The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design Languages >> Verilog-AMS >> Interpolating From Function
https://designers-guide.org/forum/YaBB.pl?num=1367133417

Message started by warprider on Apr 28^th, 2013, 12:16am

Title: Interpolating From Function
Post by warprider on Apr 28^th, 2013, 12:16am

Hi everyone! I'm interested in simulating alternative FET-style devices (CNTFET, TFET, UTBSOI, etc), and I'd like to make Verilog-AMS models for their DC characteristics.

Right now, I have a series of lookup tables with IDS given VDS, VGS in increments of 10mV, and I interpolate between these points. It works out pretty well, as I am just doing simple circuits, such as inverters, transmission gates, etc.

My problem with this current method is that I want to do a Monte Carlo simulation with Vth variation with a large number of trials. Even though I have fitting equations and can recompute the lookup tables easily in Matlab, or even Verilog-AMS, I still need to write to a file, then read from that file using $table_model, so running tens of thousands of simulations like this seems very inefficient.

I originally tried to simply use the fitting equations as IDS equations for the device, but this led to convergence errors many times. I haven't had nearly as many problems with the lookup tables. I imagine because of the leeway given by the interpolation?

Can any of you suggest a quicker way to do this, or how to give more leeway to the fitting equations so I don't have to precompute these gigantic tables?

If you have any more questions that may help you give more detailed advice, I'd be happy to answer.