CPU On NF7-S v2 Started Overheating But Fan Speed OK ?!

Here's a funny thing ... over the last 2 days my NF7-S-v2-based homebrew
system CPU has begun overheating, after 2 years of trouble-free
operation. The weird thing is that the CPU fan is running fine as
reported by the BIOS "PC Health Status" screen.

"MBProbe" always reports the CPU fan speed as zero RPM (the fan is
rotating, so I've always assumed MBProbe can't detect the fan for some
silly reason), and it agrees with the BIOS about CPU & system temperatures.

System spec: Athlon XP 2600 (Barton), Akasa AK827 heatsink & fan (claims
to be good up to Athlon XP 3400), 1Gb PC3200 RAM, FX5200 graphics
(passive cooling), 40Gb IDE HD, 2 x 160Gb SATA HD, IDE DVD-ROM, IDE
DVD-RW, no overclocking, aluminium case, "Hiper" 350W power supply.
Normally runs so cool I don't even need the two case fans.

I've made *no* recent hardware changes to the system.

Symptoms: Five times in 2 days the CPU temperature has risen slowly (by
about 1 degree every 3 or 4 minutes) from about 38C at system startup
until it reached the warning temperature (60C) I have set in the BIOS,
at which point the two-tone alarm sounded (and I shut the system down).
I tried removing the case cover, which slowed the temperature rise
down, but it still reached the warning temp eventually. Normally the CPU
temp never exceeds ~48C. The CPU fan speed always has been, and still
is, around 2600RPM.

I haven't been working the beast any more than usual - mostly just email
& web.

ODDLY: today the CPU temp is hovering at around 53C - still higher than
normal, but not getting any worse.

UK weather yesterday and the day before that was torrential rain; today
there's no rain, but no sun either. Nothing unusual.

The north-bridge fan has been making a nasty noise for a few weeks, but
I don't suppose that's relevant. [I know it's the north-bridge fan cos
I tried disconnecting it, and the noise went away]

I updated the system BIOS revision from 2.4 (as purchased) to 2.7 about
3 months ago in a futile effort to resolve SATA problems when I added a
second SATA drive (but that's another story).

Given that the CPU fan seems to be okay, I can only think that the
thermal paste between CPU and heatsink has somehow failed or degraded -
but *why* ? (and Googling does not reveal anyone else reporting such a
thing).

Can CPUs just fail all by themselves, in such a way they start
generating more heat ?

Or does anyone know of any other possible cause ?

Cheers
Nick Boyce

I would remove the Akasa...clean it out maybe dust inbetween the fins.
I would remove the old thermal compound from both the CPU and the
heatsink...apply new thermal compound and re-seat.
Then check temps............
I also would re oil that NorthBridge fan......3in1 machine oil works
nicely..but use sparingly and wipe off excess.
peter
"Nick Boyce" <> wrote in message
news:f53u6q$9bh$1$...
> Here's a funny thing ... over the last 2 days my NF7-S-v2-based homebrew
> system CPU has begun overheating, after 2 years of trouble-free
> operation. The weird thing is that the CPU fan is running fine as
> reported by the BIOS "PC Health Status" screen.
>
> "MBProbe" always reports the CPU fan speed as zero RPM (the fan is
> rotating, so I've always assumed MBProbe can't detect the fan for some
> silly reason), and it agrees with the BIOS about CPU & system
> temperatures.
>
> System spec: Athlon XP 2600 (Barton), Akasa AK827 heatsink & fan (claims
> to be good up to Athlon XP 3400), 1Gb PC3200 RAM, FX5200 graphics
> (passive cooling), 40Gb IDE HD, 2 x 160Gb SATA HD, IDE DVD-ROM, IDE
> DVD-RW, no overclocking, aluminium case, "Hiper" 350W power supply.
> Normally runs so cool I don't even need the two case fans.
>
> I've made *no* recent hardware changes to the system.
>
> Symptoms: Five times in 2 days the CPU temperature has risen slowly (by
> about 1 degree every 3 or 4 minutes) from about 38C at system startup
> until it reached the warning temperature (60C) I have set in the BIOS,
> at which point the two-tone alarm sounded (and I shut the system down).
> I tried removing the case cover, which slowed the temperature rise
> down, but it still reached the warning temp eventually. Normally the CPU
> temp never exceeds ~48C. The CPU fan speed always has been, and still
> is, around 2600RPM.
>
> I haven't been working the beast any more than usual - mostly just email
> & web.
>
> ODDLY: today the CPU temp is hovering at around 53C - still higher than
> normal, but not getting any worse.
>
> UK weather yesterday and the day before that was torrential rain; today
> there's no rain, but no sun either. Nothing unusual.
>
> The north-bridge fan has been making a nasty noise for a few weeks, but
> I don't suppose that's relevant. [I know it's the north-bridge fan cos
> I tried disconnecting it, and the noise went away]
>
> I updated the system BIOS revision from 2.4 (as purchased) to 2.7 about
> 3 months ago in a futile effort to resolve SATA problems when I added a
> second SATA drive (but that's another story).
>
> Given that the CPU fan seems to be okay, I can only think that the
> thermal paste between CPU and heatsink has somehow failed or degraded -
> but *why* ? (and Googling does not reveal anyone else reporting such a
> thing).
>
> Can CPUs just fail all by themselves, in such a way they start
> generating more heat ?
>
> Or does anyone know of any other possible cause ?
>
> Cheers
> Nick Boyce

