• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Stability problems - I can't figure it out!

S

Swanny

Diamond Member
Hello All,

Yesterday I put together a new computer to be my Linux rig. It is an Epox 8RDAE with an Athlon XP 2500+ and 512Mb of DDR (2x 256Mb Corsair XMS 3200 LL). It is running Mandrake 9.1 with KDE 3.1. I am currently posting from it.

My problems start when I leave the computer. If I leave it for any period of time (over 15 min aprox.) it is hard locked when I come back. I've turned off all power saving (at least I think I have), so it is just frozen on the desktop. The only thing the system is doing is running SETI. All settings in the BIOS seem good and I have no overclocking going on. I am running the latest BIOS. My CPU is running at 50-55 C (hot I know, but it shouldn't be causing problems). The case temp is about 35-40 C.

I have two thoughts on possible causes. For one, the passive heatsink on the northbridge is very hot to the touch. Almost to the point of burning. Let's just say you definately know if you finger is on it!

The other is the CPU. I bought it as a retail CPU, but I'm using a Thermalright SK-7 and a YSTech TMD fan to cool it. The problem with the CPU is that it came with a bent pin. I was very suprised by this and at first I thought I had somehow bent it. On closer inspection, I found it had some green stuff on it. The green stuff is the same green color as what's on the CPU. I bent the pin back so it would fit in the socket, but I left the green stuff on it. It was not totally covered, so I figured it would have adequate electrical contact. I'm not sure if this could be effecting stability at all, but it seems very strange to me in any case.

I'd appreciate any input you have on what could be causing this!

EDIT: One more thing I forgot. There was one time when it crashed while I was at the computer. I was wondering about temps, so I had taken the case side off. I then went into the PC health section of the BIOS and was just sitting there watching the temps. I wasn't pressing any buttons, just watching. Finally I decided I was done watching, so I went to exit the BIOS. Nothing. It was hard locked at the health screen. This makes me think that it isn't anything temp related since being in the BIOS doesn't put much load on the system and the temps were as low as I had seen them.

Thanks,
Swan
 
Ok, here's an update. Out of curiosity I logged out of KDE and left X at the login screen. Then I dropped into a command prompt screen (not sure what the technical term is, but it's what you get with ctrl+alt+F1). From there I ran SETI. It was running for an hour with no problem. I only quit it to come back into X so I could post here.

So my new guess is that I have something weird in KDE that is making my system freeze after a while. Does anyone have any guesses?


Thanks,
Swan
 
Power supply?

Or maybe ram?-->Run memtest86

Btw you're using thermal paste or AS, not the yucky gum pad right?
 
Originally posted by: Ionizer86
Power supply?

Or maybe ram?-->Run memtest86

Btw you're using thermal paste or AS, not the yucky gum pad right?
Click to expand...

I'm using AS2 for the TIM.

Power supply is a 400W, but a brand I've never heard of. Mustang is what it says. It came with the case (a cheap one from Newegg, Foxconn maybe?).

Like I said earlier, the RAM is Corsair XMS 3200 LL running at 2700 speeds, so I don't think it could be a problem. Also, if it runs fine with SETI (very memory intensive) in the command screen, the RAM is pretty much ruled out.
 
Another update:

I just came back from leaving the computer for about an hour. It was left in the command prompt, which I had thought to be stable. Anyway, it was crashed when I got back, so I'm back to thinking it's a hardware problem. I think I'm going to request a RMA for the CPU and drop in my 1800+ to test it out.
 
Probably way to obvious, but double check closely to see that the heatsink is properly mounted on the cpu. Sometimes, even with experienced users, the heatsink catches on the lip of the bracket and you don't get full connection. The artic silver instructions webpage shows how this happens. It's probably not the case for you, but it's worth double checking.
 
its not software im tihnking because it crashed in bios and thats way before the OS which makes me think your ram or psu is bad.
 
