Flaky P2B - suggestions requested

[P2B]

One of my venerable P2B/Tualatin systems is misbehaving, and I'm at a
bit of a loss as to what to try next.

It's a P2B-S with a P3-S 1.4 on a Slot-T adapter and 512MB of Micron
PC133. Processor is running stock (1.4Ghz), mainboard overclocked to
133Mhz FSB, stock PCI clock. I was clocking it at 150Mhz FSB initially,
but it's the Cottage PC (no aircon there) and it was locking up on hot
days. It has been 100% stable at 133Mhz for 2+ years, even during the
July heatwave. No recent hardware changes.

Two weeks ago it froze - monitor would not come out of power save mode.
I pressed reset, no POST. Powered off and on, no POST. Cleared CMOS, no
POST.

Since I have little in the way of diagnostic tools or spare parts at the
cottage, and the case was fairly grungy inside, I decided on
preventative maintenance as the next step - removed and cleaned all
components, cleaned all edge connectors with iso alcohol, replaced CPU
and northbridge thermal compound, serviced fans, visual inspection of
mainboard, mounting posts, capacitors, cables etc. No obvious problem
found, system POSTed and ran fine after reassembly.

I thought I'd (inadvertently) fixed it, but 8 days later... same problem :-(

I disconnected everything but CPU, RAM, and video card - no POST.
Cleared CMOS, no POST. Left it overnight (without standby power or CMOS
battery), no POST. OK, this time it's failed hard, I'll drag it home to
my workshop where I can diagnose and fix it properly. Reconnected
everything - still no POST - and carted it down to the boat.

Dang thing works fine at home - been running benchmarks for 2 days
without a glitch. I have reset and powered off multiple times, it POSTs
every time. Checked supply voltages on the oscilloscope, all are well
within spec and free of ripple while the system is under load. I've even
tried mild percussive maintenance, since transporting the system
appeared to 'fix' it, but it won't fail here!

I'm heading back to the cottage tomorrow, and expect the system will
still work when it gets there but will probably fail in the same manner
sometime soon. Frustrating!

Any suggestions?

Thanks,

P2B

 

[David Maynard]

One of my venerable P2B/Tualatin systems is misbehaving, and I'm at a
bit of a loss as to what to try next.

It's a P2B-S with a P3-S 1.4 on a Slot-T adapter and 512MB of Micron
PC133. Processor is running stock (1.4Ghz), mainboard overclocked to
133Mhz FSB, stock PCI clock. I was clocking it at 150Mhz FSB initially,
but it's the Cottage PC (no aircon there) and it was locking up on hot
days. It has been 100% stable at 133Mhz for 2+ years, even during the
July heatwave. No recent hardware changes.

Two weeks ago it froze - monitor would not come out of power save mode.
I pressed reset, no POST. Powered off and on, no POST. Cleared CMOS, no
POST.

Since I have little in the way of diagnostic tools or spare parts at the
cottage, and the case was fairly grungy inside, I decided on
preventative maintenance as the next step - removed and cleaned all
components, cleaned all edge connectors with iso alcohol, replaced CPU
and northbridge thermal compound, serviced fans, visual inspection of
mainboard, mounting posts, capacitors, cables etc. No obvious problem
found, system POSTed and ran fine after reassembly.

I thought I'd (inadvertently) fixed it, but 8 days later... same problem
:-(

I disconnected everything but CPU, RAM, and video card - no POST.
Cleared CMOS, no POST. Left it overnight (without standby power or CMOS
battery), no POST. OK, this time it's failed hard, I'll drag it home to
my workshop where I can diagnose and fix it properly. Reconnected
everything - still no POST - and carted it down to the boat.

Dang thing works fine at home - been running benchmarks for 2 days
without a glitch. I have reset and powered off multiple times, it POSTs
every time. Checked supply voltages on the oscilloscope, all are well
within spec and free of ripple while the system is under load. I've even
tried mild percussive maintenance, since transporting the system
appeared to 'fix' it, but it won't fail here!

I'm heading back to the cottage tomorrow, and expect the system will
still work when it gets there but will probably fail in the same manner
sometime soon. Frustrating!

