Quote Originally Posted by my87csx481 View Post
I'm having the same problems with OCZ 1600EB 2 x 2. For the longest time I could not get 'em to run at stock speeds and timings....I'm on my third kit. I finally got 'em to memtest last night at 1600 7-6-6-24-2t for 4 passes....but with 1800 QDR with the following settings (0603 BIOS):

I previously tested with everything at stock except with memory timings, ram and NB voltage as above and it fails prime95 and OCCT. If I run CPU intensive tasks (OCCT CPU, Prime95 small fft's) = no problems. I have yet to test the above settings in windows/OCCT/Prime95....later today hopefully. I have a 2 x 1 1800 kit I need to test at stock settings as well.
I am sorry to say but I would not trust 4 error-free Memtes passes. I got some of my errors in pass 7-8. But you will find out when testing with Prime95, OCCT.

Everybody, when I replaced the TIM everywhere under the heatpipe system with MX-1, I used a very thin layer on the copper part for both the mofset areas instead of a (large) drop on each of the small chips as recommended in the guide at Asus forum: http://vip.asus.com/forum/view.aspx?...Language=en-us. I tightened the heatpipe carefully (so careful that I can not unscrew one of the screws now) so ensure good contact. Does anybody think too small amount of TIM between the voltage regulators and the copper profile can be the cause of my instability with non-default FSB and MEM speeds? I ask instead of just testing it because, as mentioned, one screw is stuck because the cross was softer than my screwdriver and...you know I am too lazy to pull off my watercooling for maybe nothing if I am RMA'ing anyway.

BTW, the Corsair RAM is fine for at least 11 passes of Memtest (7 hours) at 1333 FSB and Memory speed (Auto). Giving 1600 MHz (MEM unlinked, FSB 1333) another chance with 8-7-7-20-2T