System Restarting Abruptly

[dante12345]

Hi all!

I have the following rig -

CPU - Core 2 Duo e4500 2.2Ghz( OCed to 2.6Ghz, now running at stock again)
Motherboard - Gigabyte P35-S3G
Ram - 2x 2GB Corsair DDR2 800
GPU - MSI 8800gt 512mb Factory overclocked
PSU - Corsair VX 450W
HDD - 1x 320GB Seagate 7200rpm running Linux Mint 17.1 64bit
1x 1TB Samsung 7200rpm running Windows 7 32 bit and storage
Case - CoolerMaster RC330 with one case fan plugged into the motherboard
Keyboard - Basic PS/2 TVS Keyboard
Mouse - Microsoft Comfort Optical Mouse 3000 USB, with both the clicks not working properly. The left click has degraded in the past month or so

For the past few months, my system has been restarting abruptly while using it. First it restarted randomly and booted back up to either OS. Then slowly the problem got worse and it kept restarting more often. Then it stopped even going to the BIOS and the fans would just spin on and off. After getting the motherboard repaired, it seemed to be fine. But after a few weeks another problem popped up! The keyboard started behaving erractically and a few restarts after the system wouldn't go past the first BIOS screen. It would simply keep restarting. I got the motherboard repaired again, thinking that the previous guy didn't do it correctly. The repair guy returned my machine just this sunday after repairing it and it seems the abrupt restarts are back. They happen after fairly long usage and the system restarts back fully to the OS. I have tried the following to rule out some components -

1) Hard Disks - Both OS are installed separately on each hard disk. Suspecting it was an OS problem, I tried both the OSes but the system still restarted abruptly.

2) Graphics card - When the system was not booting past the first BIOS screen, I replaced my 8800gt with an old ATI card that can run just off the pci-ex slot. Unfortunately, it still didnt boot past the BIOS screen at that time. So I reckon the graphics card is not at fault here.

3) Memory - I will be mem-testing it today overnight. One peculiar thing I have noticed is that in the BIOS, there is a value called DDR18V which is 1.9V. In HWMonitor, VIN1 is showing up as 1.936V.Assuming VIN1 and DDR18V are signifying the current ram voltage and that my RAM is rated to run at 1.8V, could there be a problem here?

This leaves out the motherboard and the power supply.

1) PSU - I haven't tested the PSU myself yet as there is no one I know that has a decent power supply. The repair guy doesn't have a PSU with a 6pin connection for the gfx card (odd i know!)to test whether there's a fault with mine. My hypothesis here is that the PSU is damaging the motherboard over time. Is there any way to test the PSU? I can get access to a multimeter for voltage measurements,etc.

2)Motherboard - Either my repair guy is making a fool out of me or he is not diagnosing the fault correctly. He said that this time he changed the BIOS chip but the problem seems to persist. The only improvement is that the system is fully usable after the abrupt restart. It might be that something on the motherboard is not working properly which is damaging other components. What could that possibly be? How could I check it?

I have already spent 20 quid on repairing this board, but being a greenie, I dont want to simply throw it away. Recycling is pretty crap where I live but repair's cheap. Moreover, getting a decent future proof machine going to cost me atleast 200 quid which I dont have spare right now.

I have attached a HWMonitor text file in case its helpful. - http://speedy.sh/5b6Bc/AbruptRestart1.txt

Please let me know if you have any suggestions on what could be the problem and how to go about diagnosing and fixing it. Thanks!

 

[C12Friedman]

Although I would be fairly quick to point to the PSU here, there really is no way to test it's proper function. A multi-meter tells only about voltages when not under load which may not correspond to actual voltages when the PSU is under a load. What are the voltages reported in BIOS? The ram running at 1.9V instead of 1.8V shouldn't present a stability issue (unless the ram is getting hot). Another common problem would be that the PCIe slot has become faulty but without a second PCIe slot there's no way to check that hypothesis.
You might try taking the side panel off and point a fan at the open PC to see if that helps stability, check chipset for heat (ground yourself first then, place slightly moisten back of finger against chipset heatsink - you should be able to keep your finger on the heatsink with only slight discomfort, it should not burn)
In re-reading this I'm not so sure I'm being helpful...

 

[dante12345]

Thanks for your reply.

