[GRLUG] BIOS problems - Round 2

[Tim Schmidt]

Well Bob...

Good to know you have a multimeter ;)

2.99v certainly sounds reasonable.  Probably not a battery problem
then (sorry about the $3).

Forgive me if I start with the basics...  It's what I do.

- Sorry for the repetition, but remove anything not strictly necessary
to the operation of the machine, cards, excess ram, drives,
everything.  Just in case.

- You've probably already done this, but look for obvious damage to
the board, blown caps, etc.  Of course, if there is visible damage the
board is likely dead (at least pending repair).

- Have you tried shorting the 'clear cmos' jumper?  If not, please do.
 Some boards don't have one, they almost universally have a space for
one though, and shorting solder pads works just as well as shorting
the jumper itself.  Some boards are extremely picky, try shorting with
and without the battery inserted, with and without the power supply
connected and all possible combinations of the two.  It's work, but
I've had one or two boards require power from one source but not the
other in order to clear properly.

- If the machine still appears to power up but nothing appears on
screen, take note of the type of graphics card in the system (AGP,
PCI, etc) and try other available types.  Often the BIOS, after
resetting to defaults, is set to initialize the type of graphics card
you don't have installed and some aren't smart enough to realize that
and move on to the next one.

- If this fails, and the system was previously working, there's not
much else to do...  check for loose screws and such between the board
and the mounting plate of the case...  Make absolutely sure every
component is seated properly (partially unseated RAM, CPUs, and
graphics cards are often culprits in non-booting machines).

- If I was to encounter this machine with no prior history, however, I
would also double-check any jumper settings on the board to configure
bus speed(s), CPU multipliers, etc.  If the board doesn't have any,
and subsequently employs some logic to auto-detect the CPU parameters,
try a slower / older CPU of similar make in the board if available. 
Older BIOS revisions may not know about newer CPUs.  I would also
cycle through different combinations of RAM and graphics cards.  Some
BIOSs just don't agree with some RAM and some graphics cards due to
timing issues or RAM configuration or buggy graphics BIOSs (or buggy
system BIOSs for that matter).

- If all else fails, and you really want to reprogram the BIOS chip
just to make sure, it can be done by finding a similar board, removing
it's BIOS and replacing it in the socket just enough so that the pins
make contact and booting it, after it's booted, remove it's BIOS chip
and replace with the one you wish to re-program (you can fully seat
this one).  And program away.  I'll dig up the links to relevant BIOS
programming software if you want.  Alternatively, if you have an exact
clone of the board lying around, you can use the same procedure with
the board manufacturer's software.

Let me know if that works or not.  I may come up with a few more
suggestions by tonight and send a second mail.  There's almost always
_something_ you can do to resuscitate a board, but there is a
diminishing return on efforts.

--tim

On 2/10/06, Bob Kline <bob.kline at gmail.com> wrote:
> I checked the battery, and it read 2.99V.   I popped in a new one
>  anyway - it read 3.27V.    With the machine powered off.
>  ( As Tim suggested,  the battery is $3.  )
>
>
>  I powered up the machine.  No POST messages.   I did see
>  the keyboard lights and the CDROM light blink.  Nothing on
>  the screen.   Now,  I had also removed the hard drive,  so it's
>  a little odd that I didn't get a POST message about that.
>
>
>  Suggestions?   The board ran for a year,  so I'm inclined to
>  think that something blew.   It had been booting now and
>  then during this recent episode,  but now nothing much happens.
>
>
>  Oh,  PS levels are all within normal range too.
>
>
>    -Bob