CPU Fried? What the ----!

[RickB1293 9]

I started getting bsd'd yesterday then lost everything. System won't POST or boot and the MB DRAM fault led stays on. I tested the memory in a different system and it is ok. ASUS product support says the CPU's memory controller probably fried. I OC'd the system to 4.8GHz with BCLK at 100MHz and the MULT set to 48. Voltage was set to 1.425vdc. The chip core temps never went over 80C during 24Hr

stress testing. Memory clock and voltage were 1600MHz at 1.65vdc, per XMP spec. I was lucky enough to buy Intel's OC protection so the i7 2600K is covered. I'm wondering if the OC had anything to do with the chip failure or if anyone else had a similar experience.

[Barnstormer1 0]

No expert but 1.425 seems extreme. I'm running 4.7 ghz @ 1.38v. Ive heard over 1.415 is flirting with disaster. Never had a CPU failure but never ran over 1.4 vcc. It;s a price ya pay for High Performance reckon.

[ramrunner800 10]

I ran my 2600K at 1.6v once (by mistake) and as soon as I ran a CPU dependent application, the temps rocketed into the 90's before I could shut down. The rig refused to boot for the rest of the day, but the next day it booted just fine. I started to have intermitent crashes a few weeks later, which were probably related, but the computer still ran. Thank goodness for Microcenter's warranty, which also covers overclocking.

 

I should also add that, from what I have read, 1.40v is the absolute red line for voltage on Sandy Bridge.

Edited May 20, 2012 by ramrunner800

[William J.S.S 173]

I hit 5 on 1.36. Not sure why you are so high. I backed it down to 4.4 and when I PMDG over orbx I bring it to 4.8 bit keep the same volt, as prime tells me after about 4 hours I am in the good.

 

I am no pro at oc but I agree with the above. I see that high a number as a positive burn setting on a system.

 

 

Sent from my iPhone using Tapatalk

[dazz 69]

Maybe just bad luck with that chip. On the bright side you are covered by Intel's OC protection program.

1.42V is more than I'm comfortable with, but 1.65V in the memory is too much for some others too... None are outrageously high volts so probably just bad luck

[Kosta 122]

I'm running my 2600K by default at 4.8Ghz with 1.5V and 1.65 on RAM as per default XMP (1866 tho). This since I've got it (over a year now). Doesn't mean your must have had too, just sayin...

[jackcnd 0]

I was just wondering if the Intel OC plan was worth it.

 

Looks like it can be purchased anytime during the 1st year, but claims take 30 days after purchase.

 

My 2700 is at 4.8 with 1.32

so I'm not too worried anyhow but it would be interesting to push it to 5+...

[Manny 1,293]

This is where you get your tuning protection plan..I believe.

 

http://click.intel.com/tuningplan/

[SpiritFlyer 987]

My 2700 is at 4.8 with 1.32

so I'm not too worried anyhow but it would be interesting to push it to 5+...

 

It does not make enough difference to really matter. After applying the 30fps internal/Nvidia Inspector 1/2 refresh rate solution there is no noticable difference between running 4.8Ghz (1.34 vcore) and 5.3Ghz (1.46 vcore). The settings keep FSX performance the same anyway, and seldom drops below a buttery 25fps regardless of anything else, at either speed. BTW, it booted at 5.4 (1.48 vcore I think) but I was too chicken to wind it up as the temps were idling in the low to mid 60s.

 

I am considering buying the extra overclocking coverage as maybe I will be a bit too brave after awhile if I get bored with a trouble free 4.8Ghz everyday setting. 5.5Ghz keeps sticking in my mind, but so far has not transcended to the bios. Can't help but wonder though. If it runs well on FSX at 5.3GHz, why not give 5.5Ghz a shot, even if it is just a short shot...and if that works, what about 5.6Ghz? Guess I better get the coverage before I do something even approaching being that stupid.

 

Apologies to the OP for getting off subject. I ran a 2600K at the same or higher levels that you have for a year or so, without any problem whatsoever. It may be just the luck of the CPU draw, or perhaps some kind of surge did it in, or another irregularity. It will be interesting if Intel gets back to you with a diagnosis of what happened. Good luck on your new replacement chip!

 

Kind regards,

[alainneedle1 97]

@ Stephen.....do I hear water cooling? If not let me know how long it took to have your CPU stable at 5.6GHz before it start smelling like burning rubber....

[ArjenVdv 190]

If your memory controller died, it might have been your memory voltage. I heard that anything over 1.58V for DRAM can damage the CPU's memory controller after a while.

[RickB1293 9]

Thanks for the input guys. After cruising the support sites I see that the ASUS P8Z68 V PRO board is known to lose the A1 & B1 banks with the install of Corsair's H100 cooler. No one seems to know why but that are literally hundreds of posts on the OC sites and at ASUS and Corsair about this problem. Apparently this is a MB problem not a chip one as reported by ASUS troubleshooting. I see 8Gb of RAM in slots A2 & B2 but of course they run in single channel mode not dual. Both the other slots are dead and any RAM installed in them, the system refuses to POST. I RMA'd the board and am waiting for ASUS to authorize the return - the board is less than six months old.

 

On the OC side - I bought an H100 which keeps core temps in the sixties and was able to crank the chip voltage down to 1.39vdc. I have had the same results as Stephen regarding FSX performance and decided to reduce the freq. to 3.6GHz. "Wise is the man that recognizes his limitations" -- Harry Callahan (paraphrased).

 

BTW, Arjen. The Corsair RAM I use is on both the ASUS and Intel QVL and runs at 1.5 vdc at 1333MHz and XMP 1.65vdc at 1600MHz. I've heard the warnings about the higher voltage many times but Corsair SWEARS that it's 1.65vdc modules are approved and won't trash the onboard memory controller on the i7 chip.

 

Thanks again for your input, love this site.

[Barnstormer1 0]

Umm. Mine are running dual with A2B2 and a Corsair H80. MB P67 may be the diff.

[RickB1293 9]

In my last post I noted the A1 & B1 slots as inop. The A1 and A2 slots, which are closest to the CPU, are the real culprits - sorry for the confusion.