PC Problem diagnosis required please...

[dern]

Having asked last week what motherboard to buy I ignored all the advice and bought what was in stock and now I have a problem

I've plugged the whole lot together and installed windows 2000. After from 1 minute to 5 minutes the pc locks solid and won't respond to ctrl-alt-delete or anything, just the reset button.

The system I have is...

Abit KD7A motherboard (Bios rev 15) (requires 2 power supplys, the normal big block and a 4 pin accessories block)
Athlon XP2400+
400DDR memory
Floppy drive
Seagate barracuda 80Gb 7200 rpm hard drive (IDE 1 master)
52 speed generic cd drive (IDE 2 master)
SumVision KY-500 ATX 400W P4 powers supply

I have no serial/usb/parallel devices plugged in and the monitor is powered from a seperate supply.

I've upgraded the latest bios (rev 15) after getting the problem but it still occurs. I've reformatted the drive and reinstalled w2k and the problem still occurs. I managed to install sp4 (it stayed working for long enough) and the problem still occurs.

The voltage readings I get from the bios are...

Core = 1.66v
ATX 3.3 = 3.28v
ATX 5 = 4.99v
ATX 12 = 11.97-12.02v
AGP = 1.55v
DDR = 2.65v
Battery = 2.97v
Standby = 4.94v

I left it off all night and booted it up and it locked a minute after w2k booted and the temps weren't high so I don't think it temp related.

Any ideas?

Mark

[Munter]

I had a system that would "Screen freeze" at any period between 1 min and a few hours. Your problem sounds similar.

I solved mine by upgrading the mainboard and processor. I'd ask to swap it with another one if you can. (Did you get it from the net or a high street shop?) Or contact support for the mainboard.

HTH

[Whoozit]

I would try the simple things first (well, they're the only ones I know )

Is everything plugged together properly? Try reseating all the components. Check the PSU is earthing properly. Is this happening if you don't start up the OS (unplug the hard drive). Try only one memory module if you have more than one. Take out any extra components e.g. graphics card (if there is onboard graphics to use as a replacement). If it's only happening on loading the OS, what's in the event logs? Uninstall or remove from Startup anything that automatically loads when the OS starts.

[plotloss]

Its a hardware fault.

Download memtest86 and burn it onto a CD rom or floppy and boot from that.

It will test each bank of memory quite quickly and that will isolate the memory.

If its not memory (who supplied it, generic or branded?) then my next move would be to try an alternate power supply to isolate that.

If neither of them turn up anything obvious then more thought is required...

[dern]

Cheers, I'll test the memory tonight and report back.

Regards,

Mark

[ultimasimon]

Highly unlikely, but just a thought. You have all the symptoms of thermal runaway - I had a customer bring his machine round after building it, and it did exactly the same thing; namely freeze after a couple of minutes.

The cause? He had left the peel-off heatsink compound label on the cooler causing the heatsink to be insulated from the cpu die. But as I said, unlikely.

Other things to check:
1) Try another PSU
2) Boot up without the hard disk, and see if it is hardware of software.

edited for:
check the underside of the mobo isn't grounding out due to screw pillars in the wrong place. So run the pc as a pile of bits on an insulator out of the case, and boot with the minimum configuration - ie cpu and fan, memory and graphics card only.

>> Edited by ultimasimon on Wednesday 25th August 15:18

[jimmyc412t2]

plotloss said:

Its a hardware fault. Download memtest86 and burn it onto a CD rom or floppy and boot from that. It will test each bank of memory quite quickly and that will isolate the memory. If its not memory (who supplied it, generic or branded?) then my next move would be to try an alternate power supply to isolate that. If neither of them turn up anything obvious then more thought is required...

I agree. Strange issues like this are almost always memory. And it doesn't mean the memory is specifically faulty either. In my experience, outfits will normally swap you memory due to the fact that some brands don't like certain boards.

If its non-branded cheap stuff then this makes it even more likely. Personally, never had a problem with Crucial matched memory. Good quality middle of the road stuff that IMHO

OH! - And check that there are no memory speed related settings. The board seems to support 400 DDR but then only states 333 CPUs. there may be something to investigate there maybe not

>> Edited by jimmyc412t2 on Wednesday 25th August 16:00

[dern]

jimmyc412t2 said:

I agree. Strange issues like this are almost always memory. And it doesn't mean the memory is specifically faulty either. In my experience, outfits will normally swap you memory due to the fact that some brands don't like certain boards. If its non-branded cheap stuff then this makes it even more likely. Personally, never had a problem with Crucial matched memory. Good quality middle of the road stuff that IMHO OH! - And check that there are no memory speed related settings. The board seems to support 400 DDR but then only states 333 CPUs. there may be something to investigate there maybe not

The memory is cheap unbranded stuff but I've had it for a year so they probably won't have it back. It worked fine in the last system right until it broke down hence the requirement for a new motherboard (symptoms were random windows program errors worsening to blue screening worsening to a complete failure to acknowledge the on button).

