Question New PC restarts the moment I launch a game

[Stg-Flame]

I'm brand new to Windows 11 and I can't find any error codes or crash logs anywhere and the Q-Code on the motherboard only says CPU post-memory initialization in the booklet. The lights on my GPU and RAM have turned off completely (they were staying lit prior to the first crash) but they all light up when I first turn on the PC then randomly turn off shortly after it boots. I ran a few different games yesterday and everything worked perfectly fine but now it just crashes.

Where do I start for troubleshooting this? The booklet doesn't give much information regarding the Q-Code and Googling the description for the Q-Code leads to the basic "remove hardware until it works and there's your problem". I was hoping I could narrow it down before tearing it apart again.

Edit: Just reseated the RAM into slots 2/4 and the second stick's LEDs went out almost immediately, but so did the lights on my GPU and the fan on the GPU stops spinning as well. The lights on my AIO worked for about thirty minutes and then they turned off and they only flash when I start the PC but they won't turn on either. The Q-Code is still the same (33/34) so that's not telling me much. Is it possible both the GPU and one stick of RAM is bad?

Edit 2: Just quickly reseated the RAM, tried each stick individually and everything seemed to work. Put them both back in and everything is working again. I was able to launch games with no problems but I can't figure out why it was crashing. I have my monitor and PC on the same power strip but I would think that would trip a breaker if it was drawing too much power and I know it wasn't overheating since I was monitoring the temps on the last crash and nothing was even close to the high temp/yellow zone.

Edit 3: Put the RAM back in the initial slots to see if I could narrow it down, but everything is working fine again.

 

[Stg-Flame]

ASUS ROG Strix Z790-E is the motherboard. When I have a bit more time and I'm not running on no sleep, I'll tear it all apart and start over. I can guarantee I'll miss something if I start that project today.

 

[BoomerD]

ASUS ROG Strix Z790-E is the motherboard. When I have a bit more time and I'm not running on no sleep, I'll tear it all apart and start over. I can guarantee I'll miss something if I start that project today.

Crap...that's one of the boards In've been considering fom my refresh.

 

[igor_kavinski]

With my 1000w PSU in my old PC, it crashed three times while playing Terraria but didn't crash once with Risk of Rain 2.

Did you use the 3090 Ti in the old PC or the 980 Ti?

 

[BoomerD]

Just the advertised maximum power draw for that CPU and video card add up to about 700 watts. Add in drives, fans, etc, you're going to be over 800 watts...then throw in the occasional transient power spike...<click>

This stuff is crazy to me. The last PC I built from scratch was an i5-760 with a GT460 OVERCLOCKED! video card...it never got over 400 watts, and most of the time, between 200-300 watts.

I still say, start from scratch out of the case and see what causes the crash.

 

[Stg-Flame]

Did you use the 3090 Ti in the old PC or the 980 Ti?

Both. First the 3090 then the 980. Both same result.

 

[igor_kavinski]

I'm beginning to think it's some sort of OS version+Nvidia driver combination+your particular games that is not compatible with each other. From your testing, all I can conclude is that it's not a hardware issue since you have tested different combination of hardware in two different PCs and still getting similar results.

If you are using Windows 10, try Windows 11 and vice versa. Download the latest Nvidia WHQL driver+Sound driver on USB flash drive and don't connect to the internet (don't even install network drivers if they are not installed automatically) when you fresh install the OS. Simply install the graphics and sound drivers and then try playing without any other installed drivers or Windows updates. That should help to identify if something bad for your tested games is getting installed by Windows Update.

The above does require you to have DRM-free version of your games, though. If your games are on Steam, you may have to block Windows Update by disabling the Windows Update service.

 

[Stg-Flame]

I'm beginning to think it's some sort of OS version+Nvidia driver combination+your particular games that is not compatible with each other. From your testing, all I can conclude is that it's not a hardware issue since you have tested different combination of hardware in two different PCs and still getting similar results.

If you are using Windows 10, try Windows 11 and vice versa. Download the latest Nvidia WHQL driver+Sound driver on USB flash drive and don't connect to the internet (don't even install network drivers if they are not installed automatically) when you fresh install the OS. Simply install the graphics and sound drivers and then try playing without any other installed drivers or Windows updates. That should help to identify if something bad for your tested games is getting installed by Windows Update.