Any suggestions?

Thanks,

P2B

Being something to check power at the cottage while you're there.

 

[Joe Bleaux]

One of my venerable P2B/Tualatin systems is misbehaving, and I'm at a
bit of a loss as to what to try next.

It's a P2B-S with a P3-S 1.4 on a Slot-T adapter and 512MB of Micron
PC133. Processor is running stock (1.4Ghz), mainboard overclocked to
133Mhz FSB, stock PCI clock. I was clocking it at 150Mhz FSB initially,
but it's the Cottage PC (no aircon there) and it was locking up on hot
days. It has been 100% stable at 133Mhz for 2+ years, even during the
July heatwave. No recent hardware changes.

Two weeks ago it froze - monitor would not come out of power save mode.
I pressed reset, no POST. Powered off and on, no POST. Cleared CMOS, no
POST.

Since I have little in the way of diagnostic tools or spare parts at the
cottage, and the case was fairly grungy inside, I decided on
preventative maintenance as the next step - removed and cleaned all
components, cleaned all edge connectors with iso alcohol, replaced CPU
and northbridge thermal compound, serviced fans, visual inspection of
mainboard, mounting posts, capacitors, cables etc. No obvious problem
found, system POSTed and ran fine after reassembly.

I thought I'd (inadvertently) fixed it, but 8 days later... same problem :-(

I disconnected everything but CPU, RAM, and video card - no POST.
Cleared CMOS, no POST. Left it overnight (without standby power or CMOS
battery), no POST. OK, this time it's failed hard, I'll drag it home to
my workshop where I can diagnose and fix it properly. Reconnected
everything - still no POST - and carted it down to the boat.

Dang thing works fine at home - been running benchmarks for 2 days
without a glitch. I have reset and powered off multiple times, it POSTs
every time. Checked supply voltages on the oscilloscope, all are well
within spec and free of ripple while the system is under load. I've even
tried mild percussive maintenance, since transporting the system
appeared to 'fix' it, but it won't fail here!

I'm heading back to the cottage tomorrow, and expect the system will
still work when it gets there but will probably fail in the same manner
sometime soon. Frustrating!

Any suggestions?

Thanks,

P2B

Hi P2B,

Since you're overclocking the system (running a 100MHz MB at 133MHz),
that would normally be the first thing I'd address. However, since
you're only having this problem at your cottage, I'd agree with David:
you better check the quality of the power there. Maybe something has
changed recently that's affecting the power.


Joe

 

[Peder]

Two weeks ago it froze - monitor would not come out of power save mode.
I pressed reset, no POST. Powered off and on, no POST. Cleared CMOS, no
POST.

Look for swollen and leaking capacitors. Common on many MB from that era.

 

[P2B]

Being something to check power at the cottage while you're there.

Dirty power is certainly a plausible explanation for the system lockups
- although there is minimal protection in the form of a surge protector
for the system, external modem, and phone line - but I would expect the
system to either recover when power is cycled, or at least when CMOS is
cleared, or otherwise remain dead due to component damage. Do you have a
theory to explain the observed behavior?

P2B

 

[P2B]

Look for swollen and leaking capacitors. Common on many MB from that era.

Indeed, but very rare on P2B series boards. This one was new old stock
when it went into service about 3 years ago, and all capacitors still
look good as new.

P2B

 

[David Maynard]

Dirty power is certainly a plausible explanation for the system lockups
- although there is minimal protection in the form of a surge protector
for the system, external modem, and phone line - but I would expect the
system to either recover when power is cycled, or at least when CMOS is
cleared, or otherwise remain dead due to component damage. Do you have a
theory to explain the observed behavior?

P2B

I've seen noisy power, and I don't mean brownouts or surges, play havok
with systems.

Old aluminum wiring is one culprit. Sometimes it can be just a loose wire
in the socket that 'appears' to work correctly with lights and motors (e.g.
vacuum cleaners, etc.). Other times it can be that new EMS or pumping
station down the block throwing trash into the AC. Which can be a bugger to
find because it might not be a regular schedule.

