Archived from groups: alt.comp.periphs.mainboard.asus (
More info?)
In article <hi5ee.3$rn6.2284@news.uswest.net>, "Michael Hobbs"
<mdhobbs@qwest.net> wrote:
> I have just finished building this computer, P4e, pc3200 memory. It reboots
> so fast I can't update or set the bios. I have encountered this before, it
> is time I understand what is going on. Obviously, I have encountered a level
> of motherboard I have not seen before. Or at least a problem not seen to
> date.
> Whatever could be going on? I believe my power supply to be okay. Can anyone
> offer sage advice on building computers with this motherboard?
> Thank you so much I am really stumped.
> mdhobbs@qwest.net
First of all, your motherboard has Vocal POST. The Lineout
connector gets error messages during the power on self test,
and those will give you limited debugging information. So,
that is one source of information.
A Port 80 PCI debug card can also be used, and it displays
two digit codes indicating what BIOS routine is currently
executing. In this case, you would be interested in whether
the same code shows before the computer resets, or whether
the code is always different. (The idea being, to see whether
the reset is software generated or is an uncorrelated hardware
event of some kind). At one time, these cards were about $100,
but can be bought new from E-bay importers for much less.
(These cards are possibly more useful for motherboards without
Vocal POST.)
In terms of test procedures, you can try the "cardboard test".
I've done this to the last two computers I've built, as part
of the build process. In other words, I tested the systems
as if they were busted, just to make sure the responses of the
system were correct and reasonable. I actually assemble and
boot the hard drive, with all the stuff sitting on the table.
Only then do I assemble the parts in the case. (I even test
the power supply, before connecting it to the motherboard,
but since I have a home made dummy load for the supply, this
is outside the bounds of what most people would bother with.)
You can place just the motherboard and an ATX PSU on your work
table. The motherboard can sit on top of a thick phone book, to
give it some support when you plug components into it. The
height of the phone book, also leaves room for AGP or PCI brackets
to hang down below the level of the motherboard.
The motherboard can be powered up without any components being
plugged into it. It should stay powered under those conditions.
When listening to your amplified speakers, you'll hear an error
message of "No CPU" or the like, telling you the Vocal POST has
detected that no processor is plugged in. (To start the motherboard,
all you need to do is short the two pins where the power switch
normally goes, for an instant, with the tip of a screwdriver. The
power pins are "momentary contact" and logic on the board latches
the start signal and holds the PSU in the ON state. No other
wires need to be connected to the PANEL header while doing the
tests. To switch off, you can just use the switch on the back
of the PSU, as it is more convenient.)
The next step would be to unplug the power supply, and add
the processor and heatsink/fan. Plug in the fan. Plug the power
supply back in and touch the power_sw pins on the PANEL header
with the screwdriver tip again, to turn the motherboard on.
This time, the Vocal POST should tell you that no memory was
detected on the motherboard, since now we have a processor, and
it executes enough BIOS code, to discover that no memory is
present in the DIMM sockets.
You add memory, video card and monitor, keyboard and mouse,
floppy drive, in that order, and listen to the error messages.
The "keyboard missing" message can take 30 seconds before you'll
hear it, so for some of these tests you have to be patient.
You can also observe the BIOS screen, once video is established,
to see if the screen contents are reasonable.
Somewhere early on in this test sequence, you are going to
get symptoms which are out of the ordinary. Perhaps even the
bare motherboard test is going to fail (i.e. if the board resets
itself over and over again with no components). The purpose of
the test sequence is to give you unique symptoms with various
components plugged into the board, which you hope will be
distinct enough symptoms to decide what is broken.
Looking at the reset logic on the reference schematic, I'm having
trouble following all the inputs to the reset logic. Certainly,
a processor can issue a reset instruction, which will reset a
board. There will be reset instructions in the BIOS code, for
the case of "Save and Exit" for example. Reset also seems to have
something to do with the power sequence logic, and signals like
Power_Good from the PSU. But right now that is about all I've
figured out.
Does the green motherboard LED on the motherboard stay steadily
lit, or does it blink off for a second ? The +5VSB supply should
stay on as long as the switch on the back of the PSU is on. If
+5VSB blinks off, I would expect the motherboard to stay off
until the power button on the front of the computer is used
again. It would take setting the power recovery setting to
restart after a power failure, for the motherboard to stay
in a "reset loop" in a situation like that, which isn't likely
to happen on its own. If the green LED blinks for even an
instant, it could mean some kind of overlood caused by the
motherboard, or a bad power supply.
Post back whatever symptoms you get during your sequential
"cardboard tests" and perhaps we can guess at what is
busted. It could be that all of this is caused by something
shorting to the bottom of the motherboard, or even stress
applied to the processor socket.
HTH,
Paul
Facebook
Twitter
Instagram