BSOD's getting more frequent on new install

Restorer · Feb 20, 2013

My last hope is some of you bright guys might be able to read what is happening here with my machine.

Current installation is less than a month old but BSODs were starting to happen often on old W7 so I have to assume problem is probably hardware related.

BSODs or program crashes now happening several times a day. Program crashes don't seem to be related to particular apps.

There are no particular symptoms except for a couple of faint isolated beeps from the case within a few minutes of boot. These are not posting-type beeps

but very quiet and vary in length.

This is what I have done so far to try to isolate what hardware might be at fault but with no luck:

1. Replaced main drive with SSD. Other 3 drives have all passed Seatools long tests - several times in fact!
2. Replaced GPU with new
3. Replaced PSU with new
4. Re-seated CPU
5. Updated to latest BIOS
6. Tested RAM to 12 hours on memtest86 with 0 errors.
7. Swapped out cables and reconnected all companents from bare motherboard up.
8. This is third installation of Windows 8 with no improvement

It's hard to know what else to try. I can't think it can be the CPU - Everest reports it's running cool (low 20's idle) and true (good voltage figures)

and I am reluctant to replace this or the motherboard until I can see they are the culprits.

Here's my full system:

CPU: AMD Phenom II X4 965 3400MHz
M/Board: MSI 790FX-GD70 BIOS v1.17
GPU: GeForce GT 640
RAM: 4GB Corsair XMS3
PSU: OCZ ZT-650W
Drives: SanDisk 128GB SSD, Samsung 1.5TB, Samsung 2TB, Seagate 3GB.

I've attached the results from the Diagnostic Tool. Just hope some of you guys can come up with some further suggestions for me to try.

topgundcp · Feb 20, 2013

All the dump files indicated various causes for BSOD's and in some instances it indicated Memory Corruption and you did run the test for the memory with Memtest86+ and found no errors but some people think this test is also unreliable with their experience.

My suggestions are as follows:

Run sfc /scannow from elevated command prompt to ensure that all system files are not corrupted.
re-run sfc /scannow if necessary until you get a message: Windows did not find any integrity violation
Make sure you have the latest firmware updated for your SSD and connect to the first Sata Port
Re-seat all memory modules for good connection and use Everest to run the tress test on the memory modules for few hours. If the test fails, get into BIOS and manually increase the voltage a little bit higher and rerun the stress test.
Update NVIDIA driver. The latest release is on 2/18/2013. Also make sure the power cable is tightly connected to the GPU.

Good luck !!!

Restorer · Feb 20, 2013

topgundcp said:
All the dump files indicated various causes for BSOD's and in some instances it indicated Memory Corruption and you did run the test for the memory with Memtest86+ and found no errors but some people think this test is also unreliable with their experience.

My suggestions are as follows:

Run sfc /scannow from elevated command prompt to ensure that all system files are not corrupted.
re-run sfc /scannow if necessary until you get a message: Windows did not find any integrity violation

Make sure you have the latest firmware updated for your SSD and connect to the first Sata Port

Re-seat all memory modules for good connection and use Everest to run the tress test on the memory modules for few hours. If the test fails, get into BIOS and manually increase the voltage a little bit higher and rerun the stress test.

Update NVIDIA driver. The latest release is on 2/18/2013. Also make sure the power cable is tightly connected to the GPU.

Good luck !!!

Thanks so much for taking the time to look.

Just checked again - Nvidia Control Panel reports I do have the latest WHQL drivers: 314.07, though I think may have only done this today. Also my new card doesn't have a power cable

. I was surprised to see how small the new ones are too compared with my old GT 8800 which took up half the case!

I will try all the things you suggest tomorrow and report back.

Thanks again - you have given me a little hope that I can get this solved without replacing the whole damned thing!

Restorer · Feb 21, 2013

topgundcp said:
All the dump files indicated various causes for BSOD's and in some instances it indicated Memory Corruption and you did run the test for the memory with Memtest86+ and found no errors but some people think this test is also unreliable with their experience.

My suggestions are as follows:

Run sfc /scannow from elevated command prompt to ensure that all system files are not corrupted.
re-run sfc /scannow if necessary until you get a message: Windows did not find any integrity violation

Make sure you have the latest firmware updated for your SSD and connect to the first Sata Port

Re-seat all memory modules for good connection and use Everest to run the tress test on the memory modules for few hours. If the test fails, get into BIOS and manually increase the voltage a little bit higher and rerun the stress test.

Update NVIDIA driver. The latest release is on 2/18/2013. Also make sure the power cable is tightly connected to the GPU.

Good luck !!!

1. The sfc scan revealed one corrupted OS file and fixed it but I think this corruption was more the result of my problems rather than the cause since I have had 2 more BSODs since it was repaired. Corruption of downloaded files is also a regular symptom I forgot to mention.

Code:

 Info CSI    000006c5 [SR] Repairing corrupted file [ml:60{30},l:58{29}]"\??\C:\Windows\SysWOW64\en-US"\[l:46{23}]"windows.ui.xaml.dll.mui" from store

2. Confirmed I'm on latest firmware from SanDisk (2.0.0). Switched the SSD to Sata Port 1 (it had been on 3).
3. Reseated memory modules (these as well as most other components have been reseated and swapped around several times over the last few months). Been running Aida64's System Stability Test with only "Stress System Memory" checked) for over 3 hours now and also had multiple browser (Chrome) windows open and clicked around between Flash heavy sites with no ill effects.
4. Double-checked GPU is well seated and has the latest drivers. Also had another good peer at the card to check it really doesn't have a power socket

.

So stress on memory or CPU or GPU doesn't seem to be connected since crashes usually occur when I am hardly doing anything. Never had a crash playing Fallout 3 or Bioshock either for example.

