Newly built PC - Shutting down after some hours of heavy load

[ImMrFilth]

Hello everyone!

 

I'm going to try to be as thorough as possible when describing my problem.

 

First of all, here's my specs.

 

Quote

OS: Win 10 with the latest drivers installed.
CPU: Intel Core i7 8700K
Cooler: NZXT Kraken x62
Motherboard: Asus Prime z370-a
RAM: 2x 8GB Trident Z RGB 3200hz CL16
GPU: Asus Rog Strix Gaming 1080 ti OC
SSD: Samsung 960 EVO 250 gb
PSU: Corsair RM 750x
Other cooling: 3x120mm exhaust. 1 in the back and 2 in the roof. The radiator is mounted at the front and acts as intake.
Case: Phanteks Eclipse p400s Tempered glass

 

The parts for the build arrived on tueday evening and I finished building the PC wednesday morning. Which means it's been fully built for, soon to be, 2 days. 
When I finished putting the PC together I proceeded to install windows and then on to the drivers. I used windows update to start off and then went on to download drivers from the manufacturers website for the motherboard and GPU. Sometime around here, when I was finished installing the drivers and started downloading essential programs for everyday use like Google Chrome etc. The computer shut down without giving me a warning. As if it had been overheated. I turned it back on 2 seconds later and didn't think much about it. 

 

However, this problem continued to happen throughout the day. It happened again when I was playing Fortnite with a couple of my friends. This time I wasn't able to turn it back on again until a few seconds later. But when I did, it turned back off. Just like it had been overheated. I got it back on a minute later and got back on to discord to talk with a friend who's been helping me with building. He gave me some instructions of what I could do and so forth. 

 

I ran a stresstest and the PC managed 15 minutes the first time around. Then it shut down. I plugged the power chord straight into the wall instead of an extension and tried to run the stresstest again to see if it had to do with the extension not providing enough power when the PC was under load. The PC lasted 23 minutes and then shut down.

 

My GPU temp under load was: 72C / 161F

My CPU temp under load was: 70-83C / 158-181F (spiking up and down)

 

So with this information I quickly told myself that it wasn't due to overheating that the PC was shutting down. The CPU should be able to handle 100C according to the intel website and the GPU somewhere in the low 90's.

I then ran a memtest86 to see if it had to do with my RAM, but the test passed 100%. I ran this test as I went to bed on wednesday. I woke up with the test finished and the PC still on. Which also gave me the information that the PC doesn't turn off while idling and only under load. 

 

Today I've been trying to troubleshoot but I'm clueless. I don't know what to do. I feel like it has to be a faulty PSU or motherboard causing the issue but I've never built a PC before so I'm in no position to even have an opinion. Maybe you guys can help me out? Any suggestions about how to proceed? 

 

 

 

Edit #1:

1. I've ran AIDA64 on full load stressing everything but the harddrive. PC shut down after 15 minutes the first time I ran the test and 24 minutes the second time. 

 

2. I've ran memtest86 to check for faulty memory. Returned with a 100% pass. 

 

Edit #2:

3. Tried resetting CMOS. Ran a stresstest after and the PC managed 2 hours before I turned it off manually. I then played 2 hours of Fortnite and watched a stream for an hour. During the stream-watching the PC shut down. Just like the other times. Off, then on, then off. Within a few seconds of each other after the first shut down. I wasn't doing anything else but watching the stream. Game wasn't even on in the background. I was basically idle'ing. 

[For Science!]

4 minutes ago, ImMrFilth said:

-

Sounds like a somewhat unstable overclock, perhaps you enabled XMP accidentally and its not behaving well? Try to reset the CMOS as a starting point, pull out the lithium ion battery in the motherboard, wait a few seconds (do this while the system is not connected to power) and the put it back in.

 

This resets the BIOS settings to factory. Try the stress tests after doing so. IIRC there may be a XMP jumper/button on the ASUS boards, you would want to make sure its disabled too.

