Sign in with
Sign up | Sign in
Your question

Intermittent shut-down problem

Last response: in Systems
Share
December 27, 2009 5:05:26 PM

Hi guys,

For the past 2 weeks I've been battling with this problem with no success and I hope someone can point me in the right direction.

Components:
Intel Core i5 750 2.66GHz Processor
2 x Crucial 4GB (2x2GB) DDR3 1333MHz/PC3-10600 Memory
Antec EarthWatts 750W Modular PSU
Western Digital WD5000AAKS 500GB SATAII 7200RPM 16MB Cache
Gigabyte GA-P55A-UD4 P55 ATX Motherboard
ATI HD 5770 1GB PCI-E Graphics Card
CM Storm Scout Case - 120mm and 2x 140mm Fans Included
Sony AD-5240S 24x DVD±RW DL Internal SATA Black Bare Drive
Windows 7 Home Premium 64 bit

Problem: After completing the build the PC worked correctly for about 4 days before the fault reared it's ugly head. The PC powers up correctly at first but at any point after switching on it will spontaneously power down without warning. This happens intermittently and often doesn't happen until after hours of operation (i.e. yesterday it was over 8 hours before it happened again) - although sometimes it happens after only seconds :-S All parts are brand new.

Symptoms:
- Random shut-downs with a warning about closing apps *if* I make it into Windows.
- Sometimes it will shut down after 2 seconds, sometimes 2 hours.
- Will only power up again if left unplugged for a minute.
- Only began to happen after 4 days of normal operation.
- Any USB devices such as portable hard drives still have power even if the pc is left off (not sure if this is normal?)
- Computer will sometimes switch on by itself!

Steps taken:
- PSU, MB and CPU have all been replaced, no change.
- All RAM (except one stick) has been removed, then swapped - no change.
- Hard disk has been separated from the system, no change.
- All case fans, LEDs and CD drives have been separated from the system, no change.
- CMOS has been reset, no change (also swapped the battery but as it's brand new i shouldnt need to).
- Checked CPU heatsink/fan was correctly seated.
- Appears to be no burning on the pcb's or melted capacitors.
- Tried swapping around PSU cables between devices, no change.
- Monitored CPU temperature, never rises above 43c.
- Unplugged all USB devices.

I think all that leaves is my video card and case, which haven't been swapped yet. I've tried breadboarding the system and removing the video card (and it stayed on for a good 20 minutes) but for all I know it was just another minute away from shutting down again.. after all it *has* stayed on before for a good 8 hours before happening again.


Any help you can give me would be really appreciated!
a b B Homebuilt system
December 27, 2009 7:31:34 PM

Just a thought - are your power and reset buttons good? Your Breadboarding did not produce random re-boots.

Try disconnecting the reset connector wire from the motherboard's front panel header, leave the power connector alone for now.

Let the system run to see if it re-boots. If not, replace the reset switch on your case and reconnect the reset connector wires. If it does randomly re-boot, disconnect the power connector wires from the font panel connectors on the board and replace with the reset connector wires from the case front panel.

Let it run again. If it continues random re-booting, replace the power switch on the case. If it does not randomly re-boot, we have a different problem.

I basically replaced my entire computer last year because of random intermittent shut-downs and re-boots, using my favorite case. It was a bad power switch - ~$2.00 momentary switch from radio shack fixed it.
December 27, 2009 9:05:31 PM

Thanks very much I'll try that and get back to you in the morning :)  It did occur to me that it might be something to do with the power/reset switches but since they're built quite tightly and securely into the case I thought I couldn't test them.. obviously I wasn't thinking straight! I'll give it a go!
Related resources
December 27, 2009 9:20:39 PM

On second thought I've just managed to get it logged into Windows and I started typing some gibberish into a Word document (in order to produce a warning about closing an application with unsaved data) and I tried pressing the power and reset button to see what would happen:

- Power button: Brings up the identical warning about losing unsaved data but it doesn't go anywhere beyond this screen (i.e. it doesn't shut down). When the shutdown occurred in the past this screen appears but not for more than a second before the system shuts off.
- Reset button: Does a straight system reboot.

