WIN10 BSODs and unstability - Memory is the suspect

micz81

Prominent
Apr 15, 2017
7
0
510
Hi! I hope the info below is sufficient, and you will find time to help me out here :)

PROBLEM:

I'm recently getting lots of BSODs. I was searching about that and some clues from similar cases around the web lead to the memory being the cause of all that instability. I get BSODs, freezes, problems with shutting down and waking up from sleep. Sometimes it happens under stress (rendering stuff from Adobe After Effects), and sometimes without any visible reason (for example while browsing the web).

INFO:

STEPS MADE:

    1. Memtest:
    • no errors
    2. System Stability Test in Aida64 Extreme:
    • - CPU temperatures seems too high
      - and there's a problem with CPU Throttling (which - as I read on the web - should stay on 0%, right?)
      - Here's a link to google folder with few screenshots from the test.
      - The test went for about half an hour
      - besides those temperatures - does the screenshots tell you something is wrong?
    3. System Stability Test in Aida64 Extreme - only the stress memory test
    • - (it seemed for me, that it haven't started at all in the initial test when all the tests available were picked by me to run - so I decided to run it separately)
      - outcome after 9 minutes: CLOCK_WATCHDOG_TIMEOUT BSOD
      - details: BlueScreenView full raport (18.04.2017 22:19:12 event)
      - screenshots: in this google folder
    4. Another System Stability Test in Aida64 Extreme - only the stress memory test
    • - (I have run another test just for the sake of error reproduction)
      - outcome after 51 minutes: KERNEL_SECURITY_CHECK_FAILURE BSOD
      - details: BlueScreenView full raport (19.04.2017 02:36:58 event)
      - screenshots: in this google folder
    5. again - stress memory test - but only Patriots
    • - (I have removed pair of Kingstons from the motherboard and run the test again)
      - outcome after 3 hours 40 minutes: no BSOD!
      - screenshots: in this google folder
    6. again - stress memory test - now only Kingstons
    • - (I have removed pair of Patriots from the motherboard and run the test again)
      - outcome after 5 hours: no BSOD!
      - screenshots: in this google folder
    17, 18, 19. stress memory test - Patriots+Kingstons with XMP enabled
    • - After few other tests in AIDA64 - by which I was trying to find out which component triggers CPU Throttling - I have enabled XMP Profile 1 in BIOS and run the test again with all the four modules installed. I have done this test 3 times for error reproduction.
      - outcome in test 17: it was OK for 30 minutes, so then I have stopped the test.
      - outcome in test 18 and 19: no BSOD, but AIDA64 got red, saying: "Warning: Hardware failure detected! Test stopped"
      - screenshots test 17: in this google folder
      - screenshots test 18: in this google folder
      - screenshots test 19: in this google folder
    20. I have bought this 4x8GB kit: CMY32GX3M4A1600C9
    • - now I'm waiting for the shipment and I will post here tests from AIDA64 when I get them
    21, 22. Stress memory tests - 4x8GB Corsair Vengeance kit: CMY32GX3M4A1600C9
    • - I have enabled XMP Profile in Bios and run two stress tests in AIDA64 - one only for the memory and one for memory and CPU
      - outcome after 3 hours: no BSOD!
      - screenshots - only memory
      - screenshots - CPU+memory

      So it looked like everything is OK and problem is solved, but... next day when I have booted the system I got black screen. Computer was on - fans and disks were spinning - but the screen was black, there was no signal on HDMI cable.
      - The debug LCD on motherboard was viewing POST codes and it seemed that the POST procedure is going trough a loop ending on code 32 - "CPU PEI Initialisation"
    23. replacements and boots:
    • - 1. I have replaced the Corsair kit with two Patriot kit (2x8GB), and boot again - everything went good and Windows loaded
      - 2. then I have changed the modules for Kingstons - no boot, and error 32 loop
      - 3. then I have changed again for two Patriots... - no boot this time either and error 32 loop

      So... it looked like there is no problem with the memory modules, but rather with the mainboard?