[Doomed83]

If I had to guess given the information above, I would say try a different power supply.

 

You didn't mention anything about overclocking so I'm assuming you haven't?

[ImMrFilth]

@For Science! Yea, this is something I've heard from other threads of people having similar problems. Do you reckon this is a good starting point though, or are there other options I can try first? I'm a bit scared to touch anything on the motherboard as I'm new to all of this. 

 

@Doomed83 This is the solution my friend told me to try. I have an old power supply from the computer I'm upgrading from. But it's pretty much ancient. It's the "FSP-GROUP" FSP750-80APG. It's 750 watt, but I'm not sure if it'll work with my 1080ti? And no, no intentional overclocking has been done!

[For Science!]

2 minutes ago, ImMrFilth said:

-snip-

I do think it is a good starting point. It is very safe to do as long as the system is turned off and you are not supercharged with static electricity. There is a jumper pin on the board, but that would involve sticking metallic objects towards the board which I think is more risky than removing the battery.

 

A photo of your build may also tell us of any glaring errors (Although I suspect none since it actually does turn on).

[Doomed83]

If the old power supply has the right connections for your graphics card and motherboard it's worth a shot to test it. 750 watt should be enough for a 1080ti

 

Edit: I looked up that power supply and it does have multiple 12 volt rails which means that you would need to make sure that the 2 power connections going to the gpu are from different rails.

[ImMrFilth]

@For Science! Alright! I'll try doing that as my first move then!

 

I also forgot to mention that I ran a stresstest earlier today after changing some settings around. The PC managed to run for an hour and 10 minutes. I actually turned the test off manually after that because I thought that the problem was fixed. The PC then turned off a few minutes later instead while I was waiting for the GPU and CPU to cool off. I increased the fan speeds on my radiator and the exhaust while doing this. Not sure if that has anything to do with why it crashed, but thought I'd include it.

 

@Doomed83 I'll try running the build of my old PSU then! I'll try For science's solution first though!

[ImMrFilth]

13 minutes ago, Doomed83 said:

-

I saw your edit now. Different rails, what does that mean more specifically?

[Doomed83]

It means that they split the 12volt output to separate "circuits" in a sense and each has a maximum current that can be safely pulled through it without causing damage to the unit. For that unit I believe that each rail is 18 amp which means that a 1080ti at full load can pull more than the rail supports. Generally speaking the higher end power supplies use a single 12 volt rail to prevent this kind of issue.

[ImMrFilth]

@Doomed83 so since the 1080ti at full load can pull more than the rail supports. Does that mean I can't use this PSU? Sorry if I'm asking stupid questions, I'm just very new to all this. 

[Doomed83]

it means that you need to make sure that each of the 8-pin pcie connections to the card come from different rails. The label on the side of the PSU explains the cable color scheme for which cable come from which rails.

[ImMrFilth]

@Doomed83 Oh! Okay! I'll make sure I do that then! 

 

@For Science! I reset the cmos and have now started a stresstest. How long should I let it run before I know if it was a success or not? 

[For Science!]

3 minutes ago, ImMrFilth said:

@Doomed83 Oh! Okay! I'll make sure I do that then! 

 

@For Science! I reset the cmos and have now started a stresstest. How long should I let it run before I know if it was a success or not? 

At stock, you really should not be crashing at all. I would recommend to use a program like Realbench that puts a load on GPU, CPU, and memory all at the same time (but not as insanely as some other programs) it emulates a multi-load very well. 

 

I would say if you're realbench stress test is stable for at least 2 hours then you're probably fine. The load is repetitive, so if you survive one cycle, you're likely to survive the rest. But yes, longer the better.

 

In case you can't find the link to Realbench:

https://rog.asus.com/rog-pro/realbench-v2-leaderboard/

 

First link on top of the page "Download Realbench"

[ImMrFilth]

6 minutes ago, For Science! said:

-

I was using Aida64 before, but I'll download RealBench instead! I'll get back to you once I've let it run for at least 2 hours then. Unless I crashed before that mark.