If a faulty switch was causing the problem would this method cause it to happen like any other time? Just in case I should clarify that my fault is a straightforward full shutdown, as opposed to a reboot.

Thanks again for your help!
a b B Homebuilt system
December 27, 2009 9:41:36 PM

Heres a thought: TEST YOUR HARDWARE

MEMTEST, PRIME95, LINPACK, ATITOOL, FURMARK, WD DIAG TOOL ETC
a b B Homebuilt system
December 27, 2009 10:41:21 PM

mkroberts said:
On second thought I've just managed to get it logged into Windows and I started typing some gibberish into a Word document (in order to produce a warning about closing an application with unsaved data) and I tried pressing the power and reset button to see what would happen:

- Power button: Brings up the identical warning about losing unsaved data but it doesn't go anywhere beyond this screen (i.e. it doesn't shut down). When the shutdown occurred in the past this screen appears but not for more than a second before the system shuts off.
- Reset button: Does a straight system reboot.

If a faulty switch was causing the problem would this method cause it to happen like any other time? Just in case I should clarify that my fault is a straightforward full shutdown, as opposed to a reboot.

Thanks again for your help!


Replacement switches are cheap compared to new hardware - you might want to try the steps I outlined earlier to confirm the switches are ok. Based on your tests, you might want to start with the reset switch connector.

I had the BIOS on my system set to power-on after power fail, so it re-booted after shutting down. The power and reset switches are momentary switches, they close the circuit and then immediately re-open it. My bad reset switch would randomly close the circuit, shutting down the system. I have no idea what caused it to do so. Once I isolated the shut-down and re-boot to the reset switch, I replaced it. Before that, I had double-checked everything numerous times, and finally replaced the entire system into the same case. Since the case was the only common factor, it had to be something with the case - that left me with the front panel connectors.

January 6, 2010 9:16:11 AM

I've attempted following the instructions you gave me but it had no effect. What's more annoying is that for the past 4 days its ran fine and only just started happening again. :( 
January 6, 2010 2:06:06 PM

maybe its some windows problem.....try reinstalling the windows, or upgrading the bios.....also check the temperature levels....
a b B Homebuilt system
January 6, 2010 6:14:04 PM

Truly a head-scratcher, so I'm thinking out loud. If you had a CPU, RAM, or driver-related hardware problem, it would either not boot or would produce a BSOD. And you replaced the PSU, so it's not a shorted or loose power cable between the PSU and board.

Your statement that it randomly shuts down...and has to be off more than a minute before re-starting sounds like something has to cool; maybe an electrical connection closes a circuit from heat expansion and has to cool to condense the contact points enough to re-open the circuit and allow the system to startup again.

Possible causes of an electrical short that would power-off your system: improper grounding of the PSU, compression, vibration, or heat.

Is your system properly grounded - power cord through a surge protector to a properly grounded electrical outlet? Reliable electricity, no routine brown-outs or spikes?

Is your CPU heatsink assembly properly attached to the board - firmly, not loose, but not too tight? If you have a backplate, is it metal and insulated from all metal points of contact on the board's backside? Did you over-tighten the screws?

Are the 24-pin and 8-pin power connectors firmly seated into the board's connectors - not loose, and properly aligned?

Are the board's electrical connectors firmly attached to the board, not loose?

I'll try to think of additional things to check.

January 13, 2010 11:42:48 AM

Sorry for the epic delay in posting! It wouldn't be a software problem - it still occurs even if the hard drive is unplugged and sometimes even when it's been switched on for literally seconds. System is properly grounded for certain. CPU heatsink is properly seated as far as I can tell. All connectors are tight. It's got me stumped :D 

Funnily enough it doesn't seem to like the cold very much. It's been snowing here and it's quite cold on the floor with the system. One morning the problem occured again but as soon as the house warmed up a little it went away. Haven't seen it again for days.

