Ripjaws Z F3-14900CL9Q-16GBZL 8x4GB problem

[Wojszwillo]

Hello forum,
I'm not novice in PC, but stuck in problem with PC stability and can't solve it.
System:
W8 Pro
Asus P9X79 Deluxe
i7-3930K
G.Skill RipjawsZ 2 kits of F3-14900CL9Q-16GBZL or 8 x F3-14900CL9-4GBZL, i.e. 8*4GB

System runs stable only at 1600.
When trying put to X.M.S. profile 1 or 2 x 1866 - problems with DIMM's (usaly one), system hangs almost allways.

When put to X.M.S. profile 1 or 2 x 1866 and manualy set values:
1. DRAM CAS# Latency - CAS# Latency (tCL) - 9
2. DRAM RAS# to CAS# Delay - RAS# to CAS# (tRCD) - 10
3. DRAM RAS# PRE Time - RAS# Precharge (tRP) - 9
4. DRAM RAS# ACT Time - tRAS - 28
5. DRAM Command Mode - Command Rate - 2
6. DRAM RAS# to RAS# Delay - tRRD - 5 or 6
7. DRAM REF Cycle Time - tRFC - 150
8. DRAM WRITE Recovery Time - tWR - 14 or 15
9. DRAM READ to PRE Time - tRTP - 7 or 8
10. DRAM FOUR ACT WIN Time - tFAW - 24 or 26
11. DRAM WRITE to READ Delay - tWTR - 7 or 8
12. DRAM WRITE Latency - tCWL - 7
system works only on "hot" system, i.e. restarting, shutdown and turn on - all OK.
On "cold" system - i.e. after couple of hours of system shutdown - system hangs, problems with DIMM's again.

On 1600 - have'nt any problems at all, but i'm not satisfied with speeds:
1600 1866
Memory read 16.850 18.500
Memory write 14.720 14.720
Memory copy 16.170 16.730
and overall - having 1866 RAM is not acceptable run them on 1600...

Any help would be very appreciate.

[Tradesman]

You may have two sets that don't want to play together (which can happen with any two sets, even same exact model, but possible problems can be compounded by total # of sticks and total amount of DRAM, which is why sets are available in basically any amount of total DRAM and in number of sticks, from 1 to 8)....that being said, I'd make sure you have the latest BIOS. Input the base timings and CR manually as XMP won't work with multiple sets (XMP is programmed by the individual set and in the advanced timings the more sticks you have the values will normally go up, i.e. with 4 sticks you may need a tRFC of 170, then with 8 sticks it may require a tRFC of 310) then try raising the DRAM voltage + 0.05 to + 0.10, raise the VCCSA to 1.25 and raise the MC (Memory Controller) voltage to 1.2 (believe CPUVTT on your mobo)

[Wojszwillo]

Solved!
Despite i'm using two the same 4*4GB G.Skill RipjawsZ F3-14900CL9Q-16GBZL kits (total 8 modules), they have different SPD on X.M.P. profile:
First kit, modules produced February, 2012, have SPD:
X.M.P. Profile 1 - 9-10-9-28 (CL-RCD-RP-RAS) / 37-150-2-6-15-8-8-26 (RC-RFC-CR-RRD-WR-WTR-RTP-FAW);
X.M.P. Profile 2 - 9-10-9-28 (CL-RCD-RP-RAS) / 37-150-2-6-14-7-7-26 (RC-RFC-CR-RRD-WR-WTR-RTP-FAW).
Second kit, modules produced December, 2012, have SPD:
X.M.P. Profile 1 - 9-10-9-28 (CL-RCD-RP-RAS) / 37-150-2-5-14-8-8-24 (RC-RFC-CR-RRD-WR-WTR-RTP-FAW);
X.M.P. Profile 2 - 9-10-9-28 (CL-RCD-RP-RAS) / 37-150-2-5-14-7-7-24 (RC-RFC-CR-RRD-WR-WTR-RTP-FAW).
At beginning, i have leave 4 modules from first kit on slots A1, B1, C1, D1 (as they were on 16GB RAM system setup) and put modules from second kit to slots A2, B2, C2, D2 and system became unstable at different settings (by the way - i don't like idea to increase voltages, i.e. to overclock).
When i put 4 modules from first kit (slower) to slots A1, A2, B1, B2 and 4 modules from second kit (faster) to slots C1, C2, D1, D2 - all is going fine at standard settings with X.M.P. profile enabled (1866).

[Tradesman]

Unless you want to experiment, would leave it as is as long as stable

[Wojszwillo]

You think system can became unstable in time becouse of ASUS Bios "adoptation" / "learning" to RAM function?

[Tradesman]

Hopefully not, but at times when they update a BIOS, they may make changes to how XMP is implemented and that can throw off a set of sticks that were perfectly stable....i.e. I had 32GB on an ASROCK P67 mobo and all was fine, then with one of their BIOS updates they put a limit on the advanced timing tRFC of 255, with the way I had the rig set up, it required a tRFC of 272, so it became unstable, did some voltage adjustments and had to also change a couple of the other advanced timings to get it smooth again with the OC I was running