Re: 3v drivers on 5v busses (CORRECTION)

["Andy Ingraham" <ingraham@wrksys.ENET.dec.com>]

I made a statement in my reply that might be misinterpreted, so I've
re-worded it.  Sorry about any confusion.


>                                       The question is, has
> anyone succeeded in making a PCI compatible pad cell for 33MHz
> 5V busses with 3V only  drivers?

Sure.  There are plenty of them.

> The part will plug into both 33MHz and 66MHz buses so it will 
> be exposed to both 3V and 5V busses.

Here you have to be a bit more careful.  PCI cells that work in both
3.3V and 5V busses have been around for some time as well (my company
makes chips with them), but you can't make a PCI cell that will work
correctly in both 3.3V and 5V busses without at least one electrical
change.  The cells MUST clamp to 3.3V in the 3.3V bus, but must NOT
clamp to 3.3V in the 5V bus.

3V drivers on a 5V bus can be somewhat more difficult for signal
integrity, because the rising edge won't overshoot enough to dissipate
energy in the 5V clamps (if present).  So bouncing and ringing back into
the threshold region are more likely.

And of course, the 3V OUTPUTS (not just the inputs) must be 5V tolerant. 
The usual CMOS driver doesn't work.

Regards,
Andy Ingraham


˜ØÆ