Sign in with
Sign up | Sign in
Your question

AthlonXP 2000 Overheating - Help!

Last response: in CPUs
Share
March 8, 2004 12:48:10 PM

Hey

I'm running an AthlonXP 2000 on an Abit KD7 board with a Global Win CAK4-88T heatsink / fan installed using arctic silver. Nothing is overclocked. I've had this system for over a year now but just recently it has suddenly become very unstable during processor-heavy tasks (eg, DivX encoding has really showed it up majorly). Taking a look with MBM5 shows that the CPU is running very hot (espcially considering the heatsink) with an idle temp of around 80c. Earlier I ran a Prime95 torture test and the temperature reading peaked at 102c after just 5mins or so before things started to crash and I removed the cover to let things cool down. I'm making the assumption that these readings are accurate judging by the instability that seems to link in.

Now ideally, I'd like to be able to check the fan speed to make sure that the CAK4 is still working as it's supposed to, however I can't work out what fan type & divider settings I need to enter in MBM5 in order to get a true reading - so if anyone can help with that it'd be a great help.

If that's not it (or I can't find out for sure) then where do I go from here? I was thinking it would probably be worth taking everything apart, cleaning up the heatsink / cpu and re-applying the arctic silver. This computer does get moved around a fair bit so I guess things could have shifted a little and broken the 'bond'?

Anyway, if anyone's got any suggestions on what would be the best thing to try then it'd be much appreciated. Oh, and I think I read somewhere that the ideal idle temp to aim for would be around 40-50c, is that correct?


Thanks in advance,

Mark
March 8, 2004 4:55:03 PM

Those temps are WAY too high, I've got an XP 2000 and a Barton 2500+. I'm under 40 degrees at idle, and about 40 at load.(Edit: these are socket temps, not diode temps)

I don't think the temp readings are accurate. I think AMD says 80 degrees is the limit. IF 100+ degrees was correct, you likely would have fried the chip

Does your BIOS show temps? If so, does it confirm MBM5 temps?

Really high temps imply that the heatsink is not seated properly.

Barton 2500+
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
512MB Corsair Twinx 3200LL
9600 Pro
Enermax Noisetaker 420 watts
Win98SE<P ID="edit"><FONT SIZE=-1><EM>Edited by Coyote on 03/08/04 01:56 PM.</EM></FONT></P>
March 8, 2004 6:35:20 PM

the bios temps match up i'm afraid... it also gives the correct reading for the cpu fan (which shows it running at around 3500rpm)

oh, and amd give the max temp for this chip as 90c... and i can certainly confirm that things get very unstable at around this temp

but yeah i guess i'll take the lot apart and re-install the heatsink then see how things look
Related resources
March 8, 2004 8:10:25 PM

Maybe the fan on your hsf is dying and not generating sufficient CFMs anymore. It Could also gotten loose. Try to reinstall the HSF and DO NOT USE TOO MUCH ARTIC SILVER!!!

Mobile Barton 2500+ @ 2420mhz 11x220 1.7v
Asus A7N8X Dlx 440 FSB
1gb Geil GD pc3500 Dual Channel (2-3-3-6)
Segata 80gb SATA 8.5ms seek
ATI Radeon 9800 Pro(420/720)
March 8, 2004 8:55:55 PM

Maybe he got a XP2000+ with Palomino Core?
March 8, 2004 11:19:24 PM

Re-seating sounds good to me. If your CPU frys it could take your board with it. I assume you've checked your PS and Bios to make sure it's not over-volting. Also, I'm not familiar with Global, but is 3450 the correct rpm for this fan? It would have to be 50% oversize because most run around 5000 rpm.
March 8, 2004 11:29:27 PM

this <A HREF="http://www.globalwin.com.tw/products/cooler/k7_p4/cak4-..." target="_new">particular fan</A> is variable speed - it's supposed to run between 2800 and 4000rpm. so i guess that means it's not sensing the cpu temp correctly (and perhaps isn't attached properly) as it isn't running at max speed and yet the cpu clearly needs it to be??

either that or it could just be broken, guess i'll have to re-attach it to prove anything really.

and yes, the core voltage is only 0.01v off what it should be so that can't be causing a problem.
March 8, 2004 11:51:13 PM

Have you recently updated your BIOS? or motherboard drivers?
It sounds like its reporting the temp incorrectly. My GigBit mb had me bamboozled for weeks, almost months cause it was giving me wrong temp readings...just my 2c
March 8, 2004 11:59:50 PM

bios: have never flashed (i've had some pretty bad experiences updating the bios on previous boards so i try not to play with that ;) 

motherboard drivers: i do usually update to the latest via 4in1s fairly regularly... could this really effect the readings?

although to be honest, the linked in stability issues seem to make it pretty much a sure thing that i do actually have a rather major problem here
March 9, 2004 1:34:13 AM

I agree, your CPU could be ready for the undertaker! You have to eliminate everything else before you RMA it though. You could run a much faster CPU on this board. Maybe this is a GOOD thing!

