[SOLVED] How to test/determine if the PSU is faulty for computer restarting?

Mar 4, 2020
4
0
10
Greetings, id like to thank you for the help in advance.

So here goes all the story:

I had an old PC (core2duo + gts 450) where I used an ICEAGE IA5000HP80 that lasted for 3 years then It died. so I bought a corsair cx550 that died after a year, I thought maybe I had bad luck and bought a corsair cx600 to replace it, that died after a year aswell.

I was tired of having problems with PSU and I had money to start a brand new setup, so last year I built the following computer, that im using until today:

I7-7700k
Watercooler Corsair H150
GTX 1060 6GB AORUS
Mobo Asus SLI Killer
Ram 4x4GB corsair vengeance.

I made a lot of research at the time and decided to buy a seasonic s12 620W, it seemed to be the best/brand model that wouldn't cost insane amounts of money.

So i used this computer for a year with absolutely no problem, but now when I demand too much of the computer it shut down and turn on again. for example, if I try to run a heavy game like black ops4 the computer will shut down in about 5 minutes after running it, but I can still run a lighter game like apex legends on low setting without crashing, but if i run apex + discord + browser boom, shut down again.

I tried running the game using Hwmonitor + coretemp to check overheat, the maximum cpu temperature was 69C and GPU 65C, so im pretty sure its not about overheating.

So the main question is:

Is there a way to determine for sure that the computer is crashing because of the psu? i cant believe i spent a ton of money on a really good psu to not have headaches and im getting screwed like this.
 
Solution
First off, Corsair's CX PSUs have a 5 year warranty. Did you pursue a warranty replacement when they died prematurely?

My #1 culprit when I get random restarts is a faulty stick of RAM. The best way to test for this is to run something like Memtest or Prime95 Blended with only a SINGLE stick of RAM installed (run CPU at stock clocks obviously). You need to test each stick individually to locate the pig.
First off, Corsair's CX PSUs have a 5 year warranty. Did you pursue a warranty replacement when they died prematurely?

My #1 culprit when I get random restarts is a faulty stick of RAM. The best way to test for this is to run something like Memtest or Prime95 Blended with only a SINGLE stick of RAM installed (run CPU at stock clocks obviously). You need to test each stick individually to locate the pig.
 
Solution
Mar 4, 2020
4
0
10
First off, Corsair's CX PSUs have a 5 year warranty. Did you pursue a warranty replacement when they died prematurely?

My #1 culprit when I get random restarts is a faulty stick of RAM. The best way to test for this is to run something like Memtest or Prime95 Blended with only a SINGLE stick of RAM installed (run CPU at stock clocks obviously). You need to test each stick individually to locate the pig.

I live in Brazil, not possible to purse a warranty here cuz you need a fiscal bill, wich if you don't get you get like 30% off on cut taxes.

so is there any way to tell if the PSU is causing the resets? ive contacted technicians in my city, noone of them has a power supply good enuff on their store to test it out for me.
 
If you can't get a different PSU to eliminate that candidate, check the RAM to make sure it's not the RAM that's causing this.

(in addition to my RAM testing instructions, it's a good idea to remove the GPU from your system to eliminate that variable while testing)
 

King_V

Illustrious
Ambassador
It's also worth noting that Corsair's CX series improved starting in 2017, and I believe that's when they started getting 5 year warranties... used to be 3 years before that, I think.

The S12 is an older design, but a rock solid PSU . . if you have the S12II (roman numeral 2). On the other hand, the S12III (roman numeral 3) is not even made by Seasonic, and I have no idea why it exists, but I'm given to understand that it's an absolutely TERRIBLE one.

The Seasonic Focus is the modern design, and would be better if available, but I don't know what the availability is like in Brazil.
 
Under load, a graphics card will draw more power so a psu is the prime suspect.
Unfortunately psu testers are only good to confirm a dead psu, but can not test for proper operation.
If you can, borrow a known good psu of 550w or better to test with.
If that is not possible, buy such a psu from a store with a good return policy.
Expect to pay a 15% restocking fee if you need to send it back.

The graphics card could be another suspect.

On ram, memtest86 is the gold standard.
You should be able to complete a full pass with NO errors.
It does not use windows.
 

bignastyid

Titan
Moderator
It's most likely the psu - but what most people won't tell you is that the Seosonic really isn't better than the competition.
I would recommend an KDM 1000 Watt, which is made for high load gaming:
https://www.ebay.com/itm/322924234791#rwid

Thank me later.
Even Seasonics worst unit is better quality than that pile. With a weak 36a available on the 12v rails it is most definitely not made for high load gaming, it's made for rubes that don't know any better. Now I suggest you quit recommending dangerous garbage or you won't be posting here anymore.