Computer shuts off/restarts, overheating @65C?

[Grim23]

I recently upgraded my old PC, and gave it some more oomph when compared to my old rig of 4 years. Here is my Build:

Silverstone Ft02 case (the one with 3 180mm fans that blow up from the bottom)
Seasonic 850 watt power supply (850X model i think)
Asus P8Z77-V board

Intel core i7-3770 (not the K version)
with a Noctua NH-C14 heatsink with OCZ Ultra 5+ thermal compound

Two Nvdia GTX460s (from old build)
Mushkin (996995 kit number) 8gb ram
Intel 520 120GB SSD

I was playing borderlands and it reset on me, then I made a groaning noise. But I realized I had the door closed, and it was getting warm in my room. I opened the door and nothing happened since. I was playing Tribes Ascend and it did it again, although I continued to play for 4 hours and it never happened again.

So I decided to run prime 95....it crashed

So I had Core Temp up and I checked the max temp in the logs and it was only 65C.

Checked people who were overclocking their 3770k models and they were running at like 80-90C, so I was wondering is my CPU overheating? I don't think it is, So what is causing this problem? I'm at a loss here.

Currently running HCI memtest, at 8000% coverage with no errors.

 

[nna2]

65C isn't overheating, is it a blue screen?

 

[Grim23]

I didn't think so either, it isn't BSOD, it just restarts. I see the usual screens pop up and then windows loads.

 

[nna2]

strange, im guessing it was a clean install of windows and all that?

 

[Grim23]

Yea fresh install, updated the bios to the latest, all new drivers from the Asus website Nvidia etc, got the SSD toolbox.

 

[arslan94]

It seems like you must reapply the THERMAL PASTE on your processor ..... secondly check that the Heatsink is properly adjusted in the right place ..... Check whether the CPU FAN is working alright ..... Cuz it seems an Overheating problem ..... BEST OF LUCK !!

 

[nna2]

what are you talking about ? 65C is well inside the thermal limits, they can run up past 100C before shutting off

and sorry, hopefully you can find help

 

[Grim23]

I forgot to mention that's what I did BEFORE I ran the prime 95 test. I may try the stock heatsink, who knows maybe I don't know what I am doing, Although I followed OCZ instructions on the whole "grain of rice" application method.

I may try the stock heat sink and see what happens tomorrow.





****EDIT*** I used coretemp for monitoring because it didn't look like realtemp was up to date for ivy bridge processors. Can anyone shed light on this?

 

[nna2]

the heat sync won't make a difference, it's running within it's thermal boundaries

 

[smorizio]

it might not be the cpu or the cpu temp. it could be a bad cap in the power supply or on the older power supply they have the older caps that were not the higher rated 105c caps. the power supply fan itself may not be working and it over heating. there a bad device on the mb under load it over heating and rebooting. i would use open hardware monitor and watch your chipset temps and power supply voltage.

 

[HEXiT]

playing borderlands 2? its not out yet Borderlands 2 will be released on September 18th, 2012 in North America and September 21st, 2012 internationally
my guess is if its the beta your playing then thats the problem not your system.
65'c is well within temp range for that cpu although with that cooler and not being a k version i would expect the temps to be around 5-10'c cooler than that...

remove the old paste and re-apply it in a thiner layer chances are you used to much the first time.

 

[Grim23]

Is "open hardware monitor" a program? I am assuming it is the one in the BIOS? What voltages should I look at specifically?

It is a brand new seasonic supply, I hear they are supposed to be excellent. I may try my older (but hassle free) 850w Antec power supply tomorrow depending if I can get the temps down my remounting the heat sink.

__________________________________________________________________________________________

I may not have applied enough, because my mindset when going to apply the paste is "don't apply too much" So I just tried a different method of application, credit card, (worked well on my last build). And I read a review on Guru 3d where they guy spread out the paste and got excellent temperatures.

http://www.guru3d.com/article/noctua-nhc14-cpu-cooler-review/5

I meant Borderlands, but I must have borderlands 2 on the brain.

 

[smorizio]

it a free program to downlod the open hardware is free to use. the us another that free for 30 days and has more features (hardware monitor) it read the temps and fan speeds and your ps voltage.

 

[InvalidError]

I may try my older (but hassle free) 850w Antec power supply tomorrow depending if I can get the temps down my remounting the heat sink.

65C under load for Ivy Bridge is perfectly within expectations and nowhere near the CPU's 105C thermal trip limit. It is unlikely that the HSF itself is the problem so if it "fixes" it, I would be more inclined to think you may have a bad connection somewhere that got wiggled into contact - at least for the time being.

In general, the PSU is the first thing I would suspect for spontaneous reboots without crash screen.

 

[Grim23]

65C under load for Ivy Bridge is perfectly within expectations and nowhere near the CPU's 105C thermal trip limit. It is unlikely that the HSF itself is the problem so if it "fixes" it, I would be more inclined to think you may have a bad connection somewhere that got wiggled into contact - at least for the time being.

In general, the PSU is the first thing I would suspect for spontaneous reboots without crash screen.

Bad connection somewhere? I guess you are referring to the actual power connections into the mobo?

