Help troubleshooting BSOD errors

[butch84]

Hey all. I really need input on how to get my computer well again. You can see the specs of my rig in my sig. The only thing I left out is my powersupply; I believe it is a 465w Enermax. At any rate, I was getting bluescreens that would often reference nv4_disp.dll, so I thought the problem was with my 6800GT. I then ran a videocard stability test, and came back to a message that "the display driver has stopped responding" or something. Clearly that was the case, as the resolution was really low and it was only 256 colors or so.

So, I RMAed my BFG 6800GT and in the meantime I've been using my old 9600pro 256mb card. I was just playing Battlestations Midway, and my computer had a BSOD again. It was another 0x0000008E stop error again. This is apparently caused by a reference to memory that should never be accessed. I've run memtest through overnight with no errors. Nothing is overclocked.

Does anybody have any suggestions? I'm about ready to just give up on this one, and I really don't have the money to start replacing parts until I figure out what the culprit is.

UPDATE:
From below:
Bad news everybody. Now my computer bluescreens during or soon after boot. This is with either 1gb crucial ddr or 2gb corsair ddr. It also happens with either my 9600pro or 6800GT. At this point, I suspect the motherboard is bad. I don't know what else it could be. Anybody have any ideas? If it is a bad board, what mobo should I get to replace it? I still have all that ddr and an agp video card, so I'm thinking about another Asrock board...... even though that pains me.

some of the errors:
0x000000D1
0x0000008E
0x0000004E
0x0000007E

Thanks a lot,
Dave

 

[oynaz]

The most likely culprit is your memory. Memtest is by no means fail-safe. Try and replace the RAM with the sticks in your 2nd rig, and see if it helps.

 

[Lord Evermore]

Also test whether it might be the mainboard (a particular slot failing) by running with only one module at a time (or two modules if you have 4 now). Run a stress test with the module in each slot or channel and see if any fail. If that module consistently fails, obviously that's a bad one. Use Prime95 to generate lots of heat and use lots of memory, or any of the other many benchmarks, but P95 is an easy one.

One assumes before you RMA'd the video card, you tried different driver versions? It seems obvious that in this case it wasn't the card or drivers, but that's still a good idea for initial troubleshooting.

 

[butch84]

Thanks for the responses guys. I think you're right oynaz - it must be memory related. The main problem is the 775i6G is sorta screwy when it comes to memory timings since it's really just overclocking the old 865pe chipset. My ram is technically rated at 2-3-3-6 I think, but when I set it to SPD I have problems with booting and especially warm reboots. I didn't have this problem when I tried my old crucial ram from my other box. It seems that this motherboard is just a bit finicky since it wasn't designed for 266fsb. Since I don't want to re-purchase 2gb of ddr (and I shouldn't rma it because it's not the fault of the memory itself) I think I might look into a board based on the via chipset that supports ddr/ddr2 and agp/pci-e. I know it's not quite as good of a performer memory wise, but if it will be stable, I think it's worth it. It will also help me transition to ddr2 and pci-e. Has anybody else run into problems with the 77596G?

Thanks,
Dave

 

[butch84]

Bad news everybody. Now my computer bluescreens during or soon after boot. This is with either 1gb crucial ddr or 2gb corsair ddr. It also happens with either my 9600pro or 6800GT. At this point, I suspect the motherboard is bad. I don't know what else it could be. Anybody have any ideas? If it is a bad board, what mobo should I get to replace it? I still have all that ddr and an agp video card, so I'm thinking about another Asrock board...... even though that pains me.

 

[mindless1]

1) How long had the system ran in the original configuration, which we should assume (?) had worked without bluescreens for awhile (How long?) ?

2) If you had made no system changes from the point of no-bluescreens to the point where you first starting seeing them, it would tend to implicate the OS (including Windows updates if you'd applied any in the interim), power supply, mainboard, heatsinks/fans (I mean failures or dust, if the room is otherwise same temp, we can't see if your system is next to a heating duct in wintertime, for examle) or video card - though video card was changed so all that remains with it is to be sure all remnants of the old drivers are gone and perhaps reinstall DirectX. At least start by making sure cooling system is working as the problem "might" be getting worse? Memory doesn't typically get worse like this, not just sitting in the system unless the board or PSU was killing it or you had grossly overvolted it.


