Oddest computer f'up I've ever seen

[Robor]

I'm guessing motherboard. I know I've seen this issue before and while I can't say it was a bad board for sure I'm pretty sure it was. Haven't had a CPU or RAM go bad that I can remember but boards... Ugh!

 

[MDesigner]

Well, how can I troubleshoot this? Besides buying a new motherboard..

 

[trexmgd]

I was experiencing the exact same thing.

I traced my problem to a Java issue with Azureus, a bit torrent client.

 

[MDesigner]

Interesting. I was running Azureus at the time. How did you resolve the issue? Reinstall the Java runtime?

 

[MDesigner]

Did some research.. looks like it might be Azureus. I haven't run Az in days, and my machine has been totally stable. Google revealed many other people having the same prob, but I wasn't able to find a resolution yet.

 

[wetcat007]

Do you have an nvidia card? newer nvidia drivers have issues with java.

Try This:
1. Start->Control Panel->System
2. Click the Advanced Tab
3. Click the Environment Variables button
4. Click the New button
5. For Variable name:_JAVA_OPTIONS
6. For Variable value:-Dsun.java2d.d3d=false

That solves a lot of problems with java it seems, by making it simply not use direct3d for everything.

 

[MDesigner]

Done. Although I don't know how much Direct3D Azureus uses...

 

[MDesigner]

OK, so the issue is basically a system lock-up. It still happens, and I just RMA'd my 3D card and got a new one.. I thought maybe the graphics card was defective, because sometimes right before lock-up, there are weird dots or graphical glitches on the screen. Nope.. it still happens. I'm running Prime95 to test the RAM.. it's been running for about 5 hrs or so. How long do I keep it running?

 

[us3rnotfound]

how about updating your motherboard's bios?

 

[BucsMAN3K]

Originally posted by: MDesigner
The PSU is an Enermax Liberty 500W, with active PFC. I doubt there are any power issues there.

PS, let me clarify something.. my computer never goes into standby or sleep mode. I don't use those. It's just my LCD monitor that gets put into "power off" mode after 30 mins of idle time. So basically what happened here was, the LCD monitor was in standby.. and the computer locked up. So moving the mouse or hitting the keyboard wouldn't wake the monitor up.

Hmm, I wouldn't think that power supply would have a problem...BUT, what you described is exactly what happened to my old computer when my PSU was pretty much...about dead. I switched out the PSU and everything worked fine again.

Have you tried switching the PSU just to make sure?

 

[MDesigner]

I don't have a spare PSU to test. You know, I was running Prime95 all day..when I got home tonight, the machine was locked up. I'm going to keep running Prime95 every day.. if the machine locks up every single day (it usually only locks up once every couple weeks or so), then it's very likely the RAM. Though I don't know what happened. The RAM was fine in my previous computer. I hope pqi has a lifetime warranty or something like that.

 

[MDesigner]

Turns out it's the RAM. I ran memtest86 three times in a row.. locked up every time, at around 30-35 mins into the test, on the same test (test #8), at about the same position in the pass. Time to contact pqi about an RMA...

My two cents: Prime95 is useless. It ran all night and part of the next day without crashing.. didn't really tell me anything. memtest86 told me within 35 mins (times 3) that my RAM was bad.

 

[letdown427]

P95 isn't useless, it's a CPU tester mainly, that's why RAM wasn't instantly diagnosed.

memtest86 is for RAM.

Glad you found a solution.

 

[Accipiter22]

Originally posted by: MDesigner
I've built quite a few PCs, and have experienced all kinds of issues.. but this is something totally new I've never seen before.

I was out of town this weekend, and was trying to connect to my computer at home via Remote Desktop. No luck. Hmm..very weird. Even if the machine blue screen'd, it will reboot and everything will be accessible again. And the network was up & running fine.. so I got home and checked it out. The monitor was in standby mode, and the computer was completely unresponsive. All fans were spinning normally, there was no power to the USB mouse (the fancy lights were off) or any other USB devices... basically it's as if the damned thing was turned off, but it wasn't. No power output, no VGA output, nothing..just fans running as usual.

