CPU Overheat Damage?

  • Thread starter Thread starter Phil
  • Start date Start date
P

Phil

I have an Athlon XP 2600 Which has been overheating for about a month and
causing system hangs,unfortunately I thought the problems were software
related and spent a lload of time clearing rogue software while the CPU was
running hot! After being advised to chek the temp, (63C) I purchased a new
heatsink. The hottest I have seen it now is 48C but anything over 45C can
cause freezes or stalls. My question is have I damaged the CPU by constantly
running it HOT? and therefore would a new processor fix the problem?

TIA,
Phil
 
I don't think 63C is that hot, and if your system still crashes, it could be
something else going on. If you notice it crashing when it is hot then maybe
check out other things like your video card getting hot.

-Max
 
Thanks Max,
I hoped someone would point at the video card, I have a GEForce 4 MX440, I
reinstalled the newest driver driver but nothing changed! Its juzt the
system stalls if the CPU temp is 45+ and I'm running pretty graphics.
Is there a temp probe on the card or will it need checking manually?

Phil
 
I have an Athlon XP 2600 Which has been overheating for about a month and
causing system hangs,unfortunately I thought the problems were software
related and spent a lload of time clearing rogue software while the CPU was
running hot! After being advised to chek the temp, (63C) I purchased a new
heatsink. The hottest I have seen it now is 48C but anything over 45C can
cause freezes or stalls. My question is have I damaged the CPU by constantly
running it HOT? and therefore would a new processor fix the problem?

TIA,
Phil
Click to expand...

Generally a CPU will not even be instable (at it's _stock_
speed), let alone be damaged from constantly running at 63C.
IF these temps you report are accurate (I suppose we could
assume it without any reason to doubt?) then I'd ignore the
CPU completely, and look at other temps, voltages. See what
specific scenario can reproduce this and whether (windows?)
has any errors reported.

Overheating can occur when one uses a "fair", adequate
heatisnk but poor case airflow. If that is what was
happening you may have other parts now heat-stressed and
reduced in stability. You failed to mention those other
parts though, some things are common culprits like generic
power supplies.

HOWEVER, if the CPU had been overheating to the point of
instability, IF that is what did really happen (and it was
the only problem for the time) you may quite easily have
file corruption of any, many things written to the drive
during this period of time. Doing things like defragging
the drive could've made so many files corrupt that it's
pointless to even try to repair it rather than start out
fresh. If you have the means I suggest you try making a
separate, temporary clean OS install to see if the problem
persists in that environment. You might even consider
underclocking the CPU before, during, and/or after this test
to see what effect that has.
 
kony said:
Generally a CPU will not even be instable (at it's _stock_
speed), let alone be damaged from constantly running at 63C.
IF these temps you report are accurate (I suppose we could
assume it without any reason to doubt?) then I'd ignore the
CPU completely, and look at other temps, voltages. See what
specific scenario can reproduce this and whether (windows?)
has any errors reported.

Overheating can occur when one uses a "fair", adequate
heatisnk but poor case airflow. If that is what was
happening you may have other parts now heat-stressed and
reduced in stability. You failed to mention those other
parts though, some things are common culprits like generic
power supplies.

HOWEVER, if the CPU had been overheating to the point of
instability, IF that is what did really happen (and it was
the only problem for the time) you may quite easily have
file corruption of any, many things written to the drive
during this period of time. Doing things like defragging
the drive could've made so many files corrupt that it's
pointless to even try to repair it rather than start out
fresh. If you have the means I suggest you try making a
separate, temporary clean OS install to see if the problem
persists in that environment. You might even consider
underclocking the CPU before, during, and/or after this test
to see what effect that has.
Click to expand...
Cheers,
I'll try and fill in a few blanks for you but I am already suspecting the
video card, I guess I just needed reassuring on the CPU.
The temps are reported by the BIOS and MBM5 and are the same so I assume
they're correc
I'm not really sure about the voltages, these are the readings I'm getting:
Core 0 - 1.67V +3.3 - 4.07V +12.00 - 12.97V The others cant have probes!
Windows reported a corrupt video driver after one of the freezes but let me
reinstall. I do have 3 CD/DVD drives installed but only a video card there
is plenty of airflow but I've just noticed the ribbon cables are touchig the
card. I'll change the card tomorrow.,
Thanks,
Phil
 
Kony has provided an abridged list of suspects. Your
analysis should test for and eliminate each from that list.

