BSOD, Hard Locks, Kernal Power Event ID 41

[Snot Rocket]

Hello

I am at wits end with this system. I have been trying to get this thing stable for a dogs age.
This is what I have done in the past 6 months.
I have reinstalled the OS a couple times, even Zorin OS on a stand alone clean drive.
I have run Memtest86+ found 1 bad set of RAM (RMA'd)
Reinstalled the OS as bad Ram was in machine at the time of the reinstall.
I have also swapped the vid ASUS 7970 OC and sound cards.
Power Supply seems to be OK, a CoolerMaster 850w modular
CPU has an Antec Kuhler 920 liquid cooling on it.

I had it kind of stable for about week after running AMD driver remover and reinstalling the driver.
Then it started acting up again.
Went to AMD and found the AMD program at nukes all traces of amd/ati drivers off the system. Ran that, installed ONLY the latest vid driver, still crashing.

CPU liquid temp according to the program sits anywhere from 38-42 C depending on the ambient temp on silent, ramped up full blown gaming it might get to 48 C

I have never had an issue vex me for so long.
A fresh set of eyes would be appreciated.
Attached it the SF Diag results.

Thanks!

 

[OvenMaster]

How old is that Cooler Master PSU? One year? Two? Any more than that and (in my experience) I'd be looking hard at it. Whenever I have had Kernel Power Event 41's, it's been either a PSU issue (solved by replacement) or wall socket power issues.

 

[Snot Rocket]

Ok, I perhaps have discovered the solution.

The AMD Phenom™ II X4 955 Black Edition Deneb Quad Core seems to picky with ram timing.
I should have picked up on this sooner, but I have been out of the tech game since 08 when I last worked in the field, I guess the rust is showing.

I was reading that the Deneb cores are picky with their ram timings.

Here is brief history of the diag process.

I had some ADATA 2000 that I originally got with the board and chip, but I was always having intermittent issues with this system. I moved to the corsair vengeance 1600 as I could get the ram in 2x4gb.
Was more stable, but not totally with the 1600, but started to get worse. Memtest86+ showed I had a bad pair, sent them back, expected this to correct this issue... so back to more searching. I thought perhaps that it was a corrupt install of Win7, reinstalled. Nope...
Checked the HDD's, and rechecked them repeatedly though out this process...still good.
Eliminated the sound card and vid from the equation and still Kernal Power Errors and hard locks.
Tried Zorin OS, it was more stable but still ran into crashes.
Crashes were all Kernal Power and some pointed to Power Supply. Having a background in electronics and tech, I watched the voltages under load, idle etc, nothing that showed a deviation outside the norm tolerances. Also I only need about 550w for this system as configured, and the PS is 850W...so I have the head room to not be pushing the PS as well.

I then posted here, thinking maybe it was a driver or something messing up with the ole MS Frankencoding in Win7 that was perhaps conflicting with hardware somewhere.

Power supply was suggested, and I even started to follow through with Cooler Master but I had a nagging suspicion that I was still going the wrong way with this...as well as I have never had a Cooler Master supply die, and have only every seen a handful of dead high end power supplies over the years.
Yes, I have seen the cheap ones blow up right out of the plastic, and users who have had whole systems wiped out by a cheap supply. That is why I always pushed people to spend on the PS unless they want to replace the whole system down the road due to PS failure.
I digress.

Right now I have set the 1600 to run @ 1333 with 10 10 10 26 timing. It has been stable through all the characteristic tasks that have caused it to crash in the past...streaming video, gaming, surfing the net.
It has been 48 hours now, with no crashes and I will be starting to research and experiment with 1600 timings.
I checked the Motherboard ram compatibility list, and the ram is compatible (so was the AData stuff too)
so it is just the chip not liking the ram timing.

The reason I am going on this long diatribe is that I have seen other users in other forums posting with the exact same issue regarding AMD chips with Deneb cores.
This may not always be the case for all Deneb, but at least it is a diag step to try early on to save a lot of grey hairs.

Cheers!

I'm going to run it through the paces for a couple days and then mark as solved if there are no further hiccups.

 

[OvenMaster]

As one who's had mobos, optical drives, keyboards and mice destroyed by power supplies dying, you do see my concern. RAM timing never entered my mind, as I've never altered mine from the defaults on any of my rigs. If fixing it so that your machine stays stable when it would otherwise fail, it does look like you've found your culprit. Fingers crossed that's all! Nice bit of detective work.

 

[Nacon]

Right now I have set the 1600 to run @ 1333 with 10 10 10 26 timing. It has been stable through all the characteristic tasks that have caused it to crash in the past...streaming video, gaming, surfing the net.
It has been 48 hours now, with no crashes and I will be starting to research and experiment with 1600 timings.
I checked the Motherboard ram compatibility list, and the ram is compatible (so was the AData stuff too)
so it is just the chip not liking the ram timing.

The reason I am going on this long diatribe is that I have seen other users in other forums posting with the exact same issue regarding AMD chips with Deneb cores.
This may not always be the case for all Deneb, but at least it is a diag step to try early on to save a lot of grey hairs.

Cheers!

I'm going to run it through the paces for a couple days and then mark as solved if there are no further hiccups.

Ok, I might have some details for you from what I've learned.
It seem I was experiencing the same issues you had. I have the same CPU but different RAM sticks from Mushkin. By default, my mobo would run this ram at 1333hz on 9-9-9-26 with auto voltage (forget what was it). It would post and run, but would crashes when gaming or surfing web at random moment after long period of time.

Cue grey hairs here.

I came across your post and decided to try pushing it to 1333hz 10-10-10-26 with auto voltage setting. It worked. But... well. Long story short version, I had to reseach far and long to get the proper details on my Ram for the timming and voltage. Turned out I had to make sure mobo run the ram at 1600hz 8-8-8-24 on 1.80v. Which is actually overclocking by my mobo's standard.

Now I have a hunch that this CPU is not stable on 9-9-9-26, no matter what ram stick it's from. Maybe you can try overclocking your RAM to 8-8-8-24 @ 1600hz? Just don't jump straight to 1.80v though. I don't have your RAM's model number.

And of course, thank! If it wasn't for your post, I wouldn't have fixed my problem. Thank again!