Re: [pci] Need some layout tips

["cfk" <cfk@pacbell.net>]

Dear Nico:
    Most likely you will tend to use either +5v or +3.3v but not favor both.
If that is the case, I would favor the predominant one with a plane and just
connect the non-predominant one with fat traces. Fat traces are also
appropriate for +/-12V which is probably not used at all. You should have no
+2.5V to the PCI interface.

    You get to decide if your interface is a +5V card, a +3.3V card or
universal (both +5V & +3.3V). I would suggest +3.3V as the IO on the FPGA
will have trouble with +5V (they arent rated for +5V). Program them for
PCI3.3, use +3.3V on the IO, not +2.5V.

    On routing, I have routed a board using the PCI bridge on only outer
layers. The reason I did that to allow the proverbial cut and jumper if
required which is much easier on the outer layers.

    So, you can route with a 4 layer board, component side, solder side,
power and ground. If you have more layers, the routing gets easier, but it
can be done with 4.

    Hope this helps.

----- Original Message -----
From: "Nico Weling" <eedniwe@eed.ericsson.se>
To: <pci@opencores.org>
Sent: Monday, May 05, 2003 7:15 AM
Subject: [pci] Need some layout tips


> Hello,
>
> I'm desinging a PCI board which will contain one FPGA (X2S200 FG456) for
the
> PCI-Bridge one FPGA for a 16 channel FFT and 8 Stereo 24bit/96kHz Codecs.
>
> As base for the PCI stuff I have one Memec Board. Could someone please
give me
> some informations on what I shoult take care of for the Layout of the
Board?
>
> Do I need one Layer for each Powerstage (+5V,+3.3V,+2.5V,+12V,-12V)?
> How many layers would make sense to not kill the Autorouter?
> Which would be the best order for the Layers?
(Signals,GND,+3.3V....,Signals)
> If I don't use 2.5V IO's do I need to supply the core with 2.5V or is 3.3V
OK?
>
> Thanks in advance for your help.
>
> Regards,
>
> Nico.
>
> --
> To unsubscribe from pci mailing list please visit
http://www.opencores.org/mailinglists.shtml

--
To unsubscribe from pci mailing list please visit http://www.opencores.org/mailinglists.shtml