Anyone ever seen this?? What the heck causes this and how do I even begin to troubleshoot this? It happens totally randomly, it seems.

I had the EXACT same problem...I had to replace every single F'ing component on the computer before it worked again......try this first though, as the problem with mine was that the mobo was grounding out.


1. Remove all unnecesary components from mobo, then take it out, leaving it still hooked up to the power source and monitor
2. set mobo on floor,
3. try booting up
4. what happens now?

I did this after I had replaced everything, and it started up fine..turns out a screw under the mobo and my network card were grounding out the comp for some reason...now this wasn't the original problem, but I probably could've saved on shipping costs of replacing EVERYTHING, and only replaced a few components

 

[bob4432]

Originally posted by: MDesigner
Not a RAM issue, most likely, because these two sticks were used in my previous machine which ran with absolutely zero issues. Now they're in my new machine.. same sticks. Still, though, if I were to run a RAM test on these two 1GB sticks, which RAM test should I run? memtest86 is fine but.. geez, it takes like a good 2 days to run a test, I don't have time for that!

just run memtest overnight and see what the shows up

 

[zachtos]

random HARD lockups are usually faulty RAM.
random reboots are usually heat or drivers
random hard lockups or hardlocking only while gaming, usually video card

memtest tests ram, prime95 will test cpu/fsb, futuremark will test video,everest can see mobo temps/voltages in windows, a digital multimeter will test a power supply for voltages but it is not 100% accurate because these can change under full load.

The hardest to solve crashes I've ever had were a combo of a bad PSU which damaged my mobo. So both caused instability. the mobo gave bad voltages on the 5V rail which caused it to crash when the video card started running full load in games but not cpu only intensive apps. very tough one that was.

 

[wetcat007]

Originally posted by: MDesigner
Done. Although I don't know how much Direct3D Azureus uses...

Java otherwise uses direct3d for everything it seems, and azureus uses java.

 

[gsellis]

If the problem does not go away after replacing the faulty RAM, this is where you are. ACPI, incorrectly implemented at the BIOS level, will not wake from suspend correctly. You would flash your BIOS to the latest version, and then in the power settings in BIOS and in XP, turn off power management (while on AC in XP). If the BIOS does not correctly detect keyboard or mouse wakeup, it will not and appear as it is.

A telltale for lockup (bypassing keyboard/mouse) is if the system does not respond to the power switch. The power switch also is tied into ACPI. A push of the power switch also send a wakeup or suspend/off message. By holding the power switch for 5 seconds, the off is forced through the BIOS (and that is a surer sign that it locked up.)

A faulty HDD might not wakeup also (ignores or does not get the start signal).

Also, it is not recommended that you "turn off" ACPI. ACPI APIC adds the benefit of better IRQ handling. If you really can force ACPI "off", you may run into more frequent issues with IRQ conflict resolution (does not mean they will happen, just increases the chances.)

Sidenote - Vista Install DVD now has a memory tester as one of the startup options. MS Online Crash Analysis team made their own as a significant portion (meaning less than 20%, but more than 5%) of crashes can possibly be memory issues.

 

[MDesigner]

Yikes. It turns out it's not the RAM after all.. it's the motherboard's DIMM slots that are bad!! How do I know this?:

OK, I've got four DIMM slots.. 0, 1, 2, and 3, we'll call them. I have two sticks of pqi RAM, and I happen to have two sticks of backup Kingston RAM. With the pqi RAM in slots 2 and 3, I get lockups. memtest86 causes a lockup after 30-35 mins. So I took the one module out of slot 2, leaving just one module in slot 3. memtest86 goes through several passes, NO problems. Totally stable. I put both pqi sticks in slots 0 and 1.. now memtest86 actually gives me errors highlighted in red.

