The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> Circuit Simulators >> gear2 , conservative , and currents in Spectre
https://designers-guide.org/forum/YaBB.pl?num=1245796920

Message started by kevinvinv on Jun 23^rd, 2009, 3:42pm

Title: gear2 , conservative , and currents in Spectre
Post by kevinvinv on Jun 23^rd, 2009, 3:42pm

I've been reading the "Designers Guide to Spice and Spectre" a bit since I've been seeing some heavily ringing (but smooth) device currents in a recent circuit. The ringing pretty much goes away if I choose Conservative (Spectre) and gear2only.

What is the general thought- is it risky to assume these rings are then numerical in nature or should I search the circuit more for a possible real-life source.

Title: Re: gear2 , conservative , and currents in Spectre
Post by ACWWong on Jun 24^th, 2009, 3:45am

You'll need to check for possible real causes. For example, does the ringing a have a defined frequency ? Does your small signal stability analysis stretch far enough to capture that frequency ?
It also depends on your circuit type and where/when the ringing occurs as to whether the judgement can be made to discount it. For example ringing on a digital level after switching isn't something you might care about unless it grows.

Title: Re: gear2 , conservative , and currents in Spectre
Post by Marq Kole on Jul 3^rd, 2009, 12:34am

A simple test in this case can be to extend the simulation time or change the maximum step size in the transient analysis - if the frequency of the ringing is sensitive to these it is more likely to be a numerical artifact. If only the amplitude changes then the frequency is probably determined by your circuit components and the origin is more likely to be teh design itself.

Marq

Title: Re: gear2 , conservative , and currents in Spectre
Post by byang on Jul 4^th, 2009, 12:51pm

To verify whether it is Trapezoidal ringing, there is one thing one can check. One can plot the waveform in "point" style. The waveform needs to look like :
* *
* * *
i.e. up and down and no point in between. This is numerical artifact.

Since the waveform looks "smooth", maybe it is not Trap ringing.

Gear2 method has numerical damping while Trap method has no numerical damping. Gear2 may dampen small oscillation even if it is real.

Baolin Yang
Gemini Design Technology, Inc.