Also everyone keeps referring to the 105C trip limit (in other posts as well), but where does this info come from? I can't find this on Intel's website. Anyone have the link?

I am going to bed now, but I will try the power supply switch in the morning.

_____________________________________________________________________________

Did you know with a P8Z77 mobo and the seasonic 850watt powersupply you cant connect 2 460 GPUs, they require four 6 pin PCIe power cords. And the way seasonic sent their 8 pin mobo power cord, you can't plug them all in unless you modify the cable. By modify I mean cut off the stupid empy 4 pin platic part.

Now I wouldn't have to do this if there wasn't a contact right beside my mobo connector.

See pictures.

I called seasonic and they said they did this on purpose. Why I can't fathom. Why they made their PCIe power ports able to support a CPU cable when they specifically put a CPU cable port on their power supply anyway. They said "The PCI ports double as a CPU port".


Empty plastic pin
http://postimage.org/image/h9012ivtj/f6ffeb7c/

Board compatibility issue ( the R20 'thing' would prevent this plug from going in if it still had the 4 empty slots on it)
http://postimage.org/image/h8swjym01/ae8c37a1/

A picture illustrating the problem with the 8 pin connector. you can thee the retention clip would prevent it from going into the CPU slot on one end if two PCIe plugs were already taken up.
http://postimage.org/image/hit7lvv7f/a298f7b2/

Finally a pic of the CPU plug using a PCI slot, mind you its useless for me. Silly design, I even called them to confirm you could plug it in this way.
http://postimage.org/image/5qx5xqduf/dca1456d/

Stupid modular design that allows only one end of the 8 pin power cable to fit into the power supply, in the designated CPU spot, but if you turn it around you can put it in the PCIe/CPU spot. I mean why rate this power supply for SLI, and then give me some stupid 8 pin connector with 4 empty pins that wouldn't be able to go into the PCI slot anyways, because they are taken by the SLI!

Well hopefully somone understands what I said there. It has taken me a while to type this up/take pictures, so I am heading to bed now, and I will try stuff tomorrow.

 

[swifty_morgan]

in the BIOS........ there's a section where you can set max temp at which machine will shut down. look for it. ........ also those processors are nothing like the prior. something about the thermal liquid or paste used under the cap...... proc. may be defective.

 

[InvalidError]

also those processors are nothing like the prior. something about the thermal liquid or paste used under the cap...... proc. may be defective.

If there was a problem with the thermal paste between the CPU and IHS, the OP's temperatures would be 90-105C even at low/moderate load, not 65C. 65C full-load for Ivy Bridge well within normal range, albeit a on the high side for stock-clock with NH-D14.

 

[InvalidError]

Bad connection somewhere? I guess you are referring to the actual power connections into the mobo?

Also everyone keeps referring to the 105C trip limit (in other posts as well), but where does this info come from? I can't find this on Intel's website. Anyone have the link?

By "bad connection" I mean bad via in the motherboard, bad solder joint, contaminant in socket contact points, conductive debris bridging a signal somewhere, etc.

As for the 105C TJmax, it is in Intel's Ivy Bridge thermal design specifications. The 67C in the marketing documents is TCase.

TJmax is the temperature within the CPU cores themselves while TCase is the IHS temperature.

 

[swifty_morgan]

If there was a problem with the thermal paste between the CPU and IHS, the OP's temperatures would be 90-105C even at low/moderate load, not 65C. 65C full-load for Ivy Bridge well within normal range, albeit a on the high side for stock-clock with NH-D14.

i was kinda grasping.......... the temp setting in the BIOS would be something to look at though........ this guy come back ?

 

[Grim23]

It isn't a NH-D14 its NH-C14, either way its still a monster heatsink.

SOrry for the delay I just woke up, as soon as I get my old 850 watt power suplly hooked up to my mobo (had to take it out last night to reinstall the heatsink), I will take a look at that motherboard setting.


As for a "bad connection" I just RMAd a stupid ASUS board because it wouldn't let me into the BIOS. If I have to RMA it again.......All I can say if it is the Mobo, ASUS is one crappy company becuase this would make my second RMA for a motherboard. Anyone know any other good board manufacturers, that they use, instead of Asus? Last time I used evga, and never had a problem.

 

[InvalidError]

If I have to RMA it again.......All I can say if it is the Mobo, ASUS is one crappy company becuase this would make my second RMA for a motherboard.

I have had Asus boards in all my PCs since my P90 and have yet to have any failure unless you count my P233MMX which is unable to boot Windows NT/2000/XP due to a BIOS glitch.

 

[Grim23]

Well I took my computer in to the only people I trust to give me honest service (over the last 15 years) and we will see what they come up with. Will post a conclusion when I find out.

 

[swifty_morgan]

uh oh, geek squad........ LOL

 

[Grim23]

No a small local business who gave me good advice on my computer that I first built, a shop with a couple local computer enthusiasts. I would never go anywhere else. And they won't try and completely gouge me for "repairs".

 

[Grim23]

Well here was the problem:

I had enabled XMP in my bios so the 1600 Mhz ram I bought would operate at its rated frequency, for some reason this was making my processor shut off. After XMP was disabled and my ram was back at 1333 Mhz everything was fine, and they ran stress tests for 24+ hours.