Yesterday my PC shut down while playing games.

System Specs:

 

 

CM H500M Case

RM1000i PSU (set in multi rail OCP, default) brand new, only three months usage.

Corsair H115i Pro with extra LL fans at the top + 4 corsair ARGB strips.

Prime X370Pro MOBO

Ryzen 1800x at 4Ghz

64GB RAM

Gigabyte 1080ti waterforce xtreme (2035 max clock, 5900Mhz memory)

2 HDDs

3 SSDs

 

Yesterday while playing Starcraft II the PC shut down completely, none of my component was overheating either, in fact my PC has a very good airflow. That game also uses only 50% GPU even at 4K. The MOBO was still on and I needed to completely cycle the machine by turning the PSU off until the MOBO LEDs went off. Then the machine could start again (heart pounding minutes)

 

I'm running all my equipment (PC and other things) through a Furman M-10x E power conditioner. By the time the PC shut down everything else was still on which makes me believe the power conditioner is not the problem.

 

In fact I have been gaming and rendering, in short pushing the system without any problems while the PC had my old 980ti card.

 

With the new card the PSU is around 465w max PSU power sum (HWinfo) during GPU stress. And I played hours of Shadow of the tomb rider without any issue.

 

Yesterday after the PC shut down I started wondering if I needed to set my PSU to Single rail mode or that is unrelated.

 

I think it was either that, or the Furman M-10x E cannot handle that load. Although the unit was not the one shutting down.

 

Recently I installed 4 LED lights on a lamp (also connected to the furman unit), maybe that is drawing a little bit more causing the shut down. But again, the Furman unit didn't shut down during all this so I'm not sure.

 

Another explanation would be a short lack of power from the wall. short enough to shut down my PC but not short enough to turn everything else down....rather unlikely though.

 

Do you guys have any thoughts in this regard?

Tried CPU + GPU stress for 5min.

 

Total usage 650W

PSU Fan was running at 550RPM at the time. The PSU temp was 35c

 

No power shut down.

 

No idea what happened last night then. :S

I ran the logger from iCUE with all the PSU data while I was running the CPU+GPU stress test and later on while gaming. Found nothing that could indicate a PSU problem.

 

The GPU stress test is the Furry MSI logo from Kombustor which uses 99 GPU (puts the clock and memory all the way up) and the CPU stress was Mandelbulb using 6 of my 8 cores so the PSU was being pushed substantially. I would say enough to expose any PSU fault. Maybe I should try for more than 5min.

 

I could even see the fan spinning as the output was 650w.

 

The PSU had the fan set to fixed 700RPM at the time of the shut down.

 

I doubt it but perhaps the PSU shut down because something (a rail?) triggered a warning because the fan was fix at 7000RPM and the PSU freaked out.

 

I doubt it cause the fan was set to fixed days ago and I played games every day without issues.

 

If it happens again I already decided to run the PC from the wall then the rest of my sound equipment through the conditioner. Which is rated for 10A but I'm not sure what that exactly means (model Furman M-10x E) since the unit has allowed me to run everything from it without any issues for about 2 years.

 

It can still be it. Since I recently added the 1080ti and the 4 LED lights on the lamp. But the unit doesn't get warm at all and also has a LED light and a breaker mechanism, which remained on during the shut down.

Run your PSU as a single rail so that power can go where it needs without concern

 

Make sure the machine is clean and no dust is fouling anything

 

check power cables to be sure they are snug too

Power cables are fine. Checked just after it happened.

 

The machine is dust free and has a rather overkill airflow. Even more with a water cooled GPU. The case temps are rather ridiculously cold at load. :)

 

Now. I've read about single/multi rail and I agree it seems a better choice to put it in single rail but I'm still not sure how much power delivery each rail on a RM1000i gives out and how, also how many. I checked the manual and specs and is not clear to me.

 

Should also things be plugged in a certain way on a modular PSU to take full advantage of multi rails? Say the CPU power and the GPU power on different levels in the connectors. (I think I have it that way anyway)

 

I've also read there's no problem using your PSU in single rail, but there's still a small chance you can damage things if there's a short.

Thanks a lot for chiming in. I remember I saw that video some months ago but I had no clue what they were talking about until now.

 

So I assume each rail can easily support a 1080ti running at max clocks/ram

 

I used a calculator:

 

1080ti: 12 V DC and 250W gives 20.3A

 

Is that correct?

 

From the video I infer the RM1000i has say, 8 rails each capable of 40A (?)

 

My PC has not shut down since that incident, even during some GPU heavy F1 2018 session.

Here's a screenshot of HWinfo during F1 2018 benckmark (Australia) at 4K maxed out + Reshade.

 

I think by reading at the numbers there is plenty of juice left on each rail. No reason for the PSU to shut down.

 

I'll keep it in multi rail and see. If it happens again I will plug the PC directly to the outlet and the power conditioner on a separate outlet.

Here's a pic of the specs of the power conditioner.

 

It says rated for 10A which seems very low for my PC

 

But also says 12.000A peak impulse current.

 

I suspect I should not use it for my PC. But I'm not sure what those specs really mean.

 

And, why the power conditioner is not tripping immediately if is rated for 10A only? It shouldn't be able to handle my PC even at low loads.

 

The power conditioner has 11 outlets, that makes me think it can handle quite some power but I'm certainly not sure of any of this.

Whatever. After reading my conditioner outputs 10A I better be safe than sorry and I connected the PC to the wall and the rest through the conditioner.