Computer stability problems

CP5670

Diamond Member
Jun 24, 2004
5,697
798
126
I have been having some major problems with my computer for a while now, which could possibly be related. I thought someone here might be able to give me some ideas; I have been reading AT articles for a while but haven't come here before. This was my system configuration before all the problems started: (it's still the same except for the memory)
Athlon Tbird 1400mhz processor
ASUS A7N8X nF2-400 motherboard
512MB crucial PC2100 DDR SDRAM
GeForce3 (original) video card
Maxtor 160GB ATA133 hard disk (7200rpm, 8mb cache)
Windows ME

About four months ago, the hard drive started going very slow, but in an unusual way. For example, when copying a file over say 50mb, it wrote around 45mb quite fast and then just paused there for about fifteen seconds (during which period the computer hung up), came back to life and wrote another 40-odd mb very fast, paused again, and so on until the file was written. In addition to this, once in a while the computer would actually freeze completely (even when the HD was idle) and upon restarting the hard drive could not be detected at all, but shaking the drive itself a bit allowed it to be detected again and fixed the problem temporarily. I tried connecting the disk through a PCI ATA133 controller and still got the same problem, so it wasn't an issue with the motherboard's IDE unit. I probably should have paid more attention to this earlier but was busy with various other things at the time.

The problems became a lot worse about three weeks ago, as the computer kept randomly freezing up much more often, especially during games. I noticed that the processor temperature was 63C/145F when idle and went up to 72C/161F under only medium load, so there was definitely something wrong. After some head scratching, I realized that I hadn't bothered to remove the fan from the heatsink (the cooler is a Vantec FCE-62540D) when cleaning out the computer and it turned out that there were some large and very thick pieces of dust stuck in there. They were so solid that at first glance I thought they were some kind of rubber padding. :p After I cleaned those out, the temperature dropped quite dramatically, to about 51C/124F idle, the computer suddenly started working fine, with proper Windows loadups and no crashes even after playing games for several hours, and it seemed that everything was okay. The problem returned the next day though; the computer again randomly crashed and/or restarted itself quite often, even though the temperatures were now quite acceptable.

The problem got steadily worse over the next week, as the frequency of the crashes increased and eventually both Knoppix and Windows failed to load up at all; the computer either froze up or restarted itself early in the loading sequences. I didn't have any spare, working components lying around to debug
the machine, so I gave it to Best Buy's diagnostic service. They said that at least one of the memory modules was almost certainly bad and the hard disk and motherboard were possibly so as well. So I went out and got some new memory (two generic PC3200 256mb modules) and I could now get the computer to start up (in both windows and knoppix), but this time, many programs that had worked fine before refused to load at all, giving various protection fault errors and BSODs (any games with the unreal or unreal warfare cores, some control panels) and those that worked at all froze up or gave such errors within half an hour. In most of the freeze cases, if I waited for about 40 seconds, it suddenly came back to life and then worked fine for a while longer, although occasionally it had locked up for good. This looked like the hard drive acting up (the original hard drive pauses were still there) and I was getting the same problem with Knoppix, so it wasn't Windows-specific either. I therefore removed the HD altogether and the crashes have become much less frequent (in Knoppix), but they are still occurring every now and then, and as far as I can tell, the machine doesn't just come back to life again.

So that's where things stand now and I am not really sure where to go from here. Any suggestions on what I should do? I think there is almost definitely something wrong with the hard disk. It's still under warranty and I managed to back up everything I needed, so I could easily get it replaced, but it looks like there is some other problem too. I have heard that the Athlon 1400 does not react well to heat and I don't know long it had been running at those temps, so could the processor have been permanently damaged because of that? I have already had one of these inexplicably die out on me about two years ago, but it was under warranty back then. I could not see any obvious burn marks on my current one, but it's hard to tell. I was going to upgrade the processor anyway (along with the video card) but there wouldn't be much point if it's something else that is damaged.

Any help here would be greatly appreciated.
 

oldman420

Platinum Member
May 22, 2004
2,179
0
0
flash the bios that may help also are you at stock speeds?
welcome to the forums!
 

Fern

Elite Member
Sep 30, 2003
26,907
174
106
Go to HD manufacturer's site and d/l their free diagnostic proggies. Those should tell you if the HD is screwed.


Download and run Prime 95 to check your cpu (I use the "torture test" portion). Sorry, I don't have the link handy. Just use google, the site will be Merrsenne. org or similar.

If it fails Prime 95, your cpu may be bad, the other thing c/b ram. Can rule out ram by running Memtest86 (also a free d/l). Run P95 at least I2hrs

Finally, your OS is prolly fubar'd, will need reinstall after above prob (cpu, HD) are fixed/ruled out. I wouldn't trust the system restore in ME at this point.

EDIT:
Free d/l of Memtest86 here
Get free d/l of Prime95 here
 

CP5670

Diamond Member
Jun 24, 2004
5,697
798
126
Originally posted by: huesmann
Clilffs Notes? :)

yeah, that turned out to be quite an essay, didn't it? :p Since the problems had been going on for a while, I thought I might as well give the full story in case anything looked suggestive.

Originally posted by: oldman420
flash the bios that may help also are you at stock speeds?
welcome to the forums!

Thanks. I checked the BIOS version and it's the same as the latest one on ASUS's website (C1007). I had flashed it several months ago when I got the board. Is there any point in doing it again?