The motherboard recognises the memory as DDR400 if you set it to SPD (auto-detect) and if you put it on manual there are quite a few settings which I don't understand yet and have no way of knowing what would be the correct ones for my memory stick (can't recall seeing a handy sticker listing them when I plugged it in the other day).

I'll run the memory tests later this evening over a beer and report back.

Thanks,

Mark

[plotloss]

Ahh, with that nugget of information in mind (blue screening etc on the last machine) and the memory is the same I would say that I am 99.9% sure that its the memory.

Still, on the upside you have a spare motherboard that works...

[dern]

ultimasimon said:

Highly unlikely, but just a thought. You have all the symptoms of thermal runaway - I had a customer bring his machine round after building it, and it did exactly the same thing; namely freeze after a couple of minutes.

It doesn't run under windows for long enough for me to install anything to monitor the temperature while running windows but if you leave it in the bios monitoring bit the temps rise gradually and then settle so that seems ok.

The cpu is a 2400+ XP and I bought it as a set with an AMD approved heatsink and fan. Admitedly I have seperated the sink and the cpu in order to install the cpu in the new board but the heat dissipating gunk was still all there so I didn't bother putting any new stuff on (mainly because I couldn't find it).

Mark

[dern]

plotloss said:

Ahh, with that nugget of information in mind (blue screening etc on the last machine) and the memory is the same I would say that I am 99.9% sure that its the memory. Still, on the upside you have a spare motherboard that works...

I replaced the oldest stuff which was the motherboard and the drive and discounted the memory because it's so new.

Feck... I'll give it to a mate.

Cheers,

Mark

[dern]

Hmm, memtest86 ran all of its tests for 6.30 hours and no errors...

Mark

[drummie]

Just a thought.
Wiring. or Conflicts.

Read on if you have the energy!

have you got the S/PDIF wired in properly. Some of these cheaper motherboards require you to "wire up" the connectors yourself.(they don't come in a block) Also, I had a similar problem with my Gigabyte GA-81PE1000 Pro2 board when I first plugged in the USB2.0 extra port. It wouldn't boot. (and this one was wired in a block for us all!! you'd think it was easy . . .)When I unplugged it and turned it on, no probs at all. Sometime, these little things make a big difference and soundcards (especially built in ones) are always a problem. I've had to disable many in my time.

All the problems I've had with memory and CPU's (I've built about 500 computers) have been pretty quick to show themselves. Boot-up time is a massive load on the system, the problems usually show themselves there, or just after. Try overclocking a CPU. It will usually(NOT ALWAYS!) crash just after boot, if it boots at all. So, if you are getting your comp up and running for say more than 30 mins, try disabling the soundcard for a day. See if the system hangs then. If it does, try disabling the USB and keep working through the system. It's gotta be something conflicting with something else. Or even itself!!
Good Luck
cheers.
peace.
D.

[lanciachris]

Make sure you install any software drivers that came with the motherboard.

I had a pc once where it would do what yours is doing because those drivers were missing - although you could make it run stably but slowly by disabling the internal caches on the motherboard.

[dern]

lanciachris said:

Make sure you install any software drivers that came with the motherboard. I had a pc once where it would do what yours is doing because those drivers were missing - although you could make it run stably but slowly by disabling the internal caches on the motherboard.

Done that, installed the latest from abit's website.

Thanks,

Mark

[dern]

drummie said:

All the problems I've had with memory and CPU's (I've built about 500 computers) have been pretty quick to show themselves. Boot-up time is a massive load on the system, the problems usually show themselves there, or just after. Try overclocking a CPU. It will usually(NOT ALWAYS!) crash just after boot, if it boots at all. So, if you are getting your comp up and running for say more than 30 mins, try disabling the soundcard for a day. See if the system hangs then. If it does, try disabling the USB and keep working through the system. It's gotta be something conflicting with something else. Or even itself!!

Basically it'll run for as long as you like in dos/bios but windows locks it up. It's a clean install of w2k withing nothing added (except the updated drivers) and on the second install sp4. The only thing I haven't installed a driver for is the onboard sound card (as I have a fortissimo to plug in) so maybe I'll give that a go.

Thanks,

Mark

[Podie]

Re-seat the memory... sometimes resolves these issues...

[dern]

Bit more info...

Seems to run ok in safe mode and in safe mode with networking. Downloaded a load of critical patches but made no difference. Disabled the onboard sound but still no difference.

What's the difference between safe mode and normal?

Mark

[goboddog]

running mem test for 6.30 hours wont always show any faults even if the mem is dodgy....i agree with a few people id say it is ur mem but could be a temp issue try running it with the case open...also try and clean any dust etcand check that any wires are not blocking vents...post the temps as well

[dern]

I figured out what this was this weekend. I took the case apart to reseat the memory and check for loose connections and so on and found that I had installed a sound card and then subsequently completely forgot about it... well it was last thing at night. Taking this card out got rid of the problem. Sadly no amount of updating drivers has enabled me to use the card (a fortissimo 3) so am currently using the onboard sound.

If I can be bothered at some point I'll start researching IRQs and all that stuff.

Sorry for the misleading info I mistakenly gave and thanks for all your help.

Regards,

Mark