[ImMrFilth]

@For Science! Okay, so.. It seems fine? I chose to run the test for 8 hours and stopped it at the 120 minute mark. No problems at all. I'm surprised. Am I alright? Here's a printscreen at the 2 hour mark. 

[For Science!]

2 hours ago, ImMrFilth said:

@For Science! Okay, so.. It seems fine? I chose to run the test for 8 hours and stopped it at the 120 minute mark. No problems at all. I'm surprised. Am I alright? Here's a printscreen at the 2 hour mark. 

It would seem so indeed. For now I suppose its safe to say that your problems are resolved. Did you buy your motherboard 2nd hand or something? (perhaps some non-standard setting was dialled in, which is why clear CMOS did anything). Once its been okay for a few days, then you can consider overclocking (if you want to).

[ImMrFilth]

29 minutes ago, For Science! said:

It would seem so indeed. For now I suppose its safe to say that your problems are resolved. Did you buy your motherboard 2nd hand or something? (perhaps some non-standard setting was dialled in, which is why clear CMOS did anything). Once its been okay for a few days, then you can consider overclocking (if you want to).

Okay, problem just came back. Rip. After the 2 hour stresstest. 2 hours of playing Fortnite and ~an hour of watching a stream, it gave up. Shut down during the stream. Wasn't doing anything special. Didn't even have the game running in the background. 

 

Computer shut down, then turned back on within 5 seconds automatically and then shut down again. I just switched the power off, on the PSU for now. Honestly, it feels so much like the PC is overheating. Like the way the shut downs expresses themselves make it seem that way. But I don't know. Why would it happen so randomly? Especially when my temps are completely fine and as I was basically idle'ing. 

 

Edit: I did not buy the motherboard second hand. It's completely new, bought from a retailer. 

[ImMrFilth]

*bump*

[DerBurner]

Since your temps are fine, and the system runs without a problem between the crashes, the only reason I could think of now is a faulty PSU. Did you checked for the GPU power cables as Doomed83 suggested, and did you tried swapping in your old PSU? The only ways I experienced PC´s just turning off without warnings were due to temps or something that tripped the fuse of the PSU, short circuits for example.

[Doomed83]

Just because I haven't asked yet.... Is your power supply mounted to intake from the bottom of the case? If so, have you made sure there isn't anything in the way of that intake?

[ImMrFilth]

@DerBurner Yea, honestly only thing I can think of would be that it's a faulty PSU as well. Maybe a slight chance that it could be the motherboard? Unfortunately I can't use my old PSU to test because it doesn't have enough 6+2 pins / 8 pins to power the 1080ti. 

 

@Doomed83 I'm not 100%. The PSU fan is aimed towards the bottom of the case where there's a hole and a dust-filter if that's what you mean! 

[Doomed83]

I believe the same thing, that either you have a faulty power supply or a faulty motherboard and I would lean toward the power supply first. I would probably ask if any friends have a good power supply to help you rule that out.

 

Basically the root of the psu fan question was. Is your pc sitting on capret where it can't breathe?

[ImMrFilth]

@Doomed83 Yeah! My friends going to lend me his so that I can check if it's a faulty PSU! Hopefully we'll get it done this weekend. 

 

Oh! No, it's standing on my desk, so that's all good! 

[Doomed83]

Good to hear you have someone wiling to help you out. Good luck!

[ImMrFilth]

@Doomed83 Yea dude, I have a few friends, all with pretty decent PC's so I'll get it done one way or the other! Thanks!

 

One thing I just noticed is that the fan on the psu doesn't even spin. I'm currently running a stresstest and it's been going for 35 minutes. I read about the rm750x in a different forum and they said the fan only activates at 40% watt usage. But surely, with my system, during a stresstest, it'll be above 40% load, right?