ADS LNA simulation example

ADS documentation Apart from the handed-out ADS tour: a lot more info on ADS and simulation tricks, know-how on [Link]

Project data

Project name: LNA_PRJ Technology: CMOS 0.25 um (included via netlist

statement)

Number of networks: 9 (plus one sub-network) Description: Shows how to simulate all important
specifications of a (common-source cascode) LNA. Output matching is not done, because after the LNA no external (50 Ohm) filter is anticipated: hence no need for matching (which cost 3 dB in gain)

Simulation of MOS characteristics

FET Curve Tracer

V_DC SRC1 Vdc=VDS

I_Probe IDS

NETLIST INCLUDE
NetlistInclude NetlistInclude1 IncludeFiles[1]=[Link]

V_DC SRC3 Vdc=VGS

MOSFET_NMOS MOSFET1 Model=nfet Length=0.25 um W idth=100 um

M eas Eqn

M easEqn M eas1 M OS_Gm=M [Link]

PARAM ETER SWEEP

ParamSweep Sweep1 SweepVar="VGS" SimInstanceName[1]="DC1" SimInstanceName[2]= SimInstanceName[3]= SimInstanceName[4]= SimInstanceName[5]= SimInstanceName[6]= Start=0 Stop=1.4 Step=0.1

DC
DC DC1 SweepVar="VDS" Start=0 Stop=2.5 Step=0.1 Other=

V ar Eqn

VAR VAR1 VGS=0 VDS=0

Set gate and drain voltage sweep limits as needed. If the transistor instance name changes (for example, M OSFET2 instead of M OSFET1), then the M OS_Gm equation must also be changed. If a library part is used on this schematic, then the M OS_Gm equation will have to be set to something like [Link].

Schematic: FET_curve_tracer

Example of simulation output

De vice I-V Curve s

40 35 30
VGS=1.200 VGS=1.400

DC Tra ns conducta nce ve rs us VDS

0.035 0.030 0.025 MOS _Gm 0.020 0.015 0.010 0.005 0.000 -0.005 0.0 0.5 1.0 VDS 1.5 2.0 m2
VGS=0.500 VGS=0.400 VGS=0.300 VGS=0.200 VGS=0.100 VGS=0.000 VGS=1.400 VGS=1.300 VGS=1.200 VGS=1.100 VGS=1.000 VGS=0.900 VGS=0.800 VGS=0.700 VGS=0.600 VGS=1.300

IDS .i, mA

25
VGS=1.100

20
VGS=1.000

15 m1 10
VGS=0.800 VGS=0.900

5 0 0.0 0.5 1.0 VDS 1.5 2.0

VGS=0.700 VGS=0.600 VGS=0.500 VGS=0.400 VGS=0.300 VGS=0.200 VGS=0.100 VGS=0.000

2.5 m1 VDS = 1.000 IDS .i=0.010 VGS =0.900000

2.5 m2 VDS = 2.300 MOS _Gm =0.003 VGS =0.500000

DC transfer curves to determine proper biasing voltage Vgs

DC and S-parameter simulation

NETLIS T INC LUDE

Ne tlis tInc lude Ne tlis tInc lude 1 Inc lude File s [1]=ge ne ric [Link]

T e rm T e rm2 Num=2 Z=300 Ohm

Port-impedance: term for Sparam simulation

V_DC S RC1 Vdc =2.5 V

I_P robe ID V_DC S RC2 Vdc =0.75 V

DC_Fe e d DC_Fe e d1

Vdra in T he S pa ra me te rs a re s imula te d to c he c k ga in a nd s ta bility.

Vga te T e rm DC_Bloc k T e rm1 DCBloc k1 Num=1 Z=50 Ohm

MO S F ET_NMO S MO S F ET1 Model= nfet Length= .25 um Width= 100 um Tris e=

S -P AR AMETER S
S _P a ra m SP1 S ta rt=10 MHz S top=3.0 GHz S te p=10 MHz Mu mu1 mu_s ourc e =mu(S )

MuP rime

Port-impedance: term for Sparam simulation Schematic: DC_and_Sparams

DC
DC DC1

Mu

MuP rime mup1 mu_loa d=mu_prime (S )

