how to set equal time step of transient simulation in Cadenc

[alard]

Hello,

The problem is as follows,

When I put 10ps for the maxstep in the option of transient simulation, I saw some steps running at 6 or 7 ps. When I set both step and maxstep to 5ps, then the simulator run at 2 ps for some step. Finally I tried with 1ps, then 50as appears....

How could define a step that I want just like in Hspice or ADS?

 

[tia_design]

Re: how to set equal time step of transient simulation in Ca

I don't think you can set equal step for simulation because of algorithm itself. Algorithm will determin the step to meet the convergence requirment. Smaller steps for fast transient than slow transient.

 

[alard]

thx a lot

 

[tshankar501]

Re: how to set equal time step of transient simulation in Ca

I can see another option called "step" apart from "maxstep" in the options for transient simulation.....did you check your simulation by inserting a value in step alone instead of inserting a value in maxstep

 

[edge_tv]

you can use strobe. The simulator still makes it own time steps but the result only has the time value with the strobe periode

 

[abcyin]

Re: how to set equal time step of transient simulation in Ca

you can use strobe. The simulator still makes it own time steps but the result only has the time value with the strobe periode

I tried the strobe period, and it works, it really provides the step what you want, but could you pls explain what's the meaning of the strobe period, and how it can determine the transient step? Thx in advance!

 

[rfsystem]

Re: how to set equal time step of transient simulation in Ca

To my Spectre knowledge you can advise Spectre only to insert additional time steps with "strobe". The others are depending on the integration error calculation.

Some sources insert also time steps. For instance pulse sources at the corners. I had some years ago an issue with a simulator vendor inserting for a sinewave an additional time step where the sinewave w/o DC crosses zero. The result was bad spectra leakage in high-Q circuits. After lengthly discussion he built it out!