Ace's Buttery Flutterball

Ace · June 15, 2006, 17:20 hrs

Since there isn't much happening downstairs, figured I could visit here. I know I'm not aloud, so I'll speak softly.

Here's the deal; I'm running into consistent crashes while running games. It started with Neverwinter Nights, and I just picked up Call of Duty (1st person shooter) and it's crashed each time, after awhile...

I'm running 1 Gig of Corsair memory now; Athlon XP1800+ CPU with GeForce TI4200 64 meg card. DirectX 9. Plenty of hard drive space, no obvious mismatches. Stop errors are 0x00000008E (hardware failure) and occasional 0x00000000A (memory; happened twice). Microsoft failure analysis lists RAM, system board, CPU and Power Supply as likely culprits.

Have contacted Monarch (builder) and Gigabyte (motherboard). Have run Memtest with no problems (4 passes...). Downloaded 3DMark06 Basic (free) video diagnostic tool; idiot thing does NO tests unless you purchase the "advanced or pro" versions. So only did what Everest Home does, and spits out what I have in place, versions, etc. That looks to be a total waste of time unless I want to fork over purchase price for the apparently "real" version.

Gigabyte offered no help, but to suggest I swap out various components to see if it helps. Naturally, don't HAVE another power supply or card or...

I suppose I'll try Memtest again, and may try playing around with RAM (remove, move) but I don't think it's RAM related. I'm limited on any attempt to upgrade the CPU (any way to test...?); only will take an Athlon XP 2400 or 2600 at 266 bus, which look to still be available at $150-$175. I'd think both games are demanding visually, since there's moving action at some depth... NOthing is crashing during any other time/event. Just the games. Don't have either one turned "way up" as far as resolution... detuning doesn't make them "go." Problem with any video card upgrade is I'm limited there, too, by the mobo voltage as newer cards take less. So, like CPU, I can only go a little higher with that piece.

Any guesses?

Ace; and don't say "I think there's a nut loose, on your keyboard..."

saoirse · June 16, 2006, 11:45 hrs

Stop errors are notoriously difficult to pinpoint at times. The fact that your machine is only problematic when playing games is of small relief. If you can attach your most recent minidumps here, it may be possible to pinpoint the culprit.

First off though, update your video card drivers from the nVidia site and it won't hurt to have a hunt around for mainboard driver updates, particularly the sound driver if this is included on the mainboard.

Ace · June 16, 2006, 18:21 hrs

I saw you have the "Ow Ouch Ow" Ace quote up; up appropriate.

I don't know how much longer I'll be with you, so wanted to say thanks to Saoirse and Johnny for their tips. Tonight I ran Memtest again, 4 passes, and just fine... then it crashed about 6 straight times on the reboot.

Error messages are the common 0x00000008E along with Page fault in no page area and IRQL glitch message and 0x0000000A and 0x00000024 CLASSNP.SYS and 0x00000005.

I finally got it started in changing BIOS to fail safe defaults; then in Safe mode. Did a driver update search from the device and found a Windows driver downloaded/install. Then crashed upon attempts to reboot... Frequently getting Earthlink error message about "something using that port" that prevents internet connect.

I'm going to try for an nVidia update; have gone there before with some glitches and wound up doing it through Windows Update. If that fails may try uninstalling drivers and going with the OEM disk version (make sense?).

Tried 3DMark 03 tonight from a new download; it crashed on first test...

Johnny, dangit, you may be right. I was able to burn all photos and docs, so will lose nothing if it goes up in flames...

Ace; gotta check for a fire extinguisher...

saoirse · June 16, 2006, 23:06 hrs

Hi Ace,

Hit start --> run --> minidump [enter]

and attach your 2 or 3 most recent minidumps to the forum so we can have a look. This might tell us what driver is causing those stop errors, and in turn tell you the hardware that's causing it.

saoirse

Ace · June 17, 2006, 08:21 hrs

I'm not finding a way to copy them, to attach here... And no view that gives info on each. I did download nVidia drivers last night (crashed on reboot...) but it did boot today and got online.