Thanks for the information and links. I ran Memtest86 and the test 4 generated tons of errors. I just got new memory, so there probably wasn't anything wrong there. I played around with the module positions a bit though and got some interesting results: (I did a few twice to make sure the results were correct)
2 modules in slots 1 and 3: 383 test 4 errors
2 modules in slots 2 and 3: 177 test 4 errors
1 module in slot 3: first try: 1748 test 4 and 1 test 3 errors, second try: 3721 test 4 and 2 test 3 errors
1 module in slot 2: first and second tries: 0 errors
2 modules in slots 1 and 2: 1342 test 4 errors

So the single dimm in slot 2 configuration doesn't give any errors, but that's only half the memory at half the speed. The random crashes now seem to be gone though. The HD pauses are still there under this arrangement, but occur at around 20mb intervals now. Does it sound like my motherboard is bad? I think it is also under warranty since I only got it last october (retail package), but I need to check what asus's coverage is.

I used Maxtor's Powermax utility as well, but that didn't find anything with at least the basic tests. I called them up though and they're sending a replacement, so I will see if that does anything.

I will leave prime95 running tonight and see if anything happens that. What memory configuration should I use for that?
 

Fern

Elite Member
Sep 30, 2003
26,907
174
106
"I will leave prime95 running tonight and see if anything happens that. What memory configuration should I use for that? "

I'd use the one module in slot two, hoping to avoid any ram problems so that P95 focuses on just cpu errors. Prolly too simplistic, but that's what I'd try.

I take it you tried both sticks individually in slot two, perhaps indicating it isn't just that one stick is bad?

Of course it would be nice if you could get the use of another mobo (borrow?) and run memtest again to help pinpoint either mobo or ram as the problem.

Just because the ram is new I wouldn't rule it out as faulty. I received too many "brand new" faulty parts. I'm more comfortable with used/known good working parts. New= suspect in my cynical little mind.
 

CP5670

Diamond Member
Jun 24, 2004
5,697
798
126
Well what do you know...you were completely right. I had actually only tried one of the modules for the slot 2 tests, which was a big mistake. I just assumed that they would both be working since they were new. I now plugged in the other module alone and got a whopping 7026 test 4 errors and all the random crashes and errors came back. Just to be sure, I inserted both modules into the other two slots (individually) and always got thousands of errors with the bad one and none with the good one.

I'm not sure what the memory brand is, but they have Samsung chips and standard "warranty void if removed" stickers. I got both of these from a local store called The Computer Place a few days ago; hope it's still possible to get the bad one replaced...

Also, I had Prime95's blend torture test running for 11-odd hours last night (with the good module only) and didn't see any problems in the morning; it was still running and CPU/mobo temperatures were well within acceptable limits.
 

Fern

Elite Member
Sep 30, 2003
26,907
174
106
Glad to see you've pinpointed the problem and that your cpu looks good. Here's to hoping your HD replacement will be an upgrade over the older one (I sent in 40gig and got back a 60 or 80giger last time) :)
 

TD77077

Member
Mar 1, 2002
150
0
0
Your hard drive issues sound like the drive is almost full or needs to be defragged. When it slows down during a large file xfer it's probably swapping. Try a disk cleanup. ME is good at fragmenting a drive, especially if you were crashing a lot.
 

CP5670

Diamond Member
Jun 24, 2004
5,697
798
126
I just got the bad memory replaced and the new one seems to be working fine, with no memtest errors. The bios detects 512MB during the startup sequence and says "memory frequency is 133mhz, dual channel mode." This is probably a stupid question but is this normal? I think it should be running at 266mhz due to the DDR specification but I just can't seem to remember what it used to say with the old memory (been quite a while since this computer was running properly with full memory :p).

Also, I noticed after I got home that the replacement wasn't exactly the same thing; it was also a 256mb PC2100 dimm but the samsung chips on it had slightly different labels, so the two modules I have now aren't absolutely identical. Not sure if this would make any difference though.

Originally posted by: Fern
Glad to see you've pinpointed the problem and that your cpu looks good. Here's to hoping your HD replacement will be an upgrade over the older one (I sent in 40gig and got back a 60 or 80giger last time)

lol that's pretty good. I guess they would only do that if the HD is out of production though.

Originally posted by: TD77077
Your hard drive issues sound like the drive is almost full or needs to be defragged. When it slows down during a large file xfer it's probably swapping. Try a disk cleanup. ME is good at fragmenting a drive, especially if you were crashing a lot.

That was my first thought also but the HD only has about 50GB of data on it (out of 137GB that WinME recognizes). Diskeeper's fragmentation check showed "moderate" fragmentation though, so I'll try running that tonight and see if it helps.

I was using ME only because it used to work with the SBLive's SB16 emulator for old dos games, but that stopped functioning when I changed the motherboard so I was will be formatting the drive to NTFS (for XP pro) anyway.
 

huesmann

Diamond Member
Dec 7, 1999
8,618
0
76
I would run SiSoft Sandra to check what speeds your CPU and memory are running at, don't just rely on the BIOS because it could be using some other way of measuring.
 

CP5670

Diamond Member
Jun 24, 2004
5,697
798
126
Oh yeah, I forgot about the memory test in that. I just ran that twice and it is coming out to 1954mb/sec integer and 1926mb/sec floating point. Does that sound accurate for PC2100/266mhz speeds? It does say "2x 134mhz (268mhz data rate)" for the speed though, so that looks good. One of the notices at the bottom reads "system bandwidth appears FSB limited," but I think that's due to the older processor.