PSpice--Monte Carlo Delay?

[Bob Penoyer]

I need to use an analog delay that has a tolerance so that the delay
varies during a Monte Carlo simulation.

TLine would be a great part but it doesn't have a tolerance parameter.

An all-pass network can't be used for the delay because it doesn't
have a linear phase response and the signal is complex.

Can TLine or something like it be made to use a tolerance that Monte
Carlo can operate on? Any other suggestions?

 

[Kevin Aylward]

I need to use an analog delay that has a tolerance so that the delay
varies during a Monte Carlo simulation.

TLine would be a great part but it doesn't have a tolerance parameter.

An all-pass network can't be used for the delay because it doesn't
have a linear phase response and the signal is complex.

Can TLine or something like it be made to use a tolerance that Monte
Carlo can operate on? Any other suggestions?

Yes. Use SuperSpice

SuperSpice allows MC and parameter sweeps on TLines. There are some
examples in the example folder.

Kevin Aylward
[email protected]
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

 

[Jim Thompson]

I need to use an analog delay that has a tolerance so that the delay
varies during a Monte Carlo simulation.

TLine would be a great part but it doesn't have a tolerance parameter.

An all-pass network can't be used for the delay because it doesn't
have a linear phase response and the signal is complex.

Can TLine or something like it be made to use a tolerance that Monte
Carlo can operate on? Any other suggestions?

I don't Monte Carlo, since chip components are ratiometric, requiring
different means to test for worst-case.

BUT, couldn't you simply parameterize TD?

...Jim Thompson

 

[Bob Penoyer]

I don't Monte Carlo, since chip components are ratiometric, requiring
different means to test for worst-case.

BUT, couldn't you simply parameterize TD?

Yes. Can I used Monte Carlo with a parameterized value? How?

 

[Jim Thompson]

Yes. Can I used Monte Carlo with a parameterized value? How?

See PSpice Reference Guide (v10.3), pp 47...50

...Jim Thompson

 

[Keith Williams]

I don't Monte Carlo, since chip components are ratiometric, requiring
different means to test for worst-case.

When I was doing this, we were given the process distributions
(including tracking) to use in the simulations. We also used end-
point distributions to converge on the "worst-case" more quickly.
Discretes were another issue though.