"I am become death, the destroyer of worlds! Now, let's eat!"
March 9, 2004 1:37:12 AM

Another thing to try is forced air cooling. Take the side off your case and replace it with a box fan for maximum air cooling. I would probably wait until after reseating the heat sink before doing this, but if it is always hot and the HSF is attatched properly and working it could be a lack of air flow in general.
March 9, 2004 5:47:33 AM

What is HSF?
March 9, 2004 5:23:28 PM

"If i'm still here what is HFS?(please don't laugh)

That should be HSF (not HFS) and it stands for "heatsink and fan)

Barton 2500+
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
512MB Corsair Twinx 3200LL
9600 Pro
Enermax Noisetaker 420 watts
Win98SE
March 10, 2004 5:08:52 AM

Sorry, about that mistake.
March 10, 2004 8:28:52 AM

ok i re-installed the heatsink (took it all apart, cleaned it up with pure acetone, carefully applied some arctic silver 3 & reassembled) but the temp is still pretty much the same, in fact it seems to be about 5c higher right now. i'm putting this down to the fact that the arctic silver probably hasnt fully 'settled in' yet, and perhaps simply that it's impossible to get as good a thermal bond on a second attempt.

anyway, it all went smoothly and i've built enough other machines (that DO run at respectable temps!) to satisfy myself that that can't be the problem.

so... what next? could the cpu itself seriously be the cause of the problem here? otherwise i guess it's gotta be the heatsink/fan as the core cpu voltage checks out to be 1.74v (which is obviously fine). are there any other voltage lines that could affect the chip like this?

i'm a bit stuck as to what to try now really, i can't exactly afford to just start buying replacement parts until i find the solution (i'm a graphics degree student without a job, what more can i say ;)  so any suggestions are much appreciated.

oh, and regarding case airflow: this case isn't great by any means, but it certainly used to be sufficient. besides, right now i've got one of the side panels off and the cpu is still at 75c
March 11, 2004 5:47:58 AM

Are you sure the HSF is aproved by AMD(and for at least that speed)? That might have caused permanent damage on the CPU. Or the CPU had a construction fault. Change it if it's still in warranty!
March 11, 2004 5:55:10 AM

global win rate this heatsink/fan for: "Socket 462(Athlon,Duron,Athlon XP3200+ up)"

the way that's worded, i'm suddenly wondering if it is actually designed to work on an xp2000?? if it can cope with an xp3200 then it must be ok with my chip right?

amd don't have it listed at all on their site though
March 11, 2004 8:13:26 AM

You have a palomino core. They run very hot. The xp3200 is much larger, so it can better dissapate heat. The palomino needs a better thermal solution.
March 11, 2004 10:03:32 AM

Look for aprooved HSF's for Athlon XP 2000+, not other version.
March 11, 2004 7:04:58 PM

I've got the same overall motherboard. Mine is the KD7A, however, both seem to have the same kind of issues with stability (it was a emergency purchase).

Your temps are way too high, but I would be curious as to what PSU your running (and the wattage - most are recommending at least 400 to 450 - no joke). Also, I would check some of the capcitors on the motherboard itself.

The motherboard in general requires alot of user attention.

I would also recommend you check all of these things (in conjunction with some of the other advice given) before getting to the processor.

Also, even though it's rated at Palomino speeds are you sure the core is 0.18 and not 0.13? If it's the later I would go copper as the die is extremely small for most alum/delta coolers. However, this latest suggestion doesn't explain why in the world your system would be okay at first then becoming a complete dog later.

This usually means either 1) component changes or 2) a dying component.

Either way I hope I've been a little bit helpful. Tell us whether you get the temps to go down.
March 12, 2004 9:24:25 AM

ok i've only got a 230w psu at the moment, obviously this isn't exactly ideal. however, the system has been stable until recently (and i've not added any hardware) so i'm reluctant to place the blame with the psu over this problem.

i'm not sure what you really mean by "I would check some of the capcitors on the motherboard itself. The motherboard in general requires alot of user attention." what exactly do you think i should look at?

and the cpu is deffinitely a 0.18 palomino... besides that heatsink is copper anyway
March 12, 2004 2:12:07 PM

The capacitors are the metal, tube-shaped components on your motherboard. If these are starting to bulge out or leak... you'll have to replace the motherboard. Don't think this would cause the heat problem with your CPU, though.

<font color=red> If you design software that is fool-proof, only a fool will want to use it. </font color=red>
March 12, 2004 4:06:23 PM

The capacitors regulate the power distribution throughout the motherboard if they are having issues this can cause a variety of side-effects (i.e instability, insufficient power to the heatsink combo (unlikely but possible with that motherboard), etc). A bad PSU can do the same thing. If push comes to shove, call AMD. They are actually pretty helpful about offering sugguestions to remedy your issue.
March 12, 2004 5:53:52 PM

some of those threads make pretty interesting reading actually... lots of stuff saying a low power psu can cause serious issues with a kd7 mobo...
!