HELP Overclock + Clear CMOS = No Post

[Fong]

So today I tried to overclock my 3930k to 4ghz. I first tried using a manual vcore of 1.35v but after instability I decided to try auto. This set the vcore to about 1.38v at load but I was still unable to run LinX. I then decided to go back to stock. However, even after going back to stock settings, I was experiencing BSODS soon after booting up into windows. I thought I would try to clear CMOS. After doing this my system isn't even posting anymore. The debug LED is stuck at b9 (or 69?). I also noticed an led has lighted up red. It is the one on the right of the label cpu_led. What should I do? Help!

[Arne_Saknussemm]

Good stuff! progress....Can you tell me the exact product number for the RAM?

Also if I were you I would download the Intel Processor Diagnostic Tool http://www.intel.com/support/processors/sb/CS-031726.htm just to check the processor is running fine.

[Fong]

Ram is gskill f3-1866c9q-16gab. Right now I have removed the 2 sticks on the left and I'm running memtest again. I don't understand why I'm getting errors at 1600 cl11 when 1866 cl9 passed memtest only a few days ago.

[Arne_Saknussemm]

Fong wrote:
I have removed the 2 sticks on the left

This does not sound right...I guess it would be better to follow the manual p2-5 in putting RAM onto the board. 2 sticks should be in B1 and D1 etc.

If I were you I would try to run the XMP according to Gskills instructions (tweak mode 1 etc) and make sure the DRAM voltage is set correctly 1.5 (could try 1.55) and VCCSA and VTT are in the range 1.15 to 1.25

In DIGI+ set VCCSA LLC to Regular, VCCSA current capability to 120%, DRAM current capability to 120% and DRAM power phase to Optimised.

[Fong]

I removed the 2 on the left because of I thought that could also check to see if the problem is with the motherboard slots. I have already tried most of the settings you suggested except for the digi+ power settings. I will try these and report back.

[Fong]

Ok memtest passed one run with 2 sticks so I tried 4 and once again I immediately get errors. I'm now trying 3 sticks with the slot on the furthest left empty. Am I on the right track or should I try something else first?

[Arne_Saknussemm]

You will have to be methodical with memtest. First this version (5.0) is for SB-E.... it is a Beta but works perfectly. http://forum.canardpc.com/threads/68001-Memtest86-5.00-Beta-available-%21-Need-betatesters-%21

You will have to run all tests at auto or at XMP......don't change until you have run all possible combinations.

You will have to eliminate bad sticks or bad slots. So run one stick in one slot. If errors run a different stick in the same slot. If this happens 4 times then the slot is bad or all sticks are bad. So try a different slot.

If the first stick passes in the first slot move that stick to another slot to test the slots for errors. If a slot gives errors try another stick in it.

If a random combination of slots and sticks sometimes give errors sometimes not then it will be something else.

When you test multiple sticks together follow the manual for placement.

Another thing you could do if one stick passes on one slot stop RAM testing and test your CPU with IPDT

But yes! You are on the right track;) just have to experiment to find out what's going on.

Is your PSU new....what model....can it be trusted?

[Fong]

I think I may have found the problem. I tested one stick on each of the slots on the left (with the 2 sticks that passed already still filled on the right) and it gave me errors immediately. I'm have now noted this stick and I'm testing the remaining stick. Does this mean this stick is faulty? If the last stick passes then I will do the IPDT. My psu is a 750w Silverstone.

[Arne_Saknussemm]

Maybe...you could test it by itself on differnt slots. You will have to be scientific with this to make sure you are eliminating sticks or slots;)

[Fong]

Ok just tested that stick on each slot and it seems to fail memtest immediately each time. Is this enough to say its a faulty stick or can there be another reason? I'm gonna get some sleep now since its 3.30am over here but I'll do more testing tomorrow. Thanks

[Arne_Saknussemm]

Yep! that's a faulty stick! Best return the kit. One kit of Ripjaws Z series is good on these boards!

Long night testing 😉 but got a result in the end!:D