HW #6HW #6 Single Stage Single Stage Amplifier DesignAmplifier DesignAmplifier DesignAmplifier Design

Ping ChenPing Chen

ECECS Dept, Univ. of CincinnatiECECS Dept, Univ. of Cincinnati

Design Objecti eDesign Objecti eDesign ObjectiveDesign Objective

Operating frequency: 8 GHzOperating frequency: 8 GHz Pick a chip level transistor that has enough gain Pick a chip level transistor that has enough gain p g gp g g

at 8 GHz: at 8 GHz: HP HBFPHP HBFP--0405 BJT0405 BJT S parameters and noise figure parametersS parameters and noise figure parameters Stability for the transistor at 8 GHzStability for the transistor at 8 GHz Input and output matching networksInput and output matching networksp p gp p g Simulate your final amplifier designSimulate your final amplifier design Layout including the bias networkLayout including the bias networkLayout including the bias networkLayout including the bias network

Choose a t ansistoChoose a t ansistoChoose a transistorChoose a transistor

cb_xx_xxxxcb_xx_xxxx --> Chip BJTs> Chip BJTs cf_xx_xxxxcf_xx_xxxx --> Chip > Chip GaAsGaAs FETsFETs

ChipChip--mounted transistorsmounted transistors --> best performance; problem> best performance; problemChipChip mounted transistors mounted transistors > best performance; problem > best performance; problem –– No layout in ADSNo layout in ADS

pb xx xxxxpb xx xxxx --> Packaged BJTs> Packaged BJTs pb_xx_xxxxpb_xx_xxxx --> Packaged BJTs> Packaged BJTs pb_xx_xxxxpb_xx_xxxx --> Packaged > Packaged GaAsGaAs FETFET

Packaged transistors Packaged transistors --> Layout available, need bias > Layout available, need bias circuitcircuitcircuitcircuit

sp_xx_xxxxsp_xx_xxxx --> Give S Parameter at specific bias point> Give S Parameter at specific bias pointSP model SP model --> Easy to start with, no bias needed> Easy to start with, no bias needed

HP HBFP0420 BJTHP HBFP0420 BJTHP HBFP0420 BJTHP HBFP0420 BJT

2V 5mA2V 5mA

2V 20mA2V 20mA

HP HBFP0405 BJTHP HBFP0405 BJTHP HBFP0405 BJTHP HBFP0405 BJT

2V 2mA2V 2mA

2V 5mA2V 5mA

HBFP0405: SOT-343 Package 3-terminal, NPN Pdiss=54mW, Vce(max)=4.5V, Ic(Max)=12mA, Vce(typical)=2V, Ic(typical)=2mA, Hfe=80, Ft=25GHz

BJT t aceBJT t aceBJT Curve Tracer

BJT tracerBJT tracerGenerates IV curves, computes Gm versus bias,and optimal bias for Class A operation.

pb_hp_HBFP0405_19980529Q1C

C1

V_DCSRC1Vdc=VCE I DC

I_ProbeIC

V_ACSRC3Freq=freq

C1C=1 F

DCPARAMETER SWEEP

I_DCSRC2Idc=IBB

Set base current and col lector voltage sweep limits and frequency at which the transconductance Gm will be calculated, as needed.

VARVAR1

VCE Stop=4 5VCE_Start=0

EqnVar

DCDC1

Step=VCE_StepStop=VCE_StopStart=VCE_StartSweepVar="VCE"

PARAMETER SWEEP

ParamSweepSweep1

SimInstanceName[5]=SimInstanceName[4]=SimInstanceName[3]=SimInstanceName[2]="Sweep2"SimInstanceName[1]="DC1"SweepVar="IBB"

BJ T_c urv e_trac erSub_pal ibX1

AC

VCE_Step=0.1VCE_Stop=4.5

ParamSweepSweep2

SimInstanceName[4]=SimInstanceName[3]=SimInstanceName[2]=SimInstanceName[1]="AC1"SweepVar="VCE"Step=20 uA

Stop=200 uAStart=20 uASimInstanceName[6]=

ACAC1Freq=8 GHz

Step=VCE_StepStop=VCE_StopStart=VCE_StartSimInstanceName[6]=SimInstanceName[5]=

[ ]

