Input on this problem. May be RAM related

[Stoanhart]

Hi,

First off, I'm not sure this is in the correct category, so mods, move away I posted it here because the RAM is one of my main suspects in this problem, and this forum gets a lot of viewers.

Alright, this is regarding my brothers computer. He has:
-Soltek SL-75FRN2-L NForce2 200Mhz SocketA motherboard
-AthlonXP 3200+
-512MB Samsung PC3200 RAM (just cheap value stuff)
-Built by ATI Radeon 9500 Pro
-WinXP Pro SP2

His computer was working quite stable until he recently clicked on a virus that he had caught over MSN. The virus caused his computer to freeze, and upon rebooting, windows would no longer load. It would just reboot during the boot sequence. I tried restoring to last known good configuration, and to boot into safe mode, but no luck. So next, we decided to reinstall windows. We did so, and booted just fine. I though all was well, but soon my brother came to me with random reboots.

My first suspect was a poor PSU. Then I remembered he was using the same as me, a ThermalTake PurePower 420W. Not exactly high-end, but not generic cheap stuff either, and mine's been great. Anyways, I set up motherboard monitor to log every second, and told him to use his computer and told him to call me for the next reboot. He did, I checked the log, nothing. All voltages fine. 3.3v=3.38 5v=5.05 12v=12.27.

Next I realised that since he had just reinstalled, that windows was probably set to autoreboot on BSOD. I disabled that feature, so we started getting BSODs. Further proof that the PSU is OK. The BSODs pointed to various drivers. A different one on each reboot. So I decided, a full format was in order. However, while running booted off of the XP CD during the format, we got a BSOD. Hmm, couldn't be the drivers or a corrupt XP install then.

My next guess was heat. His CPU is hot, but his HSF is rated for a 3200+, and he is below AMD max temp. His on die temp is 60-65 MAX. I had a spare 120mm fan laying around, so I jurry-rigged it onto his PC with elastic bands )) and pointed at the CPU HSF and the chipset, which is passively cooled. He was able to get through a format no problems now, and I though great, problem solved, we'll get a new HSF.

However, it wasn't that easy. He was still getting unstable preformance, and more than normal crashing. I decided to run a Memtest86, since it is cheapo RAM. After a full night, it had found five errors. Hmm. Not a huge amount, but there should be zero. I thought maybe it wasn't quite able to handle PC3200 speeds at stock voltage so I went to 2.7. Ran for another night and got the same 5 errors. I now have a primary suspect. I have some more RAM coming soon, so I will be able to try that out, as I will have a spare 512 module.

Anyways, he tried installing NHL2005 today. The install went well, but in the game, there are several graphics glitches. The players will kind of flicker, or turn into black shadow shaped at the wrong angle to the ice every once in a while. The shadow shapes get bigger and smaller. I doesn't really hinder gameplay, since it only happens every once in a while. However, after about 2-3 minutes, or at the most 10, the game will either freeze, crash to desktop, or cause the computer to reboot. Normally I would say this is probably the RAM problem, but this video card has had a traumatic experience. I posted this in the "What's the stupides computer mistake you've ever made" thread:

I was testing out a video card for someone. It was a 9550 pro, so I devided to use my brother's computer since his video drivers wouldn't need to be changed. I put the computer on its side, took out the 9500pro, and laid it on the drive cage. I put in the new card, and turned on the computer. Everything boots and works well. Good, I thought, now I'll put the old one back in. I look down, and to my horror realise I had forgotten to remove the MOLEX plug from the side of the card! It was laying on the metal drive cage, all solder points short circuited together, POWERED! I went to carefully pick it up, and the second I moved it even one millimeter, the whole thing erupted into spaks. I pulled the power cord from the PC as quickly as I could, reinstalled the card and turned it on. It worked just fine since.

So, with that in mind, I moved the 120 mm fan so it was covering the chipset, and the video card, because I really don't think CPU temp is an issue. No change. I updated the BIOS and restored BIOS defaults, no change.

So, if you're still reading, to recap:
1) RAM, likely, shows errors
2) GPU, likely, but less so. Has worked well since "incident"
3) Temp, unlikely, since it has never been an issue before.

