Computer unstable

[echelonphoto]

I have a two year old custom built with an AMD 3400+ processor running two gigs of ram
on an Asus A8v MB with a pretty beefy power supply. This is a work station for photography and has a raptor main drive with two IDE storage drives also in the case. It has been a rock, but this week its going crazy...it shuts off and it runs thru the reboot process, but never get to windows, it would just keep rebooting over and over. My son checked the bios and saw pretty high processor temps, way over 50 C, so he lowered the voltage and underclocked and I was able to run it again. Two days later , same problem...we tried another processor known to be good, but same problem..we changed the ram and that didn't do any good. I finally put in the old processor with the new ram and a better cooling fan and it still kept rebooting at stock clocking...so we lowered the clocking again and now it runs pretty well actually and the temp stays at 38 C. I am afraid to shut it off in fear of it not starting again and I would really like to clock it a stock speed.

Could it be the MB or the power supply? Any tests I can do?

 

[robisbell]

if you are running XP, you've been hit with the SP3 bug.

 

[Swampster]

With due espect to robisbell, if you were running into the SP3 bug, changing CPU clock speeds would have no effect. <FYI>, that bug us usually more common on a factory built system (HP primarily, but I have seen it on others) and has to do with how they load their "generic" installs of XP on a computer that is running an AMD processor.

Based on your symptoms, the first thing I would do is remove the CPU heatsink and clean it and the mating surface on the CPU carefully and completely and install a good even coating of thermal paste.

Item next would be to go into Control Panel / System Properties / Advanced tab / in the Startup and Recovery section, remove the check mark from Automatic Restart. Click Apply and OK several times until it closes System Properties.

The purpose of this is that it will then show you a BSOD instead of just rebooting. In that window, you want to write down the statement at the top that is in caps separated by underlines. Many times this may seem like a nonsensical phrase, but it means something in geek-speak. Next look for something that says STOP: and gives a series of numbers . . . write them all down. Then look towards the bottom for a mention of a file name or names and write this down. tFrom there, you have a basis to do some legitimate diagnosis instead of guessing. The only reason I am suggesting a CPU cooler repair before you try this is because it has been off and on several times, and I want to make sure it has a good thermal contact as this will cause overheating which will trigger a forced restart.

If you still have a fairly frequent power down or reboot problem and you are not seeing the BSOD, then it is almost always heat related.

If, with your reduced clock speeds, it is down to where it is only an occasional problem, download and run memtest86 (from Microsoft) and set it run through completely at least seven times. Usually that is enough to show any faults in the memory modules.

Next, you might get a good HDD tester (I use SpinRite) and do a complete diagnosis. I usually use a level 2 test as it is a good compromise between speed and throughness and because I have never had the more through level find anything that level 2 didn't find.

If you still haven't been able to pinpoint the problem, check back here as I will be clicking on Subscribe which means that I will get an email notice of any further posts to this question.

 

[echelonphoto]

Thanks for the info.

The machine was running OK for several days and now I have a problem again.

Here is the blue screen info:


DRIVER_IRQL_NOT_LESS_OR_EQUAL



STOP 0X000000D1 (0x4ADF7049,0x0000000,0x00000001,0x80006802



Hopefully that will give a clue

 

[echelonphoto]

BTW...it seems to run OK in safe mode

 

[Pulsar]

Stop 0xD1 messages can occur after installing faulty drivers or system services. If a driver is listed by name, disable, remove, or roll back that driver to confirm that this resolves the error.

Otherwise, try updating all your drivers and see if that resolves it. If not, try removing the cards in your slots until the probelm goes away. Then you will know which driver is causing the issue.

 

[Swampster]

Yes, it definitely helps to know the exact error message <G>.

Pulsar is absolutely correct in that it is a driver error. In my experience, this most commonly happens with the video driver, although it could happen with any of them.

Just as a matter of practice, I never update a video driver beyond the original update from the manufacture's site at the time I build the system unless there is a particular problem.

Remember, everything you do goes through the video driver in one form or another. That coupled with the fact that there is an infinite combination of hardware modules serving an infinite number of software combinations, operated by an infinite combination of users, and you start to see why it is so critical . . . plus why you it is not a good idea to update it just because there is a newer version available.

 

[echelonphoto]

Well, I have a Matrox G550 AGP video card...I never updated the drivers. I did try system restore and that didn't solve the problem...maybe I should go back further. Should I reinstall the stock driver?

 

[Pulsar]

Start from installing the stock driver, yes.

 

[echelonphoto]

We reinstalled the MB and video card drivers....the machine will run at a multiplier of 8 and lower voltage...but crashes again when we try to clock it at stock speed. It runs ok at the lower speed...I guess I'll just have to live with it until I build a new machine.

 

[echelonphoto]

Update:

The computer crashed while running a file operation...I tried restarting and got the blue screen again with more info. I addition to the drivers phrase and all the zeros, this is the last line:



sbp2port.sys-Address BA91d227 base at BA91800 Date stamp 410765A

 

[echelonphoto]

Anyone have any thoughts...I am really frustrated!

 

[RebateMonger]

Originally posted by: echelonphoto
sbp2port.sys-Address BA91d227 base at BA91800 Date stamp 410765A

That's a FireWire driver.

It never hurts to start with the basics, the hardware. Insert a CD with Memtest86+ and let it run overnight. That will test your memory and, effectively, test your CPU and memory controller, too.

Then test your hard drives with their manufacturer's diagnostics program.

If the basic hardware is perfectly stable, then you can go on with something like BART PE or a Linux Live CD and see if your other hardware is stable with "standard" drivers. This could be a lot of things, including failing capacitors in your power supply or motherboard.