unexpected hard resets - maybe a problem with RAM?

Hi there, I recently put together a mostly new PC and i've had two unexpected system resets while playing Final Fantasy XIV. I'd say both times it was after 45-60min. In both cases there was no perceivable warning. The first time, both screens went black, RGB stayed on, went to boot screen and then back to windows. No errors, no recovery mode, just back to normal. The second time was the same except only the memory RGB went completely off, and stayed off. I ran windows memory diagnostic and no issues were found. I did a proper shut down yesterday and started up today and the ram rgb is now back.

 

The second time the reset occured I was keeping a pretty close eye on my temps because I have an issue with my H100i (thread here). cooler temp was consistently around 35c, core temps were around 40-50c, gpu was consistently around 79c, but I wasn't really looking at the ram temps. I checked immediately after the reset and the ram temps were all around 50c. Today after the system was off all night the ram temps started around 30c and are now idling around 40c.

 

I'm thinking I will reseat all modules today and do a little more research, but if anybody has advice i would appreciate it. I'm still worried that there might be a problem with my PSU due to its age, and it seems like it's been making more noise than usual, so that might be something I try to look into as well. Thanks for taking a look.

quick related question regarding module placement. I just looked through the manual again for my mb and noticed that the modules are left to right "B1 - B2 - A1 - A2" which kind of confuses me. does that mean I should have placed the modules in order of "3 - 4 - 1 - 2" where 1-4 are the module numbers incrementing up? Does it even matter if I have 4 modules of the same make/model/speed? The way I placed them was the way I used to always place them which was "1 - 3 - 2 -4" but maybe that's wrong? Then again I don't see why this would cause a hard reset.

Change ram slots, change voltages, change fequency bla bla blaa... These are all b....t. You are cheated just like me. Return your ram and take your money back if possible.

Change ram slots, change voltages, change fequency bla bla blaa... These are all b....t. You are cheated just like me. Return your ram and take your money back if possible.

I mean my system is working pretty well overall. I'm not even convinced this is a memory problem yet... but yea building pcs is a bit of a nightmare. I've enjoyed building for a long time but I think I'm kind of over it. Probably just going to get pre-built in the future... sorry for your trouble mate good luck.

I re-seated the modules in the order i thought made most sense based on the gigabyte manual. I also removed some cantilevered power adapters I was using for the 24 pin power header and front panel usb header as i thought they might be loose. I also did a general double-check of connections.

 

re-started and played FFXIV for a about 40-50 minutes with no trouble, then right as I was teleporting from one zone to another, the PC reset again. This is pretty interesting because the last reset happened right as I was teleporting as well, and the first reset happened when I was effectively doing the same thing - moving to a different zone. It seems like this could be a pretty significant clue that there's an issue with the memory over time where unloading/loading of data into memory incurs a fatal error or something. I would have thought something would show up in the memory diagnostic if that was the case though. It also makes me wonder if it has to do with the drive that the game is installed on though, which is an older disc drive. I think I'll reinstall the game and test some other games.

 

Also, this time instead of the RGB going out on all modules, it went out on the right 3 modules, which may be relevant.

I reinstalled the game and played for a long time yesterday and didn't have a single problem. Also, i enabled a frame rate cap to 60fps which seemed to help gpu load and temp considerably.

 

BUT unfortunately I just played for about an hour today and had a reset just like the others, except this time I was watching an IGC... not quite matching the criteria from the other occasions.

 

This time the RGB is out on the 1st, 3rd and 4th modules. So that's interesting.

 

I've looked around online for problems with unexpected reboots but wasn't able to find much. I looked at more power stuff and have some things i'm going to try - moving the power off the power strip and/or to a different outlet. I still think the power supply might be struggling, but it's still a toss-up between that or a problem with the RAM.

Maybe get HWinfo or some other software monitor your voltages while you play. Look if the 12V dips too low?

2070 at almost 80C, ram at 50.. your case seems to be pretty hot but that's still far from problematic temperatures when it comes to crashes.

I'd first check voltages during gaming

Thanks for the suggestion, I haven't used HWinfo before. Looking at their sensor mode now I can see the +12V row under mainboard, just idling right now it's saying min is 12.168V. I'll try playing for a while and see what it reads.

 

I also was feeling like the system was a bit hot. It definitely doesn't help that one of the PWM fans isn't functioning, but I have 2x 120mm exhaust fans on the ceiling that I would have expected to be keeping the ram cooler. For the GPU, I've seen some people saying that high 70s up to 80c is pretty normal for 2070, but maybe that's heating up other components.

 

I went ahead and ordered a new 750w PSU (EVGA SuperNOVA G3), and a 1000VA UPS for good measure since I get occasional power loss in this house. I don't know if it will solve this problem anymore but figured it couldn't hurt.

 

I've looked around a bit and I've been starting to lean towards the GPU being a possible culprit. If you look around online for restarts affecting specific games, people often boil it down to random NVIDIA problems, which is familiar to me. I've heard from a friend who had similar issues in the past and was able to minimize crashes/restarts by downclocking the gpu by ~100mhz. So I have EVGA Precision X1 now and might try experimenting with that if the problem occurs again.

 

 

edit 1 - just started up the game and the +12V dropped down to 11.952 while just standing around in a dense area

edit 2 - dropped down to 11.88 and it will stay at these levels for a while. I'm not sure what an acceptable range is so i'll just try and see how low it will go.

edit 3 - never saw +12V below the 11.88 but encountered another system reset while testing. It happened right as I was transitioning from one zone to an adjacent zone. That's 4 out of 5 times now that have seemingly occurred during an event that could potentially be increasing load across multiple components in the system.

Edited July 11, 2020 by ohnomelon
more info