Found another event of suspicion, based on a Corel Barista print error (due to using old WP8 version from the Quantex, back when...). Hopefully knocked out all references in Registry for that. Will uninstall the whole package if needed.

One Stop error last night was specifically "caused by video device driver". Most common has been the hardware failure, but as noted I'm getting a nice variety... I'll probably see what I could do as a video card replacement; I know the motherboard voltage will keep me from anything really current.

AC; here today...

Ace · June 19, 2006, 19:57 hrs

After having downloaded current nVidia drivers (not Windows Certified...) the computer is stable except when trying ANY game. Tonight I attempted Links Golf, which is usually stable and not troublesome. Crash once game started that was a "page in non page" dealy. Diagnosed as "device driver." Don't know anything else WITH a driver except sound, on mobo... and why THAT would suddenly glitch is beyond me.

I can either:
try orginal drivers, off OEM CD
Try Windows Update to go back to a Windows Cert driver
leave the video card alone
Try a sound driver update (did once, with Windows, which did not work.. had to roll back)
Memtests have been fine, on RAM

Is there a way to sense if power supply is going? Would that lead to a "driver" error? Any suggestions?

Just want to stabilize it to shove it off...

JA; I have never in my life attempted to locate a battery or mess with jumpers, so I'd have to study the layout for an hour to even have a clue what you're referring to. And would that have anything to do with these errors?

Ace · June 20, 2006, 06:25 hrs

J; how does that involve device driver errors?
A

Ace · June 20, 2006, 08:09 hrs

I figured it was like shock paddles to the heart. Well, brain. The Monarch isn't dead (I am using it, whilst we speak). So that's why I don't know what that would do as far as its inability to get through a video game... I'm not having any disruptions doing anything else (wp, cd burning, internet, etc.). Just gameplay.

I'm doing an audio driver download now from Gigabyte. I did the "Windows Update" audio driver download install again, and again the device would not work after... So instead of rolling back on that will try anew.

Does Virtual Memory have any correlation with crashes, I wonder?

AE

pat · June 20, 2006, 20:01 hrs

One thing I have done in the past to help game play for games that would crash was to lower the sound acceleration using the DirectX diagnostic tool.

Don?t know if it would help in your situation or not, but may be worth a try.

Start-Run-dxdiag-enter-Sound tab, adjust slider down a notch or two and see what happens.

If that doesn?t help, set it out under an Oak tree??????..

Ace · June 21, 2006, 06:32 hrs

Thanks; I moved the sound slider. While I was in there I figured I'd try the DirectDraw and 3D tests... the last one cancelled (twice) saying "test was cancelled because you touched a key"; I didn't touch anything (?).

Ace; "whatever" as my niece would say

Ghost · June 21, 2006, 20:30 hrs

Hey Ace,

Well I took some time to read the several posts and it seems to me a heat issue. Why you might ask, well like you said everytime you load it up with a game which for the most part are resource happy the machine reboots or you get the blue screen. Check the bios tempertures if there listed, look to see if your cpu fan is coming on. If nothing comes from that, format, install game or games and see if it happens right after a fresh install, if it does it seems obvious that it's hardware related. Oh, run a cpu test as well I'm pretty sure there's some free one's out there. I'm thinking its to do with the cpu from the symptoms you describe.

Mr. Sam · June 22, 2006, 00:52 hrs

I checked out the recommended requirements of Call of Duty, and here's what I got:

RECOMMENDED

Pentium? III 1000 or Athlon?processor or higher
Graphics NVidia Geforce/ FX or Ati Radeon = 128 mb or higher
English version of Microsoft? Windows? 98/98SE/ME/2000/XP
256MB of RAM
52x Speed CD-ROM drive
100% DirectX? 9.0a compatible 16 bit sound card and drivers
100% Windows? 98/98SE/ME/2000/XP compatible mouse, keyboard and drivers
DirectX? 9.0a (not included)

The usual "rule" with games when dealing with requirements is that you should have three times the minimum requirements and double the recommended requirements. Having said that, you seem to be ok in all areas except your CPU and video card. They simply just don't seem to have enough "juice." Since these are some of the most vital parts, I'd say you'd get all your games running just fine if you upgraded those significantly. Either that or ja's first post says it all ;).

