The P4 system problem from hell (long)

I've been building and fixing PC's for nearly 20 years, and this one has
me stumped. Let's see if anyone here can help nudge me toward a solution.

Original config (config0 from here on): Cheapo case, 450 watt Anitec PSU,
ASUS P4P800, 3.0 GHz Pentium 4, 512MB of PC3200 RAM, 2 IDE HDD, 1 IDE SONY
DVD R/RW+- drive, 1 IDE Panasonic CD ROM drive, 1 FDD, USB mouse (700-
series
Logitech wireless), PS2-attached IBM-brand keyboard, old Maxtor dual-head
board.

I built this system right after the P4P800 came out, and it ran perfectly
since
day one, until about a week ago.

I had been using it for a few hours when the system locked up solid--no
mouse cursor movement, nothing. Rebooted, and it ran fine. It did it a
couple more times over the next 2 or 3 days, same symptoms. I opened the
case, reseated the memory and all cables, and the problem disappeared for
about 4 days. Then one morning it wouldn't boot--no BIOS screen, no
speaker
beeps, no signal to the monitor at all. The PSU and CPU fans ran, and the
drives had power, but it refused to boot. Pressing RESET does nothing. I
can tell it's not booting, as there's no HDD noise after the initial power-
on head reset. (Hereafter, I'll call this state "no boot".)

After numerous attempts I did manage to get it to boot into the BIOS set
up screen about 4 different times, but each time it locked up solid after
just a few seconds.

I reseated everything I could find, "no boot".

I pulled the memory completely, "no boot".

I replaced the motherboard (put in a brand new P4P800 SE) and the power
supply (put in a brand new 350W Anitec), "no boot".

I tried a different graphics board, in both config0 and the new MOBO
and PSU, "no boot".

I went back to config0, "no boot".

I replaced the CPU with an identical 3.0 GHz P4, "no boot".

I stripped the entire system down--pulled all disk drives, the USB hub,
everything except MOBO, CPU, memory, keyboard, mouse, and graphics board.
I tried booting about 50 times, and got "no boot" in all but about 3 or 4
trials, and in those it locked up after about 10 to 30 seconds.

During this whole time I used an old Athlon system with the same monitor
I've been using on the now-broken system since day one. Tonight, after
trying the stripped-down step mentioned above, I gave up, plugged in the
Athlon, and it exhibited the "no boot" symptom(!!!) for about 5 boot tries,
then started working, and I'm using it now.

As I see it:

It's not the motherboard, the CPU, the BIOS, the power supply, the disk
drives, the memory, or the graphics board. These were all replaced with
known-good units (or simply removed) without any change in the symptoms.

I thought it might be some kind of weird grounding issue, but the problem
is so unpredictable that I don't see a pattern that would suggest that's
it.

The Athlon system not working is almost wierd beyond belief, and the only
way I could see it being connected to the failure of config0 (and all the
ensuing test configs) is if there were some sort of low-voltage situation.
My wife and I did move into a new house in 2004, about 6 months ago, but
the house is only three years old and I run all systems on a combination
UPS and surge protector. I even tried running config0 without the UPS,
with no success.

Aside from bad Bermuda Triangle and haunted house jokes, I would REALLY
appreciate any help. I'm about one step away from burying this system
in my back yard during the next full moon.



Thanks in advance,
Lou

Lou Grinzo wrote in message ...
I've been building and fixing PC's for nearly 20 years, and this one has
me stumped. Let's see if anyone here can help nudge me toward a solution.

Could it be something as simple as a bad (intermittently open) power
cord, and you inadvertently used the same cord for the Athlon ? Others
have already suggested replacing RAM, checking for a dirty power
supply line and trying to boot with only the UPS. I think a lot of us
would like to know the outcome of those things and how you finally
solve the problem.

If you do bury it under the full moon, let me know and I'll come over
to do a few chants.

Hi Lou,
Did u try another casing?
Maybe there is some kind of current leakking through the casing?
Regards and good luck
Boldy
"Zotin Khuma" wrote in message
om...
Lou Grinzo wrote in message
...
I've been building and fixing PC's for nearly 20 years, and this one has
me stumped. Let's see if anyone here can help nudge me toward a
solution.

Could it be something as simple as a bad (intermittently open) power
cord, and you inadvertently used the same cord for the Athlon ? Others
have already suggested replacing RAM, checking for a dirty power
supply line and trying to boot with only the UPS. I think a lot of us
would like to know the outcome of those things and how you finally
solve the problem.

If you do bury it under the full moon, let me know and I'll come over
to do a few chants.

"boldy" ¦b¶l¥ó
l.nl ¤¤¼¶¼g...
Hi Lou,
Did u try another casing?
Maybe there is some kind of current leakking through the casing?
Regards and good luck
Boldy

Or try not using a case at all!

The "cheapo" case seems to be the counterpart. Check all screws and any
connections the motherboard make with the case.


"boldy" wrote in message
l.nl...
Hi Lou,
Did u try another casing?
Maybe there is some kind of current leakking through the casing?
Regards and good luck
Boldy
"Zotin Khuma" wrote in message
om...
Lou Grinzo wrote in message
...
I've been building and fixing PC's for nearly 20 years, and this one has
me stumped. Let's see if anyone here can help nudge me toward a
solution.

Could it be something as simple as a bad (intermittently open) power
cord, and you inadvertently used the same cord for the Athlon ? Others
have already suggested replacing RAM, checking for a dirty power
supply line and trying to boot with only the UPS. I think a lot of us
would like to know the outcome of those things and how you finally
solve the problem.

If you do bury it under the full moon, let me know and I'll come over
to do a few chants.