Thanks for all your help so far!
March 4, 2011 6:22:53 AM

Can I just say a BIG THANK YOU!

I've had an intermittent total Freeze - then power down/restart fault on my machine for some time.
I scratch-build Orac (my i7 860) lovingly in a Thermaltake Lanbox case.
When I first built it I had to scrounge the bits, I spent where I couldn't avoid it (ASUS EVO motherboard for example)
But cash constraints meant that I made do with a bargain PSU (Possibly the worst component to scrimp on)

When I first powered it up, it used to switch off occasionally - no warning, no shutdown just BONK... and then a restart.
I assumed it was static, though I had earthed myself and been meticulous with the build.
I considered this 'proven' when, after very little use the machine settled down and ran fine for over a Year.

I did notice that if I switched it off, the fault would come back for a short while and then disappear again.
This I thought was either the earth-return not being perfect, or possibly a dry joint in the dodgy £12.50 PSU.
-it was an EZcool 650w but smelled a little gluey and I knew needed replacing - eventually.

I gave Orac an extra 4gb of OCZ ram the other week and Yes, u guessed it - fault came back.
Rather than be sensible, I decided a bit of destruction testing was in order.
So I ran a Flat-out Nvidia demo and went out, letting the whole thing run nearly roasting :o )

I came back home to a fused Ground-floor circuit and a totally dead Orac, the room stinking of burned insulation.
Well, it was kill or cure. I figured the extra strain on the PSU would either cook the dry joint out, or destroy it.
By now I was quite convinced that my dead PSU was indeed the culperit (sp?) and promptly bought a new one.

I have a Gtx285 card, so paid particular care to choose a PSU with 2 high-amp 12v rails and lots of good reviews.
I got a Powercool 650w modular from Ebuyer for £46 plus £9 for Saturday delivery (Total Bargain if you ask me)
The 550w Powercool has tonnes of great reviews, the 850w the same (tho this is massive overkill for my lean 1156)

Once again I carefully routed cables, checked and double checked connections, and put him back together.
Very happy that Orac was now even cleaner and quieter than he had been to start with.

and then the fault came back :o (

So it wasn't the PSU, Bugger... I ran him for a while hoping it was static once again and a few earth-outs would fix it.

But this time it was much worse, sometimes staying on for days, sometimes resetting every few minutes.
I tested EVERYTHING, ran earth resistance checks, diagnosed every frikkin component took every part off one by one.
Including the case header and STILL - after a little while... BONK... and a restart.

So eventually I start asking my pals and trauling the internet for answers.

The irony is we build these machines to cope with the heat...
I was even saying to myself, 'you must need discrete earthing, it's the high power signature in such a compact case'

WRONG

I built Orac with a straight air-path, even removed the hard drive cage to keep it that way. (HD in the redundant Floppy bay)
Near perfect air-flow in the front filter, out of the rear case fans so clean so sweet, so quiet, so Cool.

But do you know what? He HATES being cold, even a 21degree room isn't warm enough.
And every ounce of hot air gets sucked straight out, sure if I was tiny I'd be uncomfortably warm sat inside him.
But for Metal? Very cool indeed...

So I warmed him up, by running 7 copies of Red Squirrel Risc-os emulation at once (64% commit, cores at 80degrees)
The effect is to increase the GPU temperature by 10degrees while it's standing still.
I have to assume that whichever tiny contact isn't connecting when he is cold, is quite happy now.

So no, I haven't 'found' my fault and Yes, I do have to leave 7 copies of Matron-RS tucked away on my bar eating up cycles.

But you know what? At least he stays ON now...

I'll just have to nip onto Ebay and find myself a CPU WARMER - or just use it as an excuse to unplug the few case fans I have.
I think, what I will actually do is get some switch cleaner and wash the CPU pins next time I have him open...
For now he is fine chewing up enough Mips to run a small planet just so he stays snuggly - lol

So THANK YOU... Thank you for helping MkRoberts and Thank you for helping me 'fix' Orac.

Kudos Everyone, R.James - from Uchooze.com
!