Sign in with
Sign up | Sign in
Your question

System locks up at anything except very low temps. Motherboard or CPU faulty?

Last response: in Systems
Share
August 9, 2013 12:20:56 PM

I've been having problems for a while with my system freezing at anything except very low temps (i.e. unless I remove the side and point an air-con unit at it! Has to be on cool - fan function of air-con isn't enough).

Every advice I've had about this so far is that its a PSU fault (*). This didn't make much sense to me, but I finally tried a new one - no difference.

What I've tried so far:


  • Disconnect hard disks. The problem still occurs in the BIOS screens. So not a disk, OS or driver problem.
  • Change graphics card. No effect.
  • Change PSU. No effect.
  • Try RAM modules one at a time. No effect. (So either OK, or all dodgy)


  • Which only really leaves the CPU and motherboard. Obviously I could swap both, but would rather only replace the faulty one if I can work out which it is. Does anyone have any clues on how to distinguish between a dodgy CPU and a dodgy motherboard?

    I've a couple of other problems that I've had for a while (months) which may or may not have the same cause, I don't know if these give any further clues:


    • Have not been able to run RAM at 1600Mhz for a while. (BIOS says 'overclocking failed', if it powers on at all. (This isn't an overclock - its the speed the RAM is supposed to run at)
    • Occasional random reboots (not frequent enough to make the system unusable, can be weeks between them, but sometimes twice a day).



  • (*) That one of the people claiming this was the 'technician' at PC World didn't exactly give me confidence it was going to be the right answer.
    August 9, 2013 2:54:41 PM

    werner123 said:
    Hi, when you say that you not have been able to run RAM at 1600Mhz for a while, do you mean the memory was running at 1600 before but not now? that could be that the IMC (integrated memory controller in the CPU) is faulty. Anyway, to be sure you need to test with another CPU. What motherboard, memory and CPU do you have? are you using the stock CPU cooler?


    Yes, I've been able to run with the RAM at 1600MHz in the past. I'm not sure exactly how long I haven't been able to for though - I've occasionally had this setting get lost and had to reset it (which I've not always done promptly).

    Motherboard is an ASUS PT6 Deluxe V2
    CPU is a Core i7-960 (3.2 GHz, not overclocked)
    RAM is 6GB Corsair Dominator 1600MHz DDR3
    CPU cooler is a Titan Fenrir (I've never had the stock cooler)

    Obviously this is an older system, so I'd like to be as sure as I can before swapping either the CPU or motherboard. Sticking with the older 1366 socket makes sense if I can swap only one to save money, but if I end up swapping both then I'd be better off upgrading at the same time.
    m
    0
    l
    August 9, 2013 6:20:31 PM

    I may have found (and cured) the main problem. I had a suggestion elsewhere it may be a problem with the soldering on the motherboard. 8 minutes in the oven later and things seem much better. I can't be certain yet, I still need to assemble the system properly (currently on its side with half the case connections undone and the CPU cooler balanced in position) and give it a workout to be sure. I'll leave that until tomorrow though, its rather late here!

    I'm still not able to get the RAM speed right, but I can live with that if the system is stable again.
    m
    0
    l
    Related resources
    Can't find your answer ? Ask !
    August 10, 2013 3:34:58 AM

    werner123 said:
    Well CPU's are very robust and it's very unlikely for a CPU to go bad, unless it's been overclocked for a long period of time and running to hot, but in your case that's not applicable. 9 out of 10 times it's the motherboard, and i would think in this case it might just that. There could also have been a small metal object or the like that some how got into the case on the board causing a short. The oven..Hmmm, interesting to say the least and i highly doubt, that, that may have solved the issue. Is the board updated with the latest bios? Have you tried clearing the CMOS?


    BIOS is up to date (well as up to date as a board of this age gets). Has been reset to default settings before, but of course I went as far as to remove the battery before cooking it. Could that have cleared it more thoroughly?

    I'm less convinced by the small piece of metal theory. The system has been moved (stood up, lying down, repeat) a fair few times while I've been experimenting. I'd expect any debris to have shifted before now. Talking of moving the system I have moved house recently, so if it was a solder problem then its perhaps my own fault for leaving that massive cooler mounted during transit?
    m
    0
    l
    August 11, 2013 6:21:03 AM

    werner123 said:

    Now this sounds interesting, did the issues occur after moving?

