Pinning down memory issues

warhammer3025 · Mar 29, 2014

For the past several months, my PC has been having random bluescreens, all related to memory issues. BAD_POOL_HEADER, MEMORY_MANAGEMENT, etc. The bluescreens appear truly random; sometimes they happen multiple times in a day, sometimes it's a week or more before the next issue. MEMTEST 86 has not revealed any issues, nor has the Windows memory tester included in Windows 7.

So far I've been unable to pin down the cause. I've updated drivers and my BIOS, fiddled with settings in the BIOS, etc.

I've not handled the hardware yet, because I don't have the money to replace any parts if they do turn out bad - and I'm hoping that it won't come to that.

MOBO - MSI Z68A-GD55 (G3)

MEMORY - 4GB Micron DDR3 1066 1.5v 5-5-5-15-20
4GB Crucial DDR3 1066 1.5v 5-5-5-15-20

Help in tracking down the cause would be most appreciated. If I need to include a dump file or something, please let me know.

f-14 · Mar 29, 2014

warhammer3025 :

i think your trc is a bit too tight.

try 5-5-5-15-23
trrd @3
trc@23
twr@5
twtr@9
trf@7.0us
this was the 1066 ddr2 crucial ballistix tracer settings that worked best for me, i had @1.875v-1.9v despite 2-2.1v recommendation and the 2nd channel always seemed to burn out the memory similar to the problem your having. i am wondering if you are experiencing a similar problem that matched alot of what i was experiencing before concluding the 2nd channel just liked to burn out ram after i had stopped using that pc for over 6 months. i think it took a static hit some how.

i know your running different stuff but this was the ddr3 stuff they put into ddr2 before discontinuing ddr2.
reference if you need it http://www.evga.com/support/manuals/files/123-YW-E175-A1.pdf skip down to page44

warhammer3025 · Mar 29, 2014

Thanks for the tip. I'm not sure which setting will control that, though.

The MSI Click Bios 2 utility seems to have different names for the settings than I'm used to seeing.

Things like tCL2, tRCD2, tWR2, tFAW2, and tCKE2.

EDIT: Strange thing though - CPU-Z reports two different memory timings. Checking under "memory" tab, it says the memory is running at 7-7-7-20.

Checking under the "Spd" tab, each memory stick is reported running what I reported before, 5-5-5-15-20 in the JEDEC 1 column.

Am I reading these reports wrong? Or is CPU-Z confused?

warhammer3025 · Mar 31, 2014

Bump.

Tradesman1 · Apr 1, 2014

So 2 4GB unmatched sticks? Try going tinto BIOS and manually set the the freq and base timings, then set DRAM voltage to 1.55 and raise VCCIO or CPUVTT (can't remember what they call the MC (memory controller) voltage on that mobo) to 1.16

warhammer3025 · Apr 2, 2014

Actually it's 4 2GB sticks. Two from Crucial, and two from Micron Technology.

Tradesman1 · Apr 2, 2014

Try as mentioned, may have to take DRAM voltage a bit higher, 1.575

warhammer3025 · Apr 2, 2014

Okay. manually set the timings to 7-7-7-20, and changed the DRAM voltage to 1.550v. Was unable to locate memory controller settings - heard that "System Agent Voltage" might be the same thing, but didn't want to risk messing with random settings
. hopefully it doesn't explode or something.

Weird bit - the AUTO setting put DRAM voltage at 1.472v yet both sets of sticks are listed as being 1.5v

warhammer3025 · Apr 2, 2014

Booted into windows and immediately had BAD_POOL_HEADER bluescreen.

Booted second time, same thing. Am reduced to using tablet. Please advise.

Tradesman1 · Apr 2, 2014

Have the latest BIOS? System Agent Voltage might help + 0.05

warhammer3025 · Apr 2, 2014

BIOS is E7681IMS v23.7

Cannot determine what setting System Agent Voltage is at currently. AUTO is the only information I can see.

warhammer3025 · Apr 2, 2014

http://www.hardwaresecrets.com/printpage/Understanding-All-Voltage-Configurations-from-the-Motherboard/995

If I read this page correctly, System Agent is indeed the setting I needed to replace CPUVTT.

changed System Agent to 1.16. windows has not yet crashed

Tradesman1 · Apr 2, 2014

Sounds good!

warhammer3025 · Apr 2, 2014

Crashed again. should I try to find the dump files for all these bluescreens?

warhammer3025 · Apr 3, 2014

Started playing a round of BF4. Computer crashed and entered infinite-boot loop.

Powered off, removed all RAM, and did the "check each stick and slot" routine. One stick failed when inserted into slot 2 (infinite boot loop), but passed when inserted into slot 4 (normal boot).

I'm very confused now.

Tradesman1 · Apr 3, 2014

Can you post your BIOS screens

warhammer3025 · Apr 3, 2014

I truly appreciate all the time and effort you're taking to help me.
Let me know if there's any other info I can provide.

Shot 1
http://s13.postimg.org/colrlv6dz/Ms_Stup1.png

Shot 2
http://s1.postimg.org/wz2f4grm7/Ms_Stup2.png

Shot 3
http://s10.postimg.org/k713f4nll/Ms_Stup3.png

Something I just noticed when looking at the screenshots - it reports my memory as being 4GB. I have 8GB! Checking CPU-Z shows that my computer isn't recognizing DIMM slots 3 or 4! Could the problem that caused the infinite-loop have actually damaged my motherboard?

warhammer3025 · Apr 3, 2014

I had assumed that the LiveUpdate utility that MSI provided would keep the BIOS updated (it even says one of the things it scans for is outdated BIOS), but it didn't. I updated the BIOS to the most recent version just now (from 23.7 to 25.8), and re-applied all the custom timing and voltage settings. The motherboard still only recognizes slots 0 and 1 as being occupied. I'm increasingly worried that actual hardware damage has occurred.

warhammer3025 · Apr 4, 2014

Computer was left on all day, and I've been doing some memory-intensive things like playing a match of BF4. So far, the computer hasn't crashed, but still suffers from only reading 4 of my 8GB of RAM.

I have no answers except that whatever was causing the instability seems to have been related to either RAM slots 2 or 3, and/or to the RAM sticks that occupied those slots. Now that the only slots being recognized are 0 and 1, it's running at reduced capacity, but otherwise "normal".

Of course, who knows how long this will last. When one bit of hardware fails, other failures can't be far behind. I fear that purchasing replacement parts is the only option left.

Tradesman1 · Apr 4, 2014

Might try loosening the CPU cooler then snugging it back down, if too tight in a corner or side may have the CPU slightly unbalanced and not getting ful connection to all the pins, that can throw off the MC (memory controller)

warhammer3025 · Apr 4, 2014

I can't imagine how the CPU might have shifted inside the securing bracket, or how the cooler might have been jostled, but okay. I'll give it a shot.

Tradesman1 · Apr 4, 2014

Like you, I'm trying to think of things it could be

warhammer3025 · Apr 4, 2014

I did as recommended, stopping short of actually re-setting everything (I have no thermal paste to re-seal the connection with). No luck.

Pinning down memory issues

Distinguished

Distinguished

Distinguished

Distinguished

Legenda in Aeternum

Distinguished

Legenda in Aeternum

Distinguished

Distinguished

Legenda in Aeternum

Distinguished

Distinguished

Legenda in Aeternum

Distinguished

Distinguished

Legenda in Aeternum

Distinguished

Distinguished

Distinguished

Legenda in Aeternum

Distinguished

Legenda in Aeternum

Distinguished

Share this page