Re: FPGA for PCI based servo control board

["ERIC KELLER" <eekeller-at-psu.edu>]

Most boards you buy nowadays are what are called "universal" boards, i.e. they
will work on both 5v and 3.3v.  You can tell if they have 2 slots in the
motherboard connection fingers. As I understand it, they check to see if they
are running at 3 or 5v and then start up as appropriate.  Almost all of the
pc's I've seen have had 5v pci.  The dell I'm using has 3.3v 64bit slots and 5v
32bit slots. 3.3v slots have the tab closer to the back of the computer.  The
connector for 32 bit 3.3v slots are just the 5 v connectors backwards.

I had thought about the issue of driving higher voltage outputs.  I've seen
numerous app notes about this issue.  My experience is that 5v logic will not
drive an opto22 module reliably, I always tie the upper control voltage to
power and pull the lower leg to logic ground with the port.  Same scheme goes
for opto isolators, assuming the fpga can sink adequate current. 

Of course, I've always been fond of boards that have socketed ttl buffers that
are hooked up to the outside world.  Makes it easier to replace them when the
careless mistake happens.

Eric

On Tue, 01 Apr 2003 04:47:32, John Sheahan wrote:

> 
> On Tue, Apr 01, 2003 at 01:52:57AM -0500, Jon Elson wrote:
> > 
> > 
> > 
> > ERIC KELLER wrote:
> > 
> > >I looked at a lot of fpgas to use for PCI interfacing.  I came down to the
> > >notion that only the Xilinx Spartan II made sense.
> > >The Spartan II has 3.3 and 5 v i/o, and thus is 5v pci compatible.
> > >
> > That is NOT true 5 V I/O.  It is 5V TOLERANT I/O.  If you want to pull 
> > up above
> > 3 V, you need low-value pull-up resistors external to the chip.  If you 
> > need really fast
> > positive-going transitions to +5 V, you can't get them.
> > 
> > Jon
> > 
> 
> 
> true - but 5v pci only needs to be 5v tolerant. 
> driving 3v3 is in specification. 
> 
> is there much 5v PCI any more?
> john
> 
> 
>