Had exactly the same problem last year and solved it the same way.

Wow lots of replies, thanks everyone.

Okay let's try to do a bulk reply-ish type thing to some of these:

Firstly just about everyone seems to think it's more than likely the my PSU the 2013 Corsair CX750M.

It occurs to me that whilst i have no spare parts to swap in and out i can almost definitely buy a PSU and then send it back if it doesn't fix the problem.
That's certainly the cheaper option because whilst i keep looking at a new 20 series GPU let's be fair a 970 is still perfectly adequate for 1080p Ultra and some 1440p gaming and i'm sorry but ray tracing, dare i say it, isn't quite there yet (don't hate me).

I'm fairly happy that with the amount of stress testing, logging and monitoring i've been doing that I probably don't need to do anymore, i've used plenty of different tools hwmonitor/gpu-z/etc etc to log data and record temps and non of them seem to show any spikes or drops in voltage or power to the hardware in the final milliseconds before the actual power drop out and i constantly monitor my temps anyway just out of habit and nothing has shown up that would even start to worry me.

Thanks to everyone again but in particular to AB2012 just because i really enjoyed reading your post, i love learning things like that.

One more thing....

I'm aware this may cause a lot of replies due to varying opinions buuuuut, what do people think of the new EVGA 750w G3 with a 10 year warranty?

Be very happy if your cheapish PSU goes bad or dies without taking other components with it. Total destruction of a motherboard and processor taught me to never buy a cheap PSU just because it is the only model sold in town. I can go to another town or order over the net and so can you.

I had this problem a few years ago. I replaced the power supply and it still happened. Turned out the motherboard was dying.

Quick hint: Go with the PSU, try to find out if that is the problem. I guess that will be it.

However, if it turns out not to be: I had a very similar issue years ago and it was not even my machine, it was my electricity wiring in my appartment.
The only way to find that out is to bring the machine somewhere else, or connect it to a different outlet in the house which you are sure is not on the same mains as the previous one.

// edit: for clarification: computers are very sensitive to shifts in voltage and such things. if you suck up more energy than the (maybe faulty) wiring can provide, you might get an outage for a fraction of a second, enough for it to restart. by "mains" i basically mean the ELCB/circuit breaker. same house is fine, as long as you are on a different circuit.

Same here. I was getting hard resets as well on my old machine. Swapped out three different PSUs and two different GPUs. Ended up that the old AMD mobo was just shot. Hardware goes bad sooner or later. Inescapable truth of computers.

Pro tip - keep your shit clean. Take your covers off once in a while and give it a good cleaning with a cloth and/or Qtips. Getting dust and dirt out will add years to hardware life.

Like others already said, the PSU is probably the culprit in your case. If not that, then the motherboard is my second guess.

The EVGA 750 G3 is very good, because it is actually made by Super Flower. SeaSonic and Super Flower are the only PSU manufacturers I can personally recommend.

If you want to buy a PSU from other popular brands that don't manufacture power supplies (like Corsair, Cooler Master, EVGA, NZXT, Thermaltake, Fractal Design, etc.), then look on this site who actually makes the respective model (the OEM Plataform column). If it's SeaSonic or Super Flower, you're good to go. If it's another one (like FSP/Fortron, Sirtech, HEC, CWT /Channel Well Tech, Andyson, ATNG, etc.), then I would personally skip it.

jonnyguru is a very good site for PSU reviews.

You can safely get a 650W PSU for your system. The OuterVision online power supply calculator is very good for finding out what a system's Load Wattage will be. Good luck!

I ran that same PSU for over 5 years and it never gave me issues, even when running an overclocked R9 290x and a (very) overclocked X4 760K an an Asus A88X Pro board.

I did, however, have identical issues when running the previous motherborad, an ASRock Fata1ity board. It was caused by insufficient northbridge cooling. The northbridge heatsink on the Fata1ity board was poorly secured at best, leaving airgaps between the heatsink and the chipset. My solution was to go to the Asus board which uses active heatpipes for northbridge cooling.

It is probably caused by a stuck reset switch pull the plug on that reset switch!

Oh and NEVER buy a "Antec Edge 750W 80 Plus Gold Power Supply" I had one and it's inside MELTED with white plastic everywhere the Shop said OH SH*T Aaah... we'll give you a store credit to the amount you spent on it and so I got a "Corsair CS650M Modular 80 Plus Gold Power Supply" I have had 0% Problems with it so far and that was on 11-may-2015 and today it's still good!

Also NEVER buy a PSU from Silverstone it fried my whole PC once.

That's interesting, plus thanks because you just reminded me that i missed out my motherboard from my spec.
I like jonnyguru it's a good site.
That other site is really interesting though, it's great to find out that sort of thing.
I always liked it on shows like Top Gear when they'd look at a Porsche or something and say that it actually had a normal VW motor with an Audi chassis and i dunno skoda suspension or something crazy.
Anyway i'm babbling thanks.

If you run a game in windowed mode so that you can observe your hardware monitoring sensors you can watch for any anomalies.

I just remembered having a similar but brief issue I had a while ago. I had a case fan that was on it's way out and was drawing a ton of power and shutting off the system. It would only happen when the fan was running at full speed.

I was using the Asus smart fan control, so the issue only occured under heavy load when the program kicked the fan speed up to max. It didn't happen often (maybe once every couple of months) because I wasn't ever doing much to push the hardware. I discovered it when I was showing off the fan controls to a friend. Every time I would turn the fan speed to max the computer would shut off. I unplugged the fans one by one until I found the culprit. I took it out, and then smashed it for good measure.

Interesting, I didn't know that fans could cause such problems O_o
Hardware issues are always a PITA to identify correctly.

from my guess, it could be one of the rails has developed a fault. Look at the voltages with CPUID hw monitor and keep an eye while gaming. If any rail changes noticeably, shut the game down and probably the PC because you have a problem on that rail (ie the 12 volt should read around 12. If it drops to about 10.6, the PC does what you see. I would FIRST remove the overclock from the GPU and see if it still powers down. If it does with your current PSU, you need a new one.

Do you dust out your pc once in a while (half-yearly, yearly)?