A spec for an Athlon 2600 says it works 100% perfectly even
at 90 degrees C. Where did you even approach that
temperature? Therefore no damage.

Load up comprehensive hardware diagnostics. Test
(diagnostics run without Windows - which is necessary to test
only hardware) everything in the system. Then repeat that
same test selectively heating everything as hot as hairdryer
can make it. Heat locates defective hardware. Hairdryer will
never overheat any semiconductors (because it does not burn
human skin). Any semiconductor uncomfortable to touch is
having a tropical vacation - and loving it. But intermittent
semiconductors will become obviously weak under minimal heat
so that a diagnostic should find the defect. Use heat to
determine which 'warm' component then starts computer
failure. And get those hardware diagnostics that are
available for free.

Heat is a diagnostic tool. Others instead try to fix a
failure by adding more fans and heatsinks. Foolishness. We
call it curing symptoms. That Athlon (and you are expected to
confirm the spec number) should work and very happy even up to
90 degree C. In the meantime, Kony has listed other
suspects. Recommend you first confirm those other suspects
are innocent. Speculation is not sufficient. You must have a
specific criteria to verify integrity of each hardware
suspect.
 
The essential voltages are 3.3, 5, and 12 volts. The
motherboard is not sufficient to measure those voltages. A
known standard such as the multimeter is necessary to measure
voltages and to calibrate the voltage measurement hardware on
motherboard. These voltages are on red, orange, and yellow
wires from power supply.

Get the diagnostic from that video card manufacturer. If
the video car is having problems, the manufacturer's
diagnostic should report that problem. If the video card is
intermittent, then heating video card with a hair dryer on
high will help the diagnostic to find that defect.
 
Cheers,
I'll try and fill in a few blanks for you but I am already suspecting the
video card, I guess I just needed reassuring on the CPU.
The temps are reported by the BIOS and MBM5 and are the same so I assume
they're correc
I'm not really sure about the voltages, these are the readings I'm getting:
Core 0 - 1.67V +3.3 - 4.07V +12.00 - 12.97V The others cant have probes!
Windows reported a corrupt video driver after one of the freezes but let me
reinstall. I do have 3 CD/DVD drives installed but only a video card there
is plenty of airflow but I've just noticed the ribbon cables are touchig the
card. I'll change the card tomorrow.,
Thanks,
Phil
Click to expand...


You need a multimeter to take accurate temp readings but
your initial report of the "4.07V" and "12.97V" is typical
of a system with an insufficient or failing power supply.
You should not run the system unless absolutely necessary
until this is resolved.

You mentioned the video card- a failing video card could
potentially draw more current and thus be additional load on
the power supply, but usually they don't cause this kind of
problem when failing. If you had a spare video card (ANY
card, even a very old PCI card) you could swap in for a
test, that might be revealing. Note that a different card
that uses significantly less power might seem to alleviate
the problem merely because the power supply is then under
less load, but if the power supply didn't return to MUCH
better voltages it would be the prime suspect, not the video
card.
 
Is your video card really overheating? Like another poster said you have to
scientifically eliminate all the possible causes. Easiest thing is to swap
out parts. I have quite a collection of old parts I just keep so I either
have a backup or can troubleshoot my computer.

But my best guess is that it's the video card and the motherboard chipset
not liking each other. I doubt it's even heat. I was able to resolve
crashing a lot of times by "turning down" some of the AGP settings in my
BOIS. Look for your motherboard model number and do a search for any known
issues. Flashing your BIOS could even make the problem worse with a
particular piece of hardware, especially older hardware.

If nothing you try seems to fix, the easiest thing is to first buy a newer
model of video card, then think about replacing mobo, CPU, RAM to newer
models. Maybe go with a pure Intel mobo. PSU's are cheap to replace as well.
Don't buy the baddest and fastest, just find like last year's model that
will be half the price now. Of course other PCI cards could cause a crash.

