I have a very odd problem with my new build

[Andaho]

I bought all my bits from Scan in August.

This post is longer than I was expecting, so I've laid it out to make for easy reading...

The basic specs are:

E6750
Corsair TwinX 2GB 800MHz
Gigabyte GA-P35-S3 motherboard
2x WD5000AAKS SATA hard drives
Seaonic S12II PSU
X1950Pro PowerColor graphics

Firstly, some background on the hardware:

This may or may not be related to the cause of the problem:
When it all arrived, the heatsink on the motherboard was hanging loose at one end and could see the spring-loaded clip had broken and was loose in the bag. Not wanting to send it back for something that appeared superficial; it was easy to repair by using cable ties: One through the heatsink and board and held with the head of another cable tie on the back. This appears to hold the heatsink in place perfectly. I would like assurance that this won't have voided the warranty, as it can easially be snipped off to leave it in it's original condition - I have kept safe the 2 parts of the broken clip together with the spring.

To the problem:

I'm finding this problem very hard to diagnose because the PC functions perfectly most of the time; I'll describe how I have noticed it in steps:

The first symptom of the problem:

I first noticed that the hard drive activity led stays on permanently after I have been using the PC for a while. Although the pc still works fine, it was driving me nuts me not knowing what was causing it, and being new to vista I started a thread thinking it was something vista was doing. With this happening, I haven't noticed any performance problems. I found that restarting the PC would stop it happening sometimes, but not if I had the PC on for a long time.

After ruling out Vista causing the hd led to be on permanently, I ran the Western Digital HD Diagnostics to test both hard drives. No errors.

The second symptom of the problem:

When I have been using the PC for a long time, the hd led will be permantly on. When I restart the POST screens take a lot longer than usual: It hangs on "Memory Runs At Dual Channel Interleaved" for around 20-30 seconds (that normally only hangs for 2-3 seconds), then at "Detecting IDE Drives . . ." (that normally happens in a split second) It hangs for vairying times, from 5 seconds to over a minute! Windows will then boot normally but the HD LED is still permanently on during all this. However, doing a hard reboot, the problem goes away.

The third symptom of the problem, and most serious:

I've just recently installed and started playing a few games, and I've found this brings on the problem very fast. Also, it's introduced a new symptom that I'm sure is related; When I've been playing for only a short time, the led will permanently stay on. After a little longer, at random times on any game, the game will hang for a few seconds upto several seconds, the sound will repeat in a fast loop, just as if the game is going to crash, or as if you'd expect a BSOD to appear, but then the game continues as normal. It happens regularly, the more often the longer I play for, until it happens so often that I have to do a hard reboot - then the problem soon comes back again. This is what has prompted me into finding and fixing the problem which could be related to the motherboard?

Other notes on my system:

I am dual booting Vista and XP. I have my games installed on XP and only really use XP for games, most of the time I use Vista. The symptom is not related to the OS as it happens on both. They are installed on the same HD on seperate even partitions and a western digital diagnostic on the hd shows it is healthy.

I'm very confident this is not a virus problem as I rarely even come across them (I have nod32 antivirus), and the problem has been happening since clean installation, happening more regularly over time.


Please, has anyone got any ideas what is causing the problem?

[TooNice]

Any overclocking?

[edp33]

Not sure if the HDD led is always on from your post. If it is, check the LED connections, HDD, Reset switch etc from the case to mobo and also try different sata ports and power cables. Also try just testing one drive at a time, unplug the power and data cable to say the Vista drive and try the XP drive in different ports with different cables etc.

Is the main power cable from psu to motherboard fitted correctly as well.

Also check event view in Vista and XP see if any HDD event errors are being registered there.

On another note the HDD led might be a diversion, the real problem could be the memory if it's getting stuck on the dual channel memory interleaved message. Try a slight increase in the memory voltage.

[Andaho]

I don't overclock. The hd led only comes on permanently after a little time being switched on. Everything is fitted correctly.

Thanks for the suggestions: I will disconnect my second hard drive and I'll plug my main HD into a different port with a different cable, then I'll clear the event log in XP, replicate the problem by playing a game, and see if there are any reports in the event viewer.

[peterb]

I would be concerned about the heatsink clip on the CPU. The clips exert quite a bit of pressure to ednsure good thermal contact between the surfaces - more than I suspect you could achieve with a cable tie. The consequence could be that the thermal compound has not spread itself evenly and there are hot spots on the CPU die, and/or elevated temperatures. Whether that is the cause of the subsequent problems I can't say, but it would be worth fixing that first.

