Parametric Sweep for PSS type analysis and MATLAB interface

You got that error because you missed out the leading quotation mark before the analysis name. I can see the same error if I do that myself.

If you do:

RFout=cds_srr('/home/manas/work_cad/TB_DIFF_NC_LNA.raw','sweeppss_pac-sweep','RFout')

(I've added the quote before sweeppss_pac-sweep which was missing in what you posted), then it should work.

Regards,

Andrew.

Dear Andrew,

Sorry it was that silly mistake. I couldn't check that before sending it to the forum.

Can you please tell:-

I actually  want to do a Monte-Carlo simulation on this PSS/IIP.

Can you please tell is their any way to do Monte-carlo in spectre through the MATLAB interface.

Kind Regards

The matlab "interface" doesn't control the simulation - it merely reads the results (assuming you're talking about the Spectre/Spectre RF Toolbox for Matlab). So I don't think I really understand your question...

Andrew.

 Dear Andrew,

Thanks a lot for your reply.

a) OK. The Matlab interface can only read the result.

Whether it can read a Monte Carlo simulation run by SpectreRF ?

b) Is it possible to do through this interface ( I mean reading all types of simulated data) what Ocean script does ?

In other words whether it has the capability to replace ocean script.

c) Is it possible to run / control the Spectre Simulation through MATLAB. ( I am not meaning the Spectre-Matlab Interface. I got the point that this is only meant for reading the data)

Answers to your questions:

a) Yes, it can. You'd probably want to use the option to save family data so you get the curves for every iteration in the monte carlo run, and then you'll be able to read the data in Matlab (it will have "iteration" as a sweep variable. It probably wouldn't be terribly difficult to write some code to read the mcdata/mcparam scalar files that are also produced during monte carlo (there's a solution on Cadence Online Support to do this - the "abStats.il" code shows how to do that in SKILL)

b) You can read all the simulation results, so presumably you could write functions to do whatever post-processing you need. Of course, there are convenient functions which directly operate on the waveform objects already existing in SKILL - you'll have to implement your own in Matlab to ensure they can work with family data and unevenly spaced points, etc, etc - but in theory there's no reason why you couldn't do all the post-processing in Matlab if you preferred.

c) Yes - there are functions in Matlab to invoke commands - so you'd simply need to invoke spectre from Matlab. We don't provide anything specific to do that, but it wouldn't be hard.

Andrew.