On Sat, Dec 30, 2006 at 06:36:06PM -0500, Timothy Miller wrote:
> As our own PCI controller matures, we're going to want to test it in
> situ. The trouble will be when it fails to work properly, and the
> solution will be to use some device to watch the PCI bus activity and
> let us observe them. Such "PCI bus analyzer" devices to exist, but
> they cost a fortune. Then it occurred to me that perhaps another OGD1
> board would be ideal for this. It's got plenty of memory to store a
> large window of bus activity, and with the user I/O pins, we could rig
> up an RS232 port on it that we could use to communicate with the
> device. Moreover, it had occurred to me that we could make some money
> selling such a product. Of course, we couldn't sell it for the same
> amount as competing products (because it's open source... you can just
> buy and OGD1 board and download the FPGA code), but we could sell it
> pre-installed with support at a premium. The only thing about this
> that's a major drawback is that the market is waning rapidly as PCI
> Express takes over.
>
> Thoughts?
You can configure it to have an internal UART? Of course you can,
it's an FPGA. I should have thought of that. So then it needs an adapter
cable to bring the serial port signals from the accessory header out to a
DE-9S connector. However, they're at low logic levels, and may not work
reliably with a terminal or user PC that expects RS-232 levels. The obvious
solution is a "tool" board, with an RS-232 hardware interface on it. That's
one chip and a litle discrete glue to provide suitable voltages and I/O
overvoltage protection. If this goes ahead, let me know, and I can probably
whip up a circuit design and board layout instructions in short order.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
