Re: Mfactor proposal

Kevin Cameron wrote:
>
> I presume you would mention "mfactor" in the text along with the {} - or
> whatever other syntax is used - so it's not going to be hard to find
> with a search. I've got to say "hard to find in the LRM" is the lamest
> excuse for not doing something I've heard in a while.

I think the look-up problem was intended from the standpoint of:
I've got this instance line that has {5} in it, and I don't know
what that 5 does, how do I figure it out?

> While I can see how "m" bears on simulation, what do the others do and
> how do they affect simulation?

n is a series multiplicity; a panelist from National Semi
mentioned it (though he called it "S"), and I've heard it
comes up in extracted parasitics.

The others are for mismatch simulations, in which one assumes
say a gradient of TOX across the wafer/chip, and then the
position (x,y) is used to compute TOX for each transistor;
one can also mirror the transistor (using hflip and vflip)
or possibly rotate it some angle. Analog designers go to
a lot of trouble to "cross-quad" devices to try to cancel
out linear gradients, and it would be good to allow them
to simulate the effects.

-Geoffrey
Received on Thu Jun 17 09:55:19 2004