Follow these steps:1) Move marker m2 to the knee of the I-V curve. ThisDev ice IV Curv es, Load Lines,

d M i DC Di i ti C

0.015

0.020

0.025

IBB=200.uC.i,

A

m2

max

ne1) Move marker m2 to the knee of the I V curve. This sets the maximum collector current during AC operation.2) Specify maximum allowed VCE, VCEmax. The optimal bias point values are determined from the load line between marker m2 and the (IC=0, VCE=VCEmax) point.3) Specify maximum allowed DC power dissipation, PDmax, in Watts. 4) Position marker m1 at some other bias point, if desired. (Must be less than VCEmax.)5) DC power consumption, average output power

i li ti DC t RF ffi i t k 1

and Maximum DC Dissipation Curv e

Equations are on the "Equations"

0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 00 0 4 5

0.005

0.010

0.000IBB=20.0uIBB=40.0uIBB=60.0uIBB=80.0uIBB=100.uIBB=120.uIBB=140.uIBB=160.uIBB=180.u

DC

.IC

m1

ICmlin

Eqn VCEmax=4.5 Eqn PDmax=0.054

in linear operation, DC-to-RF efficiency at marker m1 bias point are all calculated.

Optimal Class A bi i t l

page. Gm plotsare on the"IV and Gm vs.Bias" page.

m1VCE=DC.IC.i=5.246mIBB=0.000080

2.000 m2VCE=DC.IC.i=11.77mIBB=0.000200

500.0m

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.0 4.5

VCEVCEvals

5.883m 2.500 40.00

DC-to-RF Efficiency,%

Optimal VCE

DC P C ti

Output Powerat Optimal BiasWatts dBm

Rl d t

bias point values.

7.696

339.941 5.883m14.71m

250

300

Beta v ersus IBB, at ICEspecif ied by marker m1

Optimal ICE

DC Power Consumption at Optimal Bias

Rload atOptimal Bias

Marker m1 bias point values, (Assuming Class A, AC current limited to marker m2 value and AC voltage nohigher than VCEmax.)

10.49m

3.934m 37.50

230.02140.

60.

80.

100

120

140

160

180

20.

20050

100

150

200

0

Beta

DC-to-RF Efficiency,%

DC PowerConsumptionRload

Output PowerWatts dBm

5.949

0u 0u 0u

0.u

0.u

0.u

0.u

0.u0u

0.u

IBB

SS pa amete sim lationpa amete sim lationSS--parameter simulationparameter simulation

sp_hp_HBFP-0405_2_19980529SNP1

Frequency="{0.10 - 10.00} GHz"Bias="Bjt: Vce=2V Ic=5mA"

TermTerm1Num=1

TermTerm2Num=2

DisplayTemplateTempDisp

S-PARAMETERS

sp_hp_HBFP-0405_1_19980529SNP2

Noise Frequency="{0.90 - 10.00} GHz"Frequency="{0.10 - 10.00} GHz"Bias="Bjt: Vce=2V Ic=2mA"

Z=50 OhmNum=1

Z=50 OhmNum=2

p y pdisptemp1"S_Params_Quad_dB_Smith"

TempS_ParamSP1

Step=1 GHzStop=10.0 GHzStart=1.0 GHz pb_hp_HBFP0405_19980529

Q1

Matching networks Matching networks –– use use ggtransmission linetransmission line

Matching networks Matching networks –– converted to converted to ggmicrostrip linemicrostrip line

Bias ci c its (afte optimi ation)Bias ci c its (afte optimi ation)Bias circuits (after optimization)Bias circuits (after optimization)2.50 V -5.56 mA

2.50 V

2.50 V

2.50 V566 uA 5.00 mA

V_DCVCCVdc=2.5 V

5.00 mAI_ProbeICE

2.00 V

RRB1R=2845.17 opt{ 100 to 4000 }

RRC1R=100.859 opt{ 10 to 800 }

890 mV

890 mV

00VCE

890 mV76.1 uA

5.00 mA

pb_hp_HBFP0405_19980529Q176.1 uA

I_ProbeI_Probe1

890 mV

490 uARRB2R=1818.2 opt{ 100 to 5000 }

-5.08 mA

Q1

DCDCDC1

DC

Ve if ing SP modelsVe if ing SP modelsVerifying SP modelsVerifying SP models

Comparing Representations: Device Data vs. Model

Measured S-parameters on thousands of active devices are available in the S-parameter Llbrary. Select the Library browser icon, and scroll down to "S Parameter Library (No Layout)".No license is required to access these libraries.