    There was a gap of a couple of months this time. Though the previous issue with random resets first occurred very soon after a previous move (I initially suspected the power supply in the rented flat, but since a UPS didn't fix it decided it must be something else).

    werner123 said:

    Look for obvious physical damage on the board, CPU, bent pins in CPU Socket in particular?

    The processor has been in and out a couple of times in the last few weeks, pins seem fine. Likewise the board looks OK, but I suspect that minor damage to joints from accidental flexing wouldn't really show.

    werner123 said:

    the cooler might have loosened and was obviously not making decent contact with the CPU that could have caused overheating and probably damaged the CPU over time without you noticing it.

    If it wasn't all working again then I'd suspect that too. I did try re-doing the thermal paste a couple of weeks back, which made no difference. Conversely while I was testing to see if the oven had killed/fixed the motherboard I actually just had the CPU cooler resting on top of the chip (old paste, no bracket). While I'd not want to run Prime95 or try playing games like that temps seemed OK just doing undemanding stuff like posting on here.

    werner123 said:

    When a PC works and suddenly have issues, you should always ask yourself, What did i do before the issues started, did i change new hardware? did i install new software? etc. and go from there.

    Well lock-up issue started a few weeks back when we had a heatwave here in the UK. I initially suspected my cooling wasn't quite up to coping with the the unusual weather, hence the experiments with the air-con unit. Unfortunately it needed really chilling, not just a better breeze (even now the weather has cooled off slightly). So I don't think it can be explained by normal overheating. Whether the unusual was what pushed an already borderline system over the edge I can't be sure.

    Anyway, it seems to be fixed now (1600MHz RAM aside), whether it was the oven or something else I'm not 100% sure, but at least its working for now (hopefully it stays that way).

    P.S. Sorry for the delay in replying, the PC has been busy running memtest86 and/or being practically rebuilt (decided to fix my card reader and redo the cable management since I'd got the lid off anyway).
    m
    0
    l
    August 16, 2013 4:34:38 PM

    werner123 said:
    Keep me up to date, good luck.


    Well a week later and the system seems completely stable. No lockups, no random resets. I've been pushing it quite hard as a test (3 days flat out folding on CPU+GPU, playing games that really warm up the graphics card). It seems more stable now than it was when I first built it.

    Whether it was the oven or something else something seems to have worked.


    EDIT: Guess what happened 10 minutes after I posted that? :( 

    Though I was also giving 1600MHz RAM another go, which did actually boot this time (hasn't done that for a long time). Of course the obvious thing was to revert the settings back to how they were. Which didn't work. Something is still iffy, and it seems that trying the different settings for the RAM has pushed it back over the edge.

    EDIT 2: I've disassembled etc again and seem to have it back to stable. I'm going to leave it alone this time, the RAM can stay at the lower speed this time.
    m
    0
    l
    August 26, 2013 8:42:22 AM

    So much for hackish fixes. It seems to be properly dead this time. Time to go shopping :( 
    m
    0
    l
    August 26, 2013 8:42:27 AM

    So much for hackish fixes. It seems to be properly dead this time. Time to go shopping :( 
    m
    0
    l
    August 26, 2013 10:27:00 AM

    werner123 said:
    Ah sorry to hear about that, if you need my help recommending a board just give me your budget and online retailer you will would be buying from.



    Thanks, much appreciated (as is all your input on this topic).

    Currently seeing if I can get a straight replacement cheap from ebay (depends how the bidding goes). If that falls through then I'll start shopping with a somewhat larger budget on Thursday.

    One question for now though - what is Intel's current socket? Retailers seem to be selling boards/processors for 4 different sockets at the moment, knowing which one I should be looking at would narrow things down rather a lot! I think 1150, but not 100% sure.

    Also any thoughts on Intel vs AMD? I've no particular loyalty to either (I've actually switched between the two at every past upgrade (Intel 8Mhz 8086, AMD 40Mhz 386, Intel 166Mhz Pentium, AMD 1.2GHz Athlon, Intel 3.2Ghz i7).

    m
    0
    l
    !