Stats

  • Locked Locked
  • Replies 3
  • Subscribers 126
  • Views 15520
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Convergence difficulties in Verilog A while using equivalent differential equation instead of laplace_nd function

Quilon

Hi

I was using a laplace_nd function to realize a CTLE block as follows

V(vp) <+ laplace_nd((0.5*(V(INP)- V(INN)))*rl*gmn , hn , hd);
V(vn) <+ laplace_nd((0.5*(V(INN)- V(INP)))*rl*gmn , hn , hd);

whre hn, hd are declared as "real hn[0:2], hd[0:3];"

It worked well without any sort of convergence difficulties in bot AC and transient simulations. But since I wanted to change the values of hn and hd in between the transient simulation, I changed the laplace_nd function into equivalent differential equation as given below

V(vp) <+ ((0.5*gmn*rl*(hn[0]*V(vip)+hn[1]*V(vipd)+hn[2]*V(vipdd)))-hd[1]*V(vpd)-hd[2]*V(vpdd)-hd[3]*V(vpddd))*(1/hd[0]);
V(vn) <+ ((0.5*gmn*rl*(hn[0]*V(vin)+hn[1]*V(vind)+hn[2]*V(vindd)))-hd[1]*V(vnd)-hd[2]*V(vndd)-hd[3]*V(vnddd))*(1/hd[0]);

where

V(vip) <+ V(INP)- V(INN);
V(vin) <+ V(INN)- V(INP);
V(vipd) <+ ddt(V(vip));
V(vipdd) <+ ddt(V(vipd));
V(vind) <+ ddt(V(vin));
V(vindd) <+ ddt(V(vind));
V(vpd) <+ ddt(V(vp));
V(vpdd) <+ ddt(V(vpd));
V(vpddd) <+ ddt(V(vpdd));
V(vnd) <+ ddt(V(vn));
V(vndd) <+ ddt(V(vnd));
V(vnddd) <+ ddt(V(vndd));

The modified model showed exactly same AC response without any convergence difficulties. But it is showing convergence difficulties in transient simulations of same conditions (blowup parameter of nature Voltage was increased to accommodate high values coming to third derivative terms in the differential equation). How can I make the differential equation as convergable as the laplace_nd function.

Thanks and Regards

Quilon

Parents
  • Andrew Beckett

    Quilon,

    You probably need to either scale the higher derivatives or better still define new natures and disciplines for them which have got different abstols and blowup values to avoid this problem. Since the higher derivatives aren't really voltages any more, this would give the simulator a chance...

    For ideas on how disciplines and natures are defined, take a look at: <SPECTREinstDir>/tools.lnx86/spectre/etc/ahdl/disciplines.vams

    Regards,

    Andrew.

Reply
  • Andrew Beckett

    Quilon,

    You probably need to either scale the higher derivatives or better still define new natures and disciplines for them which have got different abstols and blowup values to avoid this problem. Since the higher derivatives aren't really voltages any more, this would give the simulator a chance...

    For ideas on how disciplines and natures are defined, take a look at: <SPECTREinstDir>/tools.lnx86/spectre/etc/ahdl/disciplines.vams

    Regards,

    Andrew.

Children
No Data

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.