I just find it strange that there were no problems before the virus! Can modern viruses actually cause hardware damage like some were able to years and years ago? I was under the impression that that wasn't a concern with modern viruses.

Thanks in advance for any input!

 

[David]

Sounds like a RAM issue to me. If it works with different RAM then there is your problem.

If an area of RAM is dodgy it will often corrupt whatever is stored in that area of RAM. Which can vary randomly depending what is stored in the faulty area at a given time. On boot this is likely device drivers.

 

[Stoanhart]

True. I just think it's weird that it was stable before the virus. Coincidence that it died at the same time, or can a virus actually damage ram?

 

[Ironsave]

Did You just format the drive, and then install? Or did u use a program like Darik's boot and nuke?( It writes 0's and 1's to the drive, govt. standard). I have had virus' that survive a reinstall before after just reformatting the drive. It has happened to me, and the only solution was shredding the drive before reinstalling, using a program like the one above.

 

[Stoanhart]

No, I just formatted it. Maybe I'll give that a try if the RAM doesn't fix it, and I can convince my brother to format...again. Thanks.

When my brother ran the virus, it send itself to everyone on is MSN list. All the people that ran it didn't have this severe of a problem. They just updated their AV program and it got cleaned out.

 

[Stoanhart]

Well, here's what's happened since.

Not RAM. Got my new sticks, tried em out, same problem
Not Videocard. Swapped it out, same problem.
Not Cooling. Tried pointing 120mm fans at it @ 12v, didn't fix anything.

The problem is, the CPU. Of that I am certain. Prime95 quits in about a half a second, even at "stock" speed. The reason "stock" is in quotes is right here.

I think the chip really is a 2500+ running at 3200+. I have upped the voltage to try to stable it out, but I can't hit 3200+ speeds. I drew the line at 1.75v, at which point prime would last maybe 10 seconds. Lowering the clock speed by 90mhz and keeping the voltage has just about stablised it. It ran 39 minutes in prime, but then my motherboard monitor alarm went off because the on-die temperature hit 70c. It did produce an error after 39 minutes, but I'm not sure whether that is heat related, or it is still clocked too high.

I think what we will do is get an xp120 HSF and strap on a nice, high-flow 120mm fan, and just try to get the thing as fast as possible in order to try to get my brother as much of his money's worth as possible. Of course he did pay $100 less than retail on ebay, so it's not a total rip. Still, frustrating.

I find it weird that it would be stable for all these months until he got a virus, though. I never prime95'd it, because I thought it was stock, but it never crashed.

oh well, 1 bad ebay transaction out of many, many good ones.

 

[bobmanfoo]

xp120 doesn't fit on AXPs, temps shouldn't get that high, have you tried reseating the heatsink?

 

[kukyfrope]

Are you certain that it's not a motherboard problem? Like have you tried all of his removable components on another board (RAM, Video Card, Processor, in that order) ??? It sounds like a general instability for almost all components which I think could be the result of an unhappy mobo.

70C is way way too high for any CPU chip, regardless.

Try some components in another system and let us know the results.

 

[Stoanhart]

It sounds like a general instability for almost all components which I think could be the result of an unhappy mobo.

It seems like it's the CPU to me. I mean, none of the other parts affect the problem, where are uppingthe voltage and lowering the clock rate has brought the computer from 'crash within 2 minutes' to 'my brother can reliably finish a couple of games of NHL2005'

Tonight or tomorrow I will remove the HSF, check the tags and look for bridge mods, as stated in the other thread, and then reseat the HSF to make sure it is getting propper contact. 70c it quite too high, I agree. It only does that under Prime, though. When my brother plays NHL, it goes to 55c, which I can accept.

The motherboard is brand new, and he just bought it to fix a different problem with his old motherboard. I sure hope that the mobo is not the problem. I don't have any other 200fsb SocketA motherboards, though.

 

[kukyfrope]

Don't forget that changing the FSB of the system often ups the rate at which all other components will run as well, with the exception being your PCI/AGP devices, granted the lock on your board works. Keep in mind that if the board itself cannot handle these higher speeds, you will still be getting instability. It could possibly be the CPU, but as you said, without another board/CPU it is very hard to diagnose. Unfortunatly new motherboards don't always function perfectly out of the box either

 

[maelstromracing]

have you tried the CPU in another computer?