The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> Serious convergence problems using HB
https://designers-guide.org/forum/YaBB.pl?num=1191085521

Message started by schultzc on Sep 29^th, 2007, 10:05am

Title: Serious convergence problems using HB
Post by schultzc on Sep 29^th, 2007, 10:05am

Hello!

I have to analyse a rather complex oscillator circuit for phase noise. The core is build as a colpitts type oscillator. The output buffer is build as an inverter chain for creating a rectangular output signal for driving digital logic.

Using SpectreRF PSS simulation convergence could be achieved. But when I switch to ADS HB analysis I can't get any convergence. (Comparing both simulator's results to each other is a key issue of my task).

It seems, that the quasi-digital output buffer is the main source of trouble, because I don't have any problems if i remove the output stage.
As setting I use 20 harmonics and 4 times oversampling. Tolerances are set to weak. I use the transient assisted HB. Observing the transient results, it seems that a near-steady-state is reached (although the simulator doesnt recognize one...).
Has anyone suggestions what I can try for solving my problems?

Regards
Christoph

Title: Re: Serious convergence problems using HB
Post by pancho_hideboo on Sep 29^th, 2007, 11:18am

Hi.

"SpectreRF PSS" you mean is shooting PSS or FB(HB) PSS ?
For oscillator analysis, I recommend shooting PSS.

BTW, what version of ADS do you use ?

Although you can do oscillator analysis without oscillator probe in current ADS,
it could often cause convergence failure if you choose node selection method instead of inserting oscillator probe.

If you don't use oscillator probe, try to use it.

Title: Re: Serious convergence problems using HB
Post by schultzc on Sep 30^th, 2007, 1:30pm

Yes with PSS I mean the shooting PSS algorithm. The oscillator converges using this method, but each simulation is taking a very long time.

The ADS version used is 2006a. I tried both kind of oscillator simulations (node select and oscprobe). The result is practically the same. As long as I abandon the output stages simulation takes it's time but converges. After adding them to simulation again, I can't get any results.

Title: Re: Serious convergence problems using HB
Post by pancho_hideboo on Sep 30^th, 2007, 9:22pm

Hi.

Try to insert oscprobe other than output of buffer.
I recommend insertion of oscprobe into low impedance node.

Anyway you should try to insert oscprobe in various node.

Title: Re: Serious convergence problems using HB
Post by Frank Wiedmann on Sep 30^th, 2007, 11:41pm

You could also try to use more harmonics, 20 might not be enough for your type of output circuit. In general, shooting (time-domain) pss is probably much better suited for this kind of circuit than harmonic balance.