Yes, I'm about to take it all apart to check the heatsink. Temps are a little higher than I think they should be. I just checked the voltages in the BIOS. They are:

VDD: 1.6
Vcore: 1.63
Vdimm: 2.7
+5: 4.97
+12: 11.79
Vagp: 1.53
-5: 4.99
VDat: 3.12
5VSB: 4.89

The one that seems a little strange is the +12V rail. Should it really be that low? I don't have much experience with power problems since I've always bought name brand PSes. This is the first "cheapo" PS I've owned.
 
Originally posted by: Ionizer86
My 12W is approaching 13; it looks a tad low.
Click to expand...

Perhaps I should consider getting a better quality power supply. This one was working fine with my XP 1800+ on an ECS K7S5A.

I'm still curious about the bent pin on my CPU. Has anyone had experience with something like that effecting stability?
 
Originally posted by: Swanny
Originally posted by: Ionizer86
My 12W is approaching 13; it looks a tad low.
Click to expand...

Perhaps I should consider getting a better quality power supply. This one was working fine with my XP 1800+ on an ECS K7S5A.

I'm still curious about the bent pin on my CPU. Has anyone had experience with something like that effecting stability?
Click to expand...

12v on all SocketA mobos I've ever seen have been low, but all your 5v are low, a new PS seems to be a good idea.

re: bent pin, as long as it didn't break it'll be fine. I have bent pins before and never had a problem with the cpu.
 
I'm running the memtest now. It's passed one loop already with no errors. I've leaving the house for a while, so I'll let it run.
 
Ok, I'm back from being gone. To my suprise the computer hasn't crashed. It's still running memtest. The important figures are as follows:

Memory 512Mb 815Mb/s
Wall Time: 5:24 (hours and minutes)
Pass: 19
Errors: 0

So it's been running for about 5 and a half hours and passed 19 tests. I think we can rule out memory as a problem.


I am suprised that this hasn't crashed it. My remaining possible problems are power supply and CPU. I presume that memtest doesn't stress the CPU very much. Because of this the PS probably isn't working it's hardest either.

When running SETI, which stresses the CPU and the RAM (and hence the PS), the system does crash. This still doesn't help me as to if the PS or CPU is bad. I'd rather not spend the money on a new PS if I don't need it. Perhaps tomorrow I'll have time to swap this CPU with the one in my main rig. That should be able to solve the puzzle once and for all.


 
Doesn't Mandrake have list serves for this kind of stuff. This doesn't sound like a hardware problem to me. If it was a powersupply problem then the memmory would be the first thing (depending on which voltages were out of wack) to go. This sounds like a bad instal problem to me. Check your list serves.
 
Does Epox have an update for your BIOS? If so... flash away!

I put together an AOpen NF2/2500/2xCorsair and had a very similar problem. After flashing the BIOS the problem went away.
 
Yes, the BIOS has been flashed up to the latest release. Just another little update. I dedided to let memtest run all night to see if it crashed. It hasn't:

Wall Time: 9:48
Pass: 35
Errors: 0
 
Yes, the BIOS has been flashed up to the latest release. Just another little update. I dedided to let memtest run all night to see if it crashed. It hasn't:

Wall Time: 9:48
Pass: 35
Errors: 0
 
Ok, another update.

The computer crashed on me again. I was in Konquerer surfing around the forums. Out of nowhere it hard locked and wouldn't respond to anything. I'm about to swap CPUs with my main computer to see if the CPU is the unstable part.

CPU temp when I restarted was at 52C.

Thanks,
Swan
 
Originally posted by: wetcat007
Have u tried another OS, to see if that goes any better?
Click to expand...

Not yet. Right now the chip is spending some time in my main computer. That is a Win XP Pro machine that has been completely stable with my other 2500+. I figured I'd let the chip in question run a while in it to see if it crashed.
 
Originally posted by: wetcat007
Have u tried another OS, to see if that goes any better?
Click to expand...

Ive had this issues before. If its messing up at BIOS level, it well could be the PSU. My uncle's computer did that too, this time it was distinctive as I could hear a noise from the PSU too 😀

Personally try another cpu also, given the green substance on the pin, but my bet is with both, actually 🙂
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top