3) To rule out windows, make a temporary clean install of the OS. Add nothing else yet, but for older known stable drivers. The goal being to make sure it is NOT modern, up to date, etc, to rule out potential software/OS/driver changes.

4) Underclock memory, higher mem timings, and FSB temporarily. Does the problem persist?

5) Unplug PSU for several minutes then examine motherboard for failed caps. If none found, leave PSU uplugged and open, examine it too.

 

[butch84]

1. I'd say about 2 months
2. I checked the cpu, videocard, system, and psu fans. All are working.
3. I reinstalled windows about a month ago - it made no difference. It was after this that I rma'ed my video card, because of BSOD's referencing nv4_disp.dll.
4. Since the problem occurs with 2 different kinds of ram, I conclude the ram isn't the problem. Running at 3-4-4-8 timings didn't help. Since the 865 chipset wasn't designed for 266fsb originally, the ram is clocked at 177mhz (3:2) ratio, so I'd say it's already underclocked.
5. I don't see any bluging caps, but I can't say I'm an expert. I'm definitely not going to open up the psu, because I don't want to be electrocuted.

Is there anything else to check out? I'm really tempted to try and get my hands on an ASUS P5VD1-X, since it has the features I want and it's not Asrock. (although I believe Asrock is a subsidiary of ASUS or something)

 

[mindless1]

I wrote "unplug PSU for several minutes", but actually after about a dozen seconds unplugged you won't be in any danger. Some people will (ignorantly, to be blunt) take a shortcut warning like "but it has capacitors in it". OK except it has a minimum of two different ways it will drain them in seconds so a little information out of context that it has capacitors isn't so useful.

Bulging caps aren't hard to spot, providing you can see their tops pretty well since a vent out the bottom would still cant the cap at a funny angle. Tops of all will look similarly flat and roughly parallel plane to the board (not crooked) if they're ok. Funny smell or residue from leakage would of course be another sign.

Another good thing to check is the questionable memory in another system. Boards in general and particularly your setup could have an incompatibility or instability problem with one module, but a different one could have failed. Memtest86+ will often find that though, but memtest can sometimes miss a very slight instability, I find that often if really trying to max out memory when overclocking. It is possible (something aggressive overclockers more often see, but could conceivably happen to anyone) a configuration is barely stable, a state allowed by the best possible new(er) motherboard, but as it or power, etc, ages it does degrade slightly. Such things are part of why margins are built in to parts, but that one unlucky person somewhere will have a scenario where it was just stable enough that being close to a margin was indetectable, until some time and normal system degradation has occured. So a new motherboard (if it were the cause) could hypothetically fix a problem only to resurface again later. The best way to avoid this possiblity would be to not get the same motherboard again, but all too often that IS what happens for reducing OS reinstallation or warranty replacement reasons.

I wasn't suggesting that you out-think what I had written, when I wrote underclock I didn't mean, "So what you have now is ok because memory is downclocked from it's spec", I specifically mean to NOW underclock. FSB you can't just ignore as if memory underclock is ok. It not only reduces heat, power slightly, also more tolerance to higher temp and worse power at lower clock, and remember where that memory access is coming from - through the chipset. You are hopefully trying to isolate variables and to do so you must change things, I meant run the entire motherboard as slow as it will go, temporarily. That must seem a rude way to state it, but not meant that way, just to the point.

It is even possible you have two different problems, formerly the video card and/or it's driver, and now something else. Bluescreens are a symptom but a generic way to report several things too, you might also check Event Viewer and since it was video, if you had a legacy PCI video card, trying it just long enough to remove AGP factors.

The OS is such a large variable, you write that you did a clean install but we don't know if it stayed "clean" truely or if the first thing you did was more recent drivers, windows update (or OS did that itself), etc. As with the memory, trying the video cards in another known viable system might not hurt.

It would be quicker to just buy new parts of course, if you had already wanted a different motherboard/upgrade I might think that is a good shortcut at resolution. But more expense, and time if it wasn't the board.