Stability measures

Example of simulation output

fre q 0.000 Hz fre q 0.000 Hz

20

Vga te [Link] ID.i 4.35mA

Vdra in 1.19 V

Dc conditions

Target for S11: 50 Ohm

m2

0 dB(S (2,1)) dB(S (1,2))

gain
Reverseisolation
0.0 0.5 1.0 1.5 fre q, GHz 2.0 2.5 3.0

m2 fre q=2.000GHz dB(S (2,1))=8.982

-20

-40

-60

S-parameters

S (2,2) S (1,1)

fre q (10.00MHz to 3.000GHz)

DC and S-parameter simulation: cascode LNA

N E T L IS T IN C L U D E

Cascode LNA for improved stability and isolation and less miller effect
V_ D C SRC2 Vd c = 0 .7 5 V

N e tlis tIn c lu d e N e tlis tIn c lu d e 1 In c lu d e F ile s [1 ]= g e n e r ic 0 2 5 .lib

V_ D C SRC1 Vd c = 2 .5 V Term Term2 N um= 2 Z = 100 O hm

Vo u t
MO S F E T _ N MO S MO S F E T 2 Mo d e l= n fe t Le n g th = 0 .2 5 u m W id th = 1 0 0 u m T ris e =

S -P A R A M E T E R S
S _P a ra m SP1 S ta r t= 1 0 M H z S to p = 3 .0 G H z S te p = 1 0 M H z

V_ D C SRC3 Vd c = 1 .8 V

I_ P r o b e ID
Mu

D C _F e e d D C _F e e d1

Vd r a in

Mu mu1 mu_s ourc e = mu(S )

Vg a te Vin D C _ Blo c k D C Blo c k1

MuP rime

Te rm Te rm1 N um= 1 Z = 50 O hm

D A_ S m ith C h a r tM a tc h 1 _ D C _ a n d _ S p a r a m s _ c a s c a d e D A_ S m ith C h a r tM a tc h 1 F = 1 .9 G H z Z s = 50 O hm L Z l= ( 5 0 .0 0 - j*1 2 7 .5 ) O h m L1 Z 0= 50 O hm L= 1 .0 5 n H R=

MO S F E T _ N MO S MO S F E T 1 Mo d e l= n fe t Le n g th = 0 . 2 5 u m W id th = 1 0 0 u m T ris e =

DC
DC DC1

T h e S p a r a m e te r s a r e s im u la te d to c h e c k g a in a n d s ta b ility.

M u P r im e mup1 m u _ lo a d = m u _ p r im e ( S )

Me a s Eq n

Meas Eq n Meas 1 Vd s _ c a s c a d e = Vo u t- Vd r a in

Schematic: DC_and_Sparams_cascode

Smith chart component first disabled and shorted to see unmatched S11

Example of simulation output

20 0 -20 dB(S (2,1)) dB(S (1,2)) -40 -60

m2

m2 fre q= 2.000GHz dB(S (2,1))=9.109

m3 freq= 1.900GHz S (1,1)=0.793 / -37.940 impedance = Z0 * (0.983 - j2.579)

S (2,2) S (1,1)
0.0 0.5 1.0 1.5 fre q, GHz 2.0 2.5 3.0

-80 -100

m3

Improved isolation due to cascode stage (and thus stability)

freq (10.00MHz to 3.000GHz)

Coil in source has right value, but inductive matching network needed

DC and S-parameter simulation: cascode LNA Same schematic but smith-chart component enabled
NET LIS T INCLUDE
Netlis tInclude Netlis tInclude1 IncludeF iles [1]=[Link] V_DC S RC1 Vdc= 2.5 V Term Term2 Num= 2 Z=100 O hm Vout
MOSFET_NMOS MOSFET2 Mode l=nfe t Le ngth=0 .2 5 u m Width =10 0 um Tris e =

S -P ARAMET ERS
S _P aram S P1 S tart= 10 MHz S top= 3.0 G Hz S tep= 10 MHz

V_DC SRC3 Vdc= 1.8 V

I_P robe ID V_DC S RC2 Vdc= 0.75 V

Mu

DC_Feed DC_Feed1

Vdrain

Mu mu1 mu_s ource= mu(S )

Vgate Vin DC_Block DCBlock1

MuPrime

Term Term1 Num= 1 Z= 50 O hm

DA_SmithChartMatch_DC_and_Sparams _cas code DA_SmithChartMatch1 F= 1 G Hz Zs = (50+ j*0) O hm L Zl= (50+ j*0) O hm L1 Z0= 50 O hm L= 1.05 nH R=

MOSFET_NMOS MOSFET1 Mode l=n fe t Le n gth=0 .2 5 u m Wid th =10 0 um Tris e =

DC
DC DC1

The S parameters are s imulated to check gain and s tability.

