How to detect a dead CPU?

Buy replacement parts that will accomodate future upgrades, so there
isn't any significant money wasted, you'll be on your way towards a
faster system still.

Agreed - unfortunately, the power supply does not appear to be the problem.
I took one out from a working machine and I still get the same problem. This
is symptomatic therefore that the problem is either mobo or cpu related or
even a combination of both, I can therefore conclude?

If so, I'm going to have to buy a new base computer, because this is very
expensive territory now.

Co-incidentally, if I want to test the CPU works in another machine can I
just swap a 1Ghz AMD Athlon for a 1.3Ghz AMD Athlon? And if the CPU is in
fact damaged, could it cause damage to the undamaged machine? Or vice
versa - if the motherboard is fried, could placing a good CPU in it damage
the good CPU?

The POST won't run until after the reset (to make sure all devices are
ready to be polled). If you don't see the flash of the keyboard LEDs
then there was no reset signal to prep it or anything else. The reset
signal to the CPU is held low until the PSU voltages stabilize and then
the CPU sends a reset pulse to all devices. If you don't see the lights
flash on the keyboard, it could be because the CPU never sends the reset
pulse (which indicates it is defective) or there is a problem with the
motherboard, like it got damaged by a surge or from improper handling
(i.e., no anti-static protection). I did assume that the "normal"
keyboard you tested with is NOT a USB keyboard but instead connects to
the PS/2 or AT keyboard port. I don't know how USB keyboards respond to
a reset signal since it would be the USB controller that got the reset
and not really the USB keyboard.

Note that testing a power supply just to check its voltages is not a
sufficient test. Something that tests at +12 volts with no load might
end up being much less under load. There are PSU testers that you can
get, like at CompUSA, Antec, or FMI, that put a *slight* load on the PSU
to check its voltages. Or leave the PSU connected to the motherboard
and all devices and shove the probe tip into the backside of the 4-pin
Molex power connector to measure voltages under load. it also possible
that regulation sucks and there is so much ripple as to not actually
provide the average power and voltage that is needed, but you can only
see that with an oscilloscope.

I never asked but didn't think of it since normally a user would've
already mentioned this. When you power up the system, do the hard
drives actually spin up? That is, do you actually get power when you
think you have powered up? The ATX power supply will not actually power
up (except for a 5V line) until the PS-On signal on the motherboard's
20-pin header to which the ATX connector attaches is no longer pulled
low. If the motherboard isn't releasing the PS-ON line (pin 14) then
the ATX power supply will not turn on the rest of its power taps and
nothing gets powered up (so neither would the hard drives which you can
do not spin up). See http://snurl.com/2mt4 where I answered a question
about PS-ON.

--
__________________________________________________ __________
*** Post replies to newsgroup. E-mail is not accepted. ***
__________________________________________________ __________


"ck26" wrote in message
...
If you have no video at all, a defective video card whose BIOS won't
load will kill the POST right away. Yank out the video card and you
should get, I think, 1 long beep and 2 short beeps (for Award BIOS).

Right - I have an onboard graphics card (which I can't remove) and a
PCI
card that was the card in use at the time of the "incident". I removed
the
PCI video card and nothing - still the same - no beeps. I can't really
disable the onboard one other than by telling the bios to use the PCI
rather
than onboard AGP video - which I did when I first installed the new
card.
Either way, I'm pretty sure its not the video now.

Do you see the LEDs flash on the keyboard? That's due to the reset
signal.

I connected a normal keyboard to see any lights being displayed upon
boot.
Nope. No lights. No beeps.

See http://www.pcguide.com/ref/mbsys/bios/boot.htm for the boot
sequence.

Great stuff on the beeps - shame there's nothing about no beeps!

the CPU sends a reset pulse to all devices. If you don't see the lights
flash on the keyboard, it could be because the CPU never sends the reset
pulse (which indicates it is defective)

Hmm, possibly - could I do a straight swap with an AMD Duron 1.3Ghz? (its
currently an AMD Athlon 1.3Ghz) - they are suprisingly cheap at my local
store and worth a try.

or there is a problem with the
motherboard, like it got damaged by a surge or from improper handling
(i.e., no anti-static protection).

Nothing has been handled internally until after it happened.

I did assume that the "normal"
keyboard you tested with is NOT a USB keyboard but instead connects to
the PS/2 or AT keyboard port. I don't know how USB keyboards respond to
a reset signal since it would be the USB controller that got the reset
and not really the USB keyboard.

