Computer question for the geeks out there

So... I has a little puter problem

The thing works fine if I am browsing, etc... everyday stuff.

But, if / when I launch a game that is graphics intensive (Xcom/Witcher3/Fallout4), the machine will turn off DEAD after a minute or two.

I have checked graphics cards. This is a brand new Motherboard, no problems there. No problems with the CPU or memory.

I have been checking temps of GPU(s), mainboard and CPU. Non go critical, even in the last second or two before shutdown.
I cant get a full log, because the computer shuts down dead and leaves me with open files.

The one thing left that I am suspecting in the PSU. It is at least 7 years old, it has/had enough ooomph for everything that is in the case.

I am suspecting that once the GPU's, etc start pumping, and demanding juice, the old PSU can't handle it.

PSU is rated at 850W. Monitor states that I am using 487W (peak 670W) with all the items in the case.

Could it just be an old PSU getting grey hairs and wanting retirement check before it dies???

will post a system snapshot in a min

when you say turn off dead- you mean like a hard shutdown- everything goes blink? I agree with you on the PSU flipping out- it probably needs replacing. 7 years of regular/daily use is a pretty hard on a PSU.

sys details

Ditto, the first thing I thought of when you said "turn off DEAD" was the PSU.

Now, even though it is the first thing that came to my mind, that may not be it. But based on everything you said, I would swap out the PSU next and see if the problem persisted.

Fingers crossed that would be the problem, no idea what to do next. I once had a problem with the computer freezing, was tearing my hair out trying everything, and after I burned my hand on the CPU heatsink, figured out the thermal paste had deteriorated...

Salvage-1 said:

So... I has a little puter problem

The thing works fine if I am browsing, etc... everyday stuff.

But, if / when I launch a game that is graphics intensive (Xcom/Witcher3/Fallout4), the machine will turn off DEAD after a minute or two.

I have checked graphics cards. This is a brand new Motherboard, no problems there. No problems with the CPU or memory.

I have been checking temps of GPU(s), mainboard and CPU. Non go critical, even in the last second or two before shutdown.
I cant get a full log, because the computer shuts down dead and leaves me with open files.

The one thing left that I am suspecting in the PSU. It is at least 7 years old, it has/had enough ooomph for everything that is in the case.

I am suspecting that once the GPU's, etc start pumping, and demanding juice, the old PSU can't handle it.

PSU is rated at 850W. Monitor states that I am using 487W (peak 670W) with all the items in the case.

Could it just be an old PSU getting grey hairs and wanting retirement check before it dies???

will post a system snapshot in a min

Click to expand...

Check to be sure that all of the fans are spinning up. Reseat teh CPU heatsink with fresh compound. But it does sound like dying PSU

Something easy to do, just to clear a nagging feeling: get a memory checker (free software) that will stress test the RAM. Run it and let it hammer away for an hour or two.

Six hard drives, two gtx970 - power supply likely failure point given the failure mode.

I am doubting the stated power utilization- seems low considering all you have in the system.

Just curious, why do you have 2 cards and not run SLI?

PSU is a possibility, but I like to start diagnostics with easy stuff like software first. GPU drivers utd? Try running the system with only the drive/s with the OS and the game and all other drives dc'd and maybe even pull a GPU out to decrease the load on the PSU. So many questions... Sounds like an old build with upgraded OS (win10). Did you ever have this problem with the old OS? How long have you been having this issue?

I usually run SLI and it is back SLI now, I run two monitors (sometimes 3) off the cards so i want one monitor on each.

Tried pulling down to one card and unplug all the USB drives, etc, leaving the main system drive, and the apps drive (I have one drive for system, one for games/apps, one for media server inside the case)
One one card and 2 drives, with nothing else plugged in... same

Pulled the 970 and put the old 560 card in there... same problem

pulled 2 memory chips, ran at 16gb, same problem
swapped over a couple of times to mix/match the memory sticks so they all got a chance to work with each other, just 2 in place at a time... same

I am about to pull the CPU and re-paste it (I may have overdone it a little), I am also going to pull the PSU and give it a hosing down with canned air (love the stuff)

I have ordered another PSU, an 850W Thermaltake Toughpower. I love Amazon, the PSU gets here tomorrow!!


