Page 1 of 1

Need help finding bad hardware

Posted: Wed Jun 04, 2003 3:06 pm
by TruckStuff
SCENARIO
My computer has been acting more and more flaky in the last few weeks (lots of BSODs, the occasional reboot, etc.). I went on vacation last week (system was off the whole time) and when I got, WinXP BSODed on me literally every 5 minutes. I pulled all unneccesary hardware to eliminte FP's favorite IRQ Channel Conflict and still got BSODs. I then decided to reinstall WinXP last night and it BSODed on me once during the install and twice after the install. Since I have not installed any software or any unnecesary hardware, I am lead to believe that it is a problem with the core hardware.

HARDWARE CURRENTLY INSTALLED
Tyan Tiger MPX Motherboard w/ latest BIOS
2x Athlon MP 1600+
2x 256Mb DDR RAM
TNT2 Ultra
Adaptec 19160 SCSI controller
Seagate 15k RPM HDD (model: ST318452LW)
Toshiba SCSI CDD
HP SCSI CDRW
WinXP Pro - clean install

WHAT I HAVE TRIED
I have run Seagate's diagnostic utilities and the hard drive passed all of them. I also run a memory diagnostic util (Doc Memory recommended here) all last night; it ran for 11 hours with no errors.

POSSIBLE PROBLEMS/TESTS?
The way I see it now, there are a few things that could sitll be wrong:
1) One of the CPUs is bad. Not sure how to test this though; ideas appreciated.
2) Motherboard is bad. Again, not sure how to test this, ideas appreciated. Linux seems to run OK on this system, although I don't run it as much as I run XP, so this may rule out the CPU and mb.
3) REALLY flaky XP driver for some piece of hardware (19160 maybe?).

I'm using the default XP drivers for everything at this point. Adaptec has some beta XP drivers for the 19160, but they have not helped with this problem in the past and I have not installed them on this install.

Any direction would be appreciated on how best to proceed/diagnose this problem. TIA.

LATEST UPDATE
I spent this morning searching for answers to these questions. Amid a few more crashes, I was finally able to get windows to tell me that my memory might be bad. So I downloaded Microsoft's memory tester and my memory passed. So now my question is is it possible for my RAM to pass all of these long-term diagnostic routines and still cause crashes like this? Very frustrated at this point. :(

Hmmm..

Posted: Wed Jun 04, 2003 3:16 pm
by eGoCeNTRoNiX
I'd try, if you can, using just one stick or ram or the other and seeing what happens. I'm sure some of the other guys who have a little more exp. will be along shortly. GL eGo

Posted: Wed Jun 04, 2003 4:23 pm
by Absolut Talent
Test the ram

then test the powersupply
CUrrently, what are you using as a PSU?


as for testing the CPUs....cant you run it on the mobo with only one? or does it need both to post?
and you dont have another dual mobo you can test them on?

Posted: Wed Jun 04, 2003 5:12 pm
by BillyGoat
My experience tells me ram can pass tests and still be bad, if you have another sys or a freinds you can pull it outa to test thats great- use a program to monitor your psu voltages and look for fluxtuations-
After doing all this i would test the board with one cpu in it and vice versa, check hs contact on cpu, as amd's will scream to a halt/possible death without proper hs contact- they just get too hot....
Good luck- dont pull to many hairs out- we've all been there, and have survived :p

Posted: Wed Jun 04, 2003 6:37 pm
by FlyingPenguin
The BEST ram tester I've ever found (I use it all the time on client's systems) is SimmTester's Docmemory. It runs from DOS (NO mermory tester can possibly check your memory properly if it runs from Windows)

It's free and you can get it here: http://www.simmtester.com/PAGE/products/doc/docinfo.asp

Also check your PSU voltages, particularly the 5 and 12 volt rails. I'd be concerned anything outside of 4.85 - 5.10 and 11.80 - 12.15 volts.

Most problems like yours I see lately are either ram or a weak PSU.

Posted: Wed Jun 04, 2003 6:42 pm
by TruckStuff
I did move my memory around as suggested and things seem to have settled down a bit. Perhaps the mem slots I was using are flaky? As I stated earlier, I ran DocMemory for almost 12 hours last night without a single error. Maybe it has something to do with the way windows is reading from mem?

The Power Supply is an Antec TruePower 430 (<6 months old). I will have to install a software package to monitor my voltages. Is MotherboardMontitor still the utility of choice here?


Thanks for the ideas. :)

Posted: Wed Jun 04, 2003 9:23 pm
by two slow
I still like MBM for things like that.

Posted: Wed Jun 04, 2003 10:35 pm
by Slugbait
Interesting...you "moved" the RAM around, and it seems to have "settled down".

You have a fine PSU, that's not your problem.

Memory slots can be flaky depending on the board, but this is very unlikely. I believe you're not using identical RAM in each slot...and one of them is not necessarily bad, but just not as good as the other, and it's creating an issue.

Some old Dell GXa and GX1 systems had serious stability problems because of the RAM...memory always tested fine, but had blue screens during OS install, or IE upgrades or reboots...they sucked.

Try it with only the bank 0 chip you currently have seated...if the problems go away, you know that the second stick needs to run solo/be replaced.