still good, close to 10 dB to 30 GHz.

Figure 32 shows the EM simulation (solid) versus the original lossy MMIC element

nonuniform DA simulation (dotted). The final layout is shown in Fig. 33 of the

nonuniform distributed amplifier (NUDA). It should have very good gain

bandwidth but the noise figure is higher than desired at lower frequencies, below

6 GHz.

Fig. 29 Schematic of initial ideal uniform 4- × 20-µm DA

Fig. 30 Simulation of initial ideal uniform 4- × 20-µm DA

PORTP=1Z=50 Ohm

PORTP=2Z=50 Ohm

LLd=0.5

LLg=0.3

LLd2=0.24 LLd2=0.1375

LLg2=0.15

Wf=20

INDID=L4L=LLd nH

CAPID=C1C=12 pF

CAPID=C4C=12 pF

INDID=L3L=LLd2 nH

RESID=R1R=50 Ohm

CAPID=C2C=12 pF

INDID=L5L=LLd nH

INDID=L8L=LLg nH

INDID=L2L=LLd nH

INDID=L9L=LLg2 nH

INDID=L7L=LLg2 nH

INDID=L1L=LLd2 nH

INDID=L6L=LLd nH

CAPID=C3C=12 pF

INDID=L10L=LLg nH

G

D

QGAN15_ES_FET2_V2ID=LN_1416_4x25_2Wu=Wf umN=4VIA_count_per_pad=1VIA=No ShareBIAS=5V 100 mA/mm

INDID=L12L=LLg nH

G

D


INDID=L11L=LLg nH

G

D


G

D


RESID=R2R=50 Ohm

G

D


0 5 10 15 20 25 30 35 40 45 50 55 60

Frequency (GHz)

amp_damp_idl

-25

-20

-15

-10

-5

0

5

10

15

0

0.5

1

1.5

2

2.5

3

3.5

413 GHz9.85 dB

13 GHz2.044 dB

DB(|S(1,1)|) (L)

Damp_4x25_idl

DB(|S(2,1)|) (L)

Damp_4x25_idl

DB(|S(2,2)|) (L)

Damp_4x25_idl

MU1(2,1) (R)

Damp_4x25_idl

DB(NF(2,1)) (R)

Damp_4x25_idl

26

Fig. 31 Simulation of re-tuned ideal NUDA

Fig. 32 EM simulation (solid) of nonuniform MMIC DA

27

Fig. 33 Layout of NUDA

7. Conclusions

Multiple MMIC designs were submitted to an ARL Qorvo 0.15-µm GaN prototype

wafer option to demonstrate the performance, bandwidth, capability, versatility,

and applicability of GaN for compact, efficient microwave circuit designs—

particularly for air and missile defense radars, but also applicable to EW, network

communications, and efficient amplifiers for future radio and communications

networks.

The designs described in this report are not the only designs in fabrication, nor the

only MMIC designs targeted for radar applications. Those designers are expected

to document these other designs in current and future reports. Some of the LNAs

and PAs built upon prior work with Raytheon, which has a high-performance in-

house GaN process. Future reports will document the testing and characterization

of the designs that are expected to be fabricated and returned 3–4 months after

commitment to the mask submission, which was completed in mid-September

2019.

28

The III/V design team has developed a full reticle of creative high-performance

circuits that were submitted to fabrication in Qorvo’s 0.15-µm commercial GaN

process.2

29

8. References

1. Penn JE, Darwish A. Broadband low noise gallium nitride (GaN) amplifiers

for next-generation radars. Adelphi (MD): Adelphi (US); 2017 Nov. Report

No.: ARL-TR-8208.

2. Penn J, Darwish A, Hawasli S, McKnight K, Hawasli S. Gallium nitride high-

electron-mobility transistor (HEMT) monolithic microwave integrated circuit

(MMIC) designs submitted for Qorvo prototype wafer option (PWO)

fabrication. Adelphi (MD): CCDC Army Research Laboratory (US); 2019

Sep. Report No.: ARL-TR-8811.

30

List of Symbols, Abbreviations, and Acronyms

ARL Army Research Laboratory

CCDC US Army Combat Capabilities Development Command

DA distributed amplifier

DC direct current

EM electromagnetic

EW electronic warfare

GaN gallium nitride

HEMT high-electron mobility transistor

LNA low-noise amplifier

MMIC monolithic microwave integrated circuit

NUDA nonuniform distributed amplifier

PA power amplifier

PAE power-added efficiency

PDK process design kit

Q quality factor

31

1 DEFENSE TECHNICAL

(PDF) INFORMATION CTR

DTIC OCA

1 CCDC ARL (PDF) FCDD RLD CL

TECH LIB

11 CCDC ARL

(PDF) FCDD RLS R

(1 HC) P AMIRTHARAJ

FCDD RLS RE

R DEL ROSARIO

A DARWISH

T IVANOV

P GADFORT

S HAWASLI K KINGKEO

K MCKNIGHT

J PENN (1 HC)

J WILSON

FCDD RLS RW

E VIVEIROS

Documents

Low-Noise Amplifiers (LNAs) and Power Amplifiers (PAs) for