1) I have noticed that the RAM does get quite hot. I did an overnight memtest and it didn't result in any errors so I reckon that rules the RAM out.

2) I'll have a look at the motherboard chip the next time I run the system. Would HWmonitor be able to show the chip's temperature? I think the sensor list might be in the file I attached.

 

[dante12345]

So I got my motherboard repaired only to have it go bad again. Thinking it was the electrics in my house I moved the system to another house to test it overnight with prime 95. I ran the small fft and blend tests and it never restarted. But when I ran the full power consumption tests, I came the next day to find that system had restarted. When I moved it back to my place and ran it for an hour or two, it would randomly freeze while browsing the internet. I have now given the motherboard for repair again, but the repair guy says the replacement chips are not readily available so I am thinking of buying a replacement. But this makes me wonder if the power supply keeps damaging the motherboard or that the motherboard itself wasn't repaired properly in the first place. Any suggestions on how to diagnose the power supply would be much appreciated.

 

[C12Friedman]

Do you recall what was repaired on the motherboard, only the CMOS got swapped? If part of the VRM then I would be looking at the PSU as a possible source of your problem but there really is no way to properly test a PSU without some rather expensive equipment - see this article for further information http://www.hardwaresecrets.com/why-99-Percent-of-power-supply-reviews-are-wrong/ . Although written for reviewers, it explains the methods for properly testing a PSU, diagnostics is even more in-depth as this article reflects http://www.tomshardware.com/reviews/psu-repair,4147.html . I suspect your PSU is now out of it's warranty period... hmmm

I'm thinking your system is ~7 years old which is around the life expectancy range of 7-10 years, I'm further thinking your PC is on it's death bed at this point. You know, 6th generation Intel (Skylake) processors are out now (subtle hint).

The VRM (voltage regulation module) on your motherboard lacks any heatsinks so I'd be looking hard at that area of the motherboard for possible damaged components

Did I mention Skylake is out? Of course, Socket 775 motherboards are a dime a dozen now so replacement of your motherboard should be cheap, I think I'd start there.

 

[dante12345]

Thanks for the links. They are rather informative! I was looking for something like this to make sure the repair guys I approach know what they are doing. This should be helpful

Regarding what was repaired on the motherboard, to be honest the original repair guy kept making excuses everytime I asked him what he had repaired. The last time (which was the second time the mobo got repaired) he said he was replacing the BIOS chip. But he was saying that the BIOS chip is hard to find for my particular motherboard and that I should upgrade to another motherboard. If I am not mistaken, BIOS chips are available quite abundantly, it would be more difficult to find southbridge/northbridge chips. Anyway, when I got the machine back, it started restarting abruptly again, just a day or two after! When I talked to him on the phone saying that it was happening again, he made excuses saying that he only has time to check the system for 10 minutes, not hours after which the problem occurs. And that he doesn't really know what the problem is anymore. *CURSES!!* Anyway, I'm hoping the current guy would be able to tell me exactly what the problem is by tomorrow. So I'll post it here then. I'll also put up a photo of the insides of the power supply to see if anyone can pick up anything wrong with it.

I would gladly upgrade to a newer system if I had the money right now. Unfortunately, where I live, the cheapest intel quad core with ddr3 ram and a cheapish mobo costs about 200 quid which I dont have right now.

 

[dante12345]

Here are some photos of my power supply -

There's something odd that I noticed when I opened it up. In photos 3 and 4 there's a ferrite coil. When I moved that coil a little it felt looser to the touch than the one next to it. I thus took a photo of its connection to the PCB and it seems as if a connection is broken. One can see it clearly in photo3 and 4. Is this an electrical connection that I should get fixed or some kind of plastic connection to the PCB that doesn't matter as much?

 

[dante12345]

Here's the link to the photos -http://imgur.com/a/pTBXs

 

[C12Friedman]

The coil being loose could very well be the cause but the MOSFET's may have suffered for it. If you have someone willing to work on the PSU, I'd say that's a good thing - it may be that the solder joints simply need a touch up. I'd think it'd be worth it to fix if needed.

 

[dante12345]

But from the photo does it look like to you if an electrical connection broke or just a plastic connection to the board?

 

[C12Friedman]

I can't really tell from the photos but coils are held in place (most commonly) at four solder joints, if the coil is loose, it would seem to indicate that one or more solder joints has failed, that's not a good thing.