On 8/30/06, Peter TB Brett <[EMAIL PROTECTED]> wrote:
Or by having a (very) small array of 0402 resistor pads, and only populating a
few of the landpads with 0R0 shorts to set up the signals required.  Not only
would that (probably) be cheaper than placing jumpers, but it would also take
up less PCB real estate.


Every individual card has to be individually tested.  We can write a
relatively automated test suite where you plug four boards at a time
into the test box, power it up, and have it automatically run tests,
reporting the results at the end.  I've done this for Tech Source
before, using a stripped-down Gentoo install, a handful of custom
programs, and a shell script to tie it all together.

Is there any reason why we can't include a firmware choice in the
test?  One of the tests involves reading the PROM, checking to make
sure it matches the file, and if not, reprogramming it (and checking
it once more just to be sure).

At Tech Source, we even set it up to ask for serial numbers, and we
could optionally program the serial numbers into the firmware.  For
that matter, we would have choices between different firmwares for the
same card, and the tester would select a global default and have the
ability to change it on a per-run basis at the time of entering serial
numbers.  There is no simple way to ensure quality and low fallout at
the end user without individual board testing, and given potential
volumes, that testing would have to be distributed.

Possibly, Traversal would hire resellers in various places around the
world.  Resellers could receive cards directly from the manufacturer
(those who populate the board with components), and the reseller would
be responsible for programming and testing the boards.  Policies and
procedures would be in place to deal with bad boards.

Thus, in the UK, you would buy an open graphics board from a UK
supplier who was the one who programmed the firmware for that region
in the first place.

Technically, you don't need multiple firmwares.  One solution is to
have a template firmware that is modified on the fly with the video
program for that region.  Another even simpler possibility is to have
the firmware include all modes and use a single-byte flag to select;
the test software would be responsible for setting that byte correctly
for the region when programming the PROM.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to