Computer freezes up frequently

jamesman32

Limp Gawd
2FA
Joined
Jan 11, 2009
Messages
445
Just got some parts for a new build, and here are the specs, before I go into details on the problems

Phenom II 955
2x2GB PNY Optima DDR3 1333
Foxconn A7DA-S 3.0 motherboard
Corsair TX750 PSU
Dual XFX 4890's
WD 500GB HDD (used previously in another build, know it works)
Soundblaster Creative Audigy 2zs

All parts except for the HDD and soundcard are new

Now, problems started as soon as I finished the build. The rig would artifact in windows and playing games. After a day of troubleshooting, it turned out to be 1 stick of RAM. By this point I had tried numerous drivers and reinstalled windows, but it seemed to run fine after that, except every few hours, sometimes less, the system would just freeze up. No BSODing, just audio stuttering and a freeze. I couldn't CTRL ALT DEL out of it either.

Now I have new RAM on order, but before I bank on that solving the problem, is there anything i can try? I've run 3DMark06, and that ran just fine. Prime ran fine too with the good stick of RAM, and had no errors, until the system froze after 14 iterations of Prime. EVERYTHING is at stock clocks and volts. Temps on the CPU have never gotten above 60c at full load, and GPU's stay at about 70c at load. Both are within the acceptable range, I think.

Sometimes when the system freezes, but not always, I will reboot and get a CMOS checksum error, or an error telling me to reset the time. Thinking the CMOS battery could be dying, I tried to replace that as well, to no avail. Frankly, I'm out of ideas. What shall I do?
 
1. Test CPU with intel burn test
2. Test RAM with memtest (overnight).
Report results.
 
Oh I forgot to mention, I did Memtest for long time yesterday, and it came back fine. Also ran check disk on the HDD, and got no errors. The CPU has been tested with intel burn in, but like I said with Prime, it just freezes before too long.
 
Run for over 5. If it fails prime with stable RAM then that sounds like a bad CPU or something.
 
It is Prime stable. It hasn't failed a test with the bad stick of RAM out of the system. It just crashes before I can really test it for more than an hour. I don't get any errors, I don't BSOD, it just hangs and freezes.
 
in the summer time, this is a very common complaint and yeah, for the past couple years its been the 3/4/5/6 year old prescott armed with that dinky P4 cooler bogged down with 3/4/5/6 years worth of dust. The first thing I try to establish with such cases, as Joseph is trying to establish with you, is weather the problem is heat related.

Generally speaking, if it is, it'l rear its ugly head once you get every component up to TDP power consumption. My favorite way to do so is with Orthos on a custom memory test (stressing 70-80% of the system memory [doing this makes the computer nearly totally unresponsive as its now using the page file very extensively, none-the-less if numlock still turns the light on and off it means your CPU is still crunching numbers]) and furmark to get the GPU temperature up. Running those tests will get your RAM, NB, CPU, and GPU up to maximum thermal output.

And if thats where it fails I'd say theres a good chance that its heat related.
 
I've tried almost every heat producing, number crunching test I know of, and the funny thing is, the only time it has frozen the past 2 days has been when I was just browsing the web or listening to music. I could game for hours on end, loop 3DMark, run Prime, Orthos, or burn in, but it'll either keep going, or it will just freeze. I've monitored all my temps. My case ambient has never gone above 40c. Processor has hit 61c max ever, and GPU's don't touch on 75c. I'll run more tests if you suggest it, because I need the advice, and that's what you're giving.
 
The "at the same time" is the crucial bit. Get all the parts up to maximum power consumption at the same time.

That said, if your saying that your sure it isn't heat related that my next guess would be shot memory silicon, but then the custom orthos test usually grabs that in minutes, and a few hours of memtest would almost certianly get it. If you've done both its safe to say its not in memory.

So now I'm thinking a cap of some kind has blown and taken some support system on the motherboard with it. Something like a blown cap in relation to sound processing once caused similar conditions to what your talking about --that said the guy with the problem was known as the "roflcopter" because he laughs alot and the quality of his voice in skype/ventrilo was always aweful. Without an obvious sign its hard to test such things, the only real way would be to RMA the motherboard.

Then it could also be the power supply, but testing that requires a multimeter or another power supply. If you do have another power supply or wouldnt mind paying the restocking fee to newegg or whoever, popping in a new power supply might help.

Short of that the problem could be in windows in software. You might want to try using a partitioning program (after defragging your drive to move all the data to the front of it) to make a small (10-15GB would do) on the drive and installing windows to it, then trying to get it to crash again. Before you do that it might be a good idea to use some software to read the SMART (god I hate these acronyms) data off the drive. Theres two programs that do that, Speed fan is one, and theres a second and much better one, but I can't remember its name right now. A $80 program called spinright would almost certianly identify any problems with the hard drive if any exist and core logic is known good (give any program, even ones written in assembler, bad memory and they'll give you crap in return).

Problem could also be environmental. I had a guy who blew three of my PCs (or rather, the same PC three times, with three seperate PSUs) and was furious at me for suggesting that the problem was with the wiring in his house. Turned out (thanks to some unbelieveably sub-par previous contracting) that there was bare un-insulated copper contacting some soggy drywall, combined with some rough power he took out 3 of my power supplies. A multi meter and probes to the wall would find any defects in your supplied AC.
 
Aside from the post above (I think he covered just about everything)...
Have you installed drivers for everything on the motherboard (...and everything in general)? I didn't install some stuff on my laptop because I didn't need them and it gave me problems until I disabled them.
 
Thanks for all of the options! It's really very helpful, and I'm greatful you took all the time to write that out.

For environment - I've tested the PC in 2 houses, 1 of which is owned by a professional electrician, and I still had the same problem, so I'm guessing that's good as out.

For the PSU - I have a friend with the same model. I may ask him to borrow it, or something similar.

I'll install a partition of windows tomorrow morning and see how long I can run it without a crash

I'll check over the motherboard and add on cards for blown caps in a few mins, I had a few go on an old PSU once, that was fun, lol.

As for the memory, I'll run tests again for the hell of it, and I have new RAM coming in a day anyway because the other stick was known to be bad.

I had also heard I should try to flash the BIOS again and see what happens, but Foxconn isn't great about making their BIOS files readily available. :(
 
I ran Orthos this morning, custom blend. It passes 2 iterations, and froze after 4 minutes. The CPU was at 43c with 2 cores at 100% usage. I also checked the device manager for hardware problems, and noticed my soundcard wasn't working right. Trying to install drivers froze the system up, so I figured it could have been the card. Took the card out, and it's still freezing.
 
just ran into your post a couple weeks ago,ive been having the same problem ,i might even go as far as to re-install windows xp.just wanted to know if you had figure out the problem?
 
Any graphical artifacts are a sure sign of a bad video card. I assume you use crossfire since you said dual video cards. Try with just one in and stress it, if no artifacts great! Try with the OTHER card in and stress it, if no artifacts then your motherboard might be damaged. Since you already said you tried multiple drivers I am ruling out that problem. I had an old nvidia 7900gt that presented the same problem as your 4890. XFX was more than willing to replace it with a 7950 512mb...

good luck!
 
Back
Top