-Max
 
OK, I'm now ruling out heat! The system comprises an ASRock mobo K78SX, AMD
Athlon XP 2600 @ 333Mhz, GEForce 4 MX440 with Generic modem and Network
cards I am not using.
This has all worked happily togeher for two years!
The voltages confuse me what should 3.3, 5 and 12 read? 3.3, 5 and 12 I
assume.also as I have a ATI radeon 9000 to put in, what watt PSU do I need?
it's powering 1 floppy, 1 DVD reader, 1 CD Burner,
1 DVD burner and a Maxtor 80Gb 7200 rpm HD
 
Chart for ATX type PSU voltages. Note that 'actually
measured' voltages should be in the upper 3/4 of limits.
Voltages that lie between Min V and Min on meter are
suspect and probably a symptom of other problems such as
excessive ripple voltage.

Voltage Wire Color Min V Min on meter Max V
+5 V Red 4.75 V 4.87 5.25 V
-5 V White -4.75 V 4.87 -5.25 V
+12 V Yellow 11.4 V 11.7 12.6 V
-12 V Blue -11.4 V -11.7 -12.6 V
+3.3 V Orange 3.135 V 3.22 3.465 V
+5VSB Purple 4.75 4.87 5.25
!Power On Green 0.8 2.0
Power OK Gray >2.4 when power is good

If that 4 volt was suppose to be 5 (red wire), then you have
a serious power supply problem. System should not be working.
 
Thanks, the chart is just what I was looking for.
The 4V is actually the +3.3
I take it still a problem?
What Watt PSU should I get?
 
Thanks, the chart is just what I was looking for.
The 4V is actually the +3.3
I take it still a problem?
What Watt PSU should I get?
Click to expand...

Your voltage readings are so far out of spec that they're
either wrong or it's unusal that the system would be running
as well as it had. In other words, the voltage data you
have already is suspect and needs confirmed via multimeter
readings. FWIW, I doubt that your 3.3V rail is actually
outputting 4V.

What make & model is your power supply?
What's it's combined 3V+5V wattage rating?

Your prior post of the motherboard model appears to be a
typo? Your board is a K7S8X?
http://www.asrock.com/product/product_k7s8x.htm
I am confirming it because the newer generations of socket A
board sometimes use 12V for CPU and other times 5V for CPU
(derived power). The K7S8X does use 5V if picture I saw is
correct, which would tend to suggest you might see a voltage
increase on the 5V rail if you were to underclock the system
(temporarily) by merely changing the FSB to 100MHz/DDR200.

The above suggestion being made to lower 5V amperage, will
raise the 5V actual voltage on the board and would be
reflected in the BIOS and software hardware/health
monitoring voltage displays... I can only presume you don't
have access to a multimeter or feel comfortable using one
since that is the necessary method to get *accurate* voltage
readings, but at least with some further test scenarios you
might see through deduction what any particular change is
doing (relatively if not accurately) through the bios and
software.
 
kony said:
Your voltage readings are so far out of spec that they're
either wrong or it's unusal that the system would be running
as well as it had. In other words, the voltage data you
have already is suspect and needs confirmed via multimeter
readings. FWIW, I doubt that your 3.3V rail is actually
outputting 4V.

What make & model is your power supply?
What's it's combined 3V+5V wattage rating?

Your prior post of the motherboard model appears to be a
typo? Your board is a K7S8X?
http://www.asrock.com/product/product_k7s8x.htm
I am confirming it because the newer generations of socket A
board sometimes use 12V for CPU and other times 5V for CPU
(derived power). The K7S8X does use 5V if picture I saw is
correct, which would tend to suggest you might see a voltage
increase on the 5V rail if you were to underclock the system
(temporarily) by merely changing the FSB to 100MHz/DDR200.

The above suggestion being made to lower 5V amperage, will
raise the 5V actual voltage on the board and would be
reflected in the BIOS and software hardware/health
monitoring voltage displays... I can only presume you don't
have access to a multimeter or feel comfortable using one
since that is the necessary method to get *accurate* voltage
readings, but at least with some further test scenarios you
might see through deduction what any particular change is
doing (relatively if not accurately) through the bios and
software.
Click to expand...
Hi again,
I'm in a wheelchair and have to ask someone to come and open the box for me
hence I try and find out as much as I can whilst awaiting my helper. I do
have a multimeter but am unsure where to place the probes on a PSU cable,
obviously one on the end terminal but the other?