      - 4. then I have changed the Patriots for Corsairs again... but this time only two of the four piece kit - boot went OK
      - 5. then, again reboot on the same setup, and... - no boot, error 32 loop.
      - 6. then I was devastated :)
      - 7. then I took only one Corsair module and place it in the slot nearest to the CPU - boot went OK.
      - 8. then reboot - still OK
      - 9. then reboot and in BIOS I have enabled Enhanced Stability (screenshot1, screenshot2) and shut down. I have added the rest of Corsair kit (all 4 modules were now placed on mobo) and booted again - boot went well
    24. CPU-Z screenshots - memory and SPD
    • - system was stable for few reboots, and then... crap:
      - after shut down for night, in the morning I got again black screen and error 32 loop.
      - So after reboot i entered BIOS and there a message was waiting for me - don't remember the whole text, but the idea was: "Resent configuration may be causing problems"
    25. So I have disabled XMP profile in BIOS
    • - system boot up. Now I'm gonna try enabling XMP again and increasing voltage as Vic 40 suggested
    26. Again 32 loop
    • - I had no time to set the increased voltage and test, because I had some work to do. So I booted up the system on 4x8GB Corsair and again got black screen and error 32 loop.
    27. So I have replaced 4x8GB Corsair with 2x8GB Patriot kit and boot up
    • - everything was good, the system was working for several hours. So I thought that I leave it like this, with 16GB of reliable Patriot RAM, because I had work to do and better 16GB working, than constant failures with 32GB
    28. And.... Again 32 loop
    • - After hew hours of non-problematic work on Patriot's 16GB kit I have shut down the system for night, and in the morning ... I got black screen and 32 loop again
    29. So I have removed one 8GB module and booted up
    • - everything OK
    30. So I shut down right after the system loaded up properly and I have added the second 8GB Patriot module and booted up again with 16GB on board
    • - system booted up OK (.. . . .. I'm confused).

NOW that should be a hint, that the motherboard is broken, right? I run the system on 16GB, then shut down, than power on again and I got 32 loop error. So I take out 8GB, boot up, everything's good. So I shut down, add 8GB, boot up and everything's good again. It looks like motherboard, after few hours of working and then few hours of not working, has problem with booting up again on the RAM set with which everything was OK just few hours earlier. And that applies to all three sets of RAM - Patriots, Kingstons and Corsairs (16GB and 32GB either)

Should I claim warranty check on the motherboard?


QUESTIONS:
The questions below can become a bit irrelevant, while the process of testing. But I leave them here, for others.

    1. Is it possible that those BSODs (or some of them) are caused by the fact, that I use two DIFFERENT sets of memory modules (2x8GB Patriot G2 Kit and 2x8GB Kingston Hyper Beast Kit) ?
    2. Do the info provided by CPU-Z match the specifications of those memory modules? I mean speeds, latency, etc. - maybe some of you need one look to say "hey, that frequency from CPU-Z seems not right for that model...", or something like that?
    3. Let's say I would like to make all those 4 modules of memory... the same - exactly same brand and model (regardless if it affects stability or not). Which of those two memory sets mentioned above would you recommend as a base? I should rather find second set of Patriots, and sell my set of Kingstons, or maybe the other way - sell Patriots and look for Kingstons?
    4. If I fail to find the second pair of my Kingstons or Patriots in the web (those are discontinued models, so out of stock everywhere) I would buy two identical kits 2x8GB (= 32GB in 4 modules). Which 2x8GB kit of dual channel DDR3 do you recommend in the similar budget as my Patriots or Kingstons?


Looking forward for some answers from you guys - I'm investigating those BSODs for a month, and I'm going slightly mad....

Of course if additional info is needed - let me know, and I'll deliver :)

Thanks!
 
Solution
We concentrated on ram right now,but something else,can you try another Antivirus program or disable Avira for now and see how it goes?
Other thing is doing a clean install from the gpu drivers,you can use DDU or when installing drivers go to the advanced options and choose a clean install there.Might even be worth to install a little older drivers for testing.
DDu also gives the option to block windows from updating the drivers,keep that enabled.
Looking at the Bluescreen view report right now ...

No real ways of testing a motherboard other than to replace it.You have the latest drivers for it i assume.

Another thread with your issue...

Vic 40

Titan
Ambassador
Is XMP enabled in the uefi (bios)?

Might well be that the different sticks make the bsod's happen.Look at the next for more info,
http://www.tomshardware.com/reviews/ddr-dram-myths,4155.html#p4
Look at "Just Adding More Dram".

Maybe we need to adjust timings and speed manually also probably add some voltage to make them play well together.

I assume you have no problems when using them per set? Tested that?
 

micz81

Prominent
Apr 15, 2017
7
0
510
Thanks Vic 40 for the answer!

- XMP is not enabled. I have edited my initial post and added there a link to my BIOS settings (bunch of screenshots from the BIOS hosted on google sites).
- I have run Stability Test - all added in the edited initial post.
- I have read the article which you have linked. I'm not sure if I understood everything, but for sure I get the point in general. Thanks.
- I presume, that adjusting timings and speed manually does not GUARANTEE that it gonna be all right from now? So to be SURE that the problematic memory setup is ruled out as a source of BSODs I should rather replace those 2 pairs of 2x8GB with one pair of 2x16GB, or at least with 4x8GB kit?

This of course doesn't mean I don't want your advice in terms of adjusting the speed and timings - I do!!!! :)

- I haven't tested one pair of my two pairs separately. Those BSODs are hard to reproduce - I don't know under which circumstances they appear. For example - last 3 days passed without any BSOD. So let's say I will take out Kingstons, and leave only Patriots, and everything will be all right for few days - that doesn't prove anything :/ Meaning: is there any way of testing that more precisely?

