Possible PSU issue?

[jfp555]

Hey everyone.

The following issue caused me immense amount of stress during a very crucial project. I would greatly appreciate some guidance here.

Specs:

i7 4790 (non k)
Asrock mobo
16 gb RAM (2 x 8gb sticks)
Corsair VS550 (550w bronze PSU)
MSI Gaming 4g 970
1 x SSD
1 x HDD 7200RPM
2 ext hdd's over USB

Main issue:

The entire computer would reboot whenever I would try to render out through Premiere Pro, or try to play Company of Heroes 2 but would not reboot during most other games. I tried OC'ing my 970 to 980 levels and ran GTA V for about 10 minutes and not a single hitch.

Then the rebooting stopped in most things after I added a voltage stabilizer between the PSU and wall outlet.

But, today I tried Company of Heroes 2 and the second it hits ingame (after everything loads and I can play the map), instant reboot. I then had an idea, I used MSI afterburner to set the power limit to 69% and the rebooting stopped and I was able to play an entire skirmish game of Company of Heroes 2.

Can anyone please help me accurately identify the problem? It seems that whenever the GPU and CPU are stressed together, I get reboots. But I have been running the same setup for over a year and had no issues.

Is this a mobo, GPU or PSU issue?

 

[lenjack]

Possibly heat related. How's the air flow? Do you have any heat monitoring software in place?

 

[Ketchup]

Running a 4790 and a 970 with a 550-watt psu may be pushing it.

If you want to check your CPU, run occt. Your CPU should hit 100 C and then start throttling.

 

[denis280]

Running a 4790 and a 970 with a 550-watt psu may be pushing it.

+1 it is borderline

 

[jfp555]

Running a 4790 and a 970 with a 550-watt psu may be pushing it.

If you want to check your CPU, run occt. Your CPU should hit 100 C and then start throttling.

I'm inclined to think the same way. However, the reason why I'm so curious is that I ran this exact same setup since dec 2015. Do PSUs (especially if used borderline) become less tolerant to this load over time?

 

[Ketchup]

Unless you know exactly how much wattage you were using since the end of last year, it would be hard to say if you put as much stress on the power supply before now. For example, Windows can see 8 logical processors on a 4790, but there aren't that many programs that would actually see and use them all. And if you found a program that did use them all, is it using your video card as well?

 

[mikeymikec]

PSUs do drop in efficiency over time. IIRC either anandtech or TechReport (I think the latter) did a review of one of the PSUs they use for testing when it reached about 6 years of age, and there were differences compared with its performance on day one.

Everything suffers from wear and tear in some respect.

 

[nerp]

It could be natural wear or it could be failing somehow. It does sound like a PSU issue. You could rule it out with a PSU swap. Other possibility is memory, but that might introduce other wonky issues in other places.

 

[jfp555]

Possibly heat related. How's the air flow? Do you have any heat monitoring software in place?

I'm actually using this with the casing lying on its side and the side panel removed. (methinks dust buildup MIGHT be an issue, but running MSI afterburner during most gaming sessions I have not noticed temps that would raise alarm)

Unless you know exactly how much wattage you were using since the end of last year, it would be hard to say if you put as much stress on the power supply before now. For example, Windows can see 8 logical processors on a 4790, but there aren't that many programs that would actually see and use them all. And if you found a program that did use them all, is it using your video card as well?

Premiere pro, during renders (that was when most of the crashing was occurring) hits CPU, GPU and RAM equally hard and if the project is complex (lots of effects using both CPU and GPU), it puts 100 load on everything for a while.

Similarly, Company of Heroes 2 is a very demanding game with complex physics as well as effects and AI being put to use. I have not had much time to game, but CSGO never crashed the system. So that makes me wonder about PSU load. It would be a relief if it was just the PSU and not the GPU/CPU or mobo.

 

[mikeymikec]

I'm actually using this with the casing lying on its side and the side panel removed.

Is this generally speaking or during this crisis? I'm wondering because I would be surprised if PSU cooling systems are designed with having the PSU the wrong way around in mind. While having a fan pointing at components and heatsinks sounds like it should result in said components being cooled, I wonder if having the PSU on its side might mean that there's a corner or side that doesn't get its air circulated as well as the rest. On the other hand, if the PSU gets plenty of time without being pushed, that might allow said air to cool enough for it not to be a problem.

This is only a theory that I've just come up with, it should be irrelevant if it's just for temporary testing purposes.

 

[jfp555]

Is this generally speaking or during this crisis? I'm wondering because I would be surprised if PSU cooling systems are designed with having the PSU the wrong way around in mind. While having a fan pointing at components and heatsinks sounds like it should result in said components being cooled, I wonder if having the PSU on its side might mean that there's a corner or side that doesn't get its air circulated as well as the rest. On the other hand, if the PSU gets plenty of time without being pushed, that might allow said air to cool enough for it not to be a problem.

This is only a theory that I've just come up with, it should be irrelevant if it's just for temporary testing purposes.

This is the way I've been using it. I did hackintosh the machine so that required a lot of switching around of drives .Thats the way I've been using it since pretty much always.