Re: [oc] Modular FPGA board

[Marco Antonio Simon dal Poz <mdalpoz@lsi.usp.br>]

On Thu, 23 Nov 2000, John Dalton wrote:

> > Very good idea, but is there some open core with PCI master/target
> > interface? I don't think so. Are you planning to use comercial chips for
> > PCI interface?
> 
> I'm planning to do PCI in an FPGA.  For this entire project, I would like
> to reuse any relevant work which has been done on the OCRP-2.  The
> free PCI core is not yet finished, but the availability of a board might
> provide the impetus for this work to progress.

Are you working in the development of the free PCI core? If you specify an
FPGA for PCI core testing, I could create such a board, but of course it
wouldn't be cheap. My idea was the opposite of your: I was planning to
built a board for PCI testing only when the core will be finished. But I
still have a problem: I don't have the PCI specs and personally I can't
afford it. So, I don't know exactly how should I match the impedances and
equalize all the delays between the signals.

If a PCI core is validated through simulations and it doesn't work over a
specific board, what should I do to detect the problem? Measurements using
oscilloscopes and logical analysers are not possible, because they modify
the circuit when connected into them due to cable impedances, that
generates multiple reflection. Is there any solution?

> On the subject of reusability, it shouldn't be too hard for someone to
> build each of the OCRP-2 peripherals onto a SIMM, allowing an
> OCRP-2 clone to be built. The main reasons for not just using an OCRP-2 are:
> 1) improved expandability (adding new things)
> 2) improved upgradability (replacing old things with new things)
> 3) improved customization (not having to pay for unused peripherals)

I agree completely.

> The current dilemma I am dealing with is that putting 2 FPGAs (1 PCI,
> 1 main) on the motherboard means two expensive chips, which are hard
> to remove, cannot be upgraded independently.  I'm currently figuring
> out how to make PCI a module, (along with ISA, EISA, etc.) while still
> satisfying mechanical (and electrical) constraints, and not spending a
> fortune on connectors.  FPGAs, with enough pins to do PCI, don't seem
> to fit in cheap sockets.

I don't think that an FPGA for PCI have to fit in a socket. IMHO, it has
to be soldered in a board. And if this board should fit in a SIMM socket,
the FPGA has to be a TQFP or BGA packet.

> Another argument for modular PCI is that a double sided edge connector
> is hard to do at home.

TO DO AT HOME??? Well, if I understood it right, it is much harder to do
the copper lines for the FPGA than for the edge connector. Is there any
problem if the lines of the edge connector have the right distance and
smaller width (ie, increasing the spacing)?

BTW, using some CAD tool, it's not difficult to design a PCI board, and
a 3-layer fabrication is not that expensive.

> > 1.6mm PCB doesn't fit into a 72 pin SIMM slot. And here in Brazil, 1.27mm
> > is much easier to build (BTW, I have never seen a 1.6mm thickness PCB).
> A pity.  Here in Australia, 99% of PCBs are 1.6mm (1/16") thick.  
> (Including most imports.)  1.27mm (0.05") is used for special
> applications, such as SIMMs. As a result, the electronic chain stores
> stock only 1.6mm. 1.27mm can be bought, but it is harder to get.

In AU, would be too much expensive to import such kind of boards and to
transfer the layout to them?

> > Again, here in Brazil it's becoming difficult do find SIMMs, and when
> > someone is able to sell or trade SIMMs, it costs more than the double
> > price of the DIMMs of the same memory capacity.
> Opps.  Incorrect use of terminology by me.  I meant to inquire about
> the global availability of 72 pin SIMM *sockets*.  They seem to be
> available in AU, though not from the major chain stores.

Then the situation is the opposite. SIMM sockets are easier to find than
DIMM sockets, at least in Brazil. But if you are planning to buy
components from the USA, then you should check www.digikey.com.

Greetings from Brazil!

	Marco Antonio