Re: Modeling method and Intel non-monotonic model

Reflecting more on this issue..

   Even though the data from Intel contained non monotonicity, the model
   itself worked out nicely because of the "addition" effect of the clamps.
   But there may arise cases where we may have
   non monotonocity in the operating region which may real or may be an
   artifact of the modelling method and software. Worse still , if the
   simulations do not converge, it is almost impossible to tell whether
   it is an artifact of the simulator or the real device itself is
   unstable. In such cases , which Jon and Kellee may be alluding to ,
   we may consider issuing general warning about the presence of
   non monotonic data.

- Kumar
Received on Thu Apr 7 06:53:16 1994