"Lou Grinzo" wrote in message
...
snip

NOTE: The original groups were relisted in the Newsgroups header of this
post to counteract the FollowUp-To header employed by the original
poster which directs replies only to 2 of the 4 original newsgroups.
The FollowUp-To header was also used in this post so replies go to the
alt.comp.hardware group. The OP chose to move the discussion into HIS
"home" groups. Likewise, I chose to continue the discussion in MY
"home" group. Tit for tat. ;-

Off-topic rant:

I don't know how you're doing on your problem since you decided to hide
your thread in a group that I do not visit. I don't visit the
*.*homebuilt groups; however, I do visit the alt.comp.hardware group
which you choose to include in your cross-posted message. You used the
FollowUp-To header so all replies go to your "home" group(s). That
means no one visiting the other groups you included will see their reply
in the thread or any replies from you or anyone else. That's the
pitfall of using the FollowUp-To header: you disconnect the posters
visiting THEIR "home" group from the discussion you moved into YOUR
"home" group(s). I am not subscribing to someone else's home group to
continue the discussion when they chose to cross-post into MY home
group.

If you use the FollowUp-To header, and to be polite, be sure to announce
that fact at the start of your message. Otherwise, the subterfuge and
covert use of FollowUp-To is RUDE to your respondents and is a trick
employed by malcontents and spammers who want to redirect and hide
[negative] replies off to some other group. Unless you want to join
their ranks, announce your use of the FollowUp-To header so respondents
are alerted that you are attempting to move the discussion to a group
that the respondent might not visit. Some NNTP clients obey the
FollowUp-To header; however, of those that do, they may not raise a flag
to the user noting the use of the FollowUp-To header.

On-topic discussion:

Hope you got your system working. It sounded like a real toughie to
isolate since you had noted changing almost all the components.

- Did assembling the minimal set of components outside the case get it
working? If it did then that might indicate the need for using
insulating fiber washers on the standoffs and under their screws heads.

- I couldn't tell if your 512MB of memory was one stick or two. If it
was two, have you tried switching between the sticks where you use only
one at a time? How about changing the slot in which the memory is
inserted?

- I wasn't sure if you tried two different PSUs. You mentioned using
the new one but I wasn't sure if the system worked using the old one
(provided it could supply the under-load current demand without dropping
the voltages outside of specs). Did you check the voltages under load?

- If you have an ATX mobo with the extra 4-wire connector for power then
you need to have a PSU that also provides that extra 4-wire power
connector. Otherwise, the problem could be your CPU isn't getting
enough juice and/or the voltage regulators won't be able to maintain the
proper voltages.

Presumably all of these problems with so many hardware parts getting
swapped in and out was occuring BEFORE or without any operating system
loaded so the problems were isolated from a corrupted OS, bad drivers,
conflicts in software, malware, etc.

In message , _Vanguard_
writes
"Lou Grinzo" wrote in message
...
snip

NOTE: The original groups were relisted in the Newsgroups header of this
post to counteract the FollowUp-To header employed by the original
poster which directs replies only to 2 of the 4 original newsgroups.
The FollowUp-To header was also used in this post so replies go to the
alt.comp.hardware group. The OP chose to move the discussion into HIS
"home" groups. Likewise, I chose to continue the discussion in MY
"home" group. Tit for tat. ;-

Off-topic rant:

I don't know how you're doing on your problem since you decided to hide
your thread in a group that I do not visit. I don't visit the
*.*homebuilt groups; however, I do visit the alt.comp.hardware group
which you choose to include in your cross-posted message. You used the
FollowUp-To header so all replies go to your "home" group(s). That
means no one visiting the other groups you included will see their reply
in the thread or any replies from you or anyone else. That's the
pitfall of using the FollowUp-To header: you disconnect the posters
visiting THEIR "home" group from the discussion you moved into YOUR
"home" group(s). I am not subscribing to someone else's home group to
continue the discussion when they chose to cross-post into MY home
group.

If you use the FollowUp-To header, and to be polite, be sure to announce
that fact at the start of your message. Otherwise, the subterfuge and
covert use of FollowUp-To is RUDE to your respondents and is a trick
employed by malcontents and spammers who want to redirect and hide
[negative] replies off to some other group. Unless you want to join
their ranks, announce your use of the FollowUp-To header so respondents
are alerted that you are attempting to move the discussion to a group
that the respondent might not visit. Some NNTP clients obey the
FollowUp-To header; however, of those that do, they may not raise a flag
to the user noting the use of the FollowUp-To header.

On-topic discussion:

Hope you got your system working. It sounded like a real toughie to
isolate since you had noted changing almost all the components.

- Did assembling the minimal set of components outside the case get it
working? If it did then that might indicate the need for using
insulating fiber washers on the standoffs and under their screws heads.

- I couldn't tell if your 512MB of memory was one stick or two. If it
was two, have you tried switching between the sticks where you use only
one at a time? How about changing the slot in which the memory is
inserted?

- I wasn't sure if you tried two different PSUs. You mentioned using
the new one but I wasn't sure if the system worked using the old one
(provided it could supply the under-load current demand without dropping
the voltages outside of specs). Did you check the voltages under load?

- If you have an ATX mobo with the extra 4-wire connector for power then
you need to have a PSU that also provides that extra 4-wire power
connector. Otherwise, the problem could be your CPU isn't getting
enough juice and/or the voltage regulators won't be able to maintain the
proper voltages.

Presumably all of these problems with so many hardware parts getting
swapped in and out was occuring BEFORE or without any operating system
loaded so the problems were isolated from a corrupted OS, bad drivers,
conflicts in software, malware, etc.


I once had a similar problem caused by an intermittent power switch
mechanism that should have been make & break but sometimes stuck on
make.

Hope you find the problem.
Dave

--
This is a newsgroup post , return email not accepted