New system unstable +12V shows 20.736V...

[trylo]

Hi!

 

I've built a new system, all was fine and dandy for about a week but yesterday I had a crash so I decided to test it with AIDA64 and after 7 minutes system became unresponsive. Do you have any suggestions on how to tackle this?

One weird thing that I noticed: HWMonitor is showing 20.736V on +12V. This does not seem normal or is it?

 

System:

Asus WS Z390 Pro

i9-9900K (not overclocked)

32GB Crucial Ballistx DDR4 RAM

Zotac GTX 1080Ti Amp! 11GB

Thermaltake Toughpower Grand RGB 850W

 

Thank you in advance for all your help!

[WereCatf]

I sure hope that +20V on +12V rail is a bug and not actually real, because that'd be fucking alarming.

[simson0606]

I do not expect that it's really 20v on that rail. Asus boards have pretty good over/undervoltage protection, so i'd guess it would have shut off by then. Not to say that the psu is fine, it could still be a problem. Do you have another psu to test with?

[Sakkura]

Software voltage readings are notoriously unreliable in general, and this one is clearly not a real reading.

[trylo]

12 minutes ago, simson0606 said:

I do not expect that it's really 20v on that rail. Asus boards have pretty good over/undervoltage protection, so i'd guess it would have shut off by then. Not to say that the psu is fine, it could still be a problem. Do you have another psu to test with?

This is a new power source, just when I got this system my old PSU died. So this power source is the only one I have right now.

[trylo]

7 minutes ago, Sakkura said:

Software voltage readings are notoriously unreliable in general, and this one is clearly not a real reading.

Do you know if this reading can be found in Bios or how to measure it on board with a multimeter?

[NelizMastr]

Also, which appalling cooler are you using? Those temps aren't healthy for a stock 9900K.

 

This instability can have a couple different causes, you'll have to test components individually. Start by not stressing memory, and see if it locks up or not.

If it doesn't, run memtest over night.

 

 

[trylo]

1 minute ago, NelizMastr said:

Also, which appaling cooler are you using? Those temps aren't healthy for a stock 9900K.

 

This instability can have a couple different causes, you'll have to test components individually. Start by not stressing memory, and see if it locks up or not.

If it doesn't, run memtest over night.

 

 

It's H80i with 2 Noctua fans on it. It is 6 years old though. 

Normally when I export videos temps are between 74 and 83 degrees.

I will do a test without stressing memory.

[Sakkura]

29 minutes ago, trylo said:

Do you know if this reading can be found in Bios or how to measure it on board with a multimeter?

BIOS voltage readings can be unreliable too. Some motherboards have probe points where you can check voltages easily with a multimeter, but your board doesn't seem to have those. It's more of an extreme OC thing, while your board is workstation-oriented.

[trylo]

1 minute ago, Sakkura said:

BIOS voltage readings can be unreliable too. Some motherboards have probe points where you can check voltages easily with a multimeter, but your board doesn't seem to have those. It's more of an extreme OC thing, while your board is workstation-oriented.

Thanks.

Yeah, I went for the workstation MoBo as this computer is supposed to be rendering videos all day long, so I don't care for OC, just for a stable system.

[trylo]

I did the test again without stress testing RAM and the system became unresponsive after 5 minutes.

The temps were reaching 90 degrees C.

H80i should be more than enough to keep this CPU cool, maybe I messed up the installation... too much paste or something...

 

[trylo]

This doesn't look like too much paste. What do you think guys?

[trylo]

I removed the paste from the cooler, just left what's already on the CPU and after a quick 10 minute Aida test it looks much better.

Temps under 80 degrees, maybe occasional half a second above that.

As you can see I ran it without GPU. Then I ran it just only on GPU just for 1 minute and again the screen was hardly managing to change the position of the cursor. Is this normal for GPU stress test?