Sign in with
Sign up | Sign in
Your question

CPU Overheat out of box?

Last response: in CPUs
Share
June 30, 2002 1:51:46 AM

Hey, I recently bought a new motherboard and athlon xp 1800 cpu. upon installing it, with the heatsink that came with the retail box, I find my idle tempature is around 68c!! I know athlons run hot, but not this hot... so I went and bought a larger high performance copper heatsink rated to 2100+, and installed it, but my tempatures are still 64c idle. I've properly mounted the heatsink over the processsor, put thin, thick, medium sized, and super-sized amounts of thermal paste on (I've remounted this thing about 19 times in hopes of the problem dissapearing). the motherboard I have (a abit nvidia based one) has a tempature diode under the processor that I've pressed up, down, put paste on, everything, in hopes that the readout was inaccurate....... but all this to no avail...

I've been working with computers for about 6 years now doing hardware repairs, so I wouldn't think it'd be anything stupid, but I really can't think of what this problem could be caused by@#%!@

someone have any ideas??

More about : cpu overheat box

June 30, 2002 2:01:43 AM

I did the same but with a Athlon 1Ghz, using the AMD heatsink that came in the box and it runs at 64C at full load and at 50C idle, trust me the heat sinks included in those boxes are aluminium rubbish with a fan, the word is cheap, I never replaced my fan and the system is nearly two years old and still running. If you weant cooler by a decent heatsink.

I would recommend a copper one but it must not be too heavy or it will crush the core (however I have never heard of this problem occuring). Check the AMD website it will list some approved heatsinks.

I not sure what one is best so maybe someone else can suggest a decent heatsink?
June 30, 2002 2:17:53 AM

I did go out and buy a better copper one already.. and I know the heatsink preforms well because I put it on my other computer, which is a 2000, and it runs at 45-48c idle... the problem must lie elsewhere.
Related resources
June 30, 2002 5:48:22 AM

Quote:
Hey, I recently bought a new motherboard and athlon xp 1800 cpu. upon installing it, with the heatsink that came with the retail box, I find my idle tempature is around 68c!! I know athlons run hot, but not this hot... so I went and bought a larger high performance copper heatsink rated to 2100+,


And your mistake is?


If it isnt locking up what temp it idles at is NOT IMPORTANT.


You shoudlnt have wasted your money on the new cooler, and your time messing with it.


Let it go, no temp is too hot which does not cause the cpu to hard lock.

Ignore it, trust me.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 5:50:58 AM

Quote:
I did the same but with a Athlon 1Ghz, using the AMD heatsink that came in the box and it runs at 64C at full load and at 50C idle, trust me the heat sinks included in those boxes are aluminium rubbish with a fan, the word is cheap, I never replaced my fan and the system is nearly two years old and still running. If you weant cooler by a decent heatsink.


The heatsinks are fine, the problem is users who cannot comprehend(not you btw this is a general statement)
That no matter what temp it says, if it is NOT locking up, there is NO danger to the cpu!!!

It is as simple as that, if 64C "feels" too high for you that is your issue, not the processors.

THe fud around amd and heat is amazing, everyday we try to help people who "think" their temps are too high only to find their systems are chugging along fine.

No lockups=no heat problem.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 5:52:50 AM

And the titles like.


"help my amd is overheating"
"amd cpus overheating right out of the box"
"my amd runs at 60C its way too hot help"


Only serve to continue the misinformation which exists.


64C is on the high end, but even 70C is fine, cause if your thermal probe says 70C and it isnt locking up, there is no damage being done, if it says 100C and its not locking up, the thermal probe is wrong!

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 12:19:06 PM

Well, it is still running, my point was that it does not matter how hot it runs, Athlon has a paper written on it how it can run up to 95C, which is why I do not worry.
Anonymous
a b à CPUs
a b V Motherboard
June 30, 2002 12:49:26 PM

yes but you have a performances loose @ 70°C!


<i>if you know you don't know, the way could be more easy ...</i>
June 30, 2002 1:22:41 PM

No you don't with Athlons. You DO with P4 because P4 will throttle the CPU speed down. Athlon will not. So your performance will remain rougly the same at 70 as it does at 40 or 50.

