Random Computer crashes

[jgcable]

My home PC is still acting up. Its a dual processor MP1800+ running DDR2 ram and Windows XP Pro.
For the last few months it has been randomly rebooting or crashing all together.
It is totally random.
Randomly it:
Restarts
Restarts multiple times
Blacks out.. then the monitor light starts blinking like its in sleep mode and the hard drive light stays on. I then have to hit the reset button to get it to start up again. I might have to hit the reset button several times.
It sometimes restarts and sometimes it restarts into that DOS looking screen that gives me restart options (safe mode and others)

Any ideas?
I have been told it could be the RAM or the hard drive. If it is the hard drive, how do I transfer all the data from the bad drive to one of the others in the PC (it has 3 hard drives in it). If its the RAM... how do I tell what stick it is? I have 1024M of DDR2 RAM in it.

Help...

[hippietim]

Lots of things come to mind John. Believe it or not this is where having a Dell would be kick ass. They have a diagnostic suite for their PCs that is VERY good at troubleshooting these sorts of problems.

It's doubtful that it is the harddrive. The only typical reasons for a hard drive completely bringing down a system would be if the physical location on the drive is where the OS page file is located. At that point Windows treats the system as insecure and will display the lovely BSOD (blue screen of death). Bad drivers can always cause bad juju but if nothing new has been installed then it's not likely to be the problem.

RAM failures can exhibit a LOT of different results. In many cases you get the BSOD - a critical RAM failure is pretty much by definition a security failure so Windows will kill itself. Depending on the RAM failure however, the OS may not even be in any sort of valid state so the machine can freeze on it's own. Depending on your hardware, certain RAM failures can just take the machine down all on it's own as well.

Did your system come with any diagnostics tools? Does the BIOS have an extended POST test you can run that will test the RAM thoroughly?

Have you checked on your thermal activity? Heat is always an issue in modern PCs. I'd look at the CPU fan, the chipset fan (if you have one), the fan on your video card (again if you have one), etc. Most of the BIOS' have some way of reading the CPU temperature.

[DonP]

No issues with any of my duallies thank goodness.

You can go here: www.2cpu.com and check out the forums. Depending on the MoBo and when it was made, there were "issues" like you are having due to filtering capacitors going bad on the MoBo. They will get puffy or leak electolitic (sp?) fluid. Since you are handy with a soldering iron, you might be able to replace these yourself.

If they don't look bad and one of my systems was acting up that way, I think I'd blank the hard drive and reload the O/S. Fortunately (or unfortunately?) I have many more systems than I need at home. Redundancy has it's benifits.

MP1800+ are very heat sensitive. +1 on checking on the cooling of these, especially with temps getting up there again. Clean out the dust bunnies.

[jgcable]

Lots of things come to mind John. Believe it or not this is where having a Dell would be kick ass. They have a diagnostic suite for their PCs that is VERY good at troubleshooting these sorts of problems.

It's doubtful that it is the harddrive. The only typical reasons for a hard drive completely bringing down a system would be if the physical location on the drive is where the OS page file is located. At that point Windows treats the system as insecure and will display the lovely BSOD (blue screen of death). Bad drivers can always cause bad juju but if nothing new has been installed then it's not likely to be the problem.

RAM failures can exhibit a LOT of different results. In many cases you get the BSOD - a critical RAM failure is pretty much by definition a security failure so Windows will kill itself. Depending on the RAM failure however, the OS may not even be in any sort of valid state so the machine can freeze on it's own. Depending on your hardware, certain RAM failures can just take the machine down all on it's own as well.

Did your system come with any diagnostics tools? Does the BIOS have an extended POST test you can run that will test the RAM thoroughly?

Have you checked on your thermal activity? Heat is always an issue in modern PCs. I'd look at the CPU fan, the chipset fan (if you have one), the fan on your video card (again if you have one), etc. Most of the BIOS' have some way of reading the CPU temperature.

Hey Tim, I don't think its a thermal issue because sometimes I turn the computer off for the day. When I go to turn it back on it sometimes reboots several times before it actually will stay on.
I have a Soundblaster Audigy 2 card
The Video card is a Radeon 9500 Pro
I will check to see if the Bios has an extended post test.
I have never gotten the blue screen.
The computer will just sometimes go black and the hard drive light stays on. Sometimes it reboots once or several times.
Sometimes it doesn't even beep when I push the reset button forcing me to push it several times.
I suspect I could have a bad stick of Ram. I guess the only way to check it is to remove one stick at a time and see what happens. I wish there was some kind of Ram diagnostic tool.
Its very frustrating because its random. I can't record on the computer because I never know when its gonna decide to reboot.

[rjohnstone]

All your symptoms point to a bad mother board.
You can try swapping out memory sticks, but I have a feeling that is not the issue.
Don't waste your time reloading the OS... it's not the issue either.

[Soap]

What mobo do you have, DFI, Abit?

[jgcable]

All your symptoms point to a bad mother board.
You can try swapping out memory sticks, but I have a feeling that is not the issue.
Don't waste your time reloading the OS... it's not the issue either.

I am pretty sure it is an Asus A7 (or something like that) motherboard.
Any suggestions for a new motherboard?

I got the computer from mm2002 who is a member here. He probably remembers everything about it.

[Soap]

My persnoal experiences with Asus and DFI haven't been all that good.
Both have a reputation of sending out some buggy boards.
It can't always be helped some bad boards are gonna get through.

I've been out of the looped for awhile on mobo's and all the feature rich options many have. I suggest, go to newegg and read through some of their mobo reviews to get a genreal feel for the board your considering.

