Overclocking Issues with 680i LT and Q6700

[csm81]

Hello,

My rig:

XFX 680i LT
Q6700, Xigmatek S1283
OCz 2x1GB SLi DDR2
eVGA 8800GTX
Ultra LSP 750W
Seagate 250Gb

I have recently built the previously mentioned computer and I have run into several difficulties overclocking the CPU and I believe they are motherboard related. I have not been able to find a working overclock of any magnitude. The most stable configuration I have found was 400x8 for 3.2GHz, running the ram in Linked and Sync mode 400:800. To avoid all possible errors I have left voltages on Auto, as well as ram timings. When the computer booted on these settings I ran a prime95 test for over 6 hours with no errors and max temps of 63-65C. However, upon rebooting or trying to reboot, sometimes the computer won’t even POST at this OC. This is where I begin to think my issues are motherboard-related.

This often happens with any overclock. The computer will attempt over and over to POST until it reverts to safe-mode where it downclocks to stock. With the FSB set to 1200 (1200/4 = 300) the computer will not even post regardless of other settings. At 1199, the computer boots and loads XP but crashes during a prime test after a few hours (this is common, not motherboard related) At 1300, the computer crashes during XP load. If I change the multiplier from 10 to 9 or 8 I can sometimes get closer to a boot up. Even if I find a working overclock that boots, sometimes on rebooting it won’t even POST. I have tried with and without SLi-Ready Memory; with modest memory timings and the recommended timings, with auto voltages and high voltages and low voltages, linked and unlinked FSB-RAM. I have updated my BIOS. C1E, speedstep, spread spectrums, thermal control, all disabled.

On one particular attempt, I thought I noticed that during one of the posts some text was coming up on the screen very slowly. Like as if it was fighting something. I watched where it said "Press ESC to enter set up" and the last few letters appeared as if they were being typed by a person as they were just about to die (pardon my informality). I have ran a multimeter test on my PSU with results of a constant 12.05V on the the 12V rail at all times. The other rails are strong as well.

Thank you, I appreciate your help.

[Methious]

Unlink the ram to take it out of the equation.

Start with the highest multiplier you have available.

Set all the voltages manually.

Increase the FSB in increments until you get a BSOD or lockup then increase voltages incrementally until it gets stable. Over clocking is done in baby steps and it sounds like your trying to get the prize with out working your way up in steps.

Then there is no guarantee in over clocking, you might have a bum chip that just won't do better than 3.2. Try the incremental approach with the highest multiplier ram unlinked and I'm betting you'll get better results.

[csm81]

Well thats where I had started with this, with the multi at 10x and ram unlinked. It was working okay at 2.99GHz (FSB 1199) but would not even POST at 1200. This seemed reason to be suspicious to me so I then started playing with everything else and found the 400x8 OC and several others. It doesn't make sense to completely fail during POST at 300x10 but survive for an hour of prime95 during 400x9. This was where I started looking for gaps in chips overclockability.