Surfer said:
I was away for a month and left my computer off and after I returned, it
worked fine for a few days and then I got a series of blue screen
crashes. Then it failed to even turn on.
You can see a power led lit on the motherboard before it is turned on
but when I turn it on it won't show anything on the monitor, it makes no
beeping sounds at all, and after a little while, it seems to restart
itself and try again but I get nothing at all on the monitor and it
won't go into the bios setup or anything.
I took it apart, cleaned everything and reseated the ram and all
connections, but it refuses to turn on or show anything on the monitor.
I tried putting the ram in one stick at a time and in different slots,
but it made no difference. I hooked up the hard drives to another
computer but they seem fine and I can read and copy stuff.
Can anyone please advise on which components may be causing this?
The motherboard? The ram? The cpu? The graphics card? The psu? Something
else?
I do not have a second computer to try swapping components around to see
which work.
Thanks for any advice.
I think you've done some good testing so far, and seem to understand
the principles.
The BSOD could have been RAM, but by testing one stick at a time,
you've kinda eliminated that possibility. Normally, all the RAM
sticks don't fail at the same time. (I've had one instance where
they all started throwing errors at the same time, but those errors
did not prevent boot attempts from working. Replacement sticks from
the same purchase lot, worked fine. Otherwise, failures affected only
one stick at a time, when I've had problems)
Philo has suggested power supply, and that's a possibility. You
would use a multimeter, to check the voltages on it. Look
for the voltages to be within 5% of nominal. You can probe
where the wires go into the 24 pin connector, to get voltage
samples.
(This doc has the pinout, so you can tell which wire is which...)
http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf
I'm curious how you know the thing is restarting itself. If it
was restarting itself in a controlled manner, that tells you
the CPU is likely working, and is reading BIOS code. And
something else is preventing it from displaying something
on the monitor.
The green LED on the motherboard surface is run from +5VSB, and
when the system is powered, that light should stay on steady. It
shouldn't blink or anything. Perhaps you have a LED on the front
of the case that is blinking ?
In testing, you try to pull extraneous hardware, anything
which is not "core" to making the machine work. If you had
a sound card, you'd remove that and put it in an antistatic bag.
Always turn off all the power, before changing cards or changing RAM.
You can also disconnect the hard drives, while you focus on
CPU/mobo/PSU.
The motherboard may or may not have integrated video. You
may have two video solutions. You can move the monitor cable
between those, and see if a video signal is present. And if you
did have integrated video, a test case you could try, is to pull
the PCI Express video card (put it in an antistatic bag), and
do another test.
Now, a fun test, is to test with no RAM at all. The computer case
speaker (the one that beeps), that one should beep a RAM missing
error code, when no RAM is present. If you get beeps, that tells
you the CPU is running, the chipset is in relatively good
shape, and the CPU is able to get BIOS code. The repetitive beep
pattern, is generated under CPU control. Which is why the
missing RAM test is useful. So you're not really testing the RAM,
you're testing the system response to no RAM being present. And
determining that the CPU is running. If you get no beeps, then
either CPU power is missing, power supply is dead, etc. There
are still a bunch of things that could have failed.
If you get beeps on the missing RAM test, but no beeps when
RAM is installed, that could be a motherboard problem with the
RAM subsystem.
There is a beep code for missing video, so with one stick of
RAM present, you could pull the video card, and listen for
that beep pattern. But if there is integrated video, it's pretty
hard to test for a beep pattern on that, since it'll always
be enabled by the motherboard (as the only option).
It's not always possible to narrow down the failure to
just a single component, and I don't have a "flow chart"
to do that. So at this point, you're still gathering
whatever evidence you can. And hoping that some symptom,
points at the culprit.
*******
When I had an Antec power supply fail here, the first warning
was a "sizzling sound" during the first 30 seconds of operation.
That was the leaking caps on +5V, making a noise. Eventually,
one day I got a BSOD (as the +5V was getting worse and worse).
And I even got a little puff of smoke from it, just for fun.
I managed to replace that PSU, with no damage to my system.
What I saw inside the supply, looked like this. Rusty orange
on top of multiple caps. Those are leaking, and account for
the power from the supply, no longer meeting spec. If you inspect
the supply like this, you unplug it, remove the four screws
for the cover, and *don't touch* anything inside. Just look,
for the orange crap. While the primary side cap has a bleeder
resistor on it, you should never trust the bleeder to be
functional, and always assume the primary is fully charged
to operating voltage. That'll keep you alive a bit longer.
And you don't use the screwdriver discharge technique on
the primary - that'll pit the tip of your screwdriver,
and deafen you. This is a case, where all you want to do is
have a look, then close it back up again.
http://upload.wikimedia.org/wikipedia/commons/8/84/PSU_Caps.jpg
Paul