The above does require you to have DRM-free version of your games, though. If your games are on Steam, you may have to block Windows Update by disabling the Windows Update service.

It's Windows 11 64-bit and I do recall that my PC restarted during one of the mandatory Windows updates that it starts the moment you connect to the internet. I also had to force the PC to shut down during the initial OS installation but when I turned everything back on it started installing without issue. If I'm going to dismantle it anyways and start over, it couldn't hurt to format the drive and completely start over from scratch.

 

[igor_kavinski]

It's Windows 11 64-bit

Do a Windows 10 (same license works for both Win10/11) install this time and disable the E-cores in BIOS. It is very possible that the E-cores could be confusing the games, especially if the games are multithreaded and they are expecting some calculations to be calculated in deterministic time. Threads of the games ending up on the E-cores could be returning results too slow for the P-core thread and that discrepancy in performance may be causing the crash. Remember, the game developers didn't develop their games with E-cores in mind and I doubt that they have updated their games so far.

If everything works without a hitch, enable the E-cores and test the games again.

 

[Stg-Flame]

Tried in Win10 without E-cores and it crashed quicker than before.

 

[Stg-Flame]

850w PSU, 980ti, i7-3770K installed in my new system and no crashes yet. After I was able to play Risk of Rain 2 for about two hours (mostly just restarted the first level trying to get the boss that always caused crashes and encountered that boss seven times with no crashes), I replaced the CPU with my new one and again played Risk of Rain 2 for another 2 hours repeating the same process above. No crashes. That doesn't exactly tell me much since I was able to beat Dying Light 2 in 70+ hours with only a handful of crashes. I'm going to test it more later with the new CPU then pop the 1000w back in and then the 3090 last. I did go over the CPU with a magnifier and none of the pins looked bent or damaged as far as I could tell and there wasn't any thermal paste overflowing off the CPU, though I did clean it off and reapply new paste.

At this point, I've contacted every manufacturer that's involved in my new PC and requested an RMA. If all else fails, I'm going to just send everything back piece by piece and get it all replaced and simply start over. It's honestly the only thing I can think of at this point.

 

[igor_kavinski]

That doesn't exactly tell me much since I was able to beat Dying Light 2 in 70+ hours with only a handful of crashes.

Ideally, there should be zero crashes, especially on such an expensive PC. I've gamed on an old HP office PC with Core i3-6100 CPU and a mining version of Geforce 1060 6GB and it was rock solid during the entire session. Seems manufacturers can't find decent staff for their QA/QC departments these days

And I stress tested my Zotac 3090 for a few hours with Unigine Superposition 4K Extreme textures setting, Furmark and UE5 Matrix demo with the Core i7-5775C CPU. No crashes, even though I was fully expecting it to crash due to how hot the card was running and the PSU is just a generic branded 650W.

I'm going to test it more later with the new CPU then pop the 1000w back in and then the 3090 last.

Good plan!

 

[igor_kavinski]

I did go over the CPU with a magnifier and none of the pins looked bent or damaged as far as I could tell and there wasn't any thermal paste overflowing off the CPU, though I did clean it off and reapply new paste.

You should game with an overlay showing the CPU/GPU temperatures and keep an eye on those. Disable automatic restart in Windows system settings so you can see what exact error occurred. Are you using the contact frame (Thermaltake or some other) for your 13900K CPU? It can improve temperatures by 5 degrees Celsius and may help to dissipate heat more reliably.

 

[Stg-Flame]

Ideally, there should be zero crashes, especially on such an expensive PC. I've gamed on an old HP office PC with Core i3-6100 CPU and a mining version of Geforce 1060 6GB and it was rock solid during the entire session.

My old 980ti system only crashed when my original Corsair PSU died. Once in 10+ years as far as I can remember and none of my other builds have had any issues like this one.

You should game with an overlay showing the CPU/GPU temperatures and keep an eye on those. Disable automatic restart in Windows system settings so you can see what exact error occurred. Are you using the contact frame (Thermaltake or some other) for your 13900K CPU? It can improve temperatures by 5 degrees Celsius and may help to dissipate heat more reliably.