[Andaho]

I would be concerned about the heatsink clip on the CPU. The clips exert quite a bit of pressure to ednsure good thermal contact between the surfaces - more than I suspect you could achieve with a cable tie. The consequence could be that the thermal compound has not spread itself evenly and there are hot spots on the CPU die, and/or elevated temperatures. Whether that is the cause of the subsequent problems I can't say, but it would be worth fixing that first.

It's not the heatsink on the CPU, it's the heatsink on the motherboard. The fixing that broke was a very light spring loaded plastic pin. The cable tie is holding the heatsink flat on the chip, just as the original pin/clip would have been.

[Thorsson]

I'd be worried that the broken pushpin was a symptom not a cause. A symptom of some rough handling en route. Motherboards are pretty delicate. I'd RMA it.

[s3ds]

have you checked the cpu temps just after a crash ?

[rabbid]

Or the chipset temp.
has the surround or chip been damaged from the sink bouncing around being held at one end only in transit?
pls check that the sink is mounted flush with ur cable ties, if its pulling slightly up one way then if the chip overheats it will cause instability problems. Your symptoms indicate problems under load or longer periods of operation, suggesting a temperature based issue.

[Andaho]

I'd be worried that the broken pushpin was a symptom not a cause. A symptom of some rough handling en route. Motherboards are pretty delicate. I'd RMA it.

Yeah, thats what I was meaning to say somewhere in the original post, that the mobo could be damaged. And pushpin! that's the word I was looking for

Or the chipset temp.
has the surround or chip been damaged from the sink bouncing around being held at one end only in transit?
pls check that the sink is mounted flush with ur cable ties, if its pulling slightly up one way then if the chip overheats it will cause instability problems. Your symptoms indicate problems under load or longer periods of operation, suggesting a temperature based issue.

It doesn't appear to have any visible damage on the board anywhere, and I did ensure I was mounting the heatsink flush and flat to the chip.

If I was to RMA it now, I'm worried that returns might not like that I fixed it with cable ties? Also, I'm worried that when they test it, it might not show any fault, because it's an intermittant fault.Can anyone from scan say anything about my concearns? or has all this thread been too much for anyone working to read through?

[rabbid]

have u re-applied thermal paste or used the original?

it may cause issues as you have remounted the sink with non standard fixings & used the board since August?
When its a DiT/visible defect issue you should report it right away, not use the product and where possible retain packaging for inspection.

I know thats not so helpful & we all want to use our new kit right away! You will just have to put yourself at the mercy of Scan/Mobo manufacturer.

Anyway, check the cabling to the drives regarding LED issue. Run just one drive and see if it happens, then both, try to pinpoint which causes an issue.
Use memtest to see if your RAM has any problems too.

[Thorsson]

Just be aware that errors on either memory or HDDs could be caused by the motherboard, the "glue" that holds the whole PC together.

If changing memory slots or SATA ports fixes the issue then you know it's the mobo, but I doubt you'll be that lucky. My bet is on a snarfed Northbridge.

[Giraffe]

Just one point: at even low temperatures a cable tie will soften slightly; also plastic expands more than metal does. A stainless steel tie (if not shorting anything) might do the trick, but that spring clip is really need to get the force. A tie doesn't have much elasticity.

[Andaho]

Since disconnecting my second hard drive, and changing the port of my first, I haven't been able to make it happen! And that's a great sign, because it was getting as regular as happening after having the machine on for only 10-20 mins.

As for the heatsink comments... it used no thermal paste, and the spring-loaded pushpins that are used to hold them in place are very weak and loose springs, and apply hardly any pressure... nothing like a CPU heatsink! I am 100% sure it is a flush and flat fit on the chip, and to proove that, if the mobo themostat is under the chip, it was reading only 33 degrees after running under load for at least an hour.

I'll report back with my findings after I plug in my second hard drive again (into a different port). I'm happy if it doesn't show any more signs after this. It's a little worrying knowing there's something dodgy there, but if I were to encounter any other more serious weird and hard to diagnose problems in the future, I'd probably straight away just replace the mobo at a cost of under £60.

[SHare]

Your biggest problem is that you didn't report the damage on arrival so scan are under no obligation to help you out. If you use something, you've accepted it. I believe scan will help you as they are a good company, but its just a word of warning. I'd RMA it, anything you do to fix it means you can't RMA if it fails later.

[peterb]

It's possible that the HDD connection was a bit iffy. If the problem doesn't happen again with the drive on a different port, swap it back to the original and see if it repeats. If not, then it was a dodgy connection. (If it happens with the drive on a different port, it could be the drive at fault) If it does happen on the original port, then that would indicate the mobo.