Nonlinear, large-signal models are available in the RF Transistor Library, which is also accessed

sp_hp_HBFP-0405_2_19980529SNP1

Frequency="{0.10 - 10.00} GHz"Bias="Bjt: Vce=2V Ic=5mA"

0 V

0 V

S-parameters for both devices will be calculated. Representing the2 2-port transistors using a single 4-port simulation is an easy wayto compare 2 results side-by-side. Results are displayed in ModelVerif dds

Nonlinear, large signal models are available in the RF Transistor Library, which is also accessed using the the Library Browser. A license is required to access these models.

Place S-parameter-based component here:

S_ParamSP1Start=1 GHz

S-PARAMETERS

0 A

0 A

0 ATermTerm1Num=1

0 ATermTerm2

Z=50 OhmNum=2

ModelVerif.dds.

An equivalent method would be to create 2 separate designs, one foreach representation of the device, calculate each set of 2-port paramters, and call the separate data sets in the data display.

Notice that VBE has been set at 0.788V to ensure the device willdraw ICE=10mA for VCC=8V. This value for VBE is calculated in BiasSetup.dsn. Using these values ensures that the model is operatingat the same bias that the measured data was taken at.Place packaged component here:

Lin=Stop=10 GHzStart 1 GHz

0 A

DC_Feed

Z=50 OhmNum=1

DC_FeedDC Feed1

890 mV890 mV2 V

890 mV 2 V2 V

-76.0 uA

V_DCVBBVdc=0.8903 V

-5.00 mAV_DCVCCVdc=2 V

-5.07 mA

76.0 uA

5.00 mA

pb_hp_HBFP0405_19980529Q1

DC_BlockDC_Block2

-5.00 mA

DC_Feed2

0 A

76.0 uA

_

0 ADC_BlockDC_Block1

0 A

0 ATermTerm4

Z=50 OhmNum=4

TermTerm3

Z=50 OhmNum=3

M d l V ifi ti

25

30

Model Verification

The simulation compares two mathematical representations of the performance of the Avantek AT-41411. Results show closeagreement over the frequency range from 100MHz to 2GHz.

10

15

20

25

dB(M

odel

Ver

if..S

(2,1

))dB

(Mod

elV

erif.

.S(4

,3))

odel

Ver

if..S

(1,1

)od

elV

erif.

.S(3

,3)

Problem: The SP model didn’t match

2 3 4 5 6 7 8 91 10

5

0

freq, GHz

freq (1.000GHz to 10.00GHz)MM

with the packaged BJT model.

elV

erif.

.S(2

,2)

elV

erif.

.S(4

,4)

0 10 0 05 0 00 0 05 0 100 15 0 15Ver

if..S

(1,2

)V

erif.

.S(3

,4)

freq (1.000GHz to 10.00GHz)

Mod

Mod

-0.10 -0.05 0.00 0.05 0.10-0.15 0.15

f (1 000GH t 10 00GH )

Mod

elV

Mod

elV

freq (1.000GHz to 10.00GHz)

sc_mrt_MC_GRM39C0G050_C_19960828Cout

sc_mrt_MC_GRM39C0G050_J_19960828Cin

OFFSET=10 milSMT_Pad="Pad_0603"PART_NUM=GRM39C0G120J050 12pF

Cout

OFFSET=10 milSMT_Pad="Pad_0603"PART_NUM=GRM39C0G030C050 3pF

pb_hp_HBFP0405_19980529 sl_cft_0603HS_J_19960828Lout1

PortP2N 2

SMT_PadPad_0603

L=20 milW=30 mil

SMT_Pad

Q1sl_cft_0603HS_J_19960828LoutPART_NUM=0603HS-220XJB 20.9 nHSMT_Pad="Pad_0603"OFFSET=10 mil

