Damnable unsolved i7 rebooting issue [details...]

So, what do you do when you start to experience a series of random reboots, some with no 'Blue Screen' in BOTH windows and Ubuntu?

Start to pull your hair out I'll bet!

There is no easy way to diagnose this, but I'll keep you informed as I try to fix it, with your help guys and girls.

Firstly the system:

Gigabyte EX58-UD3R V1.6 BIOS: FB
Core i7 920 @ stock, HT and Turbo both on. Vtt 1.235, others: Auto. Vtt raised to keep nice 'space' from VDIMM
Zalman CNPS 10X Extreme
6GB Crucial Ballistix [3*2] 1333 7/7/7-24 Turnaround:2 VDIMM 1.66 [Sticks say 1.65, but I can only choose 1.64 or 1.66. Crucial support say 1.66 is fine]
700W Cooler Master Silent Pro
Zotac Geforce 470 @ stock BIOS V70.00.1A.00.03
WD 1002 FAEX
Sony DVD R/RW Drive
D-Link DWL G-520 WiFi
CoolerMaster Scout Storm.

Windows 7 64Bit Pro

Symptoms are: At some, seemingly random point the system will either BSOD or just straight reset. Sometimes an application will freeze first, sometimes not. This can happen in ANY application, and does not seem to be just because of load. Once it did it as soon as I ran Prime95, once when I was doing next to nothing. Has happened a few times in Assassin's Creed, but not always.

Once even happened after I had been stressing the system with Prime95 AND Assassin's creed and a 4GB 64bit Photshop pounding and streaming a video. Seeing as it took this for half an hour I had closed them all and given up to just browse the web!

So far:

Removed the reset switch.

Checked and reseated all connections, cards and thermal paste etc.

Checked all temps and cleaned heatsinks [all temps are less than 60c, bar grfx which hits 90 under full load]

Gently shock tested the motherboard, RAM and video card [tapped them gently while running looking for a poor connection causing a reboot]

I have updated my BIOS to FI [from FB] which is 9 months and 6 versions newer, and gone through the BIOS options with a fine tooth comb.

The PSU's Rails all look OK [loaded and unloaded], bar the -12v which seems to be at -14v, but this may be because there is no load on it

Windows sometimes reports "There maybe a problem with your RAM" afterwards, but Memtest+ 4 runs fine for a loop, and has run before for a night [will test this again tonight]

I am not enjoying this one bit.

What is the next step after running Memtest+ for 6 hours I wonder? Any other suggestions?

Thanks in advance!
11 answers Last reply
More about damnable unsolved rebooting issue details
  1. I now notice someone else not far back in the posting with a similar issue and the same mobo, could this be connected I wonder? However, my 'gut 'suspicions are with the vid card at the mo. How can I test only the card in software?

    Meanwhile, bed. And for the PC: Memtest86+ for many hours [maybe 10 loops possible in this time]

  2. I know you'll hate to hear this but my buddy had 3 of these boards come back to his shop in less then a week.
    I'll talk to him and give you an update.
  3. Over 6 hours of memtest passed. Will now try prime95 - Large, in Safe Mode.

    Does no-one have any ideas? This is needed as a work machine, so any advice is greatly appreciated.
  4. davcon said:
    I know you'll hate to hear this but my buddy had 3 of these boards come back to his shop in less then a week.
    I'll talk to him and give you an update.

    OK, cheers. Meanwhile Prime seems fine for 40 mins or so, so I doubt the CPU has gone faulty. I have bumped a couple of voltages in the BIOS, as it seems to be a stability problem with the motherboard.

    Vcore is/was 1.2625
    QPI/Vtt is 1.275 was 1.175
    IOH Core is 1.3 was 1.1
    DRAM Voltage is/was 1.66 [reads 1.65 under load in SpeedFan]

    If this doesn't help I will try some graphics card tests I guess.

    Thanks again.
  5. I talked to my buddy and he said all the mobo problems he encountered with that model were Raid related.
    Crap Raid support on that board fyi.
    I'm not technically qualified enough to help with your issue.
    Although i have read that updating more then 4 bios versions at once can create issues.
    I've been told 4 max at a time.
  6. Frequently, these seemingly random faults are caused by flaky power. Try to borrow a known good PSU for testing.

    I am not a fan of CM PSU's.
  7. Cheers again, dude!

    Anyone else got any ideas?

    Is bumping the voltages as above a good idea? I see lots of people using QPI/Vtt of 1.275 or higher and IOH Core of 1.3 plus so I wonder if these are good voltages to use for signal integrity etc, and the BIOS defaults are just too low. I always like to use low voltages to save heat and stress etc.

    Not heard about the 'BIOS version jumping limit' before, but it might be just because everything has moved around and it's really hard to find it again!

    Thanks for following up, will keep trying stuff.
  8. OK, if the QPI / IOH bump doesn't help I'll find another PSU.

  9. Sorry, I should have said, PSU Rails were tested with a multimeter under max load [Prime95 and Assassin's creed running at the same time]
  10. System seems stable now, I think the bus voltages might have been the problem.

  11. System still stable with IOH adjusted lower as discussed in this thread


    FIXED! - Thanks all
Ask a new question

Read More

Homebuilt Intel i7 Systems