The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Oct 17th, 2019, 3:34pm
Pages: 1
Send Topic Print
Could the ideal delay cell from analogLib be used to run PSS simulation? (Read 2925 times)
Homer
Junior Member
**
Offline



Posts: 14

Could the ideal delay cell from analogLib be used to run PSS simulation?
Jul 16th, 2015, 8:55am
 
I used the ideal delay cell of analogLib to generate 256 delayed signals as the inputs of the circuit. Now I got convergence problem. It could be caused by other issues. But the is the ideal delay cell OK for PSS simulation? It seems the absdelay in veriloga can't be used for PSS simulation.
Back to top
 
 
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2222
Silicon Valley
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #1 - Jul 16th, 2015, 11:21am
 
No, you cannot use ideal delay in PSS.

-Ken
Back to top
 
 
View Profile WWW   IP Logged
Homer
Junior Member
**
Offline



Posts: 14

Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #2 - Jul 16th, 2015, 1:22pm
 
Ken Kundert wrote on Jul 16th, 2015, 11:21am:
No, you cannot use ideal delay in PSS.

-Ken


Thanks Ken。
Back to top
 
 
View Profile   IP Logged
Homer
Junior Member
**
Offline



Posts: 14

Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #3 - Jul 22nd, 2015, 1:31pm
 
Hi, Ken,

Is timer() function valid for PSS simulation? I used real RC delay to implement the 256 delayed clocks. It's fine but the simulation time is two long.

So I want to implement 256 clock source in veriloga. I used timer() to implement the clock. It's fine for tran. But there no clock for PSS plot. Is timer() valid for PSS? If it's not, any suggestion?
Back to top
 
 
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2222
Silicon Valley
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #4 - Jul 22nd, 2015, 7:53pm
 
Using timer() is fine. Just be sure to add a period along with the start time. Of course all timer functions should have the same period and it should be commensurate with the PSS period.

-Ken
Back to top
 
 
View Profile WWW   IP Logged
cmosa
New Member
*
Offline



Posts: 9
IIT
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #5 - Sep 18th, 2019, 5:36pm
 
Hi Ken,
I am trying to create a new delay cell that does not have hidden states so that I can use to run PSS shooting method.  Here is my code.

module delay_trans (out, in);

output out; voltage out;
input in1; voltage in;

parameter real vh = 1;                  // output voltage in high state
parameter real vl = 0;                  // output voltage in low state
parameter real vth = (vh + vl)/2;      // threshold voltage at inputs
parameter real td = 0 from [0:inf);      // delay to start of output transition
parameter real tt = 0 from [0:inf);      // transition time of output signals

analog begin
   @(cross(V(in) - vth));
   V(out) <+ transition( V(in), td, tt );
end
endmodule

Please suggest.
Back to top
 
 
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2222
Silicon Valley
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #6 - Sep 18th, 2019, 8:25pm
 
I don't know why the cross function is there.

The transition function is not designed for continuous signals and it will not give good results. In addition, the delay in a transition function represents hidden state. They can work in SpectreRF if there is no transition in the queue at a period boundary. But that would only be true with discrete events that are isolated from the period boundary.

-Ken
Back to top
 
 
View Profile WWW   IP Logged
cmosa
New Member
*
Offline



Posts: 9
IIT
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #7 - Oct 11th, 2019, 4:25pm
 
Hi Ken,
Thanks for your input.
I found a workaround for running PSS with a "delay" cell.
I ended up using an ahdlLib inverter for which we can specify the delay value. I also specified non-zero tr and tf for the ahdlLib inverter.

With this, PSS  analysis does not give me an error about "distributed element" or "hidden state". However, PSS does not converge !!!

Is it possible that the ideal ahdlLib inverter transfer function is not smooth causing PSS convergence errors?


The log file shows the following. "PSS failed to converge PSS will adjust tstab a little and try again".
finally, "Even after adjusting tstab value, PSS failed to converge"

I checked the transient waveforms at the end of "tstab" time, they look periodic and correct.

Thanks
-CMOSA
Back to top
 
 
View Profile   IP Logged
Ken Kundert
Global Moderator
*****
Offline



Posts: 2222
Silicon Valley
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #8 - Oct 11th, 2019, 6:25pm
 
The only non-hidden state is the state of capacitors and inductors. If you want to implement a delay that does not have issues with hidden state, use capacitors and inductors. Consider designing an all-pass filter.

-Ken
Back to top
 
 
View Profile WWW   IP Logged
cmosa
New Member
*
Offline



Posts: 9
IIT
Re: Could the ideal delay cell from analogLib be used to run PSS simulation?
Reply #9 - Oct 13th, 2019, 5:59am
 
Hi Ken,
Thanks, Will try CV/I kind of delay element.
Thanks
Marshnil
Back to top
 
 
View Profile   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2019 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.