Need a good diagnostic tool

Teasy · 8 Jan 2007 at 20:42

My PC has been randomly resetting for the last few weeks. Until recently it wasn't happening frequently enough to worry to much about. Now I really think it needs to be sorted out. Anyway I installed a new version of windows in case it was a simple software problem (thought I'd take this opportunity to install Vista for the first time

). Vista is great but the problem is still there. I need to try to find out which peice of hardware is dying so I can replace it. But without parts to try its difficult. What I need is a diagnostic program that can do thorough error checking on my CPU, RAM, GPU, HDD ect and if possible something to test wether or not my PSU is doing its job correctly. Any suggestions?

Teasy · 10 Jan 2007 at 14:18

Nobody have any idea's?

Dr_Evil · 10 Jan 2007 at 14:20

Ultimate Boot CD (UBCD)

t31os · 10 Jan 2007 at 14:22

Memtest or Windows Diagnostic for memory, sometimes memtest can miss things the other gets, but very rarely.

Hdd, get a utility from the manufacterers websites, graphics are normally obvious when they play up, though it wouldnt hurt to watch the temps. Also benchmarks like 3dmark01 seemed to be good at rooting out probs with graphics.

CPU and mobo there's not much way to know with swapping parts around, though cpu issues that cause reboots are usually in my experience heat or incorrect bios setting issues, or possibly a failed OC.

PSU, swap it for another, or run it on bare bone parts and see if things improve.

Teasy · 13 Jan 2007 at 19:27

I tried removing some hardware (DVD-RW, second HDD and case fans). The reboots still occured but less often. Problem is its difficult to tell for sure because the frequency of the reboots does seem to change randomly anyway. One day I'll get a couple of reboots over the whole day, another its happening every 20 minutes. Also looking at the BIOS the motherboard seems to be getting more then enough voltage from the PSU even with everything plugged in (12v rail and the rest are all getting just over what they require). Plus my PSU is 550w, so I would have thought it could handle this system. Still I'll post my specs and exact PSU and I'd appreciate anyones opinion on wether it should be ok or not.

Athlon 64 3800+ X2
Asrock 939Dual-Sata2 mainboard
Colorsit Silent 550w PSU
Connect3D Radeon X1800XT 256MB
2GB PC3200 memory
74GB Sata2 Western Digital Raptor HDD
120GB IDE Seagate HDD
NEC ND-3550A DVD-RW
Jeantech Phong Case (two 140mm case fans)

I think that's everything, appart from the odd USB device taking up power of course, mouse, XBox 360 game pad, 12-1 card reader.

Oh by the way, nothing is overclocked.

masterk · 13 Jan 2007 at 21:42

Teasy said:
Colorsit Silent 550w PSU

Install Speedfan and see what your 12 and 5 volt voltages are in Windows.

I bet they are too low, that PSU does sound poor.

Edit- Thats a 20 quid power supply!!?!?!?!?!?

Teasy · 13 Jan 2007 at 23:26

Low price doesn't always mean poor. But I hope that is the problem, I did get it long before I bought the X1800XT.

mysticsniper · 13 Jan 2007 at 23:41

Teasy said:
Low price doesn't always mean poor. But I hope that is the problem, I did get it long before I bought the X1800XT.

No it doesn't but as its not a well known PSU makers, you do have to question it as a factor. I would certainly upgrade the PSU, as if that blows it can take the rest of your system with it.

If you need to sort some tests out, check out this website

http://www.benchmarkhq.ru/english.html?/be_cpu.html

t31os · 14 Jan 2007 at 02:21

Try running it outside of the case, and i say this assuming you know what you're doing...

Teasy · 14 Jan 2007 at 03:04

masterk said:
Install Speedfan and see what your 12 and 5 volt voltages are in Windows

Installed speedfan:

VcoreA - 1.38v
VcoreB - 1.44v
+3.3v - 3.39v
+5v - 5.16v
12v - 12.16v
-12v - 1.79v
-5v - 2.55
+5VSB - 5.64
VBat - 0.00v

Teasy · 14 Jan 2007 at 03:06

t31os said:
Try running it outside of the case, and i say this assuming you know what you're doing...

Yeah I could certainly do that, but I don't think it would help. Heat definately isn't the problem.