I tried underclocking the board (K7S8X) but MBM5 and the BIOS vol;tage
readings (whichare consistant with one another) are unchanged.
Phil
 
Confirming what Kony has posted. If that 4 was your 3.3
volts, then the computer suffered overvoltage. 3.3 volt
components are either damaged or overstressed (may fail in the
future).

I posted many points. All must be considered important.
One point was that the motherboard monitor is only a monitor.
Not valid for voltage readings and must be calibrated with a
3.5 digit multimeter. Either do them all or doing nothing.
Do not selectively choose what you will and will not do; to
solve this problem. Most often, things that appear to have
absolutely no purpose (from your perspective) is a 'smoking
gun' solution. If you don't have a 3.5 digit multimeter, then
stop doing everything else and get that meter. Meter is that
critical both for your education (and your education is the #1
reason to fix things) and so that others here do not waste
time helping you.

It was a most important point that should have had your
entire attention. Motherboard hardware cannot accurately
measure voltage. The monitor first must be calibrated. No
way around that fact of computer repair. You need a meter.
Later one appreciates why the meter (and not shotgunning) is
how intelligent computer repair begins (and completes
quicker).

Sidebar - notice how many will immediately jump to blame
heat. That is, unfortunately, how technically naive so many
computer assemblers are. Heat is too often blamed - and the
solution is too often what was the joke in Tim Allen's Home
Improvement. "More Fans". Somehow too many computer
assemblers never appreciate they are promoting the joke by
hyping heat as a problem and then recommending more than one
chassis fan. The point being made here - you had no idea what
was right or wrong with the computer until numbers were first
obtained.
 
Everything was takien on board. thank you.
I have done nothing but gather what info I can so far, so I know what I'm
looking for when my helper arives.
The reason I asked about a suitable wattage of PSU
was purely to determine whether I could fit the spare one I have here.
Phil
 
I had a similar problem with an athlon 2600xp chip.
I changed motherboard, graphics, sound, memory and psu all to no avail.
I was messing in the bios one day and put the voltage up to 1.725 and bingo
no more problems.
It was actually a post I'd seen on an AMD website that prompted me to try
it.
Mikey
 
These aforementioned questions may be significant. It isn't
just the wattage your system may need that matters but
whether the PSU manufacturer or reseller rated it
accurately. Most generic "300W" PSU can't output 300W
stabily, but something like a Sparkle 300W, can.

For your system a decent 300W PSU should suffice, ideally
one with at least 200W combined 3V+5V rating due to the
motherboard's use of the 5V rail for derived CPU power
circuits. Even better to go with a name-brand 350-420W PSU
as it allows free expansion or the next large scale upgrade,
if you would ever be inclined to replace just individual
comonents like video card or motherboard/CPU/memory.

I'm in a wheelchair and have to ask someone to come and open the box for me
hence I try and find out as much as I can whilst awaiting my helper. I do
have a multimeter but am unsure where to place the probes on a PSU cable,
obviously one on the end terminal but the other?
Click to expand...

Insert probes into the back of the ATX connector while it's
still plugged into the motherboard. Primarily the 3.3V
(usually orange), 5V (usually red) and 12V (usually yellow)
readings are important in your situation and putting
postitive multimeter probe on that color and negative on any
black ground on the ATX connector will give the reading. If
it is difficult for you to get both multimeter probes on the
wires simultaneously, you might try clamping (somehow firmly
affixed to) the ground multimeter lead to the system chassis
on a metal (not painted, anodized or clear-coated section of
metal) portion, as whole system case should be, almost
certainly is grounded.

Voltages must be taken while system is turned on of course,
I suggest putting it in the bios health/hardware monitor
screen to directly compare the meter reading to the bios
reading in realtime.

I tried underclocking the board (K7S8X) but MBM5 and the BIOS vol;tage
readings (whichare consistant with one another) are unchanged.
Phil
Click to expand...

Curious, it might be a completely wrong bios report. IMO,
the low end motherboard more often have little
quirks/bugs/flaws/etc with their bios. A bios update
"might" help? Be sure system is completely stable
beforehand if you were to try to update bios, else don't
chance it.
 
Back
Top