get the difference between two wave forms after parametric sweep.

Svilen,

Almost certainly this could be done with an expression using the eval type set to “sweeps”, but could you please share the expressions used and what you are sweeping too - it’s hard to be certain precisely what you are doing and I’d like to make sure that any specific suggestion I make has been tested and is close to what you actually need.

Maybe also sharing a screenshot of the results you get in ADE and then showing which you want to get the difference between?

Thanks,

Andrew

Hi Andrew,

Thanks for replying.

I did try to attach an image yesterday but the system didn't allow me. A while back I also replied to you with the information you requested and also gave the URL to where I uploaded the image online and my reply was flagged as spam. So, I will again explain in words what I was doing and provide no visual material unfortunately. Which is sort of ridiculous not being able to do.

I have a variable in my testbench called bits which controls the delay of a clock and I want to sweep that variable and get different delays. The sweep is over a list of values which are non-uniform, say 0 1 3 7 15 31 etc. I also created another variable ind (for index) which is uniform 1 2 3 4 5 6 etc. That variable doesn't control anything in the test bench but I grouped it in a parametric set with the first one, so they go together.
The output I am interested in is VT("/out"). 
I have defined my reference with respect to which I measure the delay as ref=value(VT("/out") "bits" 0)
I have defined the family of clocks from the sweep as family=VT("/out")
Then I define the delay from the rising edge of the ref to the rising edges of each clock in the family del=delay(ref 0.4 2 "rising family 0.4 2 "rising")
Also defined index=VAR("ind")
All those are evalType sweeps.
Form the calculator I plot two things
waveVsWave(?x index ?y del)
rshift(waveVsWave(?x index ?y del) 1)
I can plot those and now I have two delay sets of points shifted with respect to each other by one index count. The idea being to take the difference and have the delay between each two consecutive edges of the clock, which ultimately I wanted to plot. However, the difference returns just 0.

Regards
Svilen

Hi Andrew,

Thanks for replying.

I did try to attach an image yesterday but the system didn't allow me. A while back I also replied to you with the information you requested and also gave the URL to where I uploaded the image online and my reply was flagged as spam. So, I will again explain in words what I was doing and provide no visual material unfortunately. Which is sort of ridiculous not being able to do.

I have a variable in my testbench called bits which controls the delay of a clock and I want to sweep that variable and get different delays. The sweep is over a list of values which are non-uniform, say 0 1 3 7 15 31 etc. I also created another variable ind (for index) which is uniform 1 2 3 4 5 6 etc. That variable doesn't control anything in the test bench but I grouped it in a parametric set with the first one, so they go together.
The output I am interested in is VT("/out"). 
I have defined my reference with respect to which I measure the delay as ref=value(VT("/out") "bits" 0)
I have defined the family of clocks from the sweep as family=VT("/out")
Then I define the delay from the rising edge of the ref to the rising edges of each clock in the family del=delay(ref 0.4 2 "rising family 0.4 2 "rising")
Also defined index=VAR("ind")
All those are evalType sweeps.
Form the calculator I plot two things
waveVsWave(?x index ?y del)
rshift(waveVsWave(?x index ?y del) 1)
I can plot those and now I have two delay sets of points shifted with respect to each other by one index count. The idea being to take the difference and have the delay between each two consecutive edges of the clock, which ultimately I wanted to plot. However, the difference returns just 0.

Regards
Svilen

Svilen,

Unfortunately there's a bug with the forum system that it intermittently disallows image updates (this affects me too). It's known and we hope that this will be fixed when we upgrade to the next version of the forum software provided by the vendor. The workaround of using an external site for images works, but external URLs are flagged as potential spam and have to be appealed, reviewed and the approved (which I did - so it's in the thread above).

I'll look into your situation and see whether I can reproduce something similar. May not be until later today or tomorrow though...

Andrew

Hi Andrew,

Thanks for unblocking that reply. I hope now with the image and my previous description it is more or less clear what I was trying to do.

Svilen