Over Heating Problems

[John Lettiere]

I recently upgraded my system, new MOBO (Abit IC7-G) hard drive (WD 200GB),
new video card (Leadtek Winfast 66OO GT 128MB AGP)
and CPU (P4 3.2 GHz Prescott).

My system now looks like this:

OS: Windows XP pro with service pack 1.

P4 3.2GHz Prescott proc, 1.5GB DDR-400 RAM, 3 HDD, all WD drives: 200, 120,
and 40GB, one old Plextor SCSI CD-RW drive,
a Lite-On DVD-RW/CD-RW combo drive, Leadtek Winfast 6600 GT video AGP,
Audigy 2 ZS, and a new Thermaltake 480 watt PSU.

Everything was going fine for about 6 weeks, then as I was playing a game
the internal thermal alarm went off.

When I checked the systems health in the BIOS it showed the procs temp at 70
degrees Celsius.

Prior to this the proc's temp was around 38-40 degrees, well within Intels
operating parameters for the Prescott proc.

The case I have is all aluminum, with 1/80mm intake fan, and 3/80mm exhaust
fans, plus the internal fan on the PSU
blowing down on the CPU.

I got the temp down to 61c by fooling around with the fan speeds, but when I
tried to reboot I got a system boot disk
error and now the BIOS reads the WD 200GB drive as only having 8400MB.

When I tried to repair the MBR with XP repair consol, Windows couldn't fined
the drive at all.

So now I'm at a total loss as to what's going on.

Has anyone else experienced anything similar, and what did you do to fix the
problem?

 

[kony]

I recently upgraded my system, new MOBO (Abit IC7-G) hard drive (WD 200GB),
new video card (Leadtek Winfast 66OO GT 128MB AGP)
and CPU (P4 3.2 GHz Prescott).

My system now looks like this:

OS: Windows XP pro with service pack 1.

P4 3.2GHz Prescott proc, 1.5GB DDR-400 RAM, 3 HDD, all WD drives: 200, 120,
and 40GB, one old Plextor SCSI CD-RW drive,
a Lite-On DVD-RW/CD-RW combo drive, Leadtek Winfast 6600 GT video AGP,
Audigy 2 ZS, and a new Thermaltake 480 watt PSU.

Everything was going fine for about 6 weeks, then as I was playing a game
the internal thermal alarm went off.

When I checked the systems health in the BIOS it showed the procs temp at 70
degrees Celsius.

Prior to this the proc's temp was around 38-40 degrees, well within Intels
operating parameters for the Prescott proc.

Ok, but, it's kinda irrelevant what it's idle temp is, you
might as well ignore that.
You should also ignore the alarm-inducing temp you report,
and run a solid lengthly stress-test in the highest ambient
(room) temp the system will ever see to get a true maximal
reading. I suggest Prime95's Torture Test, running for a
minimum of an hour.

Even so, that's to determine results more than a reason to
hesitate to act, as we can already determine that the result
of the test isn't going to be acceptible without some
improvements.

The case I have is all aluminum, with 1/80mm intake fan, and 3/80mm exhaust
fans, plus the internal fan on the PSU
blowing down on the CPU.

What case? A good (few if possible) selection of interior
shots (linked to, not posted here) would help us see what
the airflow is like.

You write of the one intake, but is this intake grill almost
completely open or blocked by stamped-in-metal grill? What
about the (plastic?) bezel in front, is it restricting the
intake for that front fan?

Rightabout now you may be realizing that all the talk about
hot-P4 CPUs was truth. There's no point in a fast CPU if it
can't run at 100% utilization and when it does that, you
have an extra 80W or so of heat to remove, plus or minus
depending on the stresses involved, ie- video card gets
hotter too.

What about exhaust fan flow? Open area or more stamped
grills impeding flow? Two completely unobstructed exhaust
fans could cool that box, but the PSU exhaust doesn't count,
it can't be considered unimpeded due to the pressure
differential of pulling it though the PSU casing even
relative to what the other case fans face.

A side-mounted fan might be a really good idea, positioned
such that half of the flow is under the video card, and the
other half above, cooling top of card as well as northbridge
& CPU area. That will also help more immediately move the
CPU exhaust heat away from it. Beware though that some
cases, due to placement of the motherboard temp sensor, will
actually seem (report) higher with this arrangement because
less flow goes past the bottom-right corner of the board.

Likewise, this will reduce flow past the hard drive bays, it
becomes more important to be sure you have sufficient front
case intake, passageways not just the fan.

I got the temp down to 61c by fooling around with the fan speeds,

Ok, but was this while gaming again, or would the temp have
dropped anyway because it then wasn't under such a load
anymore?

How how is the room the box is in? Is it somewhere (like
under a desk) where there might be lesser air circulation so
to a certain extent it's recirculating some of the expelled
air from the case? Usually this isn't such a difference to
be concerned about, but at 61C you might want to see if you
can get that temp down just a little more, especially if it
wasn't while recreating the prior alarm-inducing event.

but when I
tried to reboot I got a system boot disk
error and now the BIOS reads the WD 200GB drive as only having 8400MB.

Check the bios to confirm it's set to "auto" for the drive.
Run the HDD manufacturer's utility on it.

When I tried to repair the MBR with XP repair consol, Windows couldn't fined
the drive at all.

No, don't use windows to write to the drive while it's in
this state. Do not write to the drive at all if you have
any data that needs be gotten off of it, if that's possible
at this point (you haven't mentioned it yet, but one step at
a time, better to wait than lose data).

So now I'm at a total loss as to what's going on.

Has anyone else experienced anything similar, and what did you do to fix the
problem?

