Strange Problems

[excitron]

I've been having some strange issues with an old system I built a while back, and I'm thinking there may be something wrong with the MB, but I figured I'd get some other perspectives first before tossing this system. Here's the basic setup:

AMD Athlon XP, 2133 MHz (8 x 267) 2600+
MSI K7T Power/Pro/Pro2/Pro2-A/Turbo/Turbo2/Turbo Limited Edition
VIA VT8363A Apollo KT133A
1024 MB PC133 (SDRAM) as well as 1.5 GB PC100 (SDRAM)
FireGL T2-128 (128 MB) as well as a GeForce 5200 (256 MB)
Antec something-or-other 450 watt PS

A multitude of hard drives in Lian Li pull-out bays have been used. I had several OSs running on this, XP SP1, 2K SP2, and Suse Linux 9.2.

It started a while back when my XP SP1 drive started to reboot in the middle of startup. I'd see a quick BSOD blow by and then it would reboot and repeat indefinitely.

I changed some BIOS setting and managed to get the BSOD to stay, but it would be different every time. For instance, one was a Stop: 0x00000024, then a 0xFFFFFFFB, etc. When I tried installing an OS from scratch (either XP or 2K), I'd get a BSOD during some phase of the installation.

Thinking my memory was going, I replaced it with some PC100 ram I had. Nothing changed. Then I thought it may have had something to do with the graphics cards, so I changed the original GeForce for the ATI. No change. I directly wired an HD to eliminate the Lian Li bays out of the equation, which it did. Still had the same problem.

Now before I started hacking up the hardware, I should note that the Linux system always came up. The 2K system was also coming up, but then it too started crapping out. I have not tried starting Linux up since I changed the graphics card.

Eventually I managed to get another 2K setup up and running, and so far it's still cranking. I finally got an XP setup going, but the moment I ran SP3, off it went. BSOD.

I tried installing another 2K installation again today and got something a little different:
STOP: 0x0000000AZ (0x00000041, 0x00000002, 0x00000001, 0x8000414A)
IRQL_NOT_LESS_OR_EQUAL
Address 8000414A base at 80001000, DateStamp 381f8c6e - hal.dll

So anyway at this point I'm confused. Any ideas why this is happening? I can live without the XP setup on this box, but of course why is all this happening? Weird.

Thanks for any help with this.

 

[cutthroat]

Try running memtest86 on your known good RAM, if you get errors I'll bet the MB is going.

 

[excitron]

Considering I'm getting the same results from two different batches of RAM and a multitude of HDs, I don't really think that's where the problem lies. I can get even more RAM to use for testing, but I feel things will be the same.

Six hard drives can't all be crapping out at once either. Something else is up here. It's gotta be something on the motherboard, the CPU, or perhaps the power supply. There's not that many other places for things to go wrong!

Could I be getting a crappy data feed from the CD-ROM installing the OS (a TOSHIBA DVD-ROM SD-R5112)? I am using a 40-pin cable on that, although I figure that's not going to be a problem there.

 

[blue68f100]

My first guess would be ram, but since you have eliminated it. My next choice is the PS. I have seen some weired things with PS.

Memory86 has never found any problems for me. I like stress testing with Prime95.

 

[Grimmy]

How are the CPU temps?

Only reason I ask, is I remember reading overheated CPU could cause some problems:

IRQL_NOT_LESS_OR_EQUAL


You mentioned you switch out video cards.. what was in there originally, and what did you switch it to?

 

[excitron]

I originally had an nVidea GeForce 5200 with 256 megs of ram and switch to an ATI Fire GL with 128 megs. Same problem with both.

Yeah, I'm wondering if the PS may be at the core of the problem. I thought about CPU temps too. However, what I don't understand is why the one 2k setup runs OK, as well as the Linux setup. Both the PS and CPU temp would affect them as well, but they run fine.

As a matter of fact, I reinstalled all the hardware in another case that allowed me access to two Lian-Li pullout bays, and I can see all the drives via the 2K setup that refuse to boot up. Go figure.

At this point, I don't even think it's a motherboard issue anymore. I don't have the slightest idea why these OS setups, which used to work perfectly fine, refuse to boot up any more. As I write this, the Suse setup is up and running, after recognizing the "new" ATI graphics card (I must admit I was surprised the Suse 9.2 setup was "smart" enough to see it and run YaST).

 

[Grimmy]

Heh, I've had an opposite problem in the past.

I remember changing my AMD 1 ghz system into a Mandrake (now FC5) linux machine. Thing was, I wasn't aware my CPU was overheating. Linux did finish its installation, but I recall it missing things, after reinstalling a few times. Once I got the temp back under control (dirty hs) things ran smoother, and I was finding everything I wanted installed. :lol:

I really haven't ran into over heating problems with MS OS's.

Kind of weird how they won't boot up properly. Sorry if I missed it, but did you try getting into safe mode to run chkdsk? Or use BartPE or repair disk to get to a command prompt to see if chkdsk could correct the FS? Just wonder if it was corrupted somehow.

 

[excitron]

Well, as accurately as the MSI MB temp sensor can show, here's the CPU at idle in the 2K setup:

If it is a temp issue, I wonder if the thermal cream has dried up, as I just blew out he heat sink and fan with compressed air. Doesn't hurt to rip the sucker off and re-cream it.

I could try rendering a 3d model or something computationally intensive in the 2k setup to see of the sucker blows up. Why would a boot up crank the CPU? Not to mention the OS installs that blow up. What's up with that? I can't imagine the CPU being taxed on an install. I mean, I know Windows is bloatware and all, but I can't see it cranking the CPU on install.

Safe mode is Hang City. Goes through the motions and then just sits there forever at some point. The text on screen is different for different systems, so I don't think it's X process screeching to a halt.

WTF man? This is just too weird.

 

[Grimmy]

Well.. to share what I remember, the thunderbird I had was reaching 60C in the bios. So I can only guess that it was higher during its initial install.

Since installs wont display cpu usage, it should be under some considerable amount of usage. During a regular boot up, it would be less.

I converted my dad's old XP 2400+ (2 ghz Gigabyte MB) linux FC6, and it's CPU idle temp is 38-40C. Load temps would get 47-50C depending on the room temps. Not exactly sure what temp difference would be for your XP 2600+ (2.13ghz), but I would image it should be lower then (idle?) ->55C, unless its just MB temp values are different.

 

[ejay]

I had a MB that exhibited the same problems. Turns out, it had one or more bad caps. You can check out the following: www.badcaps.net for further info.