I use an overlay but I wasn't aware I could disable the system restart. I've used GPU-Z to log all the data going to my GPU leading up to a crash and sent that to Nvidia who then told me there wasn't any reason (related to the GPU) as to why my PC restarted. As for the paste, I'm using whatever came with my AIO but I've used ASUS Armoury Crate to monitor the temps on a second monitor as well as used the Aura function to use my LEDs as warning indicators for temperatures and during multiple crashes none of the temps even came close to warm (all LEDs were green) . I suppose it's possible temps could be an issue but when 50% of the crashes occur the moment I launch a game, it seems like it's something else.

 

[igor_kavinski]

You could try some other games, just to confirm that it's the games themselves that are not the issue.

 

[Stg-Flame]

You could try some other games, just to confirm that it's the games themselves that are not the issue.

So far I've only played a handful of games on my new hardware and I've tried all of those except for the one my old hardware can't run. Using my old CPU, PSU, and GPU, I'm not getting any crashes on any of them. I should also note that when I was getting crashes, I tried lowering the in-game settings and disabling all RTX settings when applicable. Prior to building the PC a friend of mine on Steam told me about some games like Fallout 4 simply crashing when you use certain in-game settings like weapon particles. Since Fallout 4 was the first game I experienced a crash, I made sure I turned it off before I even started playing, but while I was messing with the mods and trying to get them to work, I somehow re-enabled the weapon particles which is why I didn't really think too much when that game crashed twice. It wasn't until other games started doing it too that I started suspecting something was wrong with the hardware itself.

 

[Stg-Flame]

FINALLY SOME GOOD NEWS!

After 12+ hours of gaming/testing various hardware configurations and logging each session using GPU-Z, EVGA concluded my new 1000w PSU is faulty. They said the logs showed my hardware wasn't causing power spikes or fluctuations, but rather the PSU was randomly losing power which is why Nvidia said nothing was out of the ordinary when they reviewed the initial logs. They did offer an RMA but made it a point to tell me multiple times in their e-mail that if the PSU has any damage to it then they won't honor the warranty and I'll have to eat the full cost as well as the fact they will perform their own testing and diagnostic before either sending it back as it was or replacing the unit. They also mentioned the replacement will not be brand new and could contain cosmetic damage or signs of wear/usage. I think just for that point alone they will be getting a mediocre review just to drive home how EVGA has given up on customer support entirely. I get they have to verify that I didn't drop it or crack it open which caused the damage, but they mentioned it three different times in a single paragraph and put it in bold at the end. Plus, I bought it brand new and they have the audacity to tell me my replacement could have cosmetic damage or signs of wear is ridiculous. I'll fight them on this if they send me something that looks damaged and likely ship it back for a refund and go with a different brand. Just kinda puts the cherry on top of this whole situation knowing what EVGA has devolved into but for all I know they could send me a brand new one as a replacement. The last time I had an issue with an EVGA product, they e-mailed me a shipping label and said a brand new replacement would be on its way. No questions asked.

So, I guess I'll be using my old hardware until EVGA gets done with their testing to see what they say about the PSU. If they ship me a new one, I'll put everything back together (at this point 90% of my hardware is hanging outside of the case) and go back to more testing.

 

[Stg-Flame]

Update: Got the new PSU today and have been testing things for the past few hours. The PC will turn on and then immediately restart after pushing the power button but then everything works fine after that so I'm not sure what's going on there but it doesn't seem to crash/restart while playing games. The immediate restart doesn't happen every time, but it's been the majority of the time. I'm probably going to do yet another format/fresh install with my new hardware connected and see how it goes from there but so far everything looks good. I still have most of my hardware hanging outside of the case since some of the cables have to be twisted upside down and backwards just to have the cables hidden behind the motherboard mounting plate and that's a massive pain in the ass to do every single time.

Also, EVGA sent me a slightly different model than the one I sent to them (SuperNova P5 "upgraded" to a SuperNova P6). Only differences I found between the two is the amount of cables they come with, the style of the cables (flat vs sleeved), the P5 is 10mm longer, and the P6 is much quieter.

 

[BoomerD]

Update: Got the new PSU today and have been testing things for the past few hours. The PC will turn on and then immediately restart after pushing the power button but then everything works fine after that so I'm not sure what's going on there but it doesn't seem to crash/restart while playing games. The immediate restart doesn't happen every time, but it's been the majority of the time. I'm probably going to do yet another format/fresh install with my new hardware connected and see how it goes from there but so far everything looks good. I still have most of my hardware hanging outside of the case since some of the cables have to be twisted upside down and backwards just to have the cables hidden behind the motherboard mounting plate and that's a massive pain in the ass to do every single time.

