Pinpointing a hardware fault - Help !

Associate
Joined
20 Dec 2005
Posts
42
Hi everybody,

Right, I have a problem with my 2 Year old setup that is kind of hard to pinpoint. So I thought it might be a good idea for some of you guys to share your knowledge of what you think is dying in my PC without me spending hundreds on parts I might not need.

WARNING this 1st post might get a little long but I wanna try and cover all bases as to what the problem might be. Anyway:

Currently I am running :

Abit AN8 Ultra (latest Bios)
AMD64 +3700
Crucial Ballistix 1GB DDR400 Pc3200 (BL6464Z402)
Winfast 6800GT Ultra PCI-E
Antec Phantom 300W (Not the most powerfull supply but I like the fanless aspect :cool:)
Antec P180 Case
Seagate 300GB SATA 7200rpm HDD
Audigy2 Platinum Sound card
WinXP


Apart from a few memory problems I had with a bad batch of Ballistix memory, the system has run as smooth as ever with hardly any crashes etc.

Until last week...

I was using ITunes and a couple of IE pages open when the system turned off. Tried turning it back on but the fans just whirred a bit and I got no Video or POST beep.

So I opened it all up, took out all unneccessary components and still it would not pass the POST. Giving me an error code of 0.4. User guide mentions something about Memory error. I noticed that one of the 3 case fans was not spinning, only on higher voltages does it start to spin! I also heard the Video card's fan which didn't sound to healthy as it kept trying to spin up like a HDD and was a bit noisy.

So I swapped the Ballistix with different memory in different combinations and still had the same error. I borrowed an old PCI video card which got it to start booting Windows. It got to a chkdsk screen so I thought I would let it run the scan, however it suddenly jumped to a black screen with mouse pointer and then rebooted.

It then continuously rebooted without getting into full windows. All Safe Modes etc would reboot.

I then used a spare SATA HDD to install a fresh copy of Windows to see if I could salvage any data off the old HDD incase it got corrupted in any way. (All my data is backed up every couple of days anyway but would be nice if I didn't have to rebuild Windows yet again)

So I got to the install options for Win XP and as it was formatting, came back with an error saying that the (new, spare and fully working) drive cannot have its partition formatted as the drive is damaged !!!

I tried an old IDE drive and that booted once then rebooted and couldn't find NTLDR !

Since then I have downloaded the HDDs' manufacturors DOS utilities to perform tests and have even zero written to the first and last sectors of the SATA drive. I then formatted it with this utlity and tried to install WinXP. It copied the files and then on first reboot it came back with an Invalid Partition Table.

So after pulling all my hair out what do you guys think could be the problem?

I don't think its the HDD's although now I think there is something wrong with the partition tables. (I won't put my orginial HDD back in as I don't want it to get corrupted if there is something wrong with the PC).

I reckon my gfx card is dead anyway but have no way of knowing as I have no other rig with PCI-E.

But even when its not in the system im still having trouble. Could the main culprit be a faulty motherboard after the PC initially died? Or could it be the lame power supply which might not be powerful enough (it has worked ok so far though).

I would like to hear your view on the best way to pinpoint the problems from the symptoms above as these sort of problems are the worst!! Or even if you guys know of any good utilities I can run to test out the system from DOS woudl be helpful.

Any thoughts or help would be appreciated.

Thanks for reading !
 
Okay so I would definetly say it's not related to the hard drives or the cables. I'd doubt it's the controller for the drives themselves, but it's worth keeping in mind for the moment (ICH* stuff controlls the SATA and IDE lines from what I remember, but I'm not sure how Nvidia chipsets work in that respect).

Keep in your old PCI graphics card whilst trying to find this problem, as the last thing you want is getting confused with 2 seperate problems from the graphics and from the rest of the system. Also make sure you remove the possibly dead card from the machine, and only use your PCI card.

Firstly I would reset the CMOS, you can do this by removing the motherboard battery for around 10 minutes and then replacing it. Or you can short the jumper if the motherboard has one, location of that will be in your manual.

Make sure the system is down to barebones also, but enough for it to boot. Such as remove the sata drives and ide drives, and also as much memory as possible and only your PCI graphics card.

I would advise memtest'ing your memory if you can, since your posting this I'm guessing you have a spare machine therefore I would visit the memtest86 website, download the ISO image and burn it to a disk. Then put it in the machine and let it do about 20 passes on the 1 stick of RAM before concluding that the stick you are testing is okay. If you hang in memtest, then remove that stick and try another in a different slot and try it again. If it's still hanging in memtest then I would look elsewhere as I doubt the memory would be the cause if it's gone through 2 different sticks and on 2 different DIMM slots.

When you boot the computer, I would quickly enter the BIOS and check the temperatures of your CPU and motherboard as well, and make sure they are within acceptable limits in case they are starting high and eventually overheating (could have just happened to due a clog up of dust or something). I would also check over your PSU rails whilst in the BIOS, and make sure they are acceptable as that could be the cause. Ideally for that circumstance however I would advise if you have the facilities to test the PSU on another machine.

Overall though, I'm going to put it down to either the PSU or the motherboard, and if I had to pick between those I'd say PSU, but run through the steps anyway and work through trial and elimination.
 
Thanks for the quick and details reply Flibby.

I pretty much have everything down do barebones but have not tried Memtest yet.