Bottom line: Should I get rid of those modules and buy 2x16GB or 4x8GB kit? Is it THE RIGHT WAY to handle this situation? Is it recommended as a PROPER setup? Or maybe adjusting the speeds and timings manually will be enough to rule out memory as the cause of BSODs?
 

micz81

Prominent
Apr 15, 2017
7
0
510
I have tested the modules separately and it seems that they work better alone (see my initial post - STEPS MADE)

Vic 40, can you guide me trough adjusting timings and speeds? Now when I have kinda reliable procedure of testing the setup, I would gladly check if adjusting stuff in BIOS will do the trick and if I would be good without buying new modules.
 

Vic 40

Titan
Ambassador
Well best thing,and most easy one,would be to get one kit of 16gb of ram.

For timings look at the next picture you made from cpu-z,
https://drive.google.com/open?id=0B7nQqP_y3cgvdldqaVVBRUNobW8
look at the XMP settings,keep the speed at 800mhz (=1600mhz),make those 5 timings 9-9-9-27-38 and see what that will do. For voltage could you add 0.05 for a few times,like sett the Dram voltage at 1.55V and test,if not enough for stability can you still ad 0.05V for two more times up to 1.65V and be perfectly safe.
 

micz81

Prominent
Apr 15, 2017
7
0
510
thanks Vic 40 for the answer. But in the meantime I have ordered 4x8GB RAM kit, in hope it solve the problems. I will be back here when run some tests. And maybe then there will be time (let's hope not! ;)) for some additional tweaks of RAM settings.
 

micz81

Prominent
Apr 15, 2017
7
0
510
I have made some tests and updated initial post. I have one more question regarding the point 24 of my test list: are those CPU-Z screenshots OK? Do those settings look OK for you for my 4x8GB Corsair Vengeance kit?
 

Vic 40

Titan
Ambassador
Yes they do.

Could well be that the enhanced stability option feeds abit more voltage to the memory controller and makes it work because of that.If still having problems could you look at that.
Other thing is if still having problems in the future could you provide abit more voltage to the ram itself. Right now the sticks run at 1.5V which is stock setting,you could add voltage up to 1.65V by increments of 0.05V and see if that helps if problems arise again.I would try this first.
Just some extra info.:)
 

micz81

Prominent
Apr 15, 2017
7
0
510
Vic 40, first of all - thank you for the help, man! It's very kind of you that you manage to advice me here in this loooooong shitty journey :) You don't know how much I appreciate this - the whole malfunction is so weird for me, and so hard to explain in easy way, and so hard to reproduce errors,.... that I began to loose hope. When there is someone out there who tries to help me... - priceless feeling! :)

Back to the hell - I got another error 32 loop (see edited initial post) and now I'm gonna increase voltages as you suggest. Will be posting outcomes.
 

micz81

Prominent
Apr 15, 2017
7
0
510
Vic 40, thanks for the links and suggestions. I will get back to it after I do some work.
But... I'm a bit confused about the system's behavior, and therefore about the idea of fiddling with RAM settings, when the motherboard seems like a suspect to me right now. You understand me, I hope - let's say I will increase the voltages, and will go on the journey of testing and finding the best values, but what if possibly faulty mobo will mess around this process "in the background"? Would you please take a look on points 26-30 of initial post and the little comment just after those and tell me what do you think or what to do to rule out the mobo's HARDWARE failure?
 
Try running with just the 16GB Patriot RAM for a long time. Stress test it and then go at least a week of normal usage to see if your BSODs are gone.

I think you may have some RAM incompatibility. Sometimes when swapping RAM in and out you need to reset the BIOS for the system to get back to normal operation. This could be the case in all your testing. Start with the above test, resetting your CMOS (no overclocking), and see if this fixes your BSODs.
 

Vic 40

Titan
Ambassador
We concentrated on ram right now,but something else,can you try another Antivirus program or disable Avira for now and see how it goes?
Other thing is doing a clean install from the gpu drivers,you can use DDU or when installing drivers go to the advanced options and choose a clean install there.Might even be worth to install a little older drivers for testing.
DDu also gives the option to block windows from updating the drivers,keep that enabled.
Looking at the Bluescreen view report right now ...

No real ways of testing a motherboard other than to replace it.You have the latest drivers for it i assume.

Another thread with your issue,
https://www.google.nl/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwi7r76v5dHTAhXQYVAKHW9eB58QFggiMAA&url=http%3A%2F%2Fwww.tomshardware.com%2Fanswers%2Fid-2833545%2Fz97x-ud5h-boot-loop-stops-code.html&usg=AFQjCNEeB-yIDUh8RQlMFKGcTUonaUPbKw
it has some things you can try.one is adding a little voltage to the ram+memory controller,other has to do with the sata ontrolers on the board and how to set them in the bios.
 
Solution