16G RAM works fine. 32G = different BSOD's ? what the........

[mrblotto]

Good day all 🙂

I have a roughly 2 year old
-ASUS Z97-E LGA 1150 mobo
- I7 4790K CPU
- EVGA 980Ti GPU
-32G G.SKILL TridentX Series 16GB PC12800 RAM 1.5V
-Win10 Pro 64bit
-Auzentech Prelude 7.1 Sound Card

Lately (past 4 months or so), my machine will just freeze, then I'll get the BSOD (or frowny face of Death) and the error will be different each time. Once the BSOD's start, they keep happening until I remove 2 of the 4 DIMMS. After a few days (when I think of it lol) I'll toss the memory back in and it works fine.........until the next time.....which maybe a week to a couple weeks later.

Memtest passes each and every time I check all 4 DIMMS, which leaves me with a couple possibilities.

1. Since it started happening rather 'recently' (relatively speaking), I'm inclined to think PS, but have no idea how to rule that out (except swapping out PS's)

2. I dont think memory will go 'bad', then suddenly work fine again. Maybe I'm asking too much of the memory and it just goes boom (undervolted?)

I have done some dmp file analysis - the errors appear random

Edit: I just looked at the 5 .dmp files in the minidump folder:


7/2/17 918pm - ntoskernl.exe (IRQ_NOT_LESS_OR_EQUAL)
9/19/17 812pm - afd.sys (BAD_POOL_CALLER)
9/19/17 949pm - ntoskrnl.exe (SYSTEM_SERVICE_EXCEPTION)
9/29/17 822pm - WdFilter.sys N/A
10/1/17 621pm - ntoskrnl.exe (PFN_LIST_CORRUPT)

I wanna say it's power-related, but where - voltage or the PS. I'm thinking the obvious thing to check is the BIOS and make sure the RAM voltage is 1.5.

What say you AT?

Many thanks,
Blah-Toe

 

[XavierMace]

Your memory's voltage spec is 1.5v, but what are you actually running it at?

 

[mrblotto]

Oh, right...that may help lol.

I took a pic of the BIOS screen showing various info. Of course photobucket doesn't do embedding anymore :/
CPU:
Frequency: 4000 Mhz
BCLK: 100.0 Mhz - No idea what that is
Vcore: 1.077 V
Ratio: 40x - also, no clue what this is either

Memory:
Frequency: 1333 Mhz
Voltage: 1.500 V
Capacity: 32768 MB

Voltage:
+12V: 12.096 V
+5V: 5.040 V
+3.3V: 3.296V

Of course, this is just when the machine is basically idling. There shouldn't be much (if any) stress on components when it's basically sitting there doing nothing.

 

[RLGL]

Does it matter which pair of sticks are removed? A's or B's.
Power supply size and brand?
My thoughts are more related to the MB itself failing.

 

[mrblotto]

Does it matter which pair of sticks are removed? A's or B's.
Power supply size and brand?
My thoughts are more related to the MB itself failing.

I've only removed the 2 'closest' sticks. The HS for the CPU gets in the way of removing the other 2 easily. I think the HS is a
Zalman CNPS5X Performa or Cooler Master 212 EVO. I got it locally so I can't look it up atm.

The PS is a EVGA SuperNOVA 750 G2 220-G2-0750-XR 80+ GOLD

Prolly couldn't hurt to try removing the other 2 sticks instead

 

[Topweasel]

Try swapping between chips between the 4 and see if you can recreate the issue. I will note that I had a X79 setup that I picked up 4 8GB sticks. After a week random restarts and BSOD's happened. Removed 2 sticks it was fine. Eventually I tried 3 sticks to figure out which it was. No combination of the three sticks had an issue either. In the end it was only with the fourth stick that I had any issues. Got a different set of 4, worked perfectly fine after that. Basically what it came down to was this manufacturer for this line, had a fault tolerance on chips much looser than anyone else. That tied with a controller dealing with 4 sticks like that were too much for it.

 

[Linflas]

In case you haven't already done this you might want to check the QVL. They have some sticks that are good with 2, not good with 4 on the QVL. I ran into this with my new ASUS Maximus IX Formula. I had 2 8GB sticks and bought 2 more and suddenly my system was unstable. Turns out that according to the QVL it was fine with 2, not with 4. No idea why that would be and it didn't occur to me to check that before I ordered the 2 additional sticks. Fortunately I was able to return them for a full refund.

 

[dave_the_nerd]

Given that it was fine for two years, it's probably just slow degradation of some component that only crops up in certain load situations. Could be one of the DIMMs, could be one of the slots on the motherboard, could even be something in the CPU, although that's way less likely.

The above advice to try different pairs of DIMMs and different slots is good advice, but you may have to hobble along like this until the problem is consistently repeatable and easier to identify.

 

[piasabird]

Timings should be the same on all 4 parts for RAM. I think this is a motherboard issue. I think most power supplies will not effect RAM unless the power is fluctuating in either the house current or the power supply rail. RAM just doesnt use much power, but they might overheat. However, often Motherboards May have configuration issues or if you push them closer to the full specs, there may be issues. In the past sometimes the motherboard might have a BIOS update that might make it run better. However, if this is a motherboard design issue there may be nothing you can do.