When I first opened up the case there was quite a fair bit of dust as I probably have never given it a full dust down since building. The gfx card NEVER got a dust down and did have a fair amount on the copper and the fan grill. This card is now out fo the system as I can't get past POST 0.4. with it in.

After changing lots of HDDs and changing the memory I have reset CMOS a couple of times - even left the Battery out. The date is currently in 2005 so I know it all reset ok.

I had an idea to check the power rails from BIOS but was unsure what sort of variancies of 5V and 12 V are acceptable.

I will note down the figures and run Memtest and reply back.


Thanks for your input
 
I had an idea to check the power rails from BIOS but was unsure what sort of variancies of 5V and 12 V are acceptable.
5% deviations are fine, so:

5v: 4.75 - 5.25
12v: 11.4 - 12.6

Also remember to double check the temperature settings of your CPU and motherboard when you are in the BIOS, to make sure that your system isn't just overheating.
 
OK, my system has been on for about an hour and current temps are:

CPU - 31c
SYS - 37c
PWM - 38c (Not sure what PWM means)

Power Rails are:

CPU Core 1.44
DDR 2.62
DDR VTT 1.31
nforce4 1.59
nforce4 stanby 1.52
Hypertransport 1.22
CPU VDDA 2.5v 2.58
ATX +12v 12.32
ATX +5v 5.09
ATX +3.3v 3.40
ATX 5VSB 5.21

The only difference is that my power hungry 6800 isnt plugged in as I have a standard 2D card in PCI.

I am just downloading memtest now and will post the results in a bit.

Thanks for any replies
 
Not sure what PWM means
PWM is for the MOSFETs. They are small chips that are normally located around the CPU socket and are used for regulating volts. You can read up some more on those if you wish here.

The temps look fine, and so do the rails. So unless it's a intermittent problem with the power supply, I'd maybe point to motherboard, but lets see what the results of the memtest86 tests are (ps. the memtest86 tests will problem take about 4 hours, just let it run for 20 loops. If you start getting errors however, just stop the test and record the info and we can go from there).
 
Thanks for your input once again.

Memtest is running now. Done 1 loop with no errors so far.

I am thinking of maybe booting up the system with another IDE HDD that I have spare. Going to see if this one gets corrupted or crashes.

If that is the case I will try to get a replacement board.

Do you know if ABIT will still issue an RMA for this board as I think they only get 2 years warrenty and i have had the board for just over that.


Thanks
 
I very much doubt they will allow you to RMA the board, even if it's just over the warranty date, but there is no harm in trying :) Sods law as well for the motherboard to possibly break not long after the warranty expires.
 
I have got an older ATX one, but it hasn't got the 4 pin connector that goes into the MB to feed the CPU.

Will it still work without ?


Cheers
 
Does it have:

20 Pins

4 Pins 12V Rail

OR

Just 20 pins?

You will need to check your motherboard manual if it has the 20Pin + 12V 4pin as it might work... I presume your current PSU is ATX2? 24pin & 4pinv 12V?

Michael
 
Actually yeah you are correct. My current PSU is 24 pin + 4 pin.

My old PSU is only 20 pin but it does in fact have the 4 pin connector.

So will this old PSU work even though the connector only has 20 pins ? (I would try it now but I am still running Memtest. Currently on Pass 6 and no errors so far)

Is there any risk to damaging my system (even more) using it.


Thanks
 
Hi,

I have yet to change over the power supplies, however I just reformatted an old 20GB IDE HDD on a spare machine and made sure it was all ok. I then plugged it into my broken machine and started the Win XP installation.

It copied the files and then went for first reboot to start loading off of the HDD. However as with another drive it crashed as it started to look for the boot sector.

Does anyone else think that this is definitly a damaged motherboard?

Thanks
 
So will this old PSU work even though the connector only has 20 pins ?
No, if use a 20 pin in a 24 pin motherboard the system will not POST.

markuk80 said:
Does anyone else think that this is definitly a damaged motherboard?
From reading up on the problem from start to finish and what you have tested, and if the memtest86 report came back with 0 errors after 20 loops, then I would be looking at changing the motherboard, however I would not like to say for definite.
 
Also here is a quick shot I took of some weird characters that appear after it skips over booting from CD and starts to boot from HDD.

http://img339.imageshack.us/img339/9045/picture002gd4.jpg

Hopefully the link will work (first time I have done this)


Thanks
Ack, that doesn't look good. Some kind of corruption, and only thing I would imagine would cause that is either the CD, CD Drive, hard drive or motherboard. Hard drives are fine since the problem persisted through sata and ide drives, and not the CD / CD drive I'd imagine as the problem originated when the OS was already on the machine. So that would leave motherboard. If the CPU was dead the machine wouldn't even be booting.

I noticed you mentioned you have the latest BIOS on your motherboard, did you upgrade it very recently or a long time back?
 
Yeah, I think it is something with the mainboard as I am certain that there is nothing wrong with that drive. I only just formatted it on another machine.

Its something to do with the board's controllers or something, not being able to read the boot sector of HDD's. It can read/write other parts but falls over at the boot process.

The BIOS is version 20 or something. Date is 16/01/2006. I don't think there is any newer version out for the AN8 Ultra. I upgraded it at the beginning of this year. Have had no problems or crashes since.

Looks like its new motherboard time then. Might make a new thread later to get some specs for a partially new system. Got to learn all this new technology again..Bah!

Anyway thanks to Flibby and gamesaregood for helping me with this problem.

Cheers
 
Back
Top Bottom