Kerazee · 14 Jan 2007 at 10:37

Teasy said:
Installed speedfan:

VcoreA - 1.38v
VcoreB - 1.44v
+3.3v - 3.39v
+5v - 5.16v
12v - 12.16v
-12v - 1.79v
-5v - 2.55
+5VSB - 5.64
VBat - 0.00v

Are those at idle? run something intensive like a benchmark or prime95 and see if the 12v and 5v drop.

t31os · 14 Jan 2007 at 11:13

Teasy said:
Yeah I could certainly do that, but I don't think it would help. Heat definately isn't the problem.

I wasnt implying it was a heat issue, it could be something shorting, which is likely to cause problems, such as restarts.

Teasy · 14 Jan 2007 at 13:21

Kerazee said:
Are those at idle? run something intensive like a benchmark or prime95 and see if the 12v and 5v drop.

I just ran a Prime95 torture test and Pro Evo 6 at the same time (just to try to get both cores working). +12v went down to 12.10v, +5 and +3.3 dropped by similarly small amounts. Yet the system did reset again pretty quickly during the test, so it doesn't look like its the PSU.

Recently I've started getting an error message when Windows restarts though, so hopefully that might help to tell us what's going wrong:

"The instruction at 0x7715703c referenced memory at 0x00000037. The memory could not be read."

Of course the actual numbers are different each time.

I thought it might be a ram problem so I tried removing one of my 1GB sticks, the problem was still happening. So I put that stick back in and removed the other one, that didnt help either.

Mekrel · 14 Jan 2007 at 13:32

both sticks may be fubar though, small errors sometimes do not get picked up by memtest.

Got a friend who can lend a stick to test?

Garyj881 · 14 Jan 2007 at 17:55

Anything in event viewer?

Teasy · 15 Jan 2007 at 12:34

There's nothing in event viewer but the obvious like "Unexpected shutdown at this time on this day". However I did just open the resource monitor in the task manager. The last panel in resource monitor is memory "hard faults". I've noticed that I am getting the odd 1 or 2 faults listed here for various apps (explorer, svhost ect). When the system was working hard (I think it was doing a restore point or something) I was getting 20 and 30 faults poping up every minute or so. This is with my PC3200 OCZ premium memory down clocked at 133Mhz DDR (266mhz) as well..

Can someone who has Vista please check the memory hard faults panel in resource monitor and tell me if its normal to get such faults, especially if your PC is working hard at the time. I mean I assume it isn't but who knows if this resource monitor is reliable or if its even reporting real critical faults so I just want to be sure.

Teasy · 15 Jan 2007 at 15:48

I just performed Vista's memory test, as far as I can make our no errors were found. Though it didn't actually give me a results window once Windows restarted like it said it would (typical Windows :rolleyes:

).

So I decided to test my main HDD. I downloaded the Western Digital HDD data lifeguard diagnostic program. The first time I started it the program saw both the physical drive and the logical partition. I started the test, but unfortunately got another reset during the process. When I restarted the program would no longer see the physical drive, it just says there is no HDD to list so I can't do any tests. Yet it still lists the logical partition on the drive!! This is driving me up the f'ing wall!

My only option may be to buy new parts, start with the most likely and keep going until the problem stops... Though that's not something I can afford to do right now.

Teasy · 22 Jan 2007 at 15:55

I've just used Ultimate Boot CD to test my memory (Windows Memory Diag) and HDD (WD Data Life Guard Diag). Both passed the test with no errors. Add that to the fact that the PSU stays comfortably above where it needs to be even during stress tests according to SpeedFan and it seems that leaves either my CPU, graphics card or motherboard as the problem. Any other suggestions?

Teasy · 24 Jan 2007 at 15:55

Just used the CPU stress test from UBCD and again no problems, no errors from 10,000,000 iterations. That leaves the graphics card and motherboard which are pretty much untestable.

The strangest thing about this problem IMO is that there is no consistent pattern but at the same time its not entirely random either. I can be playing Pro Evo 6 for hours with no problem, open a Internet Explorer window and it'll suddenly reset. Restart and try again and IE will make it reset again. Then the next day I can surf fine and Pro Evo 6 will make it reset. Or I can be surfing fine and open calculator and start using it only for the PC to reset. Start it up again and try again with calculator and it'll reset again. Some days playing videos causes the problem, some days it doesn't. The cause seems to change from day to day, its absolutely mental.