Lout1PART_NUM=0603HS-220XJB 20.9 nHSMT_Pad="Pad_0603"OFFSET=10 mil

sr_ims_RC-I_0603_J_19950814RC

OFFSET=10 milSMT_Pad="Pad_0603"PART_NUM=RC-I-0603-1000-J 100 Ohm

sr_ims_RC-I_0603_J_19950814RB2PART NUM=RC-I-0603-1802-J 18 kOhm

sr_ims_RC-I_0603_J_19950814RB1PART NUM=RC-I-0603-2702-J 27 kOhm

Num=2PortP1Num=1

PO=-10 milSM_Layer="cond"SMO=0 milPadLayer="cond"L=20 mil

OFFSET=10 milSMT_Pad="Pad_0603"

_

OFFSET=10 milSMT_Pad="Pad_0603"PART_NUM RC I 0603 2702 J 27 kOhm

PortP3Num=3

S-Parameters, Noise Figure, Gain, Stability,Circles, and Group Delay versus Frequency

amp_all3X1AMP

VCCRF in RF out

V_DCSRC1Vdc=2.5 V

TermTerm1

Z=Z0Num=1

TermTerm2

Z=Z0Num=2

Set SystemImpedance Z0:

VARVar

Z Z0

S_ParamSP1

S 0 1 GHStop=10 GHzStart=1 GHz

S-PARAMETERS VAR1Z0=50

EqnVar

OptionsOptions1

Tnom=25Temp=16.85

OPTIONS

Set S-parameter analysis frequencyrange. If an S-parameter file without noise data is used, the noisesimulation results will be invalid.

Computation ofStability factorsand circles:CalcNoise=yes

Step=0.1 GHz

MeasEqnmeas1

EqnMeas

Maximum Available Gain, Associated Power Gain (input matched for NFmin, output then conjugatelymatched) and dB(S21)

Stability Factor, K Geometric stability factorsmu source and mu load

Minimum Noise Figure, dB,and Noise Figure with Z0Ohm terminations

5

10

15

20

0

25P

gain

_ass

ocM

AG

dB(S

21)

2 3 4 5 6 7 8 91 10

0.4

0.6

0.81.0

1.2

0.2

1.4

Km

u_loa

dm

u_so

urce

output then conjugately matched), and dB(S21)

1.52.02.53.03.54.0

1.0

4.5

NF

min

nf(2

)

mu_source and mu_loadOhm terminations

Equations are on the "Equations" page.

Move marker m1 to select freq point. All listings and impedances on Smith Chartwill be updated.

2 3 4 5 6 7 8 91 100

freq, GHz

2 3 4 5 6 7 8 91 10

freq, GHz

2.00G

3.00G

4.00G

5.00G

6.00G

7.00G

8.00G

9.00G

1.00G

10.0G

m1

RF Frequency Selector

2 3 4 5 6 7 8 91 10

freq, GHz

Source and Load Stability Circles; Optimal Source Reflection Coefficients for Mininum NF (Sopt), Simultaneous Conjugate Matching, and Load Reflection Coefficient for Simultaneous Conjugate Matching, and with source matched for NFmin RF F

If either mu_source or mu_load is >1,the circuit is unconditionally stable.

See also the "Gain, Noise, and Stability Circles" page, and the"S Parameters, Group Delay" page.

Sop

top

t_at

_m1

maS

_at_

freq_

ptm

aL_a

t_fr

eq_p

tm

maL

_wS

opt

e_st

abcir

[m1,

::]

_sta

bcir[

m1,

::]

Matching, and with source matched for NFmin

MaximumAvailable

dB(S11)

-3.506dB(S21)

5.930dB(S12)

-16.918dB(S22)

-8.622

Matching For Gain

8.000GHz

RF Frequency

0.923Stability FactorZsource Zload

DUT*

freq (1.000GHz to 10.00GHz) (0.000 to 0.000)

SG

amm

Gam

mG

am

indep(Source_stabcir[m1,::]) (0.000 to 51.000)

Sour

ce

indep(Load_stabcir[m1,::]) (0.000 to 51.000)

Load

_

11.424

AvailablePower Gain, dB Simultaneous Match

Zsource0.000 + j0.000

Simultaneous MatchZload

0.000 + j0.000

Power Gain with theseSource and Load

Conjugate Match Load Impedance if Source R fl ti C ffi i t

Matching For Noise Figure

SystemImpedance

50.000

Note: if the device (or circuit) is unstable at the freq point, the simultaneous conjugate matching impedances are undefined and GammaL_at_freq_pt and GammaS_at_freq_pt default to 0. Also, MAG is set equal to the maximum stable gain, | S21| /| S12| .

