Frequent crashes - thoughts (update)

[Andrew H.]

Over the past several months, my computer (AMD XP 1800+ w/ 512 M of RAM, Asus A7V266 Mobo, GeF 4 Ti4400 vid, running XP Home) has been crashing more and more frequently - I now get several BSoDs per day, with different messages, such as "Page Fault in Non-paged area," IRQL not equal to or less than," and maybe one or two other messages.

The WinXP diagnostic thing suggested that it was some sort of driver problem, but I don't think that's right, given that I first updated all of the drivers, and then went whole hog and reinstalled Windows.

So now I suspect that it's a hardware problem. I downloaded and ran memtest, which seemed to find no problems with my RAM. I'm wondering if it's a video card problem, since the crashes seem to occur more frequently when I'm playing games - less so in CM than in more video-intensive games, too. Also - I've been trying to install Kotor, but the install always crashes when the game tries to install the texture packs (which may or may not be related to the other crashes; it's conceivable that there's just a problem with that kotor disk.)

Any theories? Any way of testing this short of buying a new vid card? (Not that I would object to a new vid card, but only if it fixes the problem...)

[ May 20, 2004, 12:42 PM: Message edited by: Andrew Hedges ]

[Schrullenhaft]

It could be a variety of things, bad memory (though a diagnostic passes it), a bad hard drive (though you should see other errors typically would be my assumption), a bad motherboard, an overheating CPU (though a lockup would probably be the most common occurance with an overheating CPU rather than an error message); maybe even an underpowered power supply.

I've seen reports of the "IRQL not less than or equal to" error being related to PNP OS settings in the CMOS/BIOS setup (disabling that setting to solve the problem). There are other claims of it being related to drivers or certain motherboards (ASUS was mentioned actually). One suggestion was to clear the CMOS and reset the settings (with the possible exception of the ACPI/PNP OS). Another suggestion involved removing any USB Storage devices or maybe just any USB device or hub plugged into the computer.

One poster even mentioned turning off the onboard LAN. Another one had minor success with removing one version of an NVidia driver (which was conflicting with a particular MIDI driver).

The general course of action I would recommend is

1) Check your CPU heatsink/fan to make sure it is running, mounted properly and not too hot. Do the same for your videocard (though it probably doesn't have a problem). Clean out any dust build up (it helps a little, but probably not a whole lot). You may also want to check whatever contact material you have between the heatsink and the CPU. For 'retail' CPUs this is often a gray 'thermal pad' that you probably don't want to remove by taking the heatsink off of the CPU. If you used an 'OEM' CPU and you used a thermal paste/grease, you may want to check to make sure that the application of it is still good (it only needs a very thin layer).

2) Find the manufacturer of your hard drive and download any diagnostics and run them against the drive (the 'non-destructive' one). This should usually find any particular hard drive problems (though they aren't guaranteed to find all possible drive problems).

3) Possibly clear your CMOS settings and reset them. This is a long shot and may be more hassle than is worth, especially if you don't know what the settings are.

4) This can be a bit of a problem, but try out different (but compatible) memory with your motherboard. It's possible that your memory is overheating, mildly defective (possibly heat-related) or just isn't running well with your motherboard/chipset/CMOS-BIOS memory timings. You can possibly set your memory timings to a slower setting (higher numbers). You may also want to use just one DIMM at a time (assuming you have more than one installed currently) to see if the problem is related to only one of the DIMMs being present. Otherwise you may need to purchase some new memory (or RMA your current DIMMs).

5) Update your BIOS. You actually need board revision 1.05 (printed on your motherboard, probably need the printed model number) and BIOS v. 1002c (Palomino - core appears 'square') or v 1010 (Thoroughbred - core appears 'rectangular') for the Athlon XP 1800+. I assume you have this already.

There are other steps, but these will do for now.

[Andrew H.]

Thanks for the advice. I forgot to mention, but I already updated my BIOS and checked to see that my CPU fan was running. I'll start with the HD diagnostic and then swapping out the RAM sticks.

[ April 27, 2004, 06:18 PM: Message edited by: Andrew Hedges ]

[Andrew Kulin]

From my own horrible PC experiences here are some things to check:

- Power supply is often most overlooked culprit. It could be underpowered or showing major voltage fluctuations. You do not say but is your power supply a quality brand name (Antec, Enermax, Power PC and Cooling, etc.) or a no-name generic brand?. Try monitoring your voltages outside of BIOS - I use a program called Motherboard Monitor but there are others out there as well. I believe the rule of thumb you want is within +/- 5% of the stated voltage.

- check capacitors on mobo and all cards and look for signs of failure (leakage and/or the tops are bulging out. In my experience this caused odd crashes that increased in frequency over time and eventually led to inability to boot up at all. (Capacitors are the things that look like miniature pop cans BTW).

- get the hard-drive diagnostic software from your vendor to do a check (you may have to boot in DOS depending on the vendor). I had a Fu****su drive (two actually when you count the replacement) that caused all kinds of grief over time that got worse as the drive slowly failed. Some IBM drives are also prone to higher than normal fail rates as well.

- if overclocking turn it back down

- Are all case fans working and blowing air the right way into/out of the case. Is inside of case neat and tidy (wires tied up) or is it a messy tangle that impedes efficient airflow? Have you set up your hardware so that there is as much open airspace around them as you can manage (e.g., try to leave empty bays between hard-drives as these run hot). Have you tried running game with side cover off the case (to expose allparts to room temperature air and does it still crash when playing the game? The program Motherboard Monitor also monitors CPU, motherboard and case temperature if your CPU/board support it

[Andrew H.]

Update - I seem to have fixed my problem; thanks to everyone who helped out. The main problem seems to have been a bad ram stick - I haven't had any crashes since I took removed it, and I was able to install the first time I tried a game that caused my system to crash the previous 20 times I tried to install it. In addition to the RAM problem, I *also* had a HD problem, which I was able to find and repair with a diagnostic program.

So, finally, this seems to have had a happy ending. The moral? Not sure, although my personal lesson here will be not to buy ram from Best Buy.