And the problems continue... plz help resolving instability probs after cpu upgrade

D

DaRuSsIaMaN

[H]ard|Gawd
Joined
Apr 22, 2007
Messages
1,216
This is a follow-up or continuation of an earlier thread, but that one became way too cluttered, and also the issue has changed.

Basically, I have upgraded an old socket A system by pulling out my athlon xp 1800+ and putting in an axp 2500+ barton, which I got from friend. The computer boots fine now with normal settings (that was my problem before), but it is unstable. I experience crashes and resets when testing it in games and sometimes even in just windows. So far I have had these symptoms:
1.) just once it randomly reset on me from windows desktop (I wasn't doing much as far as I remember)
2.) I crashed to a BSOD from windows. The BSOD said "Hardware malfunction. **The system has halted.**"
3.) The most common: it crashes when trying to play games. The screen goes black, then after several seconds it resets.

I am thinking maybe it's an overheating issue? When I go into BIOS, it shows the cpu temp consistently around 60 degrees celsius. That's really high, isn't it?? When I'm in bios, the cpu is basically idling, right?

What else could it be? Too little power from PSU maybe??

Help would be greatly appreciated... I want to solve all this PC hell so I can move on with my life lol :(
 
Last edited:
yeah 60C is to high.. what heatsink are you using? because barton cores run way hotter then the other athlon xp cores.. max temp for that core is 73C which is probably why its crashing if its at 60C idle.. it should be running some where around 40-45C idle..
 
I would take it back apart, clean everything, reapply the TIM correctly and reseat the heatsink.
 
It is just the sorry-looking stock HSF which my friend gave me along with it.

Of course I cleaned it up and applied new thermal grease but maybe I did a poor job? How thin is the layer supposed to be? And you apply it only to the top surface of processor die, nowhere else, right? Despite the way hsf's come with tons more thermal grease so that it basically engulfs the entire die, when you do it with the liquidy stuff, you only want a very thin layer, correct?
 
ironic that I'm playing with my old AXP box this week as well

this is an AXP, not an AXP-M, right? and the BIOS is detecting/setting it up properly, right? (i.e: its at whatever 1733/1800 (as memory fails) mhz that it should be at)

and has sirmonkey said: 60* C is way too high, I look for 30-35* C in the BIOS (assuming an ambient temperature in the room of 20-24* C) and temperatures with gaming something in the 40's or 50's, 60's in the BIOS is a bad sign

another question: how many fans, and how big, does your case have? sometimes you can skimp a bit on the heatsink if you've got ridiculous case cooling

yet another question: what PSU do you have, if you're concerned that it may be a problem, thats probably a sign of something (i.e: do you know its a piece?), the power draw for an XP 2500+ shouldn't be that ridiculous, unless the system is loaded to hell with extras (like 10 HDs and whatnot)

and yet another yet another question: you cleaned the dust/fuzzies out of the heatsink/etc lately? might help a bit at least
 
Yeah, it is an AXP not AXP-M. BIOS doesn't say expressly what frequency it is, but it looks like it is detecting it correctly because in POST it says "athlon xp 2500+". When I run cpu-z in windows, it does show it at the correct frequency (which is 1833 Mhz btw :) ) and everything else looks fine too.

My case cooling is not ridiculous but I think it's normal. One giant 120mm Antec fan in the back and another 120mm bottom-facing fan in the PSU, and an 80mm fan in the bottom front of case as well. This case and everything else worked 100% fine with the older cpu, my AXP 1800+. There were no overheating or crash/restart issues then, so I think the case cooling should be decent enough.

The PSU, theoretically, should be plenty of power for what I have... my system is *not at all loaded with extras... just 1 hdd, a total dinosaur video card (ati 8500), you get the idea. It's a Coolmax V-500, 500W. I brought it up only because I thought back to all the forum posts I've read from other people about computers restarting for no good reason, and PSU problems seems to be the other common thing besides overheating. Am I right?

About fuzzies/dust... well I got it from my friend in a terribly dust & fuzzy-loaded state, and I bet he never cleaned it even once the entire time he used it, yet it worked for him. I cleaned it up as best as I could before popping it in but there were some left-overs in there.

I ordered a new hsf since everyone seems to agree my temps are much too high. So I think I will just not do anything until it arrives and then I will put it in and see how it goes. That will isolate the variables, and if temp decreases and if problem goes away, then I'll know for sure that it was heat and nothing else.

I'm thinking I must have done a poor job applying the thermal paste and putting on the hsf somehow so that the conductivity isn't quite what it should be.
 
I would flash to newest bios are just reflash just incase it got corrupted then apply new tim.
 
how exactly does "TIM" translate, btw? I'm betting you're talking about thermal compound but... what are the exact words?
 
Follow the instructions for applying TIM from the Arctic Silver website. ;)

Thermal Interface Material
 
Sathsayin said:
I would flash to newest bios are just reflash just incase it got corrupted then apply new tim.
Click to expand...

this.

regarding if its bad TIM application, you can get bad temps, but generally not that terrible (if you're using something like AS5 or ASC, even if you botch it, it still works fairly well)
 
Hey everyone, I just want to update this thread on the results. Turns out, it was indeed an overheating issue. I got my temps to drop by about 15C, and now it works perfectly. I got a cheap copper heatsink to replace the stock one. Then... this is very strange. I was looking around in bios because temps were still sort of high, like 50C after I got done playing a game. I suddenly noticed that my vcore voltage was 1.85 instead of 1.65... I saw that there was a setting for it and that I had it set to 1.85 for some reason (wtf??). I changed it to auto, and that immediately dropped the temp by like another 5-7 degrees. Now it's like 45C tops when idling.

I have no idea how it happened that I had the vcore voltage set to 1.85 because I don't remember ever touching that setting... by default it should be just set to Auto, I assume.

Anyway, all is well now. :)
 
You must log in or register to reply here.
Back
Top