Memory Management BSOD with no memtest errors

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
Hello everyone,

I'm experiencing a once a day Memory Management BSOD (typically 0x0000001A).

I have run 7 passes on Memtest 86+ and gotten 0 errors, and at this point I'm not sure what I should do next.

Below is a summary of installed hardware:

CPU: AMD Phenom II X4 955
Motherboard: Micro-STAR 790FX-GD70
Memory: 8GB Kingston DDR3
GPU: Radeon HD 5750

We're running on Windows 7 64 Bit as well.

In addition, I have tried increasing the CPU/NB voltage up to 1.4 based on the recommendation of a member of this forum, which seemed to help for a while but now the BSOD seems to have become more frequent.

I am at my wits end and do not know what to do next or which parts to replace.

Any help would be extremely welcome.
 

MyNewRig

Honorable
May 7, 2013
278
0
10,860
I was on your same shoes on a completely different system, Intel Haswell, Memtest86 would pass without any problems, but windows would give memory_management BSOD every other week, i was able to expose that the RAM was indeed defective by using memtestOSX (not an option in your case because you run AMD platform)

It was fixed when i replaced the RAM, the best thing to try to know if your memory is actually causing it or not is to drop memory clock and loosen timings and see if it still happens, if it goes away then replace your memory, if not then you would need to look elsewhere, your MB or CPU.

What RAM frequency are you running?

Best advice is that start replacing one component at a time going for easiest to replace to the hardest, i would start by RAM, then Motherboard then CPU at the end.

Drop your RAM frequency, loosen timings, increase voltage and report back if things become stable.
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
Could you clarify what you mean by dropping the memory clock and loosening timings?

I'm a hardware noob and need some definitions.

Last night I pulled half the memory to start seeing if memory was the problem. I got another BSOD today, this time a system service exception. Not sure if the change in error signifies anything.
 

MyNewRig

Honorable
May 7, 2013
278
0
10,860
I am not sure how to do that on an AMD platform/BIOS because I am not familiar with AMD platforms, I only use Intel recently, but by dropping clock and loosening timings I mean that for example if you have memory rated at (the speed and timings writing on the box) of 1600Mhz 9-9-9-24 for example, you would drop the frequency by making it run at 1333Mhz in BIOS and also you could make the timing looser instead of 9-9-9-24 you would go for 11-11-11-28 for example also through BIOS which makes things more stable.

I call upon the AMD experts in here because like I told you, I am not exactly sure how to do that in an AMD platform, but you can always google how to implement that, it should be fairly easy.

If by doing that, the system becomes stable, then most likely your memory is defective and needs to be replaced, if by doing that, you still get memory_management BSOD then maybe it is the board or the CPU.


Keep in mind that software like bad, conflicting or incompatible drivers can cause memory_management BSOD .. so an important part of your testing is to test on a vanilla windows (one that is freshly installed) just to eliminate any possible software causes.

You have not told me what frequency and timing is your RAM rated at? this should be written on the box or the sticks .. should look something like this 1600Mhz 9-9-9-24 or 1866Mhz 9-10-10-28 . something like that .. maybe you are running RAM speed that is higher than your CPU can handle ...

 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
I am currently at work but will post the Frequency and timing when I get the chance.

I will point out that the machine worked for 3 years with no problems and the issue manifested after we moved to a new apartment 5 hours away, I'm not sure if something could have been damaged in the move.
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
So as I was trying to test the RAM, I used two of the 4 sticks in the left two slots, which eventually crashed

The other two of the sticks in the left slots, which eventually crashed

The first two sticks in the right two slots, which wouldnt POST

And the last two sticks in the right two slots, which eventually crashed

Don't know if this means anything but thought I'd mention it
 

MyNewRig

Honorable
May 7, 2013
278
0
10,860
Of course voltage could be THE PROBLEM, even though I have not seen consumer RAM rated at anything above 1.65v because 1.7v or 1.9v are very high for an Intel Core CPU, but maybe with AMD the situation is different and the IMC (Integrated Memory Controller) in the CPU can take more than 1.65v ..

