GA-990FX-Gaming power off

[KiwiShepherd]

This one has me baffled. This is a new build, 2 weeks old. The first 1 - 3 times I turn it on in an evening it will power off, like it's been switched off, after a minute or two. Sometimes before windows has loaded. It had been showing code 40, or maybe 4D, the last couple of days. It currently shows AA, and seems to be running fine. Once it gets going it runs very well, 3D gaming doesn't seem to faze it at all. Running optimised defaults in Bios. Any help or suggestion would be appreciated.

GA-990FX-Gaming
AMD FX 9590
G.Skill DDR3 2133 8gig x 4 (quad channel kit)
Gigabyte GV-R939G1 Gaming-8GD
Cooler Master V-series 1000 Gold 1000w PSU
Cooler Master Nepton 240M
Windows 10 Home
Sony TV via HDMI for monitor

[AgentFXA]

My board shows similar problems IF I have cores unlocked in the sys bios - try it!

[KiwiShepherd]

Thanks for the tip AgentFXA. I'm not sure what setting in the bios you are referring to though. I took a quick look in what's called M.I.T. in my bios. It has all the voltage and clock speed controls but I couldn't see any setting called core lock. Could you clarify more please?

[KiwiShepherd]

In addition I am now almost certain it has something to do with LCD readout on the motherboard showing code 40. Tonight had the same issue. I had read the first reply and tried to access the bios to check for the setting mentioned. I was too slow on the del key and windows loaded. I checked the code on the motherboard and it showed 40. 30 seconds later, power off. It now shows AA and everything seems to be fine. I've checked the manual and can't find any reference to this code.

[AgentFXA]

Hello,
sorry it's called CPU Unlock - disable it if enabled ..see screenshot..
I assume you got that too?

http://pasteboard.co/1Ke5EuXs.jpg

There is very little info about your board available

[KiwiShepherd]

Thanks for the screenshot. Seems the Bios is almost identical to yours. The CPU lock was disabled.

[Tiger]

Running at Optimized Defaults.

It appears you don't know how to set the memory options in the BIOS. The motherboard has XMP settings but you are using memory with the wrong speed.

I don't know what memory you are using, but as an example lets look at  F3-17000CL9Q-16GBSR
http://www.gskill.com/en/product/f3-17000cl9q-16gbsr

With optimized defaults this memory will run at
 SPD Speed 1600MHz 
 SPD Voltage 1.50v
 
FX CPUs support this memory:
1866MHz speed with one DIMM per channel
1600MHz speed with two DIMMs per channel
Higher frequencies require overclocking

You might want to post a link to your memory or the part number.

[KiwiShepherd]

Thanks for the reply Tiger. The memory modules (and the supplier) can be found here: http://www.newegg.com/global/nz/Product/Product.aspx?Item=N82E16820231756

It runs as you said, 1600mhz at 1.5 volts. I am running optimised defaults until I can get this problem solved. I had planned on a mild over clock how ever after doing some reading about the CPU (better late than never right?) I may not bother.

I had more trouble than ever this morning. I was checking the bios settings and saw that the cpu temp was fairly high. I may have used far too much heat grease.
I removed the heat sink, cleaned off all the old grease and applied a dot a little smaller than a pea. Removed all the RAM except the stick in slot one. Removed and double checked the gfx card seating. Also tested the voltage from the PSU with a multi meter.

Results:

• 3.36 volts
• 5.05 volts
• 12.23 volts
• -12.44 volts

The only thing that concerned me was the CPU 2 x 4 pin connector. It tested -12.23 volts. I'm not sure if this is a problem or not. After putting it all back together again it started with no power offs and I am making this post with it.

I'll start adding the ram one stick at a time and post later with results.

[AgentFXA]

What exactly is the CPU temperature under load?
Is all bios settings on Auto timing? Avoid manual settings until all is stable.

[KiwiShepherd]

Yeah all timings a voltages are auto. Cautiously optimistic at this point. By "temp fairly high" I was referring to the Bios CPU temp, which I know is not very accurate indicator of actual core temp. 80 degrees C did have me a bit worried though.

After you asked about load temps I got a core temp monitor (Core Temp v1.0 RC9) and stress tester (HeavyLoad) and ran CPU stress test for 15 minutes with Core Temp running along side it.
Pretty happy with the results in the picture.

 

[AgentFXA]

.. 80 degrees C did have me a bit worried though.

I read elsewhere that 80C is *shutdown* temp for this CPU - reading your screenshot shows 46C as max temp and this is fine.

[KiwiShepherd]

Since removing the old grease and re-seating the heat sink I've had no more problems (touch wood). I took a screen shot of the stress test results and put it in my last post. Only ran it for 15 minutes but the max core temp was 46° C. Last time I was in the Bios i noted the CPU temp at 23° C. I've got 3 of the 4 ram sticks back in now and still no issues.

I plan to power it off for 30 minutes or so to let it cool off and add the last ram stick. I won't know if I have completely resolved it til I do another completely cold boot, since that's where the issue seemed to be coming up.

Wish me luck.

[KiwiShepherd]

I am now 99% sure this is a heat issue. I just tried booting from cold with all 4 ram sticks in and the thing was hot at idle. I loaded this page to post about it and it pushed it over 80° C and it shut down.

[KiwiShepherd]

The pipes coming out of the pump and heat sink are so close ram slot 4 that there is barely a gap at all.

• 1st picture with 4 sticks in: system idle at 50 ~ 70° C.
• 2nd picture with 2 sticks in: Typing this post with 2 tabs open and an explorer window 0 ~ 6° C

Both boots I opened Core Temp as fast as I could when windows loaded. 1st boot maxed at 78°. 2nd boot maxed 47°.

Is it normal to get heat spikes that big on boot? I'm wondering if the pump isn't kicking in as fast as it should maybe.
I may try turning the heat sink upside down to get the pipes away from the memory slots. The logo will be upside down but short of buying a new radiator kit I'm stuck.

[KiwiShepherd]

Solved

I turned the heat sink upside down, replaced all the ram and maybe most importantly, connected the pump to the CPU fan header. Stress tested for 10 minutes and maxed out at 49°C.

Thanks for all your help AgentFXA.

Tiger, if you think I have a legitimate concern regarding ram timings please reply at your earliest convenience.

Pix of stress test results and re-attached heat sink for anyone who's interested.