Ace · June 22, 2006, 07:46 hrs

Thanks, CG and G. Fans are working; I had checked that when I added the 512 RAM and cleaned it out... The case is in a fairly contained area, so there's not a lot of airflow. As we speak, Everest sensor has the CPU at 70 C, Motherboard at 51 (CPU fan is at 4688 rpm). I'd agree Call of Duty is stressing it, even though CPU is "in their range"... Weird that Neverwinter is glitching now, after I went THROUGH the entire game once without probs...

I agree with the last statement on Johnny's first statement, and am going to upgrade. Just tried Links golf and it actually played ok (it had crashed, when I tried it this week..). So if it seems stable on nondemanding games will just figure it will stick to that level.

I can't upgrade the CPU or video card significantly, with this mobo, as it is restricted to 3.3/1.5 v cards (AGP 2.0) circa 2002. CPU would have to stay an Athlon XP, at 2400 or 2600 (where mine's the 1800). Putting a couple hundred in it to goose those two things minimally doesn't seem worth it..

Would a power supply glitch (only..) when under game stress? That's really the only piece I could see replacing, for the cost/effort (but not if it's not an issue). As far as spare parts, I've got nothing to swap with anything...

Ace; appreciate your helps!

Ace · June 22, 2006, 10:11 hrs

I appreciate the tip on not frying my guy; it'd be bad enough to fry the stupid computer, let alone my guy.

I just downloaded Hot CPU tester from PC Mag, and ran it... computer crashed while running (device driver stop error..). I'll try it again, but that seems telling.

Ace; "Hot" might be true enough.

Ace · June 22, 2006, 11:07 hrs

You know, Everest Home says there is no battery under Power Management Properties.. is that literally true?

It is difficult and time consuming to take this apart; all plug ins barely reach to where it exits the front of the ledge it's on when I pull the case out. And without having a clue on jumpers and that it's not a simple thing for me to dig in and try to figure out what I'm looking for; hopefully I won't have to "de-case" until I have something definite to do in there...

Ace · June 22, 2006, 14:23 hrs

I unplugged. I plugged. I'm baack. Didn't affect anything.

I wish I had played with jumpers. Or in jumpers. But I haven't. Never touched one. I've got a rope outside, if that qualifies...

I did 40 the other night. I just now caught my breath.

I tried Hot CPU test again; it crashed again, at about the 4 minute mark. The CPU "Quality" was at 100% throughout, until it went black (then blue). Same IRQL not-less-or-equal stop error, which is translated into device driver failure. Now that test wouldn't seem to stress the video card, since there's not much to watch... But I'd have to figure CPU (naturally) and power supply would be stressed? And even in a video card test, it'd also be stressing the CPU (and power) true...?

Where I'm at now, I guess, is: if the CPU is at fault, would it act fine except during tests or games when it's stressed? Is it actually the power supply that's choking this thing, and not the device targeted by each test...? I'm figuring RAM is ok after always passing Memtests...

Ace; I don't want to do a stress test.

Ace · June 22, 2006, 15:05 hrs

Well, this is Indiana so our clocks are never "right"... But yeah, it's right on.

Here's something I found suspect: went into CMOS and it said "system fan speed = 0". I know both were on fine when I did the RAM install. I don't see any reading on Everest for that (just CPU fan). And for power ratings it said
+3.3 v is at 3.2 v
+5 v is at 4.919 v
+12 v is at 11.520 v
so is the 3rd rail off?