Could be that when you moved to the new place and reinstalled your computer you got a BIOS reset or you played with BIOS settings that changed RAM settings incorrectly. also by removing and reinstalling your RAM sticks for your testing, BIOS could reset the RAM settings to defaults which could be incorrect for your sticks, so you need to make sure that you reapply the correct settings.

Two things to try now, raise your voltage to 1.7v (if your CPU allows for this) make sure that you search for the maximum DRAM voltage for your type of CPU so that you do not kill the IMC with voltage that is higher than it can take

or you could leave the memory running at 1.5v but change frequency and timings by making it run at 1333Mhz or 1066Mhz also you can try timings of 11-11-11-28 ..

By doing this the sticks would require less voltage to operate because you make them run slower or require less performance (of course you would do this only for testing but eventually you would want to run at 1600Mhz for max system stability)

I think in your case things are now pretty clear, you have memory settings misconfiguration which will always cause memory_management BSOD because the RAM is not stable due to not enough voltage supplied to it or not correct settings as in frequency or timing ..

If the sticker on your RAM says 1600 9-9-927 1.7v then this is the settings you need to configure exactly in BIOS for your RAM to be stable ... anything other than that, it would be trial and error and stability is not guaranteed ..

I am sorry I am talking in general terms because I am not familiar with AMD platform at all so I can not give you very specific steps to perform but you now know what you need to know, you can google the small missing details :)
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
I went ahead and set the timings to 9-9-9-27.

I also set the ram voltage at 1.7

However in CPU-Z, I noticed that the speed is listed at 666.7 mhz, which seems odd, and the voltage is listed as 1.5 even though I placed it at 1.7 in BIOS.

Tradesman, the part can be found here http://www.newegg.com/Product/Product.aspx?Item=N82E16820104129 I was unable to find it on Kingstons website.

Could you explain what 2N means? I can try the other settings you listed.
 

MyNewRig

Honorable
May 7, 2013
278
0
10,860
the "N" in the 2N also referred to as "CR" or Command Rate but it is not as critical to stability as the frequency, primary timings (9-9-9-27 in your case) and voltages so if you can not find an N, CR or Command Rate in your BIOS you can just ignore this setting for the sake of stability testing you are doing

The 666 you see in CPU-z is just right, the effective frequency is that number times 2 so (666 * 2) is your effective frequency of 1333Mhz

I think now after you got your settings right for the RAM and CPU you will no longer get BSOD

PS: since you are now running your RAM at 1333 instead of its 1600 rated speed it should no longer require 1.7v to operare, you can lower the voltage to 1.5v or 1.6 to put less stress on the IMC and RAM, so after you confirm that it is stable at 1333 1.7v you can test lower voltages down to 1.5v
 

Tradesman1

Legenda in Aeternum
If need be can up to 1.75, with those sticks, the 2N is the Command Rate or Command Timing, when when the MC goes to access memory, it first has to grab a memory bank, oft called chip select, the CR of 2T or 2N provides an extra clock cycle to grab or latch onto a bank, sometimes 1 is is to fast and it has to start over..sort of
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
Just got another memory management crash. Was playing a game, the game crashed, I restarted, played for a minute or two, and then got the memory management crash.

It did seem to last significantly longer though?

Certainly seemed a bit more stable.

Any suggestions on where to go from here?
 

MyNewRig

Honorable
May 7, 2013
278
0
10,860
apart from the voltage increase that Tradesman1 is suggesting, which you should do, have you tried reinstalling a clean installation of windows just to eliminate software issues? keep in mind that bad drivers can also cause memory_management BSOD ..
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
We did previously try a full windows re-install, yes.

I haven't been able to make the changes suggested yet but will report back when I have some information

The good news is the reason I haven't been able to is my girlfriend has been playing Diablo 3 for like 5 hours with no crashes so hey at least there's that
 

chinchymcchilla

Distinguished
Oct 22, 2011
65
0
18,630
Hey everybody

So it's been about 3 weeks and for almost two there were no problems. This week we have had two different types of crashes.

One was some graphical glitching while playing a game (artifacts on screen) followed by a system exception, and the other was the same memory management error I've been experiencing

There's no doubt that upping the voltage and changing the timings helped but the problem isn't totally gone and I'm wondering where to go from here