Yes it was a straight no frills PS/2 keyboard.

Note that testing a power supply just to check its voltages is not a
sufficient test. Something that tests at +12 volts with no load might
end up being much less under load. There are PSU testers that you can
get, like at CompUSA, Antec, or FMI, that put a *slight* load on the PSU
to check its voltages. Or leave the PSU connected to the motherboard
and all devices and shove the probe tip into the backside of the 4-pin
Molex power connector to measure voltages under load. it also possible
that regulation sucks and there is so much ripple as to not actually
provide the average power and voltage that is needed, but you can only
see that with an oscilloscope.

I switched power supply with a working machine's - no difference
unfortunately...

I never asked but didn't think of it since normally a user would've
already mentioned this. When you power up the system, do the hard
drives actually spin up?

Yes, I'm pretty sure - my fans are quite noisy, but I'm sure I heard it
spinning. The HD light comes on anyway.

That is, do you actually get power when you
think you have powered up? The ATX power supply will not actually power
up (except for a 5V line) until the PS-On signal on the motherboard's
20-pin header to which the ATX connector attaches is no longer pulled
low. If the motherboard isn't releasing the PS-ON line (pin 14) then
the ATX power supply will not turn on the rest of its power taps and
nothing gets powered up (so neither would the hard drives which you can
do not spin up). See http://snurl.com/2mt4 where I answered a question
about PS-ON.

Fair enough - a little out of my depth, but you sound like you know what you
are talking about!

The quickest and surest way to tell would be to look at the manual or
the maker's web site for that motherboard to be sure they say they
support whatever CPU you were going to try. Unless I missed it, you
didn't mention your motherboard brand and model. I don't really need to
know since I'm sure you can browse to their site and hunt around as well
as I.

I mostly build my systems using Intels. Cost more but I prefer more
robust components in addition or in lieu of external protection as is
required for the AMDs. If the heatsink falls off an AMD, it can burn up
and be destroyed in a couple seconds. It will survive a bit longer if
just the fan stops spinning but it will eventually burn up. Newer
motherboards provide protection against overheating but they don't seem
quick enough to prevent thermally stressing the CPU and severely
reducing its longevity and reliability. Older motherboards don't have
this external protection. See Tom's review at
http://snurl.com/amd_fried (note the smoke in a couple seconds, and in
one case the system crashes before the tester can completely remove the
heatsink). The Intels reduce their duty cycle to reduce heat so it
doesn't get to a catastrophic level. The AMDs burn up without something
to try to protect them. I consider the external protection against
catastrophic overheating more to protect the motherboard from frying
than to protect the AMD processor.

The Tyan Trinity KT400 and Asus A7V8X handle Athlon XP, Athlon, and
Duron. They don't mention any special instructions when using a Duron
versus an Athlon. I would think that if it takes an Athlon then it will
take a Duron. They both fit in a Socket A. You didn't mention which
motherboard you have, but then you can visit the maker's web page as
well as I.

On Sat, 11 Oct 2003 01:08:51 +0100, "ck26"
wrote:

Buy replacement parts that will accomodate future upgrades, so there
isn't any significant money wasted, you'll be on your way towards a
faster system still.

Agreed - unfortunately, the power supply does not appear to be the problem.
I took one out from a working machine and I still get the same problem.

A machine of similar power usage, so the odds weigh in favor of it
being adequate for this sytem? In other words, a PSU of adequate
wattage and known quality?

This
is symptomatic therefore that the problem is either mobo or cpu related or
even a combination of both, I can therefore conclude?

If you feel confident you've ruled everything else out...

If so, I'm going to have to buy a new base computer, because this is very
expensive territory now.

Surely you can find some way test your parts free or inexpensively?
A lot of people do end up scrapping the whole system and starting over
with a new one... I have stacks of OEM boxes in my basement as a
result.


Co-incidentally, if I want to test the CPU works in another machine can I
just swap a 1Ghz AMD Athlon for a 1.3Ghz AMD Athlon? And if the CPU is in
fact damaged, could it cause damage to the undamaged machine? Or vice
versa - if the motherboard is fried, could placing a good CPU in it damage
the good CPU?


Usually not, but I've heard of it happening. Your present motherboard
can't run a modern AMD CPU, right? You might buy the motherboard,
memory, and try the CPU... the odds are it won't damage the board, and
if it doesn't work you need a new CPU.... might be considered an
acceptable, reasonable risk.

Dave