CPU is at 71 C/160 F; Mobo is at 51 C/124 F. Aux at 81 C/178 F. Maxtor is at 43 C/109 F. CPU fan is steady at 4688.

Ace; maybe I could attach a window air conditioner...

pat · June 22, 2006, 19:53 hrs

70c is a little on the warm side, but not terrible, I think that processor is rated to 90c. It could be your motherboard is under reporting the temp though. You could take the side of your case off and set a fan to blow in there and see if that lowers temps and you?re able to compute longer on your games.

If so, perhaps you just need a better heatsink/fan.

Ace · June 22, 2006, 20:06 hrs

ed. note: Ace regrets pushing this whole thing to Page 2, although it does pad his already absurd number of poasts beyond reasonable limits.

Oh, sure... I'm going to try to play a game while blowing on the dang motherboard...

Geez criminy, the stuff these guys suggest... I might as well freeze a cold pack and hold that on with one hand while I try to kill some stupid demon with the other one.

well, unless I'm playing Madden 2005; there aren't that many demons.

You know, today I installed Diablo II and it wouldn't even play. Had to bail each time; uninstalled it. Tried Madden, and it crashed after awhile. The Links Golf is the only one that isn't dumping. The CD (D) seems to be oblivious, but it is recognized for most stuff ok... This unit played Return to Castle Wolfenstein all the way through, and well, and also Medal of Honor which isn't much different from Call of Duty. So, historically, it used to keep up fine with some normally challenging games.

May try the air conditioning. Probably have to go in and just see if both fans still spin. I'm still thinking it's power. Monarch figured .2 range.

Ace; Power Ranger.

Ghost · June 23, 2006, 08:20 hrs

Hey Ace,

I have a suggestion that might shed some light. Start up a game and let it crash then right after, check your temperture's.

Ghost

Ace · June 23, 2006, 15:33 hrs

Will do, Ghost. This just in: I tried the Hot CPU test again, with Everest on. The CPU at boot was at 68 C. It climbed into the test, up to 74-77-78 C at which point the CPU fan showed a jump from the usual 4688 rpm to @5200 or so, at which point the pc rebooted (no blue screen, just a reboot with a MS error report request). CPU is hanging around 72 to 71 degrees C, now. Also saw the 12v rail go from 11.5 to 11.6, couple minutes into the test when it bailed... CPU quality was high, so no drop in the test readings. Just stopped.

Probably will try a new power supply, as well. Anything regular ATX (1 12v) fits, right? And rotate some RAM while I'm digging. And if I come across a battery, change it.

Ace: and change the oil, on my fingers, beforehand.

pat · June 23, 2006, 15:55 hrs

Yes, I think Best Buy and Circuit City both sell Antec, if you?re in the mood for one of those. The new Antec Smart Power design does not have a fan at the bottom of the unit to draw air up from the CPU area, didn?t matter in the cases with the 120mm fan right behide the CPU area.

You may want to look into the True Power design that has the fan on the bottom of the unit to help remove heat from the CPU area.

Bill · June 23, 2006, 15:56 hrs

Acer,
Not all power supplies will fit in all cases. I found that outr the hard way. Bought an cheap P/S got my olD Q and had to mod the case to get it in.

Might be a good idea to buy the one you would for your new machine - use it to test. If it doesn't solve the problem, you've taken the first step to building the new one.
Bill

Ace · June 23, 2006, 16:45 hrs

Thanks BP (huh.. sounds like gas). I was thinking I might just shop retail locally for that. There's not a lot of selection online for anything old style. The case I got from Monarch had a 350 watt in it. I'll probably tear into it and eyeball before shopping... I just noticed there's a 2 12v form factor style out now, which I figure ain't for me. Thing is I'll probably get a case WITH a fan installed on a new one, so as not to hassle with it...