How long have I had this issue. Hmm... looking back, about a year. Over the summer I had to open the case side and put a window fan there! I thought it was the CPU, then I thought it was the GPU's, then I thought it was the motherboard.

It started with high graphics FPS PvP games, the computer would play for an hour or two then it would turn off.
Recently it has started getting worse. Weird thing is that Kerbal Space Program doesnt cause it to shut down. XCOM / Witcher / Fallout 4 all cause the shut down.
I just ran a graphics benchmark tester. I heard a fan start up and then click... machine dead.

So...
Machine turns off when high usage, sometimes it needs a complete turn the PSU switch on/off and wait.
Everything else seems to be checked, I cant think of anything else.

I aint gonna lose anything by replacing the PSU, gods, everything else has been replaced over the last year.

Yes, this is a 'built myself' machine. I have upgraded all items over the last year, including upgrading OP system drive to SSD, etc. etc.

If your CPU and other temps are wnl I wouldn't bother with reseating the CPU for the purpose of a solution. PSU's are hard to cook, but like anything mechanical they do indeed fail. Replace PSU tomorrow and let us know how that worked out. If same problem exists I would wager a software issue. As I'm sure you are aware that even with 64 bit architecture there isn't any software out there that will utilize 32G of ram practically or theoretically so any reason you have so much?

Cpu fan

The fan that hit and then it shutdown- was the fan spinning before this? Could the fan have a short?

Can you take a clear picture of the MB and post it?

That's a very low CPU temperature - 11c. Sure that's a good report? Can you capture as you are playing games and see a difference? Reviewing your processor, the typical running temps are:
Idle Temp Normal Temp Max Temp
FX-6100 30 to 35°C 45 to 55°C 70°C

Just seems very low unless you just caught it a boot and it hasn't warmed up to normal operating yet.

Here is the irony. Part of what I do is repair computers.
I end up with all sorts of FrankenPuters, but this one is mine.
It has 32Gb because of an accidental order, someone wanted an upgrade, and then cancelled. I kept them myself, I am usually happy with 2x8gb, now I have 4x8.

From watching the temps, the fan that caused the shutdown was the second GPU fan.

I have the CPU fan and all case fans set to min 75% and max 100% in the bios
I have set Nvidia control panel to set GPU 1 fan to 100% from windows boot, and the 2nd GPU fan is automatic.

DOH!!!! as I was writing this, I kicked myself just now. One of the computers that I have in for repair has a functioning (but I have no idea how old) PSU. It is 550W but if i pull the 970's and the USB drives, and use the older 570 card, that should be plenty..

Now to check.

The 550W PSU has been in here for 30 mins now. So far, no problems doing systems checks etc. Going to start up Fallout 4 and see if I can last past my 6 minute record.

Attached cap is how it is running now, just browsing. I will try and get another when the game is running.

edit - well i screwed up ... so the pic that is here is the snapshot of the machines temps, etc... game running

Yes... it was the PSU... weird, in all the years that i have been building and fixing PC's, I don't think that I have ever once seen a PSU that has been worked to death.

Fallout 4 is quite good so far.

I had one that failed and wiped out the MB so consider yourself lucky. It was a darn nice Gigabyte with both northbridge and southbridge back in the day when that was unique....

PSU's these days seem to age faster now that more parts are made in China by the lowest bidder...
Capacitors in particular lose their capacitance over time, and as a result, a PSU's rated max wattage could, in all likelihood, be half that after 7 years.
I also have a new-ish 2-year old 600W (cheap ?? brand) where two of the sata hard drive power connectors had bad wiring, resulting in one of my hard drives receiving on-off power so often that it now has over 30,000 power cycles and perhaps as a result has a "BAD" S.M.A.R.T. status as well.
I was going to look for a new high-quality PSU, but ended up buying a new computer all together.

I recommend Antec power supplies- been the most consistent I have used.

Happy you figured it out. I had my bets on the PSU after reading the first few posts.
I always build my own (and families) PC's - I skimped on a PSU once and never again.

When you say it was rated for 850W - at what efficiency? Bronze/Silver/Gold?

I usually oversize the PSU a little - not had another one fail since the cheap one.
Currently using a Corsair Pro 750W Gold. 6 years and still going strong.