please help X58 eclipse 6gb DDR3 1333

[tic_toc]

hi all, brand new to this forum. Unfortunately my first post will be one where I ask for help

built a new desktop about 4 weeks back. i get seemingly random Blue Screens, all have this 'Memory Management 1a' error in the title. I have nothing overclocked, manages to fall over with fail-safe defaults in the bios even!

MSI X58 Eclispe bios 1.A (10th bios revision)
GSkill F3-10666CL9T-6GBNQ DDR3 Ram
i7 920, D0 revision
GTX280
old 850w coolermaster power
Windows 7 ultimate x64

in the time i have outside work I haven't been able to sort it out... like
run ram at 1066 and overvolted
raising RAM voltage to 1.65V, increasing QPI voltage
manually setting timings instead of 'auto'

i would love to get the build working properly for my family to use. any help is much appreciated

J T

[Tradesman]

Prob just need a tad more voltage to CPU and/or RAM, try setting the voltages manually, MSI not well known for effectively managing the voltages to true spec.

[tic_toc]

tradesman, thanks for the reply.

I am trying what you suggested, I added a little voltage ( 0.1V ) to the CPU and QPI, and am running the ram 1333 @ 1.65V

but seeing as the memory is rated for 1333 MHz @ 1.5V could the be any other underlying problem?

[GSKILL TECH]

Should not need such a high memory voltage for DDR3-1333. 1.60V is well enough.

Thank you
GSKILL SUPPORT

[tic_toc]

hi thanks for the advice. I would much rather run everything at stock voltages and frequencies where possible. anyway, the PC blue screened with those voltages (mem at 1333 & 1.65V).

I ran memtest x86 for 6.5 hours and it returned no errors (6 full passes). Could it be that these memory modules are incompatible with the MSI motherboard?

Has anyone heard of bad Windows installs returning 'Memory Managent ' errors?

[Tradesman]

Can you post your BIOS settings

[tic_toc]

all the advice is appreciated tradesman, my setting are as follows:

all green power (phase control) settings disabled, spread spectrum off, C1E & HT on, C-state on.

CPU > 133x20= 2666 @ 1.104V (stock Volts, Turbo Boost off)
QPI > 4.80GT with +0.12V
DRAM > 5x133=1333 @ 1.60V
Uncore > 10x133= 2666
IOH > 1.20V (+0.1V)

as for the memory timings

2N-9-9-9-24 (command rate-CL-tRCD-tRP-tRAS, manually set) along with
tRFC 74
tWR 10
tWTR 5
tRRD 4
tRTP 5
the tRFC clock setting does not show up in the bios, but shows in CPUz as 33

again, these settings allowed me to run 40 loops of an old crysis benchmark last night, as well as play stalker with no problems- and this morning BSOD 2 times in a row shortly after bootup, at desktop idle... random.

maybe I'll check the NB heatsink is seated properly... even try running in dual channel mode eh

[tic_toc]

no bios changes have helped, I actually measured gaming performance was being reduced by about 30-40% with bad stuttering and loading times.

I re-installed windows, and performance came back to normal for a day. After updating directx for the call of pripyat benchmark, I had a series of blue screens, and performance is choppy again. especailly loading a game, takes minutes to read a dvd and begin running.

I really can't be sure if this is software or a hardware problem, and havent been able to troubleshoot.

Any more suggestions greatly appreciated.

[Tradesman]

Above you show tRFC as 74, but then say it shows in CPUZ as 33, might want to try around 80. What driver are you using for your 280. Might want to try the newest, or if using the latest might want to fall back a version can find them at:

http://downloads.guru3d.com/Videocar...Vista_c32.html

[tic_toc]

that was a typo, meant to write tRC instead of tRFC, like

tRC not shown in bios, and tRC reported as 33 clocks in CPU-Z. tRFC was set at 74 and reported as 74 clocks by CPU-Z, also tried 88 clocks on this setting.

i have it set to 'fail safe defaults again.

the nvidia driver is 195.62, previous to this was running 191.07 whql. the 280 is a factory overclocked model by msi (648c/1296s/2304m)

[tic_toc]

solved

running Memtest+ on each stick individually revealed one stick of RAM is faulty. the two other sticks each ran memtest for hours with no errors, the third began spitting out errors immediatley and blue screened instantly when booting up into windows.

so thats it, finally got to the bottom of it. really confusing how it could run all day sometimes and pass memtest in 3-channel mode with a faulty stick, but I suppose it always pays to check each stick individuallly of you suspect a RAM problem.

thanks tradesman for the help- merry christmas

[Tradesman]

Same to you, should have the stick replaced by then and be cookin' on all cylinders