[comp.unix.sysv386] quack surgery on a VGA card

chip@chinacat.Unicom.COM (Chip Rosenthal) (09/14/90)

I am running ODT on a Micronics 386/20 with an Orchid Prodesigner running
in normal 640x480 VGA mode.  This board brings a trace out to the IRQ9/IRQ2
bus pin.  I am assuming it's the vertical retrace signal.  I further
believe that such a signal isn't used by the system, therefore I took an
Xacto knife to it to free up that IRQ.

Now I am seeing flaky behaviour on the ODT machine.  Every once in a
while, the display gets scrambled.  Things continue running, but it looks
like someone took a huge phospher eraser and smudged up the whole screen.
Switching to a different mscreen and back cures the problem.

I haven't seen this behaviour before.  However, there were other changes
(such as the installation of a multi-port serial card) at the same time
which might be having an effect.  It does seem like something is going
goofy with the memory system, and I'm wondering if that signal is indeed
used after all in a refresh somewhere.  Can anybody say definitively
whether I made a major goof?  I will try reconnecting that trace tonight
to see if there is any effect.

-- 
Chip Rosenthal  <chip@chinacat.Unicom.COM>
Unicom Systems Development, 512-482-8260 
Our motto is:  We never say, "But it works with DOS."

gemini@geminix.in-berlin.de (Uwe Doering) (09/18/90)

chip@chinacat.Unicom.COM (Chip Rosenthal) writes:

>I am running ODT on a Micronics 386/20 with an Orchid Prodesigner running
>in normal 640x480 VGA mode.  This board brings a trace out to the IRQ9/IRQ2
>bus pin.  I am assuming it's the vertical retrace signal.  I further
>believe that such a signal isn't used by the system, therefore I took an
>Xacto knife to it to free up that IRQ.
>
>Now I am seeing flaky behaviour on the ODT machine.  Every once in a
>while, the display gets scrambled.  Things continue running, but it looks
> [stuff deleted]
>Can anybody say definitively
>whether I made a major goof?  I will try reconnecting that trace tonight
>to see if there is any effect.

I assume you mean the B4 bus contact. I've done the same thing on several
VGA cards with no problems at all. UNIX doesn't use this IRQ. And the
VGA BIOS in the EPROM on the card isn't used under UNIX (with the exception
of VP/ix). Therefor it really doesn't matter whether this trace leads to
the bus connector or not. Currently I'm running an AST 4-port compatible
serial card on IRQ2/9 and I use an Optima/1024 VGA card which is compatible
to the Orchid Prodesigner card. This is with ISC 386/ix 2.0.2.

     Uwe
-- 
Uwe Doering  |  Domain   : gemini@geminix.in-berlin.de
Berlin       |----------------------------------------------------------------
Germany      |  Bangpath : ...!unido!fub!tmpmbx!geminix!gemini