MuP rime mup1 mu_load=mu_prime(S)

Meas E qn

Meas Eqn Meas 1 Vds _cas cade= Vout-Vdrain

Schematic: DC_and_Sparams_cascode

Use design guide to match input (I)

In schematic, ADS Menu: DesignGuidefiltersmith chart control window

Set frequency at 1.9 GHz unmark: normalized impedances

Click On Zl

Fill in S11 impedance

Use design guide to match input (II)

S11

49.1-j128.95
After entry, press enter

Use design guide to match input (III)

Select series inductance

Move around smith chart until matched

Last step: build ADS circuit

Lumped Element Low P as s Filter Des ign As s is tant Need Help? P leas e s ee the appropriate Des ignGuide Us er Manual

Va r Eqn

VAR VAR1 P arameters ="#1.9 GHz#50 Ohm#(49.10-j*128.9) Ohm#50 Ohm" L L1 L=10.756248 nH R=1e-12 Ohm

P ort P1 Num=1

P ort P2 Num=2

Click on this button

Automatically creates this sub-network

Repeat simulation of schematic:

DC_and_Sparams_cascode
m2 m2 fre q= 2.000GHz dB(S (2,1))=13.063

20 0 -20 dB(S (2,1)) dB(S (1,2)) -40 -60 -80 -100 0.0 0.5 1.0 1.5 freq, GHz

m3 fre q=1.900GHz S(1,1)=0.010 / -148.430 impe da nce = Z0 * (0.983 - j0.010)

S (2,2) S (1,1)

m3

2.0

2.5

3.0

Improved gain (4 dB) and minimum return loss due to input matching.
fre q (10.00MHz to 3.000GHz)

DC and S-parameter simulation: cascode LNA

Same schematic as previous schematic but with smith chart component replaced by coil (~ 10 nH)
NETLIS T INCLUDE
Ne tlis tInc lude Ne tlis tInc lude 1 Inc lude F ile s [1]=ge ne ric [Link] Te rm Te rm 2 Num =2 Z=100 Ohm V_DC SRC1 Vdc =2.5 V

Vout
MOS FET _NMOS MOS FET 2 Mode l=nfe t Le ngth=0.25 um Width=100 um T ris e =

S -P ARAMETERS
S _P a ra m SP1 S ta rt=10 MHz S top=3.0 GHz S te p=10 MHz

V_DC SRC3 Vdc =1.8 V

I_P robe ID V_DC SRC2 Vdc =0.75 V DC _F e e d DC _F e e d1

Mu

Vdra in

Mu mu1 mu_s ourc e=m u(S ) The S pa ra m e te rs a re s im ula te d to c he c k ga in a nd s ta bility .

Vga te Vin DC _Bloc k DC Bloc k1 L L2 L=10.6 nH R=

MOS FET _NMOS MOS FET 1 Mode l=nfe t Le ngth=0.25 um Width=100 um T ris e =

DC
DC DC 1

MuP rime

Te rm Te rm 1 Num =1 Z=50 Ohm

L L1 L=1.05 nH R=

MuP rim e mup1 mu_loa d=m u_prim e (S )

Me a s Eqn

Schematic: DC_and_Sparams_cascode_match

Me a s Eqn Me a s 1 Vds _c a s c a de =Vout-Vdra in

AC simulation for noise figure

NETLIS T INC LUDE

N e t lis t In c lu d e N e t lis t In c lu d e 1 I n c lu d e F ile s [1 ]= g e n e r ic 0 2 5 . lib Te rm Te rm 2 N u m =2 Z =1 0 0 O h m

V_ D C SRC1 Vd c =2 .5 V

Vo u t

V_ D C SRC3 Vd c =1 .8 V

M O S F E T _N M O S MOSFET2 M o d e l= n fe t Le n g th = 0 .2 5 u m W id th = 1 0 0 u m T r is e =

AC
AC AC 1 S w e e p V a r= "f r e q " S t a rt = 1 0 0 M H z S to p =3 G H z S te p =1 0 0 MH z C a lc N o is e = y e s N o is e N o d e [ 1 ]= "V o u t " N o is e N o d e [ 2 ]= "V in " S o r t N o is e = S o r t b y v a lu e I n c lu d e P o rt N o is e = y e s

