You set f1=2490MHz as small signal in PAC analysis.

Show me followings.

(1) Over Plot Results of fundamental output(f1=2490MHz, f2=2495MHz) for QPSS and PSS/PAC.

(2) Over Plot Results of IM3 output for QPSS(2485MHz and 2500MHz) and PSS/PAC(2500MHz).

If you can not do them, upload "simulation/"test_bench_name"/spectre/schematic" for QPSS and PSS/PAC.

Show me version of Cadence Spectre you use.

Also show me time waveform for Idrain, Vdrain, Iload and Vload.
These can be available in PSS results for PSS/PAC.

.

purplewolf wrote on Jun 29^th, 2010, 10:10am:
You can't understand my requests.

My request is time waveform with Pin_dBm as parameters.
See my simulation results surely.

I don't think your amplifier is Class-B operation, although I don't know what Pin_dBm value is used for time waveform you showed.

purplewolf wrote on Jun 29^th, 2010, 10:12am:
Use more new Cadence Spectre.

When I look at your earlier two-tone sweep data, it looks like there is
1. a little bit of gain expansion before compression (although it is difficult to tell from just looking at the plot)
2. the IM3 has definite dips (nulls if you like) near where the gain expansion is occurring

IF you do have some gain expansion before compression then this will definite violate the condition necessary for 9.6 dB P1dB to OIP3 relationship. The gain expansion can make P1dB happen at an input power level that is higher than that of a classic class A amplifier.

I will look at the latest plots later, but it seems your circuit is operating class AB.

purplewolf wrote on Jun 30^th, 2010, 3:47am:
Why is drain current of the following not coincident to any of Transient Analysis ?
http://www.designers-guide.org/Forum/Attachments/IVD.png
This drain current is not for Class-B operation completely.

On the other hand, drain currents in your Transient Analysis are for Class-B operation.

purplewolf wrote on Jun 30^th, 2010, 5:05am:
I think you can get same time domain waveform current as Transient Analysis, even if you use HB-PSS of Cadence Spectre.

purplewolf wrote on Jun 30^th, 2010, 5:05am:
Are your results of PSS/PAC and QPSS using Cadence Spectre truely correct ?

Try to use Shooting-Newton-PSS not HB-PSS.

Do you have other vendor's simulator such as Agilent RFDE(ADSsim), Agilent GoldenGate, Synopsys HSPICE-RF, Mentor eldoRF ?

purplewolf wrote on Jun 30^th, 2010, 6:03am:
I can't understand what is your problem in PSS, if your PSS result is correct.

purplewolf wrote on Jun 30^th, 2010, 6:03am:
Agilent GoldenGate is difficult to use correctly.
See http://www.designers-guide.org/Forum/YaBB.pl?num=1268404959/7#7

purplewolf wrote on Jun 30^th, 2010, 6:03am:
I can have access to almost all vendor's simulator, but I don't use Cadence Spectre for RF circuit design.