[SOLVED] PC crashing when full load on CPU and GPU after LED upgrade

DarioCaptain

Prominent
Jun 8, 2019
10
0
520
1
Hello all,

my PC was running fine for a long time now. However I wanted to upgrade it with some more RGB lights to use it as a mood light.
Therefore I put an LED strip on the back of my case (Thermaltake P5) and a self made LED controller inside of it. And connected both to the MOLEX connector of the PSU (can this already be a problem?).
After some time the LED‘s stopped working and I got them out of the case. One of the solder connections got bad.
I messed around with it and tested it on my bench power supply and it worked again.
I connected everything and rebooted the PC. It crashed immediately but I had no display connected to it because I usually use Remote Desktop with it.
After connecting a display I could see it crashes with BSOD’s. However I could not read the message of them because they went away so fast. They also do not produce a mini dump file.
First I thought that maybe it crashed because of the high LED current draw and windows got corrupted.
Therefore I disconnected everything which is not necessary and reinstalled Windows 10 Pro.
My components are now as followed:
  • Gigabyte X399 Designare EX
  • Threadripper 1900x
  • RTX 2070S
  • 64GB G. Skill RAM @ 3000mHz
  • 500 GB WD Black NVME SSD
  • 120 GB Sandisk Sata SSD
  • Corsair HX850i power supply
  • Windows 10 Pro
Everything seemed to be fine until I played some games. While playing the PC randomly crashes from time to time.
Because of this I downloaded different stress tests and diagnosis tools.
I have done:
  • Check the WD SSD with the manufacturers tool and run some Benchmarks
  • Checked the disks with the windows cmd tools (CHKDSK)
  • stressed GPU whith Furmark
  • stessed CPU&RAM with prime95 (smallest FFT, small FFT, Large FFT and Blend)
While doing all of this the PC is working fine. However when I run Furmark & Prime95 together the PC crashes. The BSOD goes away immediately so that I cannot read it. But filmed it and went through frame by frame and it says: “HWAE UNCORRECTABLE ERROR”
This is reproducible however it seems like the error only occurs when I use the blend or the Large FFT mode.

Now I am clueless what could be the culprit?
Did I harm my PSU when adding the LEDs and now it does not work anymore?

Any help would be very much appreciated and thank you for reading
Dario
 

DarioCaptain

Prominent
Jun 8, 2019
10
0
520
1
Finally I figured it out. Maybe it can help somebody in the future troubleshoot his/her problem.

After seeing the low voltages in the bios I suspected a problem with the PSU even more than before.
Turned out that my 24pin cable got slightly damaged so that the voltage dropped a little bit.
Using a tweezer I bent the pins of the cable back in place and now everything is working in its previous glory.

SH Gaming thank you very much for your help :)
 

DarioCaptain

Prominent
Jun 8, 2019
10
0
520
1
Update:
I am currently running a Memtest86 run and it already found 2 errors in pass 3.
Is there a way to see in which ram module it found the error?
 
Oct 2, 2020
57
8
45
1
I would think it would be easy to use one stick and test each stick individually. It will take some time but it will help you pinpoint which stick(s) it could be, at least I think.
 

DarioCaptain

Prominent
Jun 8, 2019
10
0
520
1
After finishing the Memtest86 run with errors I tried multiple stick configurations in windows.
All of them resulted in BSODs.
I tried dual channel combinations and single channel.
The only results of using fewer sticks was that the PC crashed faster.
Now it also crashed while only running Furmark.

Does this now mean my Mobo is faulty?
Or could it be possible that the PSU is the problem or something completely different?

Edit:
Is it possible to test a PSU with a multimeter?
In my Bios under PC Health it says:
  • 3V3 - 2.98
  • 5V - 4.80
  • 12V - 11.880
I do not know how much pc components rely on exact voltages but 2.98 seems to be quite a bit of from 3V3.
 
Last edited:
Oct 2, 2020
57
8
45
1
It could be the motherboard but being newer to this kind of stuff I do not ave a definitive answer to this.
I am Not completely sure if you can use a multimeter on a PSU but there is probably a way. But to my understanding it might be the Motherboard Causing the problem.
 

DarioCaptain

Prominent
Jun 8, 2019
10
0
520
1
Finally I figured it out. Maybe it can help somebody in the future troubleshoot his/her problem.

After seeing the low voltages in the bios I suspected a problem with the PSU even more than before.
Turned out that my 24pin cable got slightly damaged so that the voltage dropped a little bit.
Using a tweezer I bent the pins of the cable back in place and now everything is working in its previous glory.

SH Gaming thank you very much for your help :)
 

ASK THE COMMUNITY