Lots of problems with <2 month old PC

[colinbell]

Warning, this may be a long post...

Up until about a week ago my computer (which I bought about 2 months ago) has been very reliable and stable, with no blue screens, random resets/locks etc. However one morning, about 20 mins into playing a game I got a BSOD. Thinking it was a one-off I restarted and carried on, however the problem persisted regularly, making the computer completely un-usable.

To sum up the problems:

-Regular BSODs less than 10-20mins into playing a game (usually WoW)
-Random freezes and locks both in games and whilst idling
-Random cut-out of monitor signal

When these problems occur resetting will restart the computer, however there is no signal to the monitor.



To sum up solutions I have tried:

-Update of all drivers, patches, etc
-Complete scan for Viruses, Spy ware, etc
-Complete scan of HDD
-2 hour RAM test using Microsoft’s bootable Memtest ISO
-Over-install of Windows XP onto existing system
-Full fresh install onto new HDD
-Re-bedding of RAM (including socket swap), and video card


So far nothing has worked. Given the complete re-install of OS I would squarely eliminate any software problems. All the components were bought from SCAN (original post here) less than 2 months ago and hence will still be under warranty. As stated originally I have no interest in over-clocking, so unless any components were over-clocked out of the box everything has been running at stock. I'm really at a loss as to what do next. Any advice would be appreciated. My instincts point squarely to the graphics card however I have read stories of other forum users sending their cards back only to be told that there is nothing wrong with them. Any suggestions or advice would be appreciated.

Thanks,

P.S. Sorry if this post sounds really abrupt, it wasn't intentional. Its just quite up-setting to spend that much money, then have it fail so early

Colin

[shadowmaster]

P.S. Sorry if this post sounds really abrupt, it wasn't intentional. Its just quite up-setting to spend that much money, then have it fail so early

No worries, I'd be p****d if my pc stopped working especially after only 2 months.

My bets its either the graphics card or psu. Have you tried using the pc with a different graphics card? Also I have just had a look at your system specs and it could be the psu thats causing the issue. What make is the psu that came with your case? Is it atrix?, because I think that case is an Atrix model but I am not sure.

[zephod]

I actually had a similar problem as regards the monitor - occasinally after restart id get no signal to the monitor. Turns out it was a result of bad feedback info from the monitor stopping the system from recognising the use of a dvi-dsub dongle. Furthermore, using the same monitor, a friend of mine got BSOD using one cable but not another. turns out that particular Daewoo monitor does have known issues with the feedback it gives to OSs. Could be that is the case with your monitor and its actually the fault of monitor/cable/dongle - try swapping these and see if it solves the problem is all i can suggest - else, try someone elses monitor.

Not 100&#37; sure this is cause of any/all of your problems, but then until 4 months ago I would have said its pretty impossible for a monitor/cable to be the cause of BSOD. Turns out nothing is impossible

[Clunk]

You need to list your system specs.

At a guess, I would have said the RAM firstly. Some of your symptoms do sound like the RAM voltage isnt set right, and it is a common problem.

I havent used the microsoft memtest thing, but I would recommend grabbing UBCD and running memtest v1.70 test 5 for a few hours, and then on each stick individually in each slot for at least an hour each, and make sure that you have set the RAM to whatever the recommended voltage is.

If that doesnt work, then the graphics card and PSU next.

[colinbell]

Cheers for the speedy replies.

Have you tried using the pc with a different graphics card?

Unfortuneatly my old computer (the one i'm using atm) has an AGP graphics card, whilst the cheapest PCI-E on at *cough* PC World would be &#163;30, which is a little expensive simply to test it.

What make is the psu that came with your case? Is it atrix?

It is indeed Atrix, which I research briefly before purchasing, and found some fairly positive reviews. However I haven't eliminated the PSU yet.

Could be that is the case with your monitor and its actually the fault of monitor/cable/dongle - try swapping these and see if it solves the problem

Very interesting suggestion, and one I hadn't considered. I will try swapping and let you know if it makes a difference.

