Re: voltage and temperature in corner cases

> I'm trying to figure out how to tweak all the parameters to exercise
> worst case simulations. ...

In general, the driver has a much greater effect than the receiver; so
I'd probably tailor the simulation parameters for the driver.

I would expect (but don't know with certainty) that temperature has no
effect whatsoever on IBIS models in a simulation. The [Temperature
Range] keyword seems to be there only for documentation ... which is
good because it's so often misused. The weak and strong parameters
that are in the IBIS model, are already margined to the temperature
extremes that the model designer chose to use. So if you were to do
a simulation that had only IBIS (no SPICE) devices, varying the
simulation temperature would be a no-op. That being the case, you can
ignore it ... or set it as appropriate for your non-IBIS devices.

The case with supply voltages is similar, but with a twist. Once
again, the weak and strong IBIS models already have supply voltage
figured into them, but only partly. Varying the supply voltage
during simulation, will slide the high clamp and pullup curves up
and down in direct proportion; but their shapes would not change.
The shapes are already built into the weak, typical, and strong
parameters; and simulators probably aren't going to "reverse
engineer" the I/V curves to second-guess what they would have been,
had the supply voltage been different.

So ... while a strong model may have been CREATED using maximum supply
voltage, it can be USED at any supply voltage. The results may not be
100% correct, but the curves will track Vdd as they should.

For the driver, where you want the shapes of the curves (drive
strengths) right, choose the supply voltage appropriately for the
desired corner of the driver.

Or do several simulations.

Regards,
Andy
 
Received on Wed Jul 16 12:52:19 1997