Question PC shutting off w/ weird symptoms

[tachophobicat]

For the past few weeks, my PC has been shutting off with strange symptoms. Every so often - seemingly at random - it will shut off, all the fans + lights will stay on, and the CPU troubleshooting light on the motherboard turns on. From there, I can only shut it off and restart it by turning it off at the power supply. There is seemingly no structure to when these shut-offs happen, though I can replicate it by simply trying to pass through the tutorial of the game Killer Frequency (there's a spot where it always happens, even though it's a pretty easy-to-run game and I play much more resource intensive games regualrly). CPU temps are hovering around 60.

I recently upgraded my CPU from a Ryzen 5 3600 to a Ryzen 9 5900X about a month ago. I can't remember if it happened before I upgraded, but it's been increasing in frequency over the last few weeks. Everything else in the system was purchased in February 2020 (except for upgrading from 16 -> 32gb RAM this past winter).

Logically, it's either a failing PSU or motherboard, but that doesn't feel right. Both were purchased about 4.5 years ago, so they shouldn't be failing by now. The Newegg power draw calculator says I need about 600-699 watts for this system, and I have 650, so I'm kind of cutting it close with my recent upgrade. Every USB port is populated with a peripheral and there's a WiFi card, so the system COULD be drawing too much power.

Please let me know if anyone needs any more information, I'd love any help or suggestions I can get. I haven't tried reseating the CPU and cooler yet because my temps - while a little higher than they were before the upgrade - are fine, and this is the only issue that I'm having. Would love to hear if anyone's seen this before and can offer a perspective!

Specs:
Ryzen 9 5900X
2070 Super
B450 Tomahawk Max
32gb DDR4-3600
Seasonic 650W Gold
1 4TB HDD, 1 1TB HDD, 1 m2 1TB SSD, 1 512GB SSD

 

[punkncat]

Have you checked in Event Viewer or Reliability History to see what codes it is throwing related to the shutdown? Even 41 should be expected since you are killing with the power switch, but what else?

Did you download the new chipset driver manually? I would assume you pretty much had to do a BIOS update for it to work in the first place.
The B350 and the B450 are both considered decent middle of the road motherboards and seem to offer pretty good value and performance for the price. I wouldn't just automatically start there given its track record for both chipsets. I do understand one of the later revisions has an issue with the PCB that may or may not be corrected by now.

Are you watching actual draw from the wall? Or using some monitoring programs and adding a touch for good measure as far as wattage? Having insufficient power can cause problems.
Are you overclocking at all?
You mention temps hovering at 60*, but is that the maximum temp you see or is it idling there? What cooler are you using?

 

[tachophobicat]

Have you checked in Event Viewer or Reliability History to see what codes it is throwing related to the shutdown? Even 41 should be expected since you are killing with the power switch, but what else?

Did you download the new chipset driver manually? I would assume you pretty much had to do a BIOS update for it to work in the first place.
The B350 and the B450 are both considered decent middle of the road motherboards and seem to offer pretty good value and performance for the price. I wouldn't just automatically start there given its track record for both chipsets. I do understand one of the later revisions has an issue with the PCB that may or may not be corrected by now.

Are you watching actual draw from the wall? Or using some monitoring programs and adding a touch for good measure as far as wattage? Having insufficient power can cause problems.
Are you overclocking at all?
You mention temps hovering at 60*, but is that the maximum temp you see or is it idling there? What cooler are you using?

Thanks so much for the well-thought-out response, it means a lot.

There don't seem to be any codes related to the shut down. In the last hour (where I've replicated the issue 3x in testing), it's thrown only event 41 as a Critical Error. I can give all of the Error event IDs, but from what I can tell, none seem to be related to anything that could cause a physical shutdown.

I downloaded the new chipset driver manually before swapping CPUs, yes.
I am not overclocking at all.
It's idling around 60-70, using a Noctua NH-U12s. Never had an issue with my temps before so I doubt this is the problem.
No, I am not watching actual draw from the wall. What's a good monitoring program for specifically measuring current system wattage?

Important update: I'm able to replicate the problem every time at that spot in Killer Frequency. However, with every peripheral unplugged except for keyboard/mouse, my computer DID NOT CRASH in that spot. I have two webcams, an XLR interface w/ preamp + microphone, and my headphones, which were all unplugged. This tells me that, with all my peripherals, my system is drawing just barely too much power. Does that line up?

 

[punkncat]

If you are idling at 60-70:

First try and drop the side panel to see if temps drop. Where you might check the installation for even tightness and good paste coverage, I really feel like you need the one step larger two fan air cooler or an AIO for that R9. I would be willing to bet the issue lies in temps.

Do you have MSI Afterburner going? You can use it to monitor CPU and GPU voltages. I typically add about 100W or so on top of that for a nice round figure. I really think with this build you do need a more powerful supply, probably 750+ but I am not certain this is the sole cause of the issue when discussing those idle temps.

What case are you using and what is the fan configuration? What is ambient in your locale?

 

[tachophobicat]

If you are idling at 60-70:

First try and drop the side panel to see if temps drop. Where you might check the installation for even tightness and good paste coverage, I really feel like you need the one step larger two fan air cooler or an AIO for that R9. I would be willing to bet the issue lies in temps.

Do you have MSI Afterburner going? You can use it to monitor CPU and GPU voltages. I typically add about 100W or so on top of that for a nice round figure. I really think with this build you do need a more powerful supply, probably 750+ but I am not certain this is the sole cause of the issue when discussing those idle temps.