I_ P r o b e ID V_ D C SRC2 Vd c =0 . 7 5 V

D C _F eed D C _F eed1

V d ra in

V in D C _ B lo c k D C B lo c k 1 P _ AC P O R T1 N u m =1 Z =5 0 O h m P a c = p o la r( d b m t o w (0 ), 0 ) F re q = f re q L L2 L= 1 0 . n H R=

Vg a te

M O S F E T _N M O S MOSFET1 M o d e l= n fe t Le n g th = 0 .2 5 u m W id th = 1 0 0 u m T r is e =

N o t ic e t h a t f o r t h is f irs t n o is e s im u la t io n t h e q u a lit y f a c t o r o f t h e t w o in d u c t o r s is in if it e (o n re a lis t ic : s e e f o r a m o re re a lis t ic n o is e f ig u re L N A _ n o is e _ re a l_ Q ).

L L1 L= 1 . 0 n H R=

Coils are ideal here: no losses included yet

Schematic: LNA_noise

Example of simulation output

Eqn NF=20*log(([Link] e/(mag(Vout/Vin)))/P [Link] e)

Measurement equation used to calculated the noise figure
5 4 3

2 1 0 0.0 0.5 1.0 1.5

m1 freq= 1.900GHz NF=0.703 m1

Noise figure versus frequency

NF

2.0

2.5

3.0

fre q, GHz

AC simulation for noise figure with realistic coils

NETLIST INCLUDE
Ne tlis tInc lude Ne tlis tInc lude 1 Inc lude File s [1]=ge ne ric [Link] T e rm T e rm2 Num=2 Z=100 Ohm

V_DC S RC1 Vdc =2.5 V

Vout
MO S F ET_NMO S MO S F ET2 Model= nfet Length= 0.25 um Width= 100 um Tris e=

AC
AC AC1 S we e pVa r="fre q" S ta rt=100 MHz S top=3 GHz S te p=100 MHz Ca lc Nois e =ye s Nois e Node [1]="Vout" Nois e Node [2]="Vin" S ortNois e =S ort by va lue Inc lude P ortNois e =ye s

V_DC S RC3 Vdc =1.8 V

V_DC S RC2 Vdc =0.75 V

I_P robe ID DC_Fe e d DC_Fe e d1

Vdra in

Vin DC_Bloc k DCBloc k1 L L2 L=10. nH R=18

Vga te

MO S F ET_NMO S MO S F ET1 Model= nfet Length= 0.25 um Width= 100 um Tris e=

P _AC P ORT1 Num=1 Z=50 Ohm P a c =pola r(dbmtow(0),0) Fre q=fre q

Both Induc tors now ha ve a s e rie s re s is tor ma king the qua lity fa c tor of the c ompone nt a pprox 7

L L1 L=1 nH R=1.7

Resistance in coils set to have a quality factor of about 7

Schematic: LNA_noise_real_Q

Example of simulation output

5

NF

m1

1 0.0 0.5 1.0 1.5 2.0 2.5 3.0

m1 fre q=1.900GHz NF=1.471

fre q, GHz

Notice tha t the nois e figure ha s de gra de d s ignifica ntly due to the the rma l nois e of the s e rie s re s is ta nce s in the inductors

Example of sweeping a parameter

NETLIST INCLUDE
Ne tlis tInc lude Ne tlis tInc lude 1 Inc lude File s [1]=ge ne ric [Link] Te rm Te rm2 Num=2 Z=100 Ohm Vout
MOSFET_NMOS MOSFET2 Mode l=nfe t Le ngth=0.2 5 um Width =1 00 um Tris e =

V_DC S RC1 Vdc =2.5 V

V_DC S RC3 Vdc =1.8 V

AC
AC AC1 S we e pVa r="fre q" S ta rt=100 MHz S top=3 GHz S te p=100 MHz Ca lc Nois e =ye s Nois e Node [1]="Vout" Nois e Node [2]="Vin" S ortNois e =S ort by va lue Inc lude P ortNois e =ye s

Var E qn

VAR VAR1 s ourc e ind=1e -9

V_DC S RC2 Vdc =0.75 V

I_P robe ID DC_Fe e d DC_Fe e d1

Vdra in

Vin DC_Bloc k DCBloc k1 L L2 L=s ourc e ind R=

Vga te

MOSFET_ NMOS MOSFET1 Mo de l=nfe t Le n gth =0.25 um Wid th=100 u m Tris e =

PARAMETER SWEEP
P a ra mS we e p S we e p1 S we e pVa r="s ourc e ind" S imIns ta nc e Na me [1]="AC1" S imIns ta nc e Na me [2]= S imIns ta nc e Na me [3]= S imIns ta nc e Na me [4]= S imIns ta nc e Na me [5]= S imIns ta nc e Na me [6]= S ta rt=1e -9 S top=15e -9 S te p=0.5e -9

