My computer shuts off during a game or benchmark, this occurs with a seasonic 1000w platinum PSU.

tektechnician

Reputable
Sep 24, 2014
18
0
4,510
My computer shuts off during gaming or benchmarking a overclock. I run my 4770k at 4.5Ghz at 1.275v and both of my gtx 780s on 1.3v at 1333Mhz. It doesn't turn off if I run the default clocks, however when I do a split power system, using a 850w psu to power one card and the 1000w psu to power the rest of the system, it shuts off within a period of 10 minutes, but not instantly. Its definitely not thermal though.
I checked on a PSU calculator and it recommends me a +1300w PSU, which I find kind of ridiculous for two 780s.
Sound right for you guys?

Specs are
2 gtx 780s
4770k
1000w Seasonic Platinum & 850w psu
z97-pro
 
Solution

tektechnician

Reputable
Sep 24, 2014
18
0
4,510


That is certainly worth a try, however does this seem like a case of my psu not being capable of powering my system? everywhere I read on the internet on sudden shutdowns are either thermals or dodgy power supplies.
 

uber9000

Distinguished
May 31, 2014
208
0
18,760


It very well could be. I personally don't have experience with 2 PSUs at the same time, so that's hard for me to say. Just never seemed right to me. If it works fine with one PSU and no Overclock, then it could be the PSU, or the board/CPU sucking too much power. Have you tried it with and without the Overclock with each PSU separately? Maybe you can rule out your PSUs that way.
 

tektechnician

Reputable
Sep 24, 2014
18
0
4,510


Seems to be more stable if I drop the cpu to stock voltages and both PSUs work just fine on stock clocks on all things, so maybe I am just tripping the OCP or OLP in the PSU ? but isn't 1000w psu used for literally triple to quad gpu rigs?
 

uber9000

Distinguished
May 31, 2014
208
0
18,760


Yeah, it definitely sounds like you are. And yea, 1000 is way high for 2 gpus unless they OCd to hell
 


With a split PS setup you have a lot of possible causes why you are experiencing the crashes and the solution could be as simple as Supahos last suggestion, meaning the CPU overclock is not stable which could cause the problems for sure.

You show no load temperatures so it's hard to see what's going on there to make a possible cause from inadequate CPU cooling.

So I have some questions.

How is the second power supply initiated?

Motherboard link jump?

Are the power supplies all single 12v rail power supplies, and what are their amperage load handling capabilities?

 

tektechnician

Reputable
Sep 24, 2014
18
0
4,510
I found the reason, it had nothing to do with my CPU or PSU, both of my cards never drew more than my PSU was capable of, it was actually the power delivery system on my first GPU shorting the short circuit protection on the PSU, the bigger issue is now, my first card is dead with a hole blown in one of the power chips, what do I do?
 

tektechnician

Reputable
Sep 24, 2014
18
0
4,510


Now that I know that my gpu has died, how can I rma it? it has been flashed to skynet but I don't think that's the reason why it died, I checked the part that has blown and that part itself can handle max 40 amps into it and operate a temp of 150 degrees that is in the area of safe (r2j20658np). so it's definitely not how I overclocked it. What do you think I should do? I can easily get that part and repair it, however I do not know what else it took out on the board (maybe even the main chip, it won't recognize in windows). Or should I try rma it, even though it is flashed, or does that not really matter considering it's no longer recognized inside of windows?
Thanks for all your help though, this issue is really out of the ordinary.

 

nayrnayr1

Reputable
Aug 8, 2015
333
0
4,860
If you want, you can feed it more power, that will fix it, but I would turn down the overclock. They should work after that. It usually will shut down because of not having enough power. It will freeze because of a bad overclock.
 

tektechnician

Reputable
Sep 24, 2014
18
0
4,510


It's dead, it had nothing to do with my overclocks or voltages, a mosfet was faulty inside my gpu, it shorted and blew apart (Hence triggering of the short circuit protection in the PSU). had nothing to do with how I pushed it, the spec sheet for the chip that died said it can handle 12v 40 amps up to 100 degrees celcius.

Do you think gigabyte will RMA it? Its power delivery system from the pcie slot is dead, so no communications from the card, I doubt they can even test it. And if they don't RMA it, I'm thinking of perhaps replacing the chip, it's the Renesas R2J20658NP, Do you think that is the only thing that died or do you think it took out everything else with it?
 


Since it has actually blown a mosfet, just RMA it to Gigabyte normally and do not volunteer any information to them regarding flashing it at all.

 
Solution

nayrnayr1

Reputable
Aug 8, 2015
333
0
4,860


Gigabyte will not RMA it. It was a problem related to overclocking. You could always lie though, or just not say the details.
 

tektechnician

Reputable
Sep 24, 2014
18
0
4,510


Okay then, thanks for helping! :)