Solved! Troubleshooting Memory Issues

[MrPickins]

I built this system almost 2 years ago: AMD 5600x, Asus B550-F Gaming, 32GB (16x2) Gskill Ripjaws DDR4-3600. It's been working well, but recently I've been getting a few "Memory Management" bluescreens in Win10, so I decided to run Memtest86+, which failed immediately.

Edit: The memory (F4-3600C16D-32GVKC) isn't listed as compatible for this mobo on the Asus site, but is listed on the GSkill site, and has been stable until fairly recently.

I then tried reseating the memory and then started testing different combinations of sticks/sockets with no luck. One stick seems to fail immediately, but with the other it fails until I disable XMP and run at 2133 Mhz. Changing sockets seems to make no difference.

I've reset the bios to recommended settings, including the 2133mhz ram speed, and it seems stable with the one stick, but it seems awfully strange that both sticks would fail to work at the XMP speeds all of a sudden. This has me concerned that it's an issue with either the memory controller or Infinity Fabric on the CPU, or possibly the motherboard.

Unfortunately I have neither another DDR4 rig in the house, nor another AM4 CPU to test. At this point I'm trying to decide which components to buy to test the system. I may end up using any unnecessary parts to upgrade my daughter's pc, but I wasn't planning on doing that now.

I've been building my own PCs for 25+ years now, but this one has me a bit stumped, as it's a failure mode I've never seen. So I suppose the question is, what parts would you try first? A 5600G (the G model because the little one could really use the upgraded iGPU) another set of RAM, or a mobo swap (thinking that one is a last resort).

 

[mikeymikec]

Do you get consistent failure results with each module?

 

[In2Photos]

Have you updated the BIOS?

 

[MrPickins]

Do you get consistent failure results with each module?

One stick consistently fails immediately, the other seemed to be fine at 2133MHz,but it also ended up with a few errors after running Memtest86+ overnight (7 passes or so).

Have you updated the BIOS?

I updated to the latest BIOS using the stick that seemed to be working well.

 

[mikeymikec]

I'd class faulty memory as the most likely reason for this situation, though two failing seems unlikely. Maybe the second/other module has been problematic for some time and has got worse?

I say 'unlikely' but that doesn't mean I'd rule it out. The fact that you're getting consistent failure symptoms suggests that everything else is playing straight dice IMO.

If you want to play on the safe side, you could pick up an absurdly low end DDR4 module just for testing (thinking 4GB and the lowest speed you can find), but personally I'd just go straight ahead and buy the kind of memory to ideally replace the faulty modules. I stick with Kingston and Crucial personally, I've had a lot less problems in my line of work since I stuck with those brands.

 

[In2Photos]

mikeymikec posted right as I was typing. I agree. If changing sockets makes no difference I would say the issue is with the memory. I'd opt for a kit that is on Asus' QVL and test again.

 

[MrPickins]

I went ahead and ordered a 5700G, thinking I can use it to upgrade my kids pc if necessary (Should be here today. Thanks Amazon!), but I'll probably also order some replacement DDR4 as well.

Edit: Ordered a 16GB (2x8GB) Corsair kit as well (CMK16GX4M2B3200C16). should also be here this evening. I'd prefer 32GB, but this kit was on the QVL from Asus, so I'll use it as a test for now.

Looks like I know what I'll be doing this weekend.

 

[MrPickins]

The new set of RAM test fine. I suppose I must have had a bad batch initially. First time that's happened to me.