You need to list your system specs.

I listed the original post in which I asked for advice in choosing components. In that post is a full list of the items I bought.

Some of your symptoms do sound like the RAM voltage isnt set right

At the moment all the voltages are set to "auto". Whilst checking the voltages I found the CPU voltage was set to about 1.35v but was acutally reading as low as 1.10-1.15v. After some "careful" Google research I found that the maximum voltage of the CPU was 1.65v (IIRC), and that having too low a voltage can cause data loss. Given the expectionally low temperatures I was reading from the CPU (35-38 degrees) from both the onboard sensor and a temperature probe, I felt a slight increase might solve the problem. It didn't. I will research the RAM voltage later. The MS RAM checker is fairly good. I have been told that if a fault isn't found with a standard sweep, then the chances of an error being found on an intensive sweep is less than 0.1&#37;. Given the regular occurance of the problems, I very much doubt the RAM itself is faulty, however I will look into the voltage. Any idea what it should be set to when running at stock? (Corsair TwinX DDR2 PC5400)



I will keep you posted (no pun intended...)

Colin

[Clunk]

I listed the original post in which I asked for advice in choosing components. In that post is a full list of the items I bought. Its just easier if you add your system specs to your "my system" bit.




At the moment all the voltages are set to "auto". Whilst checking the voltages I found the CPU voltage was set to about 1.35v but was acutally reading as low as 1.10-1.15v. After some "careful" Google research I found that the maximum voltage of the CPU was 1.65v (IIRC), and that having too low a voltage can cause data loss. Given the expectionally low temperatures I was reading from the CPU (35-38 degrees) from both the onboard sensor and a temperature probe, I felt a slight increase might solve the problem. It didn't. I will research the RAM voltage later. The MS RAM checker is fairly good. I have been told that if a fault isn't found with a standard sweep, then the chances of an error being found on an intensive sweep is less than 0.1%. Given the regular occurance of the problems, I very much doubt the RAM itself is faulty, however I will look into the voltage. Any idea what it should be set to when running at stock? (Corsair TwinX DDR2 PC5400)

It might say on the RAM itself what the voltages and timings should be, and if not, have a look on the corsair website or email them.


I will keep you posted (no pun intended...)

Colin

.

[colinbell]

Its just easier if you add your system specs to your "my system" bit

Done. I didn't realise it was so simple!

It might say on the RAM itself what the voltages and timings should be, and if not, have a look on the corsair website or email them.

Nothing on the sticks themselves. On the Corsair forums, the voltages seem to vary depending on how much people over clock them... Not many people seem to run at stock. One post suggested 1.9v whilst the maximum is 2.1v. I'll give it a go at 1.9v.



Colin

[Thorsson]

Regular BSODs in games, plus freezes in idle, definitely hardware. I'd guess Mobo or Graphics. Only way to work out if it's the mobo is to swap bits in and out, e.g. try a different PSU, or try the CPU in someone else's system.

If you don't have an available "swapper" I'd speak to SCAN and look to RMA the lot, seeing as you got everything from them.

[colinbell]

Right:

-Tried adjusting the RAM voltage as suggested on Corsair website. With a P5B Mobo they also suggested disabling Legacy USB controller too if memchecker turns up lots of errors. Still crashed in under 5mins.

-Tried swapping monitors, crashed in less than 5 mins.

I also tried disabling onboard sound card, incase of conflict, etc. and it still crashed within 5mins in WoW. However this time it did something interesting: Instead of simply freezing/BSODing/Black-screening, all the characters, items, menus, sprites etc disappeared, leaving only the scenery for 2 or 3 seconds. Then came the BSOD (IRQL_NOT_LESS_OR_EQUAL). For a specific grapical failure like that I'm definetely favouring the graphics card...

