PC just switches off under load - was rock steady before moving house.

Feb 3, 2019
9
0
10
Hi All, I've used this forum a ton due to all the great response you all give.

I've finally hit a wall with my PC and need to make difficult financial decisions :(

In short, I have an oldish system that I overclocked and was running rock steady. I recently moved and now have had constant issues.

Intel i5-3770k 3.5GHz (was overclocked to 4.3)
cryorig H7 air cooler
gigabyte GA-Z77X-UD3H motherboard
Radeon RX580 Nitro+ sapphire 8GB
Corsair RM650i PSU
2x Kingstons HyperX RAM 8GB

So the story goes. After moving I set up the system and pressed power. It gave a continous short beep error and would then attempt to restart, rinse and repeat.

Ultimately, I took everything out of the case, breadboarded it and gradually added devices back. I also took out the CMOS battery and reset the BIOS (using the pins). I also reapplied thermal paste to my CPU and put in some rubber stands for the motherboard.
This seemed to work.

Except, not long into working again, and at this point I had not overclocked again. The PC started just shutting down.
Event viewer would come up Kernel error 41.

The things I've tried so far;
I used DDU to clear the GPU drivers and update them.
I reset the BIOS again.
I updated the BIOS.
I re-installed my sound driver.
I re-seated the RAM
I've tried different RAM
I re-seated the GPU
I tried underclocking the GPU and trying suggested Wattman profiles for the nitro+
All fans appear to run OK.
There doesn't appear to be any visible damage on the motherboard or bulging capacitors.
I've left it idling for 24 hours with no crashing. Although, it has sometimes crashed again shortly after 1 crash.

I have tried testing with Prime95, memtest86 and OCCT, every test usually crashes, but there doesn't seem to be any consistency about how long it will last.

I have HWInfo but temperatures don't seem to be excessive. Generally around 55-60 for the CPU and ~72 for the GPU at max load. I can't tell if its spiking before a crash as the data doesn't get a chance to be logged.

My problem now lies in the fact I don't have spares to switch out the items. Is there anything else I could try do to figure out if its the PSU or the motherboard that has failed?
or maybe it hasn't failed and there is something else I can do?

Thank you heaps for all your help.
 
Feb 3, 2019
9
0
10
So disconnected the gpu. Started up ok. But couldn't even run the occt test. First attempt band up with "error detected". Subsequent tests crashed.

This was the first tab, not CPU linpack though, which could start.
 
Feb 3, 2019
9
0
10


Hey, as far as I can tell it should be. The BIOS has been reset to defaults and I uninstalled any overclocking software.
As far as I can tell memory is OK, has passed a few memtests. Is there some other way I should check?

I also tried to run a prime95 overnight without my GPU and it crashed within the hour as has been typical.
 
Feb 3, 2019
9
0
10


Typically around 35-40 at idle. Under load its around 60-65, a couple of cores register high 60s. All this using HWinfo.

Edit: I reinserted by GPU and managed to play tabletop simulator today without issue. So it seems issues only appear under load.

I guess one option is to buy a new PSU and if that doesn't work I know its the motherboard (outside chance its still the ram but a new motherboard needs new ram as mine are DDR3)? or am I over simplifying here.
 
Feb 3, 2019
9
0
10


As mentioned earlier, its just over a year old. I purchased the GPU and PSU December 2017 to give the old girl a bit more life. I definitely think the PSU can handle the power requirements. It was working fine until moving house so I'm guessing something is damaged... just trying to figure out what as I can't really afford to buy basically a whole new setup right now.

GPU is approx 30 at idle, gets up to ~75-78 during stress testing. When I successfully get a gaming session in, its in the high 60s.
 
Feb 3, 2019
9
0
10


Just want to say thank you for responding to me so far.

Yes I took the CPU out and re-seated as well as cleaned and reapplied some thermal paste. This resolved my computer not even starting and giving a continous beep error. But now I'm stuck with stability issues.
 
Feb 3, 2019
9
0
10


I only looked at it briefly and didn't notice any bent pins. I'm not sure I would have noticed if some were missing or broken however.