What case are you using and what is the fan configuration? What is ambient in your locale?

I'm using the Phanteks P400 with 2 in-flow fans, one exhaust fan behind the CPU cooler, and one exhaust fan on top of the CPU cooler. The ambient temperature in the room is about 70 degrees F. Taking the side pane off took temps from 57 -> 54 at idle.

When the game that it's crashing on is running, temps don't rise at all, and there's no spike in the moments before it crashes.

I'm about to check to make sure, but the only correlation I can find with the crashing is the peripherals that are plugged in. Do you know if there's any solid way to get a log of my CPU temperatures that would be saved even through a crash?

Note - without the peripherals plugged in, I'm able to pass through the spot that I normally crash at in that game. Monitoring my temperatures through this, they're still hovering around 55.

 

[punkncat]

I am not aware of anything right off.

When you restart the PC, get it up and running doing nothing with it it all, can you check Task Manager to see what % usage is going with the CPU and/or even a core speed?
In theory, if you wait a few moments to let the system get up, run its scan and such, within a few minutes it should just be sitting there doing nothing. I can understand an R9 getting hot during usage but when idling and everything is "parked" I feel like your temperatures are too high even with that cooler.

Is the P400 the "A" version, or the solid front panel?

 

[tachophobicat]

I am not aware of anything right off.

When you restart the PC, get it up and running doing nothing with it it all, can you check Task Manager to see what % usage is going with the CPU and/or even a core speed?
In theory, if you wait a few moments to let the system get up, run its scan and such, within a few minutes it should just be sitting there doing nothing. I can understand an R9 getting hot during usage but when idling and everything is "parked" I feel like your temperatures are too high even with that cooler.

Is the P400 the "A" version, or the solid front panel?

I have the basic P400 with the solid front panel.

Thanks for that suggestion. I got to Task Manager about 45 seconds after booting and it was at 60% CPU usage, slowly going down to about 3% at idle. Is that indicative of a heat lock?
Core speed is about 4.3 GHz at idle after booting.

Another potentially random update...in trying to do an assignment for a college class, I've found that the gradient tool in Photoshop also makes it crash. Unplugging everything does not have an effect on whether it crashes or not. Not sure if that means anything. I've thought before that "when the CPU undergoes high usage, it crashes," but that's not the case. It doesn't happen playing resource-intensive games like Rust, it doesn't happen while rendering videos, and it hasn't happened when I've run CPU/GPU stress tests. But it happens with the gradient tool in Photoshop and the tutorial level of Killer Frequency...

 

[punkncat]

I have the basic P400 with the solid front panel.

Thanks for that suggestion. I got to Task Manager about 45 seconds after booting and it was at 60% CPU usage, slowly going down to about 3% at idle. Is that indicative of a heat lock?
Core speed is about 4.3 GHz at idle after booting.

Another potentially random update...in trying to do an assignment for a college class, I've found that the gradient tool in Photoshop also makes it crash. Unplugging everything does not have an effect on whether it crashes or not. Not sure if that means anything. I've thought before that "when the CPU undergoes high usage, it crashes," but that's not the case. It doesn't happen playing resource-intensive games like Rust, it doesn't happen while rendering videos, and it hasn't happened when I've run CPU/GPU stress tests. But it happens with the gradient tool in Photoshop and the tutorial level of Killer Frequency...

What power plan are you using? (Edit Power Plan) Do you have it set to performance?

Also, you can order the mesh front for those P400 and it makes a world of difference. It is a great looking case and was fun to build in, for sure.

 

[tachophobicat]

What power plan are you using? (Edit Power Plan) Do you have it set to performance?

Also, you can order the mesh front for those P400 and it makes a world of difference. It is a great looking case and was fun to build in, for sure.

I have it set to High Performance.

I would like to note that, in troubleshooting my Oculus Rift S not working the other week, I did a few seemingly random things. I turned off "Game Mode" (which was on by default for some reason?) and turned off XMP (just forgot to turn it back on). I don't think those would have any impact, but...

Thanks for the help. I can also just take the front of my case off and experiment to see if that'll make a difference in it crashing. I should also probably reseat my CPU cooler and recheck my temps.

With all this information, do you think it'd be a good idea to order a new PSU?

 

[punkncat]

I have it set to High Performance.

I would like to note that, in troubleshooting my Oculus Rift S not working the other week, I did a few seemingly random things. I turned off "Game Mode" (which was on by default for some reason?) and turned off XMP (just forgot to turn it back on). I don't think those would have any impact, but...

Thanks for the help. I can also just take the front of my case off and experiment to see if that'll make a difference in it crashing. I should also probably reseat my CPU cooler and recheck my temps.

With all this information, do you think it'd be a good idea to order a new PSU?

Check reply #4 on that.

I wish you loads of luck in getting some stability. That is a whole lot of horsepower to enjoy.

 

[tachophobicat]

Check reply #4 on that.

I wish you loads of luck in getting some stability. That is a whole lot of horsepower to enjoy.

I think I'm going to experiment a little more and buy a new PSU. Messing with power settings, when I set it to 'power saver' mode, it doesnt crash with the Gradient tool in Photoshop. Setting it back to 'High Performance' instantly crashes it.

Very strange symptoms. Thanks for your help. I'm going to look into my high temps while I wait for the PSU to show up. I don't think they're causing this specific problem but I didn't realize they got so high after the new CPU swap.