Evaluated waveform discontinuity

Hi All, 

It will be very helpful to get some insights on this. 

regards

Sid

Sid,

This is the second post you have chased in a few days. If it's urgent, please contact customer support - that's what it's for. Everyone who responds here (including me, as a Cadence employee) does so as a volunteer - in their spare time - and sometimes the day job gets in the way...

Especially as the first post was on a Sunday. Be reasonable!

Andrew

What is the exact expression you are using, and what does the source waveform look like for each of the three traces (I'm interested really in what it looks like up to the times that the green, yellow and red curves start at)?

Andrew

What is the exact expression you are using, and what does the source waveform look like for each of the three traces (I'm interested really in what it looks like up to the times that the green, yellow and red curves start at)?

Andrew

thanks for the reply. The source waveform is of an oscillator called "in" as described in the below expression. The tran simulation runs till 20us. I have tried with both "falling" and "rising" edge in the expression resulting in the same response. Also I tried the auto/user mode to set the threshold which in this case is 0. Using the "clip" function or not also doesn't make much of a difference.

The expression is freq(clip(VT("/in") 0 2e-05) "rising" ?xName "time" ?mode "auto" ?threshold 0)

"in" waveform looks like the below - 

"in" waveform from 0 to 1us : 

regards 

Well, there must be two transitions of the green waveform earlier (which looks likely from the waveforms you've shown). The freq function outputs its first point at the second transition (obviously it can't compute the frequency before that). The red and yellow waveforms seem to be around 0 until the signal really starts changing. I'm not sure there's a great deal you can do about this, other than clip the green signal so that it omits the beginning part, but the ramp-up is going to cross a threshold somewhere. There's certainly nothing to ensure that the other waveforms start in the same place. You could (I suppose) do something like:

freq(clip(VT("/in") 0 2e-05) "rising" ?xName "time" ?mode "auto" ?threshold 0)+VT("/in")*0

that will (I think) cause the initial frequency point to be extrapolated back to 0 (the addition of a waveform multiplied by 0 will not affect the Y values, but the X points will then be generated.

Andrew

Thanks Andrew, from the addition of a waveform multiplied by 0 the resultant output (red and yellow waveforms) has a constant value in the previously discontinuous region as shown in the below screenshot.  

regards