Power Device PSPICE Macromodel Convergence Problem

[Regnum]

Hi,

I've been working on a power device macromodel based on abm, but I got stuck in convergence issues (transient analysis) when Vdd goes over 100V
I changed RELTOL, ABSTOL, VNTOL, etc., but I wasn't yet able to have it working with higher voltages.

The model is in the attached pdf. I am using PSPICE from Cadence SPB v16.3

Any ideas?

Thanks
Regnum

 

[dick_freebird]

Macromodels with controlled sources can produce singular
or blowup results pretty easily. They lack any limiting. If you
have the ability to make function-based sources, tanh() is a
nice way to make a limiting function economically. You might
try putting large value resistors across cccs outputs, small
resistors in series with voltage sources, small finite capacitances
on nodes that have high dV/dt and so on. Timestep errors often
are about too-small capacitance while insane I, V convergence
errors are likely about the controlled sources. Try sweeping past
100V real slow and see if you can catch the beginning of the
blowup (some node or port that just looks wrong, right before
it all bombs).