The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> Cadence Convergence Error
https://designers-guide.org/forum/YaBB.pl?num=1280229946

Message started by mohammadreza on Jul 27^th, 2010, 4:25am

Title: Cadence Convergence Error
Post by mohammadreza on Jul 27^th, 2010, 4:25am

I am simulating an RF Power Amplifier in Cadence, SpectreRF environment. I am controlling the supply current by a seperate voltage applied to a transistor. For some values that I choose, I face with convergence error, both in PSS and Transient. However, when I remove my mathcihng network and replace it by a real load value that I had designed, I face no errors.

Is there any way for setting the options correctly so that I don't lose accuracy and have my expected result? I have tried the Moderate and ven liberal presets but the moderate one faces error as well and the liberal one is completely away.

Title: Re: Cadence Convergence Error
Post by mohammadreza on Jul 27^th, 2010, 5:08am

Thanks, I didn't see any HB-engine. Just shooting and flexible balance. I chose flexible balance and it worked.

About the Q, since I need to operate between 2.4GHz and 2.49GHz, I chose 300. Is that fine?
thanks

Title: Re: Cadence Convergence Error
Post by mohammadreza on Jul 27^th, 2010, 5:29am

thanks!

decreasing the Q helped run the simulation successfully with the moderate accuracy.

Title: Re: Cadence Convergence Error
Post by pancho_hideboo on Jul 27^th, 2010, 7:16am

pancho_hideboo wrote on Jul 27^th, 2010, 4:30am:

Maybe your matching network has very high Q.
High Q could cause convergence failure for time domain solver.

Use HB-engine not Shooting-Newton-engine in PSS of Cadence Spectre.

mohammadreza wrote on Jul 27^th, 2010, 5:08am:

I didn't see any HB-engine.
Just shooting and flexible balance.
I chose flexible balance and it worked.

See http://www.designers-guide.org/Forum/YaBB.pl?num=1258711290/1#1

mohammadreza wrote on Jul 27^th, 2010, 5:08am:

About the Q, since I need to operate between 2.4GHz and 2.49GHz, I chose 300.
Is that fine?

BW_3dB=f0/Q=2.445GHz/300=8.15MHz. It is too narrow.

You have to have broad BW_3dB which is far broader than 90MHz.
So set Q lesser than 27.

Also see the followings.
http://www.designers-guide.org/Forum/YaBB.pl?num=1207009725/2#2
http://www.designers-guide.org/Forum/YaBB.pl?num=1278414763/18#18

mohammadreza wrote on Jul 27^th, 2010, 5:29am:

decreasing the Q helped run the simulation successfully with the moderate accuracy.

Your case summaries are followings.

For Q>=300, Shooting-Newton-PSS with accuracy of moderate can not converge.
For Q<30, Shooting-Newton-PSS with accuracy of moderate can converge.

For Q>=300, HB-PSS with accuracy of moderate can converge.
For Q<30, HB-PSS with accuracy of moderate can converge.

Are the aboves correct ?

Title: Re: Cadence Convergence Error
Post by mohammadreza on Jul 28^th, 2010, 5:41am

Yes, that's right