Z68XP-UD5 GD2 nstability = excessive overvoltage

[Aussie Allan]

  Bingo!...... now we seem to have something stable .... this will prove the board, PSU, GPU and drive are in fact fine .... then we can throw on the last piece of the puzzle and see if it fall over.

  Aussie Allan

[Deathchant]

Ok it takes 52minutes for 1 memtestpass. I've got 4 passes without errors.
I did a reboot until the "insert bootdisk" text appeared. Guess what? Leds are out!

I know 100% sure that the leds where on in this stage of the boot process with my 2600K....

[Aussie Allan]

  Take notes!.....but keep going with the procedure we worked out .... all these little tit bits help form a picture .... looking more like CPU BUT

  The new BIOS may have played a hand ....we wont know until we finish the procedure!.....I'm happy with memtest if you want to go to the next step.

  Aussie Allan

[Deathchant]

just to be clear, I did not flash the beta in the motherboard, but version F4.

Currently when the celeron boots, the leds go out right after the post (when the pc is in searching bootdisk-status).
If I do all the tests with OCCT (is prime95 also ok?), we can have the following situations:

1: Celeron test = OK -> motherboard +CPU look OK
2: 2600K leds go out after post and cpu test = ok -> no cpu or motherboard RMA
3: 2600K leds stay on, CPU test OK -> Mobo RMA???
4: 2600K leds stay on, CPU test NOK -> 2600K RMA
5: 2600K leds stay on, but go out later than with the celeron and CPU test OK -> Mobo RMA
6: 2600K leds stay on, but go out later than with the celeron and CPU test NOK -> 2600K RMA

Let's hope we don't end up in situation 3 or 5.
Alan, when I played with the bios settings, I got a spontaneous poweroff-poweron again, this time with the celeron in it. Is it normal for Gigabyte Z68 mobo's to do that?

[Aussie Allan]

just to be clear, I did not flash the beta in the motherboard, but version F4.......Thats fine, as long as we have a constant...leave at F4

Currently when the celeron boots, the leds go out right after the post (when the pc is in searching bootdisk-status).
If I do all the tests with OCCT (is prime95 also ok?), we can have the following situations:...Prime is OK, I like OCCT for running temp Graph

1: Celeron test = OK -> motherboard +CPU look OK
2: 2600K leds go out after post and cpu test = ok -> no cpu or motherboard RMA
3: 2600K leds stay on, CPU test OK -> Mobo RMA???
4: 2600K leds stay on, CPU test NOK -> 2600K RMA
5: 2600K leds stay on, but go out later than with the celeron and CPU test OK -> Mobo RMA
6: 2600K leds stay on, but go out later than with the celeron and CPU test NOK -> 2600K RMA

Let's hope we don't end up in situation 3 or 5.
Alan, when I played with the bios settings, I got a spontaneous poweroff-poweron again, this time with the celeron in it. Is it normal for Gigabyte Z68 mobo's to do that?.........NO!...not that I'm aware of, but good you're taking notes though.....when you say " spontaneous power-off / power-on" as in it shut down and rebooted ?

 Your pretty well right with the 1 to 5 ,.... Although we need to load windows now with what was stable (Celeron) and run Prime or OCCT for a while and see what happens............if every thing runs Hunki-dorie,....we can swap in the 2600 and power up

  Don't read too much into the glitch you had while in BIOS  just yet but it's been noted .... the real test now it's up is when it goes under load for a while , ......one step at a time!

  Aussie Allan

  Update    just when I was starting to lean towards "Crook CPU" the motherboard wobbled .... it really is 50/50 until you complete those tests.

  Bloody time consuming , but necessary.

[Deathchant]

well, it just that I had saved my bios settings with F10, then it rebooted (this is normal ofcourse), but as soon as my pc did a power-reset after saving the bios, it did another poweroff and stayed off for 1 or 2 complete seconds and after that a poweron again.....so weird

[Aussie Allan]

    Now you have explained ..... not so weird ... Giga boards will quite often do a sort of double take/start when changes are made to BIOS large or medium parameter changes.....the little one some times are excepted without a second thought .... don't worry, this one I think normal

  Have had any luck installing OS and running any type of stress test with the Celeron as yet .... even if it run for one hour would be good enough for a "PASS"

  Aussie Allan

[Deathchant]

Ok thanks for the comforting words. You would expect to have an unstable PSU hehe, but ok it is normal then
well, a friend of mine is coming over with a W7 64bit enterprise cd, so I am planning to do that this afternoon or tomorrow.

[Aussie Allan]

  Good news then! ... stay in touch as you progress....

  just to let you in to my thoughts.... hopefully it will load and run and stress test with no problems under the Celeron ...this will tell us 99% that it's the CPU........but it will have to go back in at some point to prove that it still fails (the 2600).................. if you do have a major glitch with the Celeron, you're going to have to leave the system as is but swap the Video cards around again ....I'm still not totally convinced that it can be completely ruled out if your present setup falls over , ..... but one step at a time ..... we need to be 100% sure before RMAing any components.

  Aussie Allan

[Deathchant]

Well, I've installed win 7 64bits enterprise. I installed all relevant benches: 3dmark11, prime95, cpu-z, gpu-z, as-ssd.
I've been updating windows first and updating my virusscanner. I've downloaded AIDA64, but is there any other handy program that monitors the temps?
Maybe I can link everest to my logitech G15 LCD.

I've installed all the latest drivers for my motherboard from the gigabyte website. (I have had XP pro for a while and I never worked with win7 before, so it takes some time to feel familiar with it )

1 thing I have to say: my god it is fast. A complete reboot takes only around 30 seconds with my crucial M4. Imagine this is only with a Celeron CPU

I have run 2 testcycles with prime95 last night, but I was so tired, I had to go to sleep. I think I can do it tonight. I think Prime95 will exit immediately when something is wrong with the CPU right?

[Aussie Allan]

 Now where getting somewhere and all good

  Now you're enjoying why we went this route with the Celeron   .... Yes it should exit if it becomes a little unstable and BSOD if it really doesn't like it

  Now you have the luxury of having a system up and running with the worst case event, .... and that is when you finally put in the 2600 and it fails, .... RMA of the CPU might be a couple of days, .... but it could also be weeks

  there is still a chance the 2600 might fire-up, ...... after all we have made more then several changes,....(But I doubt it)

  Just run the test with the celeron and establish that it's rock solid stable.........then try the 2600

 Aussie Allan

[Deathchant]

I'm afk this whole evening, but I will activate the prime95 and see how that turns out when I come back. Tuesday = 2600K-day

I have the gut feeling that the 2600K will also be just fine.
in that case, nothing needs to be RMA'd, and I've learned an important lession on F1 bioses and how to interprete Gigabyte- LED-statuses
I will keep you up to date.

[Aussie Allan]

 OKI DOKI!

  Aussie Allan

[Deathchant]

OK, i've run the Prime95 for 1 hour and 12 minutes. The only thing it says it could not open some logfile, but other than that it went fine.

CPU load: 100%
CPU #1 temp: 39C
CPU #2 temp: 37C
vcore fluctuates from 1.056V - 1.12V
12V-rail: 12.172V

[Aussie Allan]

  Normally I would say over night! .... but under your circumstances and what you're trying to prove ........ bring on the 2600

  Aussie Allan