Stick to simplicity with mobos, the more feature rich they are, they are more likely to be potentially buggy. All the on board features and bios options can be overkill. Consider your needs and scout around before buying. Personally, I love Tyan boards. Their boards are well designed and very very rock solid stable. I don't buy boards to have the highest benchmark or bragging rights for the latest and greatest. DFI has a good variety as with Asus, Abit and MSI, but plain jane Tyan gets my bills.

[rjohnstone]

I am pretty sure it is an Asus A7 (or something like that) motherboard.
Any suggestions for a new motherboard?

I got the computer from mm2002 who is a member here. He probably remembers everything about it.

ASUS makes great mother boards, but bad ones do slip by every now and then.
I personaly only by genuine Intel mother boards. They cost a bit more in some cases, but they've been rock solid for me. The only drawback is they only support Intel processors, so this won't help you.
Look for another ASUS board or something similar. Go to egghead.com and read the reviews as well.
STAY AWAY FROM FRY'S ELECTRONICS (aka outpost.com)!!!
99% of their stuff is factory recondition crap.
And stay as far away from an ECS board as possible.

[DonP]

My persnoal experiences with Asus and DFI haven't been all that good.
Both have a reputation of sending out some buggy boards.
It can't always be helped some bad boards are gonna get through.

I've been out of the looped for awhile on mobo's and all the feature rich options many have. I suggest, go to newegg and read through some of their mobo reviews to get a genreal feel for the board your considering.

Stick to simplicity with mobos, the more feature rich they are, they are more likely to be potentially buggy. All the on board features and bios options can be overkill. Consider your needs and scout around before buying. Personally, I love Tyan boards. Their boards are well designed and very very rock solid stable. I don't buy boards to have the highest benchmark or bragging rights for the latest and greatest. DFI has a good variety as with Asus, Abit and MSI, but plain jane Tyan gets my bills.

I've heard lots of complaits with Tyan and AMD platforms, or when using the Via chipsets with Intels. Tyan is usually the first out of the gate, but being on the bleeding edge, you often get cut. I do have a Tyan Thunder100 with a GX chipset that is uber-stable, but that's an old chipset.

Asus buggy? First I've heard of that. Probably my #1 choice for speed and reliability, and my PC-DL with dual Xeon overclocked from 1.6Ghz to 2.69Ghz is a model of stability. I also have a pair of Asus Serverworks chipset MoBos that are rock solid. Maybe you are confusing Asus with Abit?

All in all, it's just like guitar makers. Sometime they get it right (the planets are in alignment), and sometimes they have a dud on their hands.

When I was building duallies, I would read lots of reviews, and go with the popular setups that most people were having the least headaches with. I stayed away from "unusual" combinations and underdogs.

[DonP]

Don't waste your time reloading the OS... it's not the issue either.

Sorry, but I disagree.

1. You can't rule out that the O/S is not the issue. A virus could have whacked some files, causing it to flake out.
2. Reloading the O/S is free - doesn't cost anything but time. During the O/S load, if there are hardware issues, they'll even be more apparent.

[krimson]

I would try removing memory sticks to see if the problem goes away. It is the easiest thing that you can do and won't cost you anything unless you have the replace a memory stick.

I have had a similiar problem on three different occasions and 2 times it was the memory and once it was the mainboard.

I will add that if it was memory stick is the issue, you may have to reload the O/S as well at least I did.

[DonP]

Another issue with duallies - power supply. Duallies suck lot's of power, and the PS generates lots of heat, heat kills components...same reason why the MoBo could be bad.

[rjohnstone]

Sorry, but I disagree.

1. You can't rule out that the O/S is not the issue. A virus could have whacked some files, causing it to flake out.
2. Reloading the O/S is free - doesn't cost anything but time. During the O/S load, if there are hardware issues, they'll even be more apparent.

I'll argue that one till the cows come home... corrupt system files will crash consistantly at the same point everytime. Unlike hardware, a file either works or it doesn't, there is no in between with software.
Hardware on the other hand can work perfectly one minute, flake out the next minute then be fine a minute later.
In this case, sometimes it starts normaly and sometimes it takes a few tries.
Sometimes it will run for a while and the crash... very random.
His problems are random and consistant with a failing hardware component.
Things like his monitor going into what appears to be sleep/suspend mode.
This would point to a possible ACPI issue,which is controled by the bios.
Or maybe the video card is crapping out and causing the system to shutdown.
Regardless, reloading the OS is the longest and hardest part to do and should only be done as a LAST resort, not the first.
Swapping out parts takes a few minutes or seconds in most cases.

One thing you may want to try... reseat every connection inside the case.
Ford calls it the "wiggle test", mechanics call it "nut and bolting".
Basicaly go through every connection and make sure it's solid.
A loose cpu or an imporperly seated memory stick can wreak havoc on any system.

[hippietim]

I'll argue that one till the cows come home... corrupt system files will crash consistantly at the same point everytime. Unlike hardware, a file either works or it doesn't, there is no in between with software.

Sorry but I would not 100% rule out a software failure of some kind. Non-deterministic software failures are very possible - I've debugged enough of these over the years to know that for a fact. However, I've never seen an intermittent software failure that wasn't dependent on hardware state. That doesn't mean hardware failure, just the state of the hardware at the time. This can be anything from random garbage in memory at startup, the timing of control lines, etc. So ultimately you do have a software anomaly that is hardware related but it does not mean bad hardware.

Having said all that it is more than likely John has an actual hardware failure.

John, the last time I looked at someone's PC for this sort of thing it was a matter of just doing a few searches for the various hardware model numbers and a few key words - odds are VERY good you are not the first person to have the problem with your hardware. I'd do the searches starting with your motherboard. If you can find a good RAM checker that'd actually be the first on my list.