I just tried Hot CPU and charted as it went; CPU was still at 72 to start; 76 at 24 sec; 77 at 30 sec; 78 at 1 min 10 sec (voltage went from 11.58 to 11.68); 79 at 2 min; fan went from 4688 to 5273 at 2 min 40 and CPU at 80 at 2 min 55; fan dropped back to 4688 at 3 min 30 but mobo from 51 to 54 degrees then; mobo at 55 and CPU at 81 at 4 min 20, fan at 4688. Died at 4 min 30.

Hung on reboot, went with start Windows normally and it rebooted. at 52 on mobo and 74 CPU on reboot. Sent in error report (Neon will like this): said "this is a problem that will be fixed by Vista in the future"... There ya go.

I will also run with panel off, but problem is it's enclosed (in a walled in space) anyway if I set it back in... I'll watch fans, too, after I open it up.

Ace; hope it doesn't suffer from tissue rejection.

Bill · June 23, 2006, 16:52 hrs

The second 12v line won't go to waste if you don't use it. Don't quote me but some/many of the newer vid cards require a separate power supply in addition to plugin into the mobo. I don't remember what voltage however.

Bill; lot'sa help on that one, huh?

Ace · June 23, 2006, 17:01 hrs

No, I see what you mean.. but I'm going to keep whatever I get in this case to keep this one functional, and then start fresh with a new one for myself... I'll give this to someone in the family who's without.

Looks like Circuit City has a better range than Best Buy; I found an Antec 330 watt low noise that sounds like it would be more "old style" than a double 12 PCI express type. Monarch showed the oldest as ATX 1.3, and the new are version 2+. Naturally, I don't want to just pop in a cheapo one and have IT die quickly... But don't want to do overkill just for this unit, either.

Big help I am, huh.

I just figure if I'm going to operate, I may as well do exploratory surgery too...

Ace; I hope I can find a jumper to wear.

saoirse · June 24, 2006, 12:36 hrs

Hi Ace.

I had a good hard look over your minidumps - time to test those memory modules with memtest. The module(s) are almost certainly the root cause. Three counts of the 0xc0000005 exception were from the memory_corruption module, which is a significant pointer. A hardware bucket was also flagged, which again points to memory, CPU or mainboard.

None of the frame addresses point to any problems with Windows drivers; although 3 page-in-page calls were made from svart.sys -- which I believe belongs to symantec (norton). You could try disabling this from the recovery console (disable svart.sys).

Also, disable any RAM features you find in the system BIOS - such as shadowing or caching.

saoirse

Neon · June 24, 2006, 20:43 hrs

Ace,

Make sure that you test the memory modules one at a time (only one module installed at a time). I have found that if only one module of an installed set of modules is bad, it is possible for memtest to pass the whole set.

If it isn't a memory problem, your results thus far suggest a possible overheating problem. Although 71C isn't outrageous for a Palomino core, it seems high. I also had an AthlonXP 1800+, and it rarely ran at such a high temp unless under significant workload, not at idle. Has yours always run near 70C? The 80C that you have observed is not good, if that is an accurate number. Something is wrong with either your motherboard's temp reporting (some are not particularly accurate), or the CPU fan, or the CPU-heatsink thermal interface, or the case airflow.

Mr. Sam · June 24, 2006, 21:38 hrs

Wow, those temps were high! :o You should probably look into some cheap(but not too too cheap) after-market cooling to bring down those temps!

Ace · June 25, 2006, 09:56 hrs

Well, you should realize the magical/mystical land of Michiana is a hot place to live.

Thanks, all. I will first do one stick per time, per memtest. I was getting game glitches with the original two 256 sticks, and that's one reason I thought adding another 512 wouldn't hurt. Will see if any are bogus.

And I'll look to see if all my fans are running... Usually they do, as soon as I poast.

I can't tell you how excited I am to go into BIOS and start tinkering. I should probably take a photo of my desktop icons while I still have them.

Should be a busy day. I was gonna go get a power supply, but will see how this goes first.

Ace; now I can tell people "hey, I've got one hot PC!" and they'll be impressed...

News:

Ace's Buttery Flutterball