More things to try:

- Clear CMOS (motherboard jumper or pull battery for a few
minutes while AC power is disconnected

- If it's a PATA drive, jumper it to single drive (not
master since it's a WD) and have it alone on the cable. If
not the boot drive, have only the boot drive on a different
cable.

- Take the drive to another system

- Put it in an external drive enclosure then try reading
it.

- Wonder if the excessive temps have damaged the power
supply and that is what led to the drive failure. Ideally
you'd take voltage readings with a multimeter, (especially
12V readings) then disconnect AC for at least 3 minutes and
open the power supply, visually inspecting it, especially
the output filter caps (which are roughly where the wiring
harness that leaves the casing, connects to the circuit
board).

 

[kony]

plus the internal fan on the PSU
blowing down on the CPU.

Hmm, is the above correct?

Your power supply has the rear fan exhausting out of the
case, AND the bottom fan is also exhausting out of the power
supply casing into the system case (down on the CPU)?

If so, that is a problem. WIth two exhausting fans but no
intake, your PSU has a low pressure, high heat retention.
The bottom fan should be an intake, sucking away from the
PSU. Ideally you would (might consider doing this) put a
shroud, ductwork-like plastic to route a lot of the CPU
heatsink exhaust towards the rear case fan rather than
directly up into the power supply. That is, if reasonably
possible.

That may not help the CPU stay cool but CPU should be cool
enough either way, and it's important to keep ALL parts cool
enough, including those with no temp monitoring. Maybe I'm
barking up the wrong tree here, only you have the box in
front of you and can check these things.

 

[John Lettiere]

First off thanks for all the suggestions Kony, I really appreciate them.

Hmm, is the above correct?

No it's the bottom exhaust fan on the PSU, my mistake, I shouldn't have used
the word blow.

If so, that is a problem. WIth two exhausting fans but no
intake, your PSU has a low pressure, high heat retention.
The bottom fan should be an intake, sucking away from the
PSU.

Funny thing is according to "Motherboard Monitor" the ambient temp inside
the case is 91degrees Fahrenheit,
well within specs for the proc.

Ideally you would (might consider doing this) put a

shroud, ductwork-like plastic to route a lot of the CPU
heatsink exhaust towards the rear case fan rather than
directly up into the power supply. That is, if reasonably
possible.

I'll think about the shroud. In the mean time I just got in a new CPU fan
and heatsink: "Vantec AeroFlow TMD",
and another 80mm fan for the front of the case, which will go in front of
one of the hard drive cages.

I have one 80mm intake fan (Not enough I know) on the inside front of the
case below one of two HDD cages, behind the front bezel, which
is slotted to allow air flow.

I installed a PCI slot exhaust fan pretty close to the proc last week,
after this problem happened, so I now have 4 exhaust fans working.

I had installed some filter material in between the case and the front bezel
to help prevent dust build up inside the case too.

In answer to your previous question as to the type of case I have, it's an
Aspire similar to this one, but without the side fan on the Plexiglas
side panel: http://www.newegg.com/Product/Product.asp?Item=N82E16811144040

That may not help the CPU stay cool but CPU should be cool
enough either way, and it's important to keep ALL parts cool
enough, including those with no temp monitoring. Maybe I'm
barking up the wrong tree here, only you have the box in
front of you and can check these things.

By rights your probably correct.

I'm probably am as much to blame as anything else for ignoring to put some
cooling compound on the proc
and using the Intel stock heatsink and fan when I upgraded.

Anyway, thanks for the help.

I'll let you know how it turns out.

 

[kony]

I'll think about the shroud. In the mean time I just got in a new CPU fan
and heatsink: "Vantec AeroFlow TMD",

That is not a very good heatsink.
Better would have a larger, thicker fan. It would have
larger copper areas and overall fin surfaces greater too.
While it might be sufficient for typical uses, given the
overheating i'd be more than a little hesitant to use it.

and another 80mm fan for the front of the case, which will go in front of
one of the hard drive cages.

I have one 80mm intake fan (Not enough I know) on the inside front of the
case below one of two HDD cages, behind the front bezel, which
is slotted to allow air flow.

I think your front intake is a large part of the problem.
It isn't always necessary to add fans to front nor increase
their fan, but the intake slots, plus the metal wall behind
the bezel can really reduce flow rate. If that area were
more open you might not even need the front fan, though now
that it's there it can't hurt providing it's not added
unacceptible noise.

I installed a PCI slot exhaust fan pretty close to the proc last week,
after this problem happened, so I now have 4 exhaust fans working.

I've never found those very effective except for
spot-cooling. Another problem is that it's reducing flow
rate through the power supply.

I had installed some filter material in between the case and the front bezel
to help prevent dust build up inside the case too.

That may be a large part of the reason it was overheating
too. With those parts and a restricted front intake area
you'd be hard pressed to get enough flow without any further
impedances.

In answer to your previous question as to the type of case I have, it's an
Aspire similar to this one, but without the side fan on the Plexiglas
side panel: http://www.newegg.com/Product/Product.asp?Item=N82E16811144040

Hmm, that's actually a pretty good case, especially in that
they had larger opening for the fan intake and exhaust.
Forget what I wrote about the front intake area so long as
you don't have so many drives stacked in one of the bays
(which would be 3, that it significantly impedes that area,
since the bays pretty much completely cover the intake
holes, IIRC. Then again I may be remembering wrong, I have
one of those and added another bay to the bottom behind the
fan.

You wrote that you only have one intake fan? Which one?
Ideally that case should have two up front, but with only
one it should be in front of the drives, not the bottom
position where the intake isn't flowing through the drives.