 

[Buffalo]

Indeed, but very rare on P2B series boards. This one was new old stock
when it went into service about 3 years ago, and all capacitors still
look good as new.

P2B

Check the power receptacle to make sure the wires are tight (don't get shocked)
and that it is properly grounded.
Try plugging in a light and jiggle and move the plug to see if it blinks,
indicating a faulty receptacle or wiring.

 

[Hackworth]

One of my venerable P2B/Tualatin systems is misbehaving, and I'm at a bit
of a loss as to what to try next.

It's a P2B-S with a P3-S 1.4 on a Slot-T adapter and 512MB of Micron
PC133. Processor is running stock (1.4Ghz), mainboard overclocked to
133Mhz FSB, stock PCI clock. I was clocking it at 150Mhz FSB initially,
but it's the Cottage PC (no aircon there) and it was locking up on hot
days. It has been 100% stable at 133Mhz for 2+ years, even during the July
heatwave. No recent hardware changes.

Two weeks ago it froze - monitor would not come out of power save mode. I
pressed reset, no POST. Powered off and on, no POST. Cleared CMOS, no
POST.

Since I have little in the way of diagnostic tools or spare parts at the
cottage, and the case was fairly grungy inside, I decided on preventative
maintenance as the next step - removed and cleaned all components, cleaned
all edge connectors with iso alcohol, replaced CPU and northbridge thermal
compound, serviced fans, visual inspection of mainboard, mounting posts,
capacitors, cables etc. No obvious problem found, system POSTed and ran
fine after reassembly.

I thought I'd (inadvertently) fixed it, but 8 days later... same problem
:-(

I disconnected everything but CPU, RAM, and video card - no POST. Cleared
CMOS, no POST. Left it overnight (without standby power or CMOS battery),
no POST. OK, this time it's failed hard, I'll drag it home to my workshop
where I can diagnose and fix it properly. Reconnected everything - still
no POST - and carted it down to the boat.

Dang thing works fine at home - been running benchmarks for 2 days without
a glitch. I have reset and powered off multiple times, it POSTs every
time. Checked supply voltages on the oscilloscope, all are well within
spec and free of ripple while the system is under load. I've even tried
mild percussive maintenance, since transporting the system appeared to
'fix' it, but it won't fail here!

I'm heading back to the cottage tomorrow, and expect the system will still
work when it gets there but will probably fail in the same manner sometime
soon. Frustrating!

Any suggestions?

I have to agree with David. Just within the past few months, I worked a
system that would just shut down shortly after booting. I took the system
with me and put it on the workbench, and I couldn't replicate the problem
even once! It was maddening. The power supply outputs under load seemed ot
be OK. I checked the voltage at the original location and it seemed to be
fine. Naturally, as soon as I'd hook everything up (trying various
receptacles) at the original location, the problem would recur.

The solution? I replaced the old power supply in the computer with a good
one (an Antec) and all was well. It's been running 24/7 since that time with
no problems whatsoever. Whatever weirdness was going on with the power
never became obvious to me but, whatever it was, that old power supply
couldn't deal with it.

 

[P2B]

I have to agree with David. Just within the past few months, I worked a
system that would just shut down shortly after booting. I took the system
with me and put it on the workbench, and I couldn't replicate the problem
even once! It was maddening. The power supply outputs under load seemed ot
be OK. I checked the voltage at the original location and it seemed to be
fine. Naturally, as soon as I'd hook everything up (trying various
receptacles) at the original location, the problem would recur.

The solution? I replaced the old power supply in the computer with a good
one (an Antec) and all was well. It's been running 24/7 since that time with
no problems whatsoever. Whatever weirdness was going on with the power
never became obvious to me but, whatever it was, that old power supply
couldn't deal with it.

I also agreed with David, so swapped the power supply with one which had
been running one of my bench test systems reliably for ages.

However, subsequent testing suggests the problem lies elsewhere - but I
just noticed that part of the discussion was only posted to
alt.comp.hardware.overclocking and does not appear on
alt.comp.periphs.mainboard.asus. I'll quote it below for those that
missed it.

Thanks for your input!