I take out the pqi, put in the Kingston modules into slots 2 and 3. memtest86 gives me errors.. and I know that the Kingston RAM is totally fine. In fact, the pqi RAM was working great in my last machine before I transplanted them to the new motherboard.

So is this possible? For DIMM slots to be "bad"? DIMM3 seems to be just fine, so I'm using just one of the pqi 1GB sticks in that slot now.

 

[dman]

Originally posted by: MDesigner

So is this possible? For DIMM slots to be "bad"? DIMM3 seems to be just fine, so I'm using just one of the pqi 1GB sticks in that slot now.

Maybe bad voltage getting to one of the contacts? or one of the traces going to the slot is scratched/bad solder/etc.

Would reset bios and test memory at lower settings just to confirm it's the board and not the configuration/settings.

--

I had similar problems with an upgrade to my wife's pc from an Intel BX to Via400 AMD system. It would flicker the display (like it was suspending) and pause / lockup for a while. The first part was fixed when I repair installed XP after I noticed that ACPI was not installed correctly (I never did a clean install).

The second problem I'm still waiting to confirm it's fixed. I changed the PSU. The old one didn't have a dedicated 12v 4pin connector and was using an adapter which I think at certain times may have been insufficient. So far, so good on that theory. [edit] Nope, did it again today, screen blanked and then a few moments later was back to normal. [/edit]

 

[MDesigner]

I will reset the BIOS and mess w/ the RAM settings. And possibly flash the BIOS too. Thanks

 

[MDesigner]

I got a replacement motherboard. Ran memtest86 for 8 hrs, everything was fine. Computer is a lot better than before.

But just the other day, I had a lockup again. And just now I did too, though the RAM tests out just fine. This wouldn't be a bad CPU, would it? Do you think it's just leftover system "funkiness" from when the RAM was giving the OS issues? I'm going to reformat this evening, I think/hope that will clear up these issues.

(OS "funkiness"..besides the lockups, yesterday all the icons on my Start menu were those default blank page icons with the Windows logo on them. Also, one time I booted up Windows, I got some error message saying something about a file in the registry was corrupt but it was repaired)

 

[gsellis]

Originally posted by: MDesigner
I got a replacement motherboard. Ran memtest86 for 8 hrs, everything was fine. Computer is a lot better than before.

But just the other day, I had a lockup again. And just now I did too, though the RAM tests out just fine. This wouldn't be a bad CPU, would it? Do you think it's just leftover system "funkiness" from when the RAM was giving the OS issues? I'm going to reformat this evening, I think/hope that will clear up these issues.

(OS "funkiness"..besides the lockups, yesterday all the icons on my Start menu were those default blank page icons with the Windows logo on them. Also, one time I booted up Windows, I got some error message saying something about a file in the registry was corrupt but it was repaired)

It could be something basic including a device driver. Check your eventvwr 's system log and see if the stop was listed. It should at least list the stop number.

 

[MDesigner]

Interesting. I posted that message a few mins after the lockup. My event log shows some weirdness at 10:25am, 4 mins before my post.. about 17 warning messages in the "system" category:

Event Type: Warning
Event Source: Disk
Event Category: None
Event ID: 51
Date: 6/16/2006
Time: 10:25:57 AM
User: N/A
Computer: SALTYDOG
Description:
An error was detected on device \Device\Harddisk0\D during a paging operation.

You know, i'm onto something now.. I Googled around and saw others having the EXACT symptoms. Thanks for mentioning this, I would never have thought of it.

 

[MDesigner]

OK, I ran chkdsk /R. It fixed a few errors. Then it was checking the files and it said they were fine. Then, last stage was to check free space. It got mostly through that (or maybe all the way) and it said something about CHKDSK unable to fix volume bitmap due to insufficient space. I have a 160GB HD with 54GB free. Is it important that I free up space and run chkdsk /r again? How much free space do I need, anyway?