[shadowmaster]

You can buy a cheap pci express card on ebay for about £16.

http://cgi.ebay.co.uk/NEW-nVidia-GeF...QQcmdZViewItem

http://cgi.ebay.co.uk/Ati-X300-128mb...QQcmdZViewItem

Or the cheaper alternative could be to borrow a pci express gfx card from a friend's pc.

[Thorsson]

I always keep an old PCI GFX card for this purpose.

You could always heavily *underclock* the GFX card. If it's more stable underclocked that would be a pretty good pointer.

BTW the "IRQL" error message isn't such a pointer as you think. CPU/Mobo/RAM errors could also lead to that message. And of course, power problems could cause any of thise to hiccup.

[Andaho]

I agree with the others that say the problem is with the Ram, Mobo or CPU. If you haven't done so yet, download memtest+ (an updated version of memtest after they stopped support) http://www.memtest.org/

It's a bootable self-run program that will stress test the ram - but errors can also be because of a faulty mobo or CPU. If you were to call up scan and tell them your problem, they'd tell you to run a test with this program.

If you can, borrow parts to test in your system, like the RAM, or test your ram with memtest in another system etc... you should be able to find the problem pretty easially.

[colinbell]

You can buy a cheap pci express card on ebay for about &#163;16.

I'd prefer not to go down that route. If I was going to buy a tester i'd rather pick it up from a shop for this purpose.

the "IRQL" error message isn't such a pointer as you think. CPU/Mobo/RAM errors could also lead to that message

Agreed, this message can be caused by anything from bad drivers, bad hardware or even anti-virues software. The BSOD error messages I get are random, often with no specific error, whilst other times its just a lock or loss of video signal. I only specified the error message incase anyone was interested. What indicates to me that it is the graphics card was the particular mode of failure this time (all the characters, items, menus, sprites etc disappeared, leaving only the scenery for 2 or 3 seconds).

You could always heavily *underclock* the GFX card

I did attempt this after reading that suggestion somewhere else. I used a program called AtiTool. However whenever I adjusted any setting...the computer froze.

If you haven't done so yet, download memtest+

I downloaded another one (Microsoft's Online Crash Analysis - Windows Memory Diagnostic, found here) and let it run for an hour. It too was a bootable ISO. It found now errors after several passes. However I will try removing one stick at a time and see what happens.



Thanks for the suggestions,

Colin

[Blitzen]

Although you are finding no errors, this has 'dodgy memory' written all over it.
Exhaust all memory tests before moving on to test other parts.

(I find it very unlikely to be a mobo or PSU problem).

[colinbell]

Interestingly after taking out 1 stick of RAM it seemed more stable for a while. WoW ran happily for about an hour, which is about the longest since this all started. Ofcourse after that it froze again...

*sigh*


Taking that as an indication that RAM may be involved, I swapped the other stick back in and am currently running the MEMTEST advised by Andaho

[DougMcDonald]

I also tried disabling onboard sound card, incase of conflict, etc. and it still crashed within 5mins in WoW. However this time it did something interesting: Instead of simply freezing/BSODing/Black-screening, all the characters, items, menus, sprites etc disappeared, leaving only the scenery for 2 or 3 seconds. Then came the BSOD (IRQL_NOT_LESS_OR_EQUAL). For a specific grapical failure like that I'm definetely favouring the graphics card...

Interesting indeed, I have experienced this problem myself and have a very similar setup to yourself. Is it only on WoW you see this problem (well obviously the dissappearing things is WoW only, but what about the IRQ not less or equal?)

I see the WoW error, mentioned it to a guild member who said he had the same thing from time to time. Mine is very sporadic and there is nothing particularly which seem to trigger it (time, heat etc)

I have seen the IRQ message on several occasions (maybe 4-5) in 3 months or so on various games, but at no other times, which makes me think GFX maybe. I did wonder if it was nVidia drivers, but i notice your on an ATi card.

I will do some tests and see what happens.