Random Crash - RAM compatibility fault?

Immo

Honorable
Sep 25, 2013
8
0
10,510
Hi, sorry for errors, I'm from Poland and ain't very knowledgable in technical English, not to mention I wrote this quite fast :)

My PC crashes randomly. Either I simply browse the Web or leave the PC for few minutes to do something else and suddenly the cursor becomes stuck and whole system crashes. It also once happened just a second after system load: loading cursor appeared for brief moment an then all system went irresponsive. There is no response from keyboard (even CapsLock etc. diodes won't turn on/off). Whole system crashes to the extent that the power button does not respond instantly (only 4sec. soft power-off works), and when I press the Reset, computer will still be working from 6-7 seconds, then it'll turn off and suddenly start again after 6-7 seconds.
I've checked the system log and there's not a trace of any error, so I presume it's caused by a hardware error. If the crash happens on 12:00, there will be log info about something that happened at 11:45 and next log entry is from 12:02, reporting that the last shutdown was unexpected. I had a plague of events like that about a week ago (Sept 17) and it turned out that the cat jumped up on the PC and made the case bounce so much that one of the DIMMs (2x4GB 1600MHz) became loosened. After that the incidents started to occur. After I fixed that and ensured the RAM modules sit tight, the problem disappeared for few days. However, it reappeared on September 20. I had RAM modules and every other hardware piece checked and they sit firmly in place. I checked the status of RAM both with utility built into the BIOS and Memtest86 - both methods inform me the RAM modules are fine and dandy. The system temperature is more than acceptable as I installed lots of fans (my previous PC nearly died in flames due to the overheat issues). I've updated my BIOS, chipset drivers, used a Microsoft Hotfix for similar problem, updated some other drivers, but it reoccurs. After I removed one of the DIMMs, the PC became more stable, but crashed after a while. The same happened with the second RAM module (in other slot).

My RAM modules have default voltage of 1,65. Sandy Bridge architecture processors don't accept RAM running on 1,65V, but BIOS automatically reduced the voltage to 1,5V. Can this voltage reduction be a reason of PC crashing?

Do I have to buy a dual channel DIMMs with 1600MHz rating and 1.5V voltage? Should I borrow a set of DIMMs like that to check the system stability? Or maybe it's something else? Motherboard (LAN adapter specifically) or power supply problem?

Few things I've noticed:
- it usually occurs while something Internet connection involved happens: Steam is running and downloading something and/or Firefox is on and I start to do something rapidly (e.g. responding to a thread on forum, sending Facebook PM or choosing an option in survey - stuff like that); starting YouTube playback was a method to achieve instant crash;
- after using the OC feature of my motherboard, when RAM was running on the same voltage (1,5V), but at higher rate (about 825Hz / DIMM), the system became a bit more stable - it even allowed YouTube playback - but it finally crashed anyways; after returning to the default setting (800MHz / DIMM) it runs just the same, with YT playback possible; it haven't crashed since, but undoubtedly it will.



SPECS:
Win 7 Home Premium 64bit
Intel Core i5 3.2GHz (2nd gen)
2x4GB Kingston RAM 1600MHz (KHX16C9T3K2/8X)
Asus AMD Radeon HD7790 2GB DCII OC graphic card
P8H77-V LE motherboard
AeroCool Templarius Imperator 750W PSU (Silver)
4 chassis fans, 1 HDD fan, 2 graphic card fans, 1 CPU fan
 
The gold standard for testing ram is memtest86+
You should be able to run several passes with NO errors.
Go to the Kingston web site and access their ram selection app for your motherboard.
If supported, you should find your kit.
1.65v on the spec says that you need to oc the ram to get it to run at 1600. Most likely, it is defaulting to a slower speed to be able to run at all, likely 1333. Try upping the voltage a bit.

One other possibility is your psu. It is a tier 4 unit(not recommended for stressful situations) on this list.
It may fail to meet ATX specs. http://www.eggxpert.com/forums/thread/323050.aspx
 

Immo

Honorable
Sep 25, 2013
8
0
10,510

My current RAM kit is NOT on the list.
I'm not an expert, I consulted the setup with some folks who theoretically were knowledgeable about that sort of things, but I guess they mislead me a bit.

On default, it ran on 1,5V and with 2x 800MHz speed, or at least it's what software shown. After the BIOS O.C. it suddenly had 2x ~825MHz with 1,5V, which was a bit strange.
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
I set the voltage to 1,65 and HWInfo64 told me each module had 800MHz. I was about to report you this, but when I typed "q" in "frequency", it died again. Now, according to your advice, I set the clock to 1333MHz @ 1,5V. We'll see if it works.


When I was buying parts for this PC, I found positive reviews of that PSU. One particular guy from a pro site said it had no problems with running at stressful settings with O.C.'d CPU and with games at max settings.
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
Died again with 1333MHz @ 1,5V setting. Lowered the freq to 1066, but I don't believe it'll save the PC from crashing again soon.
Gonna try other DIMMs tommorow.

UPDATE: No matter what settings I'd apply, PC crashes after a while.
I'm working on borrowing another PSU, Platinum rated Dell PSU @ 1300W for testing.
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
Replaced the DIMMs with Patriot Vipers @ 1,5V.
Error still occurs. I will try other PSU.
Maybe - since it all started with loosened DIMM - that it's a mainboard's fault?
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
All drivers are up to date, only the graphic card isn't. Maybe it's graphic card's fault?
I did some thinking and analysing - first malfunction was caused by loosened RAM module on September 17th, then it disappeared, but reoccurred after plugging in a new display, but not at once.

It went like this: On September 25th I first plugged in the new monitor and tried to run it at native resolution - there was some problem with it (system said it was on native resolution, but it was downsized and 3/4" of the screen from each edge was unused. I fixed this with Catalyst,. but then it reconfigured the system scheme to default without a possibility to configure it again. I tried to update the GPU driver, but it crashed and BSODed during the install.
Then I tried again and it was updated, but the driver started to be non-configurable (when I tried to set up anything, it returned me with a pop-up with "There is nothing that can be configured in Catalyst" or something like that.
So I removed the driver and reinstalled a previous version. And the crashes ensued. What the heck?

P.S. YouTube playback is - again - the most certain way to crash the PC. But games still don't seem to crash it.
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
My friend ran a full test on each and every part of the PC. Processor, PSU and graphic card are flawless, new RAM is immaculate.
Mainboard is somehow damaged and it needs to be replaced.
 

Immo

Honorable
Sep 25, 2013
8
0
10,510
OK, it's fixed. It took a while as I needed to buy few important things first. It was indeed a mobo fault.
Changed mainboard to MSI Z77-G45, rest of the hardware is the same and the PC runs perfectly. The MSI mobo itself looks and feels like a lot better designed piece of hardware, by the way.
Thanks for help, guys! :)