Sign in with
Sign up | Sign in
Your question

Ram issue (could be resolved quickly)

Last response: in Memory
Share
February 6, 2012 5:55:55 PM

First I direct you to this thread:

http://www.tomshardware.com/forum/304745-30-timing-volt...

I am experiencing the same issues. For well over 2 years I've been using Corsair XMS2 DHX 2gb sticks x 4 to achieve 8gb in my system. They were bough in pairs, but the pairs do not match. This has never presented ANY problems. The timings and voltages are the same as his -One pair at 5-5-5-18 at 1.8v and another pair at 4-4-4-12 at 2.10v.

As previously stated I had NEVER experienced any issues whatsoever with the RAM, and have not since moved or even touched the RAM. After a recent windows update though (in the beginning of january), everything's gone to hell. Frequent BSOD's with "Memory_management" as the titular cause. I have tested all sticks with memtest86 for entire nights, and on different slots finding no issues whatsoever. I can go up to 4gb using one of the pairs, but the moment i put the second pair in the computer starts freezing (seemingly at random as it is not always while it is under load).

Symptoms:

Using 8gb of memory in correctly paired installations causes computer to BSOD, using one pair of 2gb sticks does can run the computer for days without issues. Using both pairs causes computer to BSOD after a random amount of time (anywhere from 2 to 8 hours).

Now in that post the community member that offered an answer said to do the following:
---------------------------------------------------START-----------------------
Here's a mix to try:

DDR2-800 5-5-5-18 1.80v
DDR2-800 4-4-4-12 2.10v
==================
DDR2-800 5-5-5-18 1.95v ~ 2.10v

1. Boot with only the working RAM

2. BIOS:

MEM (DDR), MHz -> 800
Memory Timing Setting /
Memory Timing Setting -> Expert
tCL (CAS Latency) -> 5
tRDC -> 5
tRP -> 5
tRAS -> 18
Command Per Clock (CDM) -> 2T
System Voltages /
Memory -> 1.95v ~ 2.10v ; trial by post/boot failure, you can start high and work down to 1.95v

Save & Exit = Yes

3. Shut down

4. Install all RAM

5. Boot and keep fingers crossed
------------------------------------------------END----------------------

I've been taking a shot in the dark and thinking that what he meant was to install the pair of 5-5-5-18 1.8v sticks and then go into the bios and enter the settings manually as described in the post including the voltage. Then turning off the computer and installing the 4-4-4-12 2.1v sticks and basically hope that works.

Is this correct? or should I install the 4-4-4-12 2.1 first and run them at 5-5-5-18 (while playing with the voltage) and then after getting them stable installing the 5-5-5-18 1.8V?

Also, I am confused as to whether its not just the order of installation but the PLACEMENT that matters as well...Should I be setting one pair over the other in a particular DIMM? (DIMMA1-B1 for faster or slower chip?)

Thanks in advance and hopefully I can get this resolved fairly quickly without pulling out any more of my disappearing hairline.

PS: Seems this is a little more wide-spread --- http://pcsupport.about.com/b/2012/01/10/patch-tuesday.h... a little ways down in the article there is a small notation saying this patch may have compromised system stability and that many users seem to be experiencing issues. Go figure.
a b } Memory
February 6, 2012 6:54:34 PM

I would always get RAM that has the same voltage.

4GB is good enough for most people.

Good Luck
m
0
l
February 6, 2012 6:57:03 PM

Anyone have any idea what the proper procedure is to adjust the timing as described on the initial post?
m
0
l
a c 347 } Memory
February 6, 2012 7:35:45 PM

First, mix-matching RAM with different CAS and Voltage there's no guarantees that it will work. Next, those BIOS settings are specific to 'that OP's' MOBO; example Quick on one Expert on another - all stick must run at the same Frequency, CAS, and Voltage.

That said, say your MOBO's DIMM configuration is: CPU: | A2 | A1 | B2 | B1 | then it's best to have the slowest in A1 + B1; otherwise if you e.g. Clear CMOS then the 'SPD' picks-up on the faster CAS 4 and the CAS 5 fails.

What I described in that prior post was finding the 'LCD' (lowest/slowest common denominator) and attempting to average the voltage disparity; so the voltage range is average (low) through highest voltage set (high); if the lowest voltage stick cannot operate at the average to high then you're SOL. Again, no guarantees...
m
0
l
!