P _AC P ORT 1 Num=1 Z=50 Ohm P a c =pola r(dbmtow(0),0) Fre q=fre q

L L1 L=1.0 nH R=

Parameter sweep block included

Schematic: LNA_noise_sweep

Matching inductor is varied from 1 to 15 nH (coils have infinite Q)

Example of simulation output

Lg ~ 10.7 nH close to the optimum value

8 6

NF

m1 freq= 1.900GHz NF=0.669 s ourceind=1.100000E-8

m1
0.0 0.5 1.0 1.5 2.0 2.5 3.0

fre q, GHz

Gain and 1dB compression simulation

NE T LIS T INC LUDE

N e tlis tIn c lu d e N e tlis tIn c lu d e 1 In c lu d e F ile s [1 ]=g e n e ric 0 2 5 . lib Te rm Te rm 2 N u m =2 Z =1 0 0 O h m

V_ D C SRC1 Vd c =2 . 5 V

From sources-freq. domain: P_1 tone source. RF_power is the sweep variable
V_ D C SRC2 Vd c =0 . 7 5 V

Vo u t

HAR M O NIC B ALANC E

V_ D C SRC3 Vd c =1 . 8 V
MOS FE T _ NMOS MOS FE T 2 Mo d e l=n fe t Le n g th =0 .2 5 u m W id th =1 0 0 u m T ris e =

I_ P ro b e ID DC _F e e d DC _F e e d1

H a rm o n ic Ba la n c e H B1 F re q [1 ]=1 . 9 G H z O rd e r[1 ]=3 S we e p Va r="R F _ p o we r" S ta rt=-5 0 S to p =1 0 S te p =1

Vd ra in

Va r Eq n

VAR VAR 1 R F _ p o we r=-3 5

HB controller Meas eq.

Vin D C _ Blo c k D C Blo c k 1 P _ 1 To n e P O R T1 N u m =1 Z =5 0 O h m P =d b m to w(R F _ p o we r) F re q =1 . 9 G H z L L2 L=1 0 . n H R =1 8

Vg a te

MOS FE T _ NMOS MOS FE T 1 Mo d e l=n fe t Le n g th =0 .2 5 u m W id th =1 0 0 u m T ris e =

Me a s Eq n

Me a s E q n Me a s 1 Vo u t_ d Bm =d Bm (Vo u t[1 ])

L L1 L=1 n H R =1 . 7

Schematic: LNA_1dB_by_power_sweep

Initialization by means of Var (main menu) needed

Example of simulation output

Equation defining the line
m3 RF_powe r=-7.000 Line =8.016 m3 m4 m4 ind De lta = 0.000 de p De lta =-1.025 de lta mode ON

Eqn Line =RF_powe r+dB_ga in[0]

20 10
30

m1

Vout_dBm

0 -10 -20

m1 RF_power=-4.000 20 Vout_dBm=9.265
10

Line Vout_dBm

-30 -40 -50 -40 -30 -20 -10 0 10

-10

RF_power
-20

Voltage gain ~ 15 dB

Input power for 1 dB compression

-30

-40 -50 -40 -30 -20 -10 0 10

RF_powe r

Simulation of IIP3

NE T LIS T INCLUDE
N e tlis tIn c lu d e N e tlis tIn c lu d e 1 In c lu d e F ile s [1 ]=g e n e r ic 0 2 5 .lib Te rm Te rm 2 N u m=2 Z=1 0 0 O h m

V_ D C SR C 1 Vd c =2 .5 V

From sources-freq. domain: P_ntone source. Two tones are defined

V_ D C SR C 2 Vd c =0 .7 5 V

Vo u t
MO S FE T_N MOS MO S FE T2 Model=nfet Le ngth=0.25 um W idth=100 um Tris e=

HARMO NIC BALANCE

H a r mo n ic Ba la n c e H B1 Fr e q [1 ]=R F _ fr e q +s p a c in g /2 Fr e q [2 ]=R F _ fr e q - s p a c in g /2 O rd e r 4 m e a n s th a t F re q [1 ] w ill O r d e r [1 ]=4 b e c a lc u la te d w ith 4 h a r mo n ic s O r d e r [2 ]=4

