[Fast]

Hey.

I recently reinstalled Windows XP on my older AMD XP computer. The thing is, now I’m having major issues with the systems stability.

The computer often freezes or reboots, often during even the slightest stress. For example it can no longer run the whole 3DMark 2003 benchmark, either it will freeze during Test 2 (Space Marine or whatever its called) or during the first CPU test (WW2 fighters and bombers). I also downloaded Max Payne 2 Demo and I can not even complete the demo level without the game crashing to desktop or totally freezing up.

I have tried everything, taken out “useless” PCI modules (soundcards, network cards etc), turned off the built in AC´97 soundcard, tested the memory with Memtest 86, tried different drivers for my GeForce Ti4200, updated to the latest BIOS and so on.

I have found nothing wrong with any individual unit (memory, graphics card or processor).
The processor passed the “CPU Stability Test 5.0”, I could detect no heat issues, the processor never went over 49 degrees.

I have noticed something else though, the volt numbers seem off.

(from Everest)

CPU: 1.77v
+2.5v - [b]1.54v[/b]
+3.3v - [b]3.25v[/b]
+5v - [b]4.85v[/b]
+12v - [b]12.41[/b]

When I tried Prime95s torture test I noticed that the volt-numbers started dropping, how much they dropped is hard to tell but the numbers were constantly updating and going down.
When having the following devices connected the test failed after just a few seconds, the whole system froze.
Graphics card, Network card, Soundcard, Harddrive, 2x CD-Burners.

When I disconnected a few devices (I disconnected the Soundcard, and the two CD-burners) the system didn’t freeze but the test still failed after a few seconds, the following message was displayed in Prime95.

”Test 1 4000 Lucas-Lehmer iterations of M19922945 using 1024k FFT Length.
FATAL ERROR: Rounding was 0.4999997969 expected less than 0.4
Hardware failure detected”

According to the readme of Prime95 this could be caused by a number of things, Overclocking (my system is not OC’d), overheating CPU (not likely as the system never reached over 49 degrees when testing with other software), bad memory (Memtest 86 did not detect any errors) or the need of a better power supply.

The above shown Volt-numbers can be compared to my Pentium 4 computer.

(from Everest)

CPU - 1.50 v
+2.5v - 2.49 v
+3.3v - 3.35 v
+5v - 4.95 v
+12v - 12.06 v
VTR - 3.39 v

Quite a difference, no? I tried running Prime95 on my P4, no problems what so ever, I didn’t notice much movement in the volt-stats either.

So I ask you are all these problems caused by the PSU? I have had the same PSU for several years, no problems.. Then this happened out of the blue?
I must add that I have been using Linux with the system (on another harddrive) and I started noticing random lockups while I was doing multiple tasks in Linux. Linux performed without any problems prior to that.

The specs of my AMD

AMD Athlon XP 1900+ (1603MHz)
SDRAM PC133 512MB
GeForce Ti4200 64MB (drivers: 66.93)
HDD = Maxtor 60GB 7200rpm
Moderkort: EPoX EP-8KTA3L+
PSU: 340Watt (Chieftec HPC-340-201)

Thanks in advance.

 

[Fast]

Now I took some better values of the Voltages while Prime95 was running.

When the program crashed or the computer froze, it seems the values of the +5V was linked in some ways, perhaps also +12V.

One time +5V was as low as 4.76 (4.75 is the lowest mark acceptable as I understand it.. differing 5% of the standard value), when it got that low the PC would freeze.
As Everest is pretty slow on the update intervals (about 2 seconds) it is possible that one of the values went lower than the acceptable mark, but I think 0.01 volts over the acceptable mark is low enough.

While Prime95 was running +5v mostly hovered around 4.78 though, +12v went higher dangerously close to the highest acceptable voltage which should be 12.6 (my brother who was observing this together with me said it reached 12.6 once before Prime95 crashed).

Should the voltages really bounce around this much? From the "normal" 4.85-4.90v for +5v it went as low as 4.76v, and +12v which normally sits at 12.35v went very close if not over 12.6v .
I don’t know, I'm waiting for you guys to tell me what you think

Also now I saw the dreaded BSOD while trying to boot WinXp for this test, the computer even rebooted 2 times before even reaching the loadscreen of WinXp.

My computer is ill, very ill.

 

[Rich18144]

I'm not sure if I should be helping here (given that there seems to be an official TSF Hardware team [Spot The Forum Newbie!]) but, assuming that the only change that occured between the computer being fine and the problem beginning was as you say, Windows XP, it seems unlikely that such a change would or indeed could cause such problems.

The only thing I can think of, unlikely as it seems, is that during the install or initial run of Windows XP, the Power Supply was put under some serious stress, which might have damaged some of the hardware inside.

I'm be no means an expert on the subject of voltages, but I would agree that such erratic jumps aren't good for your system.

Is there a Power Supply you can borrow of someone briefly to check if te power is sufficient? It seems that the problems grow more severe when you load the PS with more items, so perhaps checkiung, then changing your power supply is your best option, as far as I can see it.

Still, this is loose, dont take my word as gospel until you hear it from one of the longer standing members of the forum first!! (I don't want to get in trouble the first time I help out!!)

 

[Fast]

Thanks for your reply Rich.
Yep I started noticing the problem first after Win XP had been installed, but when I think about it, the problem might have started a few days prior to that. I had Linux (Ubuntu) installed on the same computer (on a different harddrive though) before reinstalling Win XP. Linux actually froze several times while I was doing what you might call "stressing" tasks, changing between workspaces or listening to music and at the same time surfing the internet and multitasking to other programs, this had never happened before. Linux could be on for days (longest uptime about 48 hours) without any problems.

Anyway I think the problem is getting worse by the day, last night when I was to do some more Prime95 tests Windows failed to load several times, the computer would just reboot and start over. This happened maybe 4 times before Win XP finally loaded, I even saw BSOD once.

Anyway, I think I'll do as you suggested. I'm going to test my P4s PSU (Lite-ON 400W) in the AMD computer, I just have to make sure its a 20pin and not a 24 pin.

If that works I'll just order a 400Watter from Antec.

 

[dai]

also check for a bios update