variable $freq in veriloga

Robin,

There's limited support for this - from MMSIM15.1 ISR3 onwards. Not sure what version you're using.

I think it's a bit fussy as to what works and what doesn't. My initial attempts didn't work, but using this:

`include "disciplines.vams"
`include "constants.vams"

module forumdollarfreq2 (vp,vn);
inout vp,vn;
electrical vp,vn;

real r,tn;
analog begin
  //r = $table_model($freq, "./resFreq.tbl", "1CL");
  r = $freq/1000.0;
  tn = 4.0 * `P_K * $temperature * r;
  I(vp, vn) <+ (V(vp, vn) + sqrt(tn)) / r ;
end

endmodule

(either with or without the table model) worked fine.

Regards,

Andrew.

My latest version is MMSIM 15.10.803

As you say it might be a bit fussy as your example does work, but my current experiment is not working although it might have worked briefly before I made some edits.

Is there any documentation or mention in a release note for this $freq feature?

Also I haven't been able to probe veriloga internal variables in AC simulation but I can for dc or tran.

Thanks,

Robin

Hi Robin,

The fact that it is not documented tells me that it's not really fully released yet - because it only works in narrow cases. Even then, it's not going to work in time-domain analysis. I wouldn't expect you to be able to probe the variables either because that's not something that can be done in ac analysis (similarly you can't access device operating point parameters during an ac analysis). Maybe I shouldn't have given the "working" example ;-)

Probably makes most sense to open a case with customer support so that we can take a look at specifically what you're trying to achieve and then assess how ready the capability is (or isn't) with R&D.

Regards,

Andrew.

Hi Andrew,

This should give an idea of what I was trying to do - a low pass with variable roll off rate.

In my simulations it had a flat frequency response of 1.

Regards,

Robin

`include "discipline.h"
`include "constants.h"


// low pass filter with 3dB/octave roll off
//

module low_rolloff2(sigin, sigout);
input sigin;
output sigout;
electrical sigin, sigout;
parameter real F0 = 10M;   // cutoff
//parameter real rolloff = 0.5; // 1 = normal RC, 0.5 = 1/sqrt(f) etc

real FreqGain;

analog begin

    //FreqGain = 1/(1+pow($freq/F0,rolloff));
    FreqGain = 1/(1+sqrt(abs($freq/F0)));
    V(sigout) <+ V(sigin)*FreqGain;

end

endmodule

Hi Andrew,

This should give an idea of what I was trying to do - a low pass with variable roll off rate.

In my simulations it had a flat frequency response of 1.

Regards,

Robin

`include "discipline.h"
`include "constants.h"


// low pass filter with 3dB/octave roll off
//

module low_rolloff2(sigin, sigout);
input sigin;
output sigout;
electrical sigin, sigout;
parameter real F0 = 10M;   // cutoff
//parameter real rolloff = 0.5; // 1 = normal RC, 0.5 = 1/sqrt(f) etc

real FreqGain;

analog begin

    //FreqGain = 1/(1+pow($freq/F0,rolloff));
    FreqGain = 1/(1+sqrt(abs($freq/F0)));
    V(sigout) <+ V(sigin)*FreqGain;

end

endmodule

p.s.

If I make the output a current and connect output to a resistor load then I get something that looks like the expected behaviour.

For this purpose, you could also try to use the fracpole component. For more information, type "spectre -h fracpole" at the command line. The article at http://www.designers-guide.org/modeling/suite.pdf has some additional background information. I have not actually tried this myself so far, so if you try this and it works for you, I would appreciate a short note.