The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> pstb gives strange results
https://designers-guide.org/forum/YaBB.pl?num=1265101150

Message started by aaron_do on Feb 2^nd, 2010, 12:59am

Title: pstb gives strange results
Post by aaron_do on Feb 2^nd, 2010, 12:59am

Hi all,

i'm trying to run a stability analysis on a current output passive mixer which is essentially a passive mixer followed by a trans-impedance amplifier. I first tried modelling the output impedance of the switching stage (the passive mixer) as a resistance, and then ran the normal stability analysis to find the loop gain. Here i found no problems. I used a cmdmprobe to break the loop. However, when I ran the pstb analysis with the LO signal turned on, the DC loop gain dropped from 70 dB down to -30 dB. Has anybody else encountered a problem like this? The PSS and Pnoise analysis give reasonable results.

thanks,
Aaron

Title: Re: pstb gives strange results
Post by pancho_hideboo on Feb 2^nd, 2010, 1:36am

aaron_do wrote on Feb 2^nd, 2010, 12:59am:

However, when I ran the pstb analysis with the LO signal turned on,
the DC loop gain dropped from 70 dB down to -30 dB.

PSTB is no more than PAC of sideband=0.
Where do you insert probe ?
See http://www.designers-guide.org/Forum/YaBB.pl?num=1217939784/3#3

BTW, I don't use Cadence Spectre for RF circuits design.
I use interstage impedance probe or custom loop gain probe with master steady state large signal analysis for stability analysis.
Again see http://www.designers-guide.org/Forum/YaBB.pl?num=1240334878/10#10

Title: Re: pstb gives strange results
Post by aaron_do on Feb 2^nd, 2010, 7:51am

Hi,

thanks for the reply.

Quote:

PSTB is no more than PAC of sideband=0.

I'll try PAC tomorrow. thanks

Quote:

Where do you insert probe ?

I'm using an op-amp based TIA, so I inserted the probe at the input of the opamp (connected to the gate of the input transistor).

I expected there to be some difference between stb and pstb analysis, but obviously not that much. BTW in order for PAC analysis to run correctly, I had to change the parameter "version" in the BSIM model file from "version=3.3" to "version=3.2". Could the problem have something to do with that?

thanks,
Aaron

Title: Re: pstb gives strange results
Post by Andrew Beckett on Feb 2^nd, 2010, 11:06am

I doubt that the version of the model would make a difference, but it does suggest that maybe you are using a rather old version of MMSIM. What version of spectre are you using? (this should appear in the output log file).

Regards,

Andrew.

Title: Re: pstb gives strange results
Post by aaron_do on Feb 2^nd, 2010, 7:42pm

Hi Andrew,

Quote:

What version of spectre are you using? (this should appear in the output log file).

the log file says i'm using version 6.2.1.145 from Jan 10, 2008. I'll see if I can update it...

thanks,
Aaron

Title: Re: pstb gives strange results
Post by Andrew Beckett on Feb 2^nd, 2010, 10:24pm

Not that I'm sure this will fix it - but it's worth trying a newer version anyway, just in case.

Best Regards,

Andrew.

Title: Re: pstb gives strange results
Post by aaron_do on Feb 2^nd, 2010, 11:24pm

ok it seems that i wasn't using the model with version=3.2, but the one with version=3.3. Now that i've changed it to version=3.2 it works. I guess i'll try updating the spectre version later when i have more time.

thanks,
Aaron

Title: Re: pstb gives strange results
Post by Frank Wiedmann on Feb 3^rd, 2010, 12:48am

This is in fact a known problem. Please see Cadence solution 11487974 (it talks about PAC but other periodic small-signal analyses like PXF and PSTB were probably affected as well). The problem was fixed by CCR 588933 in MMSIM versions 6.2.1.345 and 7.0.1.076.

Title: Re: pstb gives strange results
Post by aaron_do on Feb 3^rd, 2010, 6:52am

Thanks a lot! I guess i'll see if I can update MMSIM sooner than I thought,

thanks,
Aaron

Title: Re: pstb gives strange results
Post by Andrew Beckett on Feb 3^rd, 2010, 12:09pm

I feel a little embarrassed now! Thanks Frank for finding something that as a Cadence AE I really should have found (my lame excuse is that I didn't actually look, and I don't actually remember the issue from 2 years ago).

Regards,

Andrew.