In article <f53u6q$9bh$1$>,
says...
<snip>

> Symptoms: Five times in 2 days the CPU temperature has risen slowly (by
> about 1 degree every 3 or 4 minutes) from about 38C at system startup
> until it reached the warning temperature (60C) I have set in the BIOS,
> at which point the two-tone alarm sounded (and I shut the system down).
> I tried removing the case cover, which slowed the temperature rise
> down, but it still reached the warning temp eventually. Normally the CPU
> temp never exceeds ~48C. The CPU fan speed always has been, and still
> is, around 2600RPM.

Checked for processes running that you don't know about?
Like maybe somebody has botted your pc and is rumming spam
out of it?

<snip>
>
> Given that the CPU fan seems to be okay, I can only think that the
> thermal paste between CPU and heatsink has somehow failed or degraded -
> but *why* ? (and Googling does not reveal anyone else reporting such a
> thing).

It's possible. Some types of paste dry out and lose their
conductivity.

It wouldn't hurt to clean and reapply some thermal paste.


>
> Can CPUs just fail all by themselves, in such a way they start
> generating more heat ?

Never heard of it. Anything is possible though.

>
> Or does anyone know of any other possible cause ?

Yes, have you cleaned out the crud that builds up between
the heatsink fins over time? It would be a good time to do
it when you redo the heatsink paste.

>
> Cheers
> Nick Boyce
>


Bill
--
Gmail and Google Groups. This century's answer to AOL and
WebTV.

peter wrote:

> I would remove the Akasa...clean it out maybe dust inbetween the fins.

Thanks - I'll do that, though I had a good look at it when checking the
northbridge fan, and couldn't see much crud in the fins ... this case (a
Lian-Li) has nice filters on the air intakes.

I'll try to do it without removing the heat sink though - it was the
worst nightmare imaginable fitting it in the first place ... the
amount of pressure I had to apply to the clamp spring thing left me in
terror I was going to slip and damage the mobo.

> I would remove the old thermal compound from both the CPU and the
> heatsink...apply new thermal compound and re-seat.

If absolutely necessary (i.e. problem keeps happening) I'll do that too.

> I also would re oil that NorthBridge fan......3in1 machine oil works
> nicely..but use sparingly and wipe off excess.

And indeed that's good advice too - I seem to recall someone posting the
procedure for that here a while back (peeling the label off the spindle
cover first IIRC).

Thanks for the response.
[CPU temp now up to 55C ...]

Nick Boyce

Bill wrote:

> Checked for processes running that you don't know about?
> Like maybe somebody has botted your pc and is rumming spam
> out of it?

Well I'll check, but it's not likely ... I'm pretty careful, and well
defended :-) The overheating occurred more than once while I wasn't
actually connected to the Net (I'm on dial-up) - though maybe it's
poorly coded malware

> It wouldn't hurt to clean and reapply some thermal paste.
> ... [and] have you cleaned out the crud that builds up between
> the heatsink fins over time? It would be a good time to do
> it when you redo the heatsink paste.

Agreed - see previous response.

Thanks
Nick Boyce

"Nick Boyce" <> wrote in message
news:f54odm$1c5$1$...
> Bill wrote:
>
>> Checked for processes running that you don't know about?
>> Like maybe somebody has botted your pc and is rumming spam
>> out of it?
>
> Well I'll check, but it's not likely ... I'm pretty careful, and well
> defended :-) The overheating occurred more than once while I wasn't
> actually connected to the Net (I'm on dial-up) - though maybe it's
> poorly coded malware

Good grief! Malware might do a lot of things, but cause your CPU to overheat
is not one of them.

>
>> It wouldn't hurt to clean and reapply some thermal paste.
>> ... [and] have you cleaned out the crud that builds up between
>> the heatsink fins over time? It would be a good time to do
>> it when you redo the heatsink paste.
>
> Agreed - see previous response.

Sounds like the most likely cause.
You can set your CPU alarm threshold higher. XP processors are rated at
something like 80 C, even if it overheats and freezes once or twice it's not
the end of the world, and you can adjust the alarm accordingly.

If you don't turn of the PC does the temp plateau?
I guess it depends on you definition of "overheat". If a PC runs at 70 C
happily, and never locks up, is it overheating? ;-)

I would have though your's was "running hotter than usual", (unless it does
lock up). Don't get me wrong, any abnormal behaviour requires investigation
also.

To my mind, overheating is when the comp no longer functions due to heat
build up.

just a thought. ;-)