V_ D C SR C 3 Vd c =1 .8 V

I_ Pr o b e ID

D C _ Fe e d D C _ Fe e d 1

Vd r a in
Va r Eq n

Vin D C _ Blo c k D C Blo c k 1 L L2 L=1 0 .6 n H R =1 8

Vg a te

MOS FE T_N MO S MOS FE T1 Mode l=nfe t Le ngth=0.25 um W idth=100 um Tris e =

VAR VAR 1 s p a c in g =1 MH z R F_ fr e q =1 .9 G H z R F_ p o w e r =- 3 5 Me a s Eq n Me a s 1 Vo u t_ d Bm=d Bm ( Vo u t[1 ]) In p u t IP 3 (IIP 3 ) To c a lc u la te th e in p u t IP 3 th e g a in o f th e LN A is n e e d e d

P _ n To n e P O R T1 N u m =1 Z=5 0 O h m F re q [1 ]=R F _ fre q + s p a c in g /2 F re q [2 ]=R F _ fre q - s p a c in g /2 P [1 ]=d b m to w (R F _ p o w e r ) P [2 ]=d b m to w (R F _ p o w e r )

Me a s Eq n

L L1 L=1 n H R =1 .7

P0

Pin

IP 3in

IP 3 in IP 3 in 1 IP 3 in 1 =ip 3 _ in ( Vo u t,1 5 ,{1 ,0 },{2 ,-1 },5 0 )

Me as Eq n

Me a s Eq n Me a s 2 to n e s =[{1 ,0 },{0 ,1 },{2 ,- 1 },{- 1 ,2 }]

Def. of tones of interest

Predefined equations

1 5 is th e s ma ll s ig n a l p o w e r g a in in d B o f th e LN A, w h ic h is ta k e n fr o m th e 1 d B c o m p r e s s io n s imu la tio n s c h e ma tic .

P0

O u tp u t IP 3 (O IP3 )
Pin

P0

Pin

IP 3out

IP 3out

Schematic: LNA_IIP3

IP 3 o u t ip o 1 ip o _ u p p e r =ip 3 _ o u t( Vo u t,{1 ,0 },{2 ,- 1 },5 0 )

IP3 o u t ip o 2 ip o _ lo w e r =ip 3 _ o u t(Vo u t,{1 ,0 },{-1 ,2 },5 0 )

Mix function in ADS

Purpose: Returns a component of a spectrum based on a vector of mixing indices. Synopsis mix(xOut, harmIndex{, Mix}) where xOut is a voltage or a current spectrum. harmIndex is the desired vector of harmonic frequency indices (mixing terms). Mix is a variable consisting of all possible vectors of harmonic frequency indices (mixing terms) in the analysis. Example: y = mix(vOut, {2, -1})

Example of simulation output

IP 3in1 1.333

ipo_lower 16.332

ipo_upper 16.333

freq 0.0000 Hz 1.000MHz 2.000MHz 1.898GHz 1.899GHz 1.901GHz 1.902GHz 3.798GHz 3.799GHz 3.800GHz 3.801GHz 3.802GHz 5.698GHz 5.699GHz 5.700GHz 5.702GHz 7.598GHz 7.599GHz 7.600GHz 7.601GHz 7.602GHz

Mix Mix(1) 0 1 2 -1 0 1 2 -1 0 1 2 3 0 1 2 3 0 1 2 3 4 Mix(2) 0 -1 -2 2 1 0 -1 3 2 1 0 -1 3 2 1 0 4 3 2 1 0

Notice tha t the uppe r a nd lowe r third orde r inte rce pt points a re a lmos t s ymme trica l (ipo_uppe r a nd ipo_lowe r). The input IP 3 (IP 3in1) is s imply 15 dB lowe r (the s ma ll s igna l ga in) tha n the output IP 3

dBm(mix(Vout,tones ))

m1
-20 -40 -60 -80 -100 1.8985

m2

m1 fre q=1.899GHz dBm(mix(Vout,tone s ))=-19.934 m2 fre q=1.901GHz dBm(mix(Vout,tone s ))=-19.939

1.8990 1.8995 1.9000 1.9005 1.9010 1.9015

freq, GHz

This is the s o-ca lle d mix ta ble of the ha rmonic ba la nce s imula tion. Numbe r 1 re pre s e nts the RF tone (with s pa cing). Ze ro me a ns tha t no tone is pre s e nt (DC). And two re pre s e nts two time s the RF s imula tion tone (of Fre q[1] or Fre q[2]).