Had two more BSODs after Step 1 which I did last night though none since. Latest Diagnostic Tool results attached in case the last two dumps reveal anything new. I noticed the last one was Bad Pool Header which I haven't seen for a while.

Just one thing I didn't mention but I have been running my 2 memory modules in slots 3 and 4 since the first two are blocked by the massive heatsink on my Fenrir Cooler. I doubt this is important though since everything worked fine in this configuration for almost 3 years. I will run the stress test again overnight in case a longer time reveals anything.

Thanks again for your suggestions. Any ideas on anything I could try next?

topgundcp · Feb 21, 2013

Suggest one more to do:

Run chkdsk /f /r C:

/f fix any bad spots on disk
/r marks any all bad spots on disk so Windows will not use them

re-run sfc /scannow

Restorer · Feb 21, 2013

topgundcp said:
Suggest one more to do:

Run chkdsk /f /r C:

/f fix any bad spots on disk
/r marks any all spots on disk so Windows will not use them

re-run sfc /scannow

OK thanks. Both checked out OK. No problems found by chkdsk and no integrity violations on the sfc scan.

Windows has scanned the file system and found no problems. No further action is required. 124674047 KB total disk space. 53703468 KB in 232456 files. 120896 KB in 48693 indexes. 0 KB in bad sectors. 401163 KB in use by the system. 65536 KB occupied by the log file. 70448520 KB available on disk. 4096 bytes in each allocation unit. 31168511 total allocation units on disk. 17612130 allocation units available on disk. Internal Info: 00 89 04 00 47 4a 04 00 6f a2 07 00 00 00 00 00 ....GJ..o....... 8b 01 00 00 1d 00 00 00 00 00 00 00 00 00 00 00 ................ 00 00 b7 0a 66 00 00 00 04 00 00 00 00 00 00 00 ....f........... Windows has finished checking your disk.

Now for fingers crossed. Still got those very faint beeps soon after boot. Usually one longish one (2 seconds maybe) followed a minute or two later by a short one. May be unrelated but I definitely don't like the sound of them !!

Longest I've been without a crash is 10 days and it has passed plenty of stress tests before and after BSOD's so I won't be putting the side back on the case just yet

.

Thanks again for all your help.

topgundcp · Feb 22, 2013

You are welcome. Hopefully there's no more BSOD

Restorer · Mar 2, 2013

I'm back again

. All went well for a few days then a couple of isolated BSODs followed by 3 in close succession this morning. There were 3 "Memory_ Management" and 2 "Bad_Pool_Header".

When SFC found 3 system files it couldn't fix, then reseated the RAM modules again and did a system restore. The result was the recent dumps were wiped but the system is stable again.

Now, I'm pretty sure as you first suspected this is connected to memory. I only have to fiddle with the modules and all is well again for days on end. But since RAM passes all the stress tests I'm thinking now maybe I have flaky DIMM slots?

This is what I propose doing before investing in any new hardware - maybe you can tell me if you think I am barking up the wrong tree or if what I am thinking makes sense?

If it happens again and when I have a spare afternoon, I will strip the thing back down and put the stock cooler back on so I can use DIMMs 1 & 2 which are currently covered by the monster Titan Fenrir heatsink. If I still have the problem I can then buy some more RAM and proceed from there.

Does this sound like a plan that makes any sense? One of the things you said to do earlier if it failed the stress test was to increase the voltage "a little bit". My BIOS offers increasing by 10 stops or 20 stops at a time. Not sure what this means in terms of volts but should I maybe try the smallest amount first? How much increase over the 1.5V would be considered safe?

Any and all suggestions gratefully received!

topgundcp · Mar 2, 2013

Looking at your MB Manual and I saw that they recommend to put the sticks in bank 1 and 2 with dual channel mode enabled. It is safe to give the RAM a little more juice, 1.6 would be OK.
Before putting your RAM in, use a piece of soft cloth to clean the connectors first. If you are not over clocking then the stock heat sink & Fan should be OK, just make sure CPU & Heat sink are clean and have adequate thermal paste. If you want a small size heat sink, this is what I use in one of my PC
Zalman CNPS9700 LED: The Reference Cooler : CPU Cooler Charts 2008: Part 4

Restorer · Mar 2, 2013

topgundcp said:
Looking at your MB Manual and I saw that they recommend to put the sticks in bank 1 and 2 with dual channel mode enabled. It is safe to give the RAM a little more juice, 1.6 would be OK.
Before putting your RAM in, use a piece of soft cloth to clean the connectors first. If you are not over clocking then the stock heat sink & Fan should be OK, just make sure CPU & Heat sink are clean and have adequate thermal paste. If you want a small size heat sink, this is what I use in one of my PC
Zalman CNPS9700 LED: The Reference Cooler : CPU Cooler Charts 2008: Part 4

Yes, I don't overclock but the reason I bought the monster cooler was because the PC lives in cupboard - though it is one with an open back, the room is south-facing and can get very warm in summer. Sounds though like your Zalman might be just as good and a quarter the size! I had been considering water cooled next time also.

I've never used dual channel mode either (can't remember why not, but maybe because the manual suggested there might be boot problems if I was using 3 & 4 in dual channel mode?). Anyway when things play up again I will try that along with bumping up the voltage a bit for the current slots. And if it's still no go I will pull the MB out yet again and re-arrange things with a different cooler.

Thanks again for keeping the ideas coming.

BSOD's getting more frequent on new install

New Member

My Computer

Well-Known Member

My Computer

New Member

My Computer

New Member

My Computer

Well-Known Member

My Computer

New Member

My Computer

Well-Known Member

My Computer

New Member

My Computer

Well-Known Member

My Computer

New Member

My Computer