Ignoring the really crap generic boards, that Asrock is related to Asus isn't necessarily significant. Distinguishing features on boards are construction quality and support, two things Asrock isn't as good at but that often make Asus desirable (in addition to the giant user base of typically-more-savvy owners).

Stop codes are often shown in abbreviated form, if your Google searchds for causes of a certain code don't turn up hits (or maybe even the first search would be better as,) try it shortened. For example 0x000000D1 as 0x0D1, revealing it is Driver related and potentially caused by some driver-like emulators too, like Daemon Tools. So many softwares now update themselves, I don't claim Daemon Tools does or does not, but this is another reason why a truely clean install is important, even refraining from adding hardware drivers if you can get away with that (if system use isn't immediately manditory, you had a backup system or problem occurs so often you have reasonable expectation the use of the system isn't gone for too long (Or of course, dual boot OS installations would avoid it too).

Another code, 0x08E does make it seem like the clean install may even allow the error, here's a MS page,
http://support.microsoft.com/kb/827663

0X7E could be boot device, driver, cable problems. It might not even be related to the other problems. If you had defragged your HDD or reinstalled windows while memory was instable you could have all kinds of impossible to resolve problems too. Therefore I would underclock the whole board, after examining PSU and checking it with a multimeter if you have one, using one memory module alone and conservative memory timings, do a clean install of windows. Buy a different board if you want to, no point in doing all this if you have the money, you can always return it for a restocking fee depending on where you buy it.

 

[butch84]

Thanks a lot mindless1, that was really in depth, and quite helpful! I think your point about some of this being os related is correct. I was trying to uninstall a game yesterday, and in the middle of that the machine bluescreened - 0x8E IIRC. After that, I started getting the 0xD1, 0x4E, and 0x7E errors. So, I think it's safe to assume that the new errors were caused the 0x8E crash that happened during the uninstall, probably corrupting system files in the process. If that's the case, I'll need to reinstall windows just to continue troubleshooting, correct? Is there a good freeware hardware diagnostics suite I can run (besides memtest)?

To be honest, if I need to reinstall windows to troubleshoot more (possibly hosing the new install in the process) I'd almost rather get a new motherboard now and save the trouble. But, if the mobo isn't the problem, I'd be really frustrated. Unfortunately, I don't have a multimeter ...... does looking at the voltage reported in the BIOS help at all? It'd be nice to know whether psu/mobo were bad, so I don't end up replacing both.

If I do reinstall with this mobo, I'll definitely underclock first. Thanks again for your help!

 

[Philippine Mango]

Its either your motherboard or power supply. Get a multimeter and attach it to the 5V rails (when the system is off), turn on the system and monitor the voltages. Turn off system and do it again for the 12V rails, report back your findings.

It Really sounds like your power supply but it could be a host of other reasons, power supplies are at fault for the majority of system problems, especially those that go undiagnosed and unresolved.

 

[Philippine Mango]

Originally posted by: butch84
Thanks a lot mindless1, that was really in depth, and quite helpful! I think your point about some of this being os related is correct. I was trying to uninstall a game yesterday, and in the middle of that the machine bluescreened - 0x8E IIRC. After that, I started getting the 0xD1, 0x4E, and 0x7E errors. So, I think it's safe to assume that the new errors were caused the 0x8E crash that happened during the uninstall, probably corrupting system files in the process. If that's the case, I'll need to reinstall windows just to continue troubleshooting, correct? Is there a good freeware hardware diagnostics suite I can run (besides memtest)?

To be honest, if I need to reinstall windows to troubleshoot more (possibly hosing the new install in the process) I'd almost rather get a new motherboard now and save the trouble. But, if the mobo isn't the problem, I'd be really frustrated. Unfortunately, I don't have a multimeter ...... does looking at the voltage reported in the BIOS help at all? It'd be nice to know whether psu/mobo were bad, so I don't end up replacing both.

If I do reinstall with this mobo, I'll definitely underclock first. Thanks again for your help!

buy a multimeter cause it will be useful in the future anyways.. preferrably digital, voltages in the bios can help bu not in your case because you need to know the baseline voltage from when it was good.