Dead NF4 Ultra-D?
I have had this board for about a week, and just today have had time to really zoom in on fine tuning. I am using a 3000+ Venice with xp-120 cooling, and 4 sticks of Mushkin Blue 512 with original BH-5 chips. The best stable settings I had been able to come up with are 2.6 GHz cpu frequency (9x289, HTT mult x3), ram at 150 (3/4) with cas set to 2-2-2-7 and most everything else on auto, with ram voltage set to 3.2 on the 3.3V rail. I had also tried a 5/6 ram divider on 3.4 and 3.5 V setting (5V rail), but the system would lock or randomly reboot when using the 5V setting, so I moved back to 3.2 V on the 3.3V rail and a 3/4 ram divider, and everything seemed stable (prime95 for 6 hours). My PSU is sufficient... it is a dual-redundant server PSU which provides [3.3v,40A; 5v,60A; 12v,30A], and I have been using all the power sockets on the board except the one near the main PCI-E slot. All the lines are very stable, even under load, when monitored.
Today I got around to checking out the DFI NF4 BIOS Memory Guide posted at DFI-Street, and decided to try some fine tuning. I made the following ram timing changes in bios, based on suggestions in that thread: tRTW 2.0, tWTR 2.0, DRAM Drive Strength 8, Idle Cycle Limit 32 clocks.
Upon saving, the system auto-rebooted and failed to post. I have since tried the "safe-boot" jumper, and clearing the CMOS, each a number of times, and the system will not post. It seems to hang with 3 diagnostic LEDs remaining lit. I have pulled all but one RAM stick, and the situation remains the same.
Does anyone have suggestions as to remaining options in attempting to resurrect the board, or should I RMA?
I posted this at DFI-Street as well, for increased community exposure.