Re: [CAD_CAM_EDM_DRO] Lament for the PC
Posted by
Alan Marconett KM6VV
on 2001-04-22 20:23:54 UTC
Hi Larry,
50%! That's pretty bad! The lights were only for quick debug, so I
probably won't fool with them again.
We did keyboard tests, lights, keys, repeat, etc. for a diagnostic
product perhaps 8 years ago, didn't seem to be too big a thing then.
I have no idea what kbd processor is in this machine, but I intend to
stay away from it. I can safely read the raw press/release codes. I
verified the the keys weren't "repeating" as I first thought, and I now
know where the "lulls" in the step streams came from. Time for a
different approach to jog.
CUL8R!
Alan KM6VV
Larry Edington wrote:
50%! That's pretty bad! The lights were only for quick debug, so I
probably won't fool with them again.
We did keyboard tests, lights, keys, repeat, etc. for a diagnostic
product perhaps 8 years ago, didn't seem to be too big a thing then.
I have no idea what kbd processor is in this machine, but I intend to
stay away from it. I can safely read the raw press/release codes. I
verified the the keys weren't "repeating" as I first thought, and I now
know where the "lulls" in the step streams came from. Time for a
different approach to jog.
CUL8R!
Alan KM6VV
Larry Edington wrote:
>
> There are some serious bugs in the keyboard microcontrollers accepting
> writes. If it's an 8042 / 8049 based controller on the system board then
> it's probably got those bugs. If it's an updated version using an 8051 or
> 8052 then those 'should' have been fixed. Depending on how interested the
> vendor is in fixing bugs.
>
> On some controllers, the documented way of doing writes just won't work.
> Unfortunately, the last time I delt with this issue was 1997. Then I did a
> test program and had the computer company I worked for's competitive
> analysis lab run it on all of the different brand computers they had. It
> worked out to about an even split 50% worked right, 50% failed. These were
> brand new 97 model units.
>
> Somewhere, in some deep dark archive I should have some 80x86 .asm code that
> worked on 100% of those tested machines. Turning on the lights was one of
> the test functions. If you are still interested in doing this I'll look for
> the code. Wasn't there a BIOS variable table entry you could write to for
> the lights? Seems like I did this on some program sometime. Maybe it's just
> a fuzzy memory though. It's just been too long since I've done low level PC
> programming. My goodness, I haven't written any 80x86 .asm in 3 years! Gotta
> fix that!
>
> later,
> Larry E.
>
> ----- Original Message -----
> From: "Alan Marconett KM6VV" <KM6VV@...>
> To: "CAD CAM" <CAD_CAM_EDM_DRO@yahoogroups.com>
> Sent: Friday, April 20, 2001 4:59 PM
> Subject: [CAD_CAM_EDM_DRO] Lament for the PC
>
> > Chris, list,
> >
> > Another lament for the PC. I tried to do a simple write to the keyboard
> > controller in a stepper program I'm writing, and BOY did the PC hate it
> > (crash)! In a windoz98 DOS window, I might understand, but booting up
> > just into DOS? Dropped that Idea. I thought the "keyboard lights"
> > would make good debug tools. Oh well, steal a bit on the parallel port,
> > hook up a 'scope.
> >
> > Alan KM6VV
Discussion Thread
Alan Marconett KM6VV
2001-04-20 16:00:11 UTC
Lament for the PC
Larry Edington
2001-04-20 17:43:26 UTC
Re: [CAD_CAM_EDM_DRO] Lament for the PC
Alan Marconett KM6VV
2001-04-22 20:23:54 UTC
Re: [CAD_CAM_EDM_DRO] Lament for the PC