0.198 / -170.180

Source ReflectionCoefficient for Minimum NF

Zopt for NFmin33.6 - j2.36 7.59743.4 + j37.0

Source and Load Reflection Coefficients

Reflection Coefficient is Sopt for Minimum NFNFmin, dB

3.004

Zopt

Conjugate match Zload if source impedance is Zopt

DUT* *DUT= Device Under Test(simulated circuit or device)

Set step sizes and number of circles here

Available Gain & Noise Circles,Source Stability CircleSource Gamma.Corresponding Load Gamma,(Bl kD t)

Power Gain Circles,Load Stability CircleLoad Gamma, GammaLCorresponding Source Move markers GammaS and GammaL to

Eqn num_NFcircles=3Eqn NFstep_size=.2

Eqn GAstep_size=1Eqn num_GAcircles=3

number of circles, here.

Stability

Eqn num_GPcircles=3Eqn GPstep_size=1

GammaS

cir[m

RF,

::]esop

tcl

escl

eMin

(Black Dot)

[mR

F,::]

ga in=11.424

ga in=10.424

ga in=9 .424

gain=8 .424

sopt

Gamma, (Black Dot)Move markers GammaS and GammaL to select arbitrary source and load reflection coefficients. The impedances, power gains,and noise figures below will be updated. The transducer power gains are invalid if the markersare moved into the unstable regions (usually inside the stability circles.)

0.923

Stability Factor, K

Outside

Source Stable Region

Move marker to desired frequency. Plots willbe updated.

mRF

RF Frequency Selector

rhos

Sour

ce_s

tabc

ga in=11.424ga in=10.424

gain=9 .424ga in=8 .424

GAc

ircle

Gam

maL

n s figu re=3 .204ns figu re=3.404ns figure=3.604N

oise

_cir

Noi

se_c

irc

rhos GammaL

Load

_sta

bcir[

GPc

ircle

sG

amm

aSo

Zsource,Source Gamma

Zload,Load Gamma

DUT*

*DUT= Device Under Test(simulated circuit or device)

2.00G

3.00G

4.00G

5.00G

6.00G

7.00G

8.00G

9.00G

1.00G

10.0G

freq, Hz

8.000GHzRF Frequency System Impedance Z0

50.000

Load Stable Region

indep(rhos) (0.000 to 2000.000)indep(Source_stabcir[mRF,::]) (0.000 to 51.000)

cir_pts (0.000 to 51.000)indep(GammaLopt) (826.000 to 826.000)indep(Noise_c irc leMin) (0.000 to 51.000)

indep(rhos) (0.000 to 2000.000)indep(Load_stabcir[mRF,::]) (0.000 to 51.000)

cir_pts (0.000 to 51.000)indep(GammaSopt) (112.000 to 112.000)

Source Impedance Transducer Power Gain, dBwhen these source and

Optimal load impedance for power transfer when source

Noise Figure (dB) withSource Impedance

Outside

3.004NFmin,dB

33.587 - j2.365

Source Impedance, Zopt, for Minimum NF

25.926 + j55.395 63.449 + j28.981 1.647

7.597

pat marker GammaS

Optimal load impedance for power transfer when source impedance is Zopt

when these source andload impedances are used

43.365 + j37.006

pimpedance at marker GammaS is presented to input

Transducer Power Gain, dBwhen these source andload impedances are used

5.411

Sou ce peda ceat marker GammaS

Equations are shown on the "Equations2" page.

See also the "NF, Gain, Stab. Fact., Matching" pageand the"S Parameters, Group Delay" page.

11.424

Maximum AvailableGain, dB (MaximumStable Gain, S21/S12) if K<1

Zsource0.000 + j0.000

Zload

0.000 + j0.000

Simultaneous Conjugate Matching(only v alid if stability f actor K >1)

1000

Noise Figure (dB) withZsource (only valid with K>1)

Transducer Power Gain, dBL d I d

Optimal source impedance for power transfer when load

Noise Figure (dB) with this optimal

, p y p g

12.945 - j11.082 7.77279.913 - j7.304

,when these source andload impedances are used

Load Impedanceat marker GammaL

power transfer when load impedance at marker GammaL is presented to output

3.866

with this optimal source impedance(at right)