Mark-

<font color=blue>When all else fails, throw your computer out the window!!!</font color=blue>
Anonymous
a b à CPUs
a b V Motherboard
June 30, 2002 1:55:22 PM

because of the heat, you lose some. (same as the memory, the gpu, the chipset or another chip)


<i>if you know you don't know, the way could be more easy ...</i>
June 30, 2002 3:32:33 PM

Quote:
yes but you have a performances loose @ 70°C!


You have no performance loss at any temprature on an athlon short of a lockup.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 3:34:10 PM

Quote:
because of the heat, you lose some. (same as the memory, the gpu, the chipset or another chip)


NO you dont, there is no performance loss from heat on a macro scale, run a benchmark at 30C@ 1.4ghz and you will get the same result(within margin or error of course) @60c assuming the chip is within lock temprature(IE stable).


Ram dosent lose performance when it heats up either, that is patently untrue.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
Anonymous
a b à CPUs
a b V Motherboard
June 30, 2002 4:51:36 PM

here an example i have found (among others...): <A HREF="http://magaland.com/cpucooler.html" target="_new">Thermoelectric CPU Coolers</A>

btw, did you know this review about heatsink & cooling: <A HREF="http://www.overclockers.com/tips263/index.asp" target="_new">Heatsink Testing Methodology</A> from Overclockers.

there are 2 formulas,
<b>C/W=(Heatsink Temp - Ambient Temp) / CPU Watts</b>
then
<b>C/W=(Thermal Diode Temp+Diode Worst Case Factor+Mb Correction Factor-Fan Intake Temp)/CPU watts</b>
to calculate the quality of a HSF, but i am wondering if it is exact/accurate.

other articles on this site about this topic:
<A HREF="http://www.overclockers.com/tips240" target="_new">Heatsink Testing</A>
<A HREF="http://www.overclockers.com/tips32" target="_new">Determining Heat Sink Efficiency Using Thermal Diode Temperatures</A>


<i>if you know you don't know, the way could be more easy ...</i>
June 30, 2002 7:40:08 PM

what temp is your house normally, and what type of airflow do you have in your computer? dual fans in the back or 1? Are the ide cables gettin in the way? Also you might try artic silver 3 instead of that white puddy paste they give you.
June 30, 2002 7:41:08 PM

id like to see some benchmarks on this, bench a cpu that is starved of fresh air with a stock cpu, maybe just take the fan off and get it hot as you can, then bench it. then bench the same setup with a high performance water cooler. id be interested to see the results.

how do you shoot the devil in the back? what happens if you miss? -verbal
June 30, 2002 8:18:32 PM

Your first link is an advertisement for tec's which they lie about heat degrading performance.

The other mention nothing about heat lowering performance.

Dont know what your getting at, wait for next post for my statement.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 8:20:27 PM

Quote:
id like to see some benchmarks on this, bench a cpu that is starved of fresh air with a stock cpu, maybe just take the fan off and get it hot as you can, then bench it. then bench the same setup with a high performance water cooler. id be interested to see the results.


You can see the results easily, run cpuid, note the mhz of your chip, run toast, during the toast run, run cpuid again, note the same mhz.

The mhz is a rating of chip speed, if the chip slowed down due to heat it would show up in its speed rating. It would benchmark exactly the same, it would do everything exactly the same.




:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 30, 2002 8:51:29 PM

what motherboard u use?? i know there are some MBs from ASUS the cpu temperature is 10c more than the actual temperature. i think a bios update may be needed.

:) 
June 30, 2002 10:00:51 PM

you of all people should realize that mhz alone is not THE deciding factor on performance. im talking about getting a baseline average of a bench mark while extensively cooled, then another heated almost to the breaking point.

how do you shoot the devil in the back? what happens if you miss? -verbal
July 1, 2002 9:46:48 AM

Quote:
you of all people should realize that mhz alone is not THE deciding factor on performance. im talking about getting a baseline average of a bench mark while extensively cooled, then another heated almost to the breaking point.


When stable yes it is.

A 1.46ghz axp will perform exactly the same @ 45C as it does at 60C, if it reaches a temp where the core begins making errors it is no longer stable.

AMD_man, you have a varying voltage fan, run some benchmarks at the low setting and high setting to show if you could.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
July 1, 2002 9:50:05 AM

BTW, it should be pointed out that analogue circtry performs better when cooled due to lowered resistance in the traces, but a cpu is a timed circuit, meaning if the cooled traces can send electricity faster it is wasted because it is forced to run at a set speed, hence my mhz comments.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
!