Also, EVGA decided to send me refurbished cables which look like they are from the early 2000s. Not even the motherboard cable is sleeved and there's noticeable wear on every single cable which I'm definitely not happy about. I can only assume the PSU they sent me was also a refurbished one though I can't confirm that and they won't tell me either.

Sending refurbished for an RMA is pretty normal. It sux, but that's the way most companies do it. I like EVGA video cards. I'm sorry to see them exit the field. I bought one of their GTX460 Superclocked cards back in...2010-2011. It died on me about 6 years later, (lifetime warranty) and they replaced it with a GTX660 Superclocked card. Refurbished, no doubt, but no obvious signs of wear or tear. That card carries the same lifetime warranty as the original. (not that it's worth a crap nowadays, but still...)

This problem you've been having has all the signs of a grounding issue, IMO. Hopefully, the new PSU will solve the problems. (and if you don't like the cables, you can always buy a nice set of sleeved cables) https://www.amazon.com/CableMod-ModFlex-Cable-EVGA-CM-PEV2-FKIT-KKR-R/dp/B089KPWFXV?th=1

 

[Stg-Flame]

This problem you've been having has all the signs of a grounding issue, IMO. Hopefully, the new PSU will solve the problems. (and if you don't like the cables, you can always buy a nice set of sleeved cables) https://www.amazon.com/CableMod-ModFlex-Cable-EVGA-CM-PEV2-FKIT-KKR-R/dp/B089KPWFXV?th=1

I have plenty of spare cables but I was just annoyed they shipped the flat ones instead of the sleeved ones. Not that it's a huge deal but I prefer the aesthetics of the sleeve over the flat and one of my concerns was whether or not I had a good PSU but a faulty cable. EVGA specifically told me not to send any cables back with the PSU so I still have the original ones, but given the fact I wasn't sure if I had a bad cable or not, if the issue persists, I suppose I can start swapping to the new cables and see if that was a factor as well.

Though since you mentioned a possible grounding issue, could this be mitigated by simply attaching a normal grounding wire to the PC itself? I work in the oilfield and deal with industrial electronics all the time so I'm no stranger to MacGyvering ground wires to instrument panels. Now that I think about it, I have an entire drawer in my toolbox dedicated to grounding wires.

 

[BoomerD]

I have plenty of spare cables but I was just annoyed they shipped the flat ones instead of the sleeved ones. Not that it's a huge deal but I prefer the aesthetics of the sleeve over the flat and one of my concerns was whether or not I had a good PSU but a faulty cable. EVGA specifically told me not to send any cables back with the PSU so I still have the original ones, but given the fact I wasn't sure if I had a bad cable or not, if the issue persists, I suppose I can start swapping to the new cables and see if that was a factor as well.

Though since you mentioned a possible grounding issue, could this be mitigated by simply attaching a normal grounding wire to the PC itself? I work in the oilfield and deal with industrial electronics all the time so I'm no stranger to MacGyvering ground wires to instrument panels. Now that I think about it, I have an entire drawer in my toolbox dedicated to grounding wires.

First...DO NOT MIX CABLES FROM DIFFERENT POWER SUPPLIES! That's a really good way to have way more problems than you've been having. There is no real "standard" as to PSU pin-outs. What works for Seasonic MIGHT work with EVGA or Corsair...or might cause you to burn down your house. If you have the original cables...and they sent the exact replacement, I'd use those.

Next...the grounding issue I was talking about is the motherboard grounding to the case somewhere. Did you ever take everything OUT of the case and build it on a piece of cardboard or other insulating surface to see if the problem replicated itself? All it takes is one stand-off touching a motherboard trace to cause all sorts of weird issues.

 

[Stg-Flame]

Next...the grounding issue I was talking about is the motherboard grounding to the case somewhere. Did you ever take everything OUT of the case and build it on a piece of cardboard or other insulating surface to see if the problem replicated itself? All it takes is one stand-off touching a motherboard trace to cause all sorts of weird issues.

I did with the old PSU and it still restarted.

According to the e-mail, EVGA determined the PSU I sent in didn't "meet their standards" and failed to state whether or not it could have caused my PC to restart.