New Build random restarts

sundodger5

Hi all, have a new build doing random restarts when gaming.
I have checked event viewer and it shows an event id 41, Kernel power error.
It is always when gaming.
Temps look ok with processor at 42 degrees and GPU at 51 degrees.
Reading on line shows many " solutions" from ram to power...
Have ran a memtest no errors.
System is s Pentium G4560 on a gigabyte B250M DS3H board, 8 gig crucial ballistix, radeon hd7870 graphics and a corsair vs 550 power supply.
Anyone troubleshoot similar issues before? If so what logical steps did you take to isolate the problem?
Thanks

ED E

Memtest: Done (though the windows one is a short test)
Prime 95: Do next - If crashes, its CPU/RAM
Furmark: Do if P95 passes, if crashes, GPU.

You definitely connected the PCIE power to the 7870 yeah?

moonlighting

if its a new build i would personally rebuild lol. check for any bad connections. sometimes people miss the cpu power connection and gpu.
next check if your power supply is enough. some websites can do maths calculations. 550w sounds on the low side for the gpu
drivers - make sure all up to date. have had issues with gpu drivers in the past.
cpu\gpu temps keep an eye on while gaming etc. thermal paste may need to be redone.

sundodger5

Update..all power connections are in place. Unplugged and plugged back in just in case.
Reseated Ram.
Updated graphics drivers.
Ran Prime95 on blend for 30 mins no problems max remp 50 degrees
Furmark for 30 mins also no crash....it reckoned my gpu temp never moved over 32....it lied.
No crash either way.
Fire up fallout 4.....restart after 10 mins.
Has me stumped to be honest. A lot of articles point towards power supply any program to log voltages? Have speedfan but it doesnt seem to see the voltages on my board.

Fysh

I'd be leaning towards low wattage on the PSU tbh, but the easiest way to confirm that would be to run it on a higher-rated PSU and see if the problem is reproducible. Which is rather tricky if you don't already have an e.g. 800W PSU kicking around...

ED E

Fysh wrote: »

I'd be leaning towards low wattage on the PSU tbh, but the easiest way to confirm that would be to run it on a higher-rated PSU and see if the problem is reproducible. Which is rather tricky if you don't already have an e.g. 800W PSU kicking around...

Sounds that way doesnt it, but I've run a 7870 and an OC'd t'feck Q6600 on CX430 no problem so 550 should be plenty.

So my moneys on a dud rather than insufficient.

moonlighting

worth checking the hard drive as-well. is it an ssd. do a full check on it. you could try a different ssd/hd and do a new install of windows.
on a different note i recently seen a mother board give same hassle and it was issue with usb3 drivers. uninstalling em solved it.
solving issues like thies can be difficult sometimes.

sundodger5

No spare power supply or at least not one with a higher wattage to hand. Would invest if i knew for sure that was it.
Sata drive that checks ok. It had a clean win 10 install when i built the rig so only 4 days old.
Should i just bite the bullet and replace it?

TerrorFirmer

There's no way it's a wattage issue. That build would literally pull about 250w under heavy, peak load.

The PSU may be faulty, but it's not an inherent wattage issue, suggestions of 800W power supplies are actually utterly ludicrous...any high quality unit of 350w+ is more than enough for that build with ample headroom.

Are we also sure it's not the card or drivers, the HD7870 is quite old and it's odd that the problem is not replicated in either Prime95 or Furmark, both of which really torture the CPU and GPU respectively.

moonlighting

try removing the graphics card.(it just may not be playing ball with the motherboard) then use the video output from the motherboad. see how stable the system is. this will rule out a lot of problems.
if it still gives hassle i would remove the cpu and check the connections on it.

sundodger5

So i finally got a fix. New power supply sorred it all out.
I had been monitoring the voltages, temps etc via saphire trixx.
Nothing showed as a consistent for the shutdowns.
The only thing i do notice having done the same with the new power supply is the 12v varies a lot less with the new supply so i assume it was a stability issue under load.

ED E

Thanks for posting the solution (many never return)