HELP - Last Ditch Effort to save my machine

[Jman13]

I've been battling increasing instability on my Athlon XP for the last two months, and I'm at my wit's end as far as solutions go:

Backstory...machine is nearly 3 years old, had an original install of WinXP the whole time. I bought a new HD (Seagate 200GB), and an upgraded 2100+ t-bred to go with it. After installing the HD and the new processor, I discovered that the T-brad wouldn't work with my mobo because of the voltage requirements (me=idiot). So I put my 1700+ back in and sent the t-bred back. In order to use my new HD, I installed SP1. Shortly thereafter (perhaps a little over a week), I started getting random crashes (reboot/blue screen). It happened with no pattern, but would crash about every 2-4 days. Previously, the machine very, very rarely crashed.

I thought it might be the hard drive, so I did thorough diagnostic testing on all my drives, which all came back fine. After exhaustive testing on data and sector by sector scans, I'm convinced my HDs are all fine.

I thought that perhaps the new HD was the final straw on overtaxing my PSU....I bought a brand new Antec TruePower 550W....absolutely no change in behavior.

As time wore on, the crashes became more frequent. Again, no pattern...sometimes while doing nothing, sometimes online, other times in games. I used memtest and scanned my RAM for errors, and none were found. I even went as far to rotate both RAM sticks through every slot, and then finally a brand new RAM stick to make sure both my sticks weren't corrupted. That too did not help.

I have run every virus scan / trojan scan / adware / spyware scan available...all found nothing.

Finally, I succumbed, thinking it absolutely had to be my WinXP install....it must not have liked the upgrade to SP1, especially on such an old install. (I held off on this so long because I had so many programs installed, I really didn't want to do a full wipe.)

So, I formatted my boot drive this week and reinstalled WinXP, slipstreamed with SP2. It made absolutely no difference. Over the past 3 weeks the crashes have been getting worse, and today they've been the worst ever....my machine has crashed 9 times this morning, and the most intensive thing I've done is browse the web.

It's quite obvious that I have something wrong with the hardware, but I have no clue where to start. If it's a cheap component, I'd love to just replace that, but I can't just spend all this money on an aging machine until I get lucky and hit on the right part.

If I have to, I'm prepared to buy a new mobo, processor and RAM (as I was planning on upgrading in the next few months anyway), but I'd really like to wait until after Christmas before I drop the $450 it will take to get the hardware I'll need. Is there any hope to postpone this, or am I screwed?

 

[mechBgon]

What motherboard do you have, and if it takes two types of RAM then which type do you use with it? You may find clues in Event Viewer, depending on the nature of the crashes, so check that out too (Control Panel > (Performance & Maintenance) > Administrative Tools).

 

[Jman13]

It's the machine in the link at the bottom of my posts. I'm using PC2100 DIMMs. I haven't had it overclocked since the problems began cropping up, so it's running at stock speeds now (and has been for several months)

What shoulud I look for in the event viewer? I do see a lot of 'hardware failure in 'wceusbsh001' which I assume is a USB related failure. Could USB problems cause massive system instability? Where do I start here?

 

[boomerang]

Clear your CMOS as a first step. I'm assuming you changed some settings to try and get the 2100 working? If you did a BIOS upgrade in an attempt to get the 2100 working, flash back to the previous version.

If that doesn't help I think you should take a close look at the heatsink/fan on your processor. These symptoms sound like they could be heat related. Maybe freshen up the thermal paste or scrape off the thermal goo and put on some thermal paste? Obviously the HS/Fan was removed to try out the other processor.

You seem to have tried everything else. Could be your mobo is going flaky on you.

 

[mechBgon]

I see you have two USB hubs. You might try removing them as a fact-finding step.

I know I would pay the $55 to escape from a KT266A-based motherboard, not to mention a three-year-old one :Q Shuttle AN35N Ultra for a basic escape route, and you can probably get more than half your money out of it on the used market when you finally go beyond the SocketA platform. You'll also need an add-in PCI hard-drive controller since you are running so many optical and hard drives, unless you can narrow it down a bit.

 

[VirtualLarry]

Originally posted by: Jman13
I've been battling increasing instability on my Athlon XP for the last two months, and I'm at my wit's end as far as solutions go ... Backstory...machine is nearly 3 years old ... As time wore on, the crashes became more frequent. Again, no pattern...sometimes while doing nothing, sometimes online, other times in games.

Bad caps on the mobo, most likely, or possibly the PSU.

Also, I might (with the machine off and unplugged, including any USB devices), try unplugging and re-plugging every single connector (aside from the CPU and heatsink assembly). It's remotely possible that something has either worked its way loose, or has slightly oxidized, or shifted due to thermal expansion/contraction. Don't forget the AGP slot too. Bad caps or a plain bad mobo are more likely, but I remember having to re-seat the RAM and BIOS DIPs on my old XT every six months or so too. It does happen. Is there a lot of dust in the case too? It's possible that heat buildup or voltage spikes could have taken a toll on the mobo's VRMs too.

If you are using a PS/2 keyboard, try a different one too, sometimes the cable gets worn and can short out.

 

[Jman13]

Thanks for the suggestions, everyone. I was halfway through a response when the machine crashed again, so here goes once more.

Virtual Larry: Inspected the motherboard, no obvious capacitor bulging, leakage or corrosion anywhere. Obviously, that doesn't rule it out, but at least I'm not leaking. The PSU is not the problem...as you'll see in my first post, that was one of the first things I replaced, and it made no difference. (and my PSU now is ROCK solid on the voltages).

Anyway, new and interesting twist: While I was on the ground inspecting the caps, I decided I'd throw my third stick of RAM in so I'd have 768 for the time I was debugging. During my RAM trials, I had noticed that the machine was slightly more stable with 1 stick of RAM, but it still occasionally crashed...I pulled the stick that it seemed had more crashes at the time. (really, it wasn't any more obvious than the others, but I was grasping at straws.) Anyway, immediately after adding the 3rd stick, I booted halfway: and crashed....tried again, and again crashed during the boot sequence. I had NEVER crashed during boot until now, and it was twice in a row. I removed BOTH older sticks of RAM now, and it booted fine (I'm running just the new 256MB stick now). I've had this setup before, but still had the occassional crash, but now it might be more pointed: I think something is screwed up with the memory controller...the more RAM to deal with, the worse it is.

I did upgrade my BIOS when I prepared to install the 2100...could a BIOS problem lead to issues like this? I'll have to give it a try and find out. I also unplugged one of my USB hubs, although I don't think it will matter since the boot crashes occurred with that hub disconnected as well. I'll keep you posted, and I'm always open to new things to try as I figure this out. I have a feeling, though, if I don't have a solution by Monday, a new A64 will be on its way...just about 2 months earlier than I expected....

 

[Fern]

Have you run Memtest86 on your ram yet?

 

[boomerang]

Originally posted by: Jman13
I did upgrade my BIOS when I prepared to install the 2100...could a BIOS problem lead to issues like this?

It most certainly can. Flash back to the BIOS you were using previously and see what happens. Be sure to follow their procedure in regards to switching jumpers and/or removing the battery. Whatever the manufacturer specifies.