Sign in with
Sign up | Sign in
Your question

New machine shuts down under full load

Tags:
Last response: in CPUs
Share
September 19, 2012 9:39:12 PM

Hello friends, I've built a new machine and it shuts down unexpectedly under full load. I have replaced the pwr supply, ram, SSD, and now the mobo.

I have tried re-seating the cpu.

components being assembled:

ram: CORSAIR Vengeance 16GB (4 x 4GB) 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Desktop Memory Model CMZ16GX3M4A1600C9B

cpu: Intel Core i7-3930K Sandy Bridge-E 3.2GHz (3.8GHz Turbo) LGA 2011 130W Six-Core Desktop Processor BX80619i73930K

mobo: Asus P9X79 Pro

pwr supply: Corsair Professional Series Gold 750-Watt 80 Plus Gold

Graphics Card: EVGA GeForce GTX 670 4GB GDDR5 PCI Express 3.0

SSD: OCZ Vertex 3 120gb

I've tried swapping out everything but the CPU. The machines completely turns off under full CPU load.

Can anyone offer insights into why this is happening? Thanks very much
a b à CPUs
September 19, 2012 10:30:07 PM

the 3930k doesn't come with a cooler, it needs a 3rd party cooler, what are you using?
m
0
l
September 19, 2012 10:35:02 PM

sorry, I thought maybe I was leaving something out.

it's a Corsair H100.

Thanks
m
0
l
Related resources
a b à CPUs
September 19, 2012 11:29:19 PM

So tell us, what are your CPU temps. under full load with liquid cooling.
m
0
l
a c 78 à CPUs
September 19, 2012 11:30:08 PM

Have you checked your power source? (The wall outlet or UPS)
m
0
l
September 23, 2012 3:08:39 AM

Thank you for your replies.

I cannot tell you what the CPU temp is right now at full load because the machine is no longer posting into BIOS.

The only component I haven't swapped out is the CPU.

Could a broken CPU prevent the machine from posting into BIOS?
m
0
l
September 23, 2012 3:33:01 AM

You should double check your power,that stuff uses a lot. trust me
explanation:at idle PC components don't work that hard right?
when at load it works overtime which needs more electricity,
get me?check all the wattage.
m
0
l
a c 117 à CPUs
September 23, 2012 3:51:06 AM

MF8461 said:
You should double check your power,that stuff uses a lot. trust me

Since his PC no longer POSTs, full-load power isn't the problem... at least not anymore. Could be the PSU, motherboard or CPU. He'll have to start the check/test list all over again since symptoms changed.
m
0
l
September 23, 2012 10:33:59 PM

I just swapped in a brand-new CPU. The problem is the same. It doesn't post into BIOS. I swapped in a smaller graphics card, removed all but one ram stick, disconnected any peripherals I could, and it still doesn't post into BIOS.

I am at a loss.
m
0
l
a b à CPUs
September 23, 2012 11:54:23 PM

its got to be ur mobo
m
0
l
September 24, 2012 6:17:47 PM

Thank you for your reply.

I would like to ask, how can you really know if it's the mobo? I've put three in so far, as well as so many other components. How could my luck be so bad that I have 3 bad mobo in a row? That seems a bit supernatural to me.

I don't mean to doubt your opinion, it's just that I'm shell shocked with this machine now and I'm questioning everything.
m
0
l
a c 78 à CPUs
September 24, 2012 9:59:13 PM

Have you checked your power source? (The wall outlet or UPS) If you have replaced all computer components, then the only common denominator would be the wall outlet/UPS.
m
0
l
September 24, 2012 10:53:06 PM

I don't think the wall outlet is an issue there on my repair bench because I was experiencing issues with this machine at the work location using a UPS.

But I probably should have a UPS there at the repair bench too now that you mention it. Thanks for the advice.

I'm still stuck with a dead machine here :( 
m
0
l
a c 78 à CPUs
September 24, 2012 11:02:17 PM

KenTanker0us said:
I don't think the wall outlet is an issue there on my repair bench because I was experiencing issues with this machine at the work location using a UPS.

But I probably should have a UPS there at the repair bench too now that you mention it. Thanks for the advice.

I'm still stuck with a dead machine here :( 


You probably have a situation where a bad component is damaging good ones.

If that's the case, the normal, systematic pluck-n-chuck process won't work for troubleshooting. For example, if you have a bad motherboard that's killing CPU's then the good CPU will die when you try it and then fail in a good motherboard etc. (because it's already damaged even though it's "new")


If I were trouble shooting a system like this at work, I would use a working system as a sacrificial lamb. Instead of trying to find the component that causes the bad system to work, I would look for the component (from the bad system) that causes the good system to fail. You can't stop at the first component you find either, because there is probably more than one now. (although I hope only one is assassinating other parts)

This is an expensive way to troubleshoot and I only use it as a last resort, even when I'm spending company money. (Though I have only had to do this a few times in >20 years of fixing electronics.)



m
0
l
September 25, 2012 1:03:49 AM

Thank you for your reply, Z1NONLY. What you are describing, a bad component damaging good ones, is my greatest fear with this build.

I am not able to provide a sacrificial lamb for this troubleshoot. If this turns out to be the case with my machine, I think I would have to purchase a new mobo ram and cpu to be sure I wasn't ruining any new components I install.

What really gets me is that there seems to be no way to isolate what component is bad without ruining other components in this 'sacrificial lamb' machine. Is there any way to determine or guess which component is the culprit? I'm just grasping at straws here.

*Question: is it possible that the graphics card could be the culprit? And also, if the graphics card is not the culprit, could it be getting damaged by the culprit? This is a $500 graphics card.

Thanks again for any advice.
m
0
l
a c 78 à CPUs
September 25, 2012 1:17:13 AM

At this point, the suspect list consists of every component that connects to another.

I would try using a cheap video card for trouble shooting. If you have another rig, you can try the card in the working rig, before and after trying it in the bad one.

If it works before but not after being used in the bad rig, I would suspect the motherboard or PSU as the "assassin". Then you are only out ~$40 for the test video card.

This should pose only minimal risk to the good rig.
m
0
l
a c 117 à CPUs
September 25, 2012 5:25:19 AM

Z1NONLY said:
If it works before but not after being used in the bad rig, I would suspect the motherboard or PSU as the "assassin". Then you are only out ~$40 for the test video card.

Since the PCIe lanes are wired directly to the CPU, there is also a possibility that the CPU if busted could kill the GPU.

Everything connected directly to 2-3 other components makes this a frustrating game of Russian roulette.

Since OP has already swapped out every component at least once, the likelihood that he got components that developed psychopathic behavior the same way as other components did multiple times in a row should be quite small.
m
0
l
September 26, 2012 1:39:45 PM

This really worries me. I hope I haven't broken every single component. This is very difficult to troubleshoot because of my ridiculous work schedule at present but I will keep at it.

This forum is very helpful and I will continue to post my results. Any and all help is greatly appreciated.
m
0
l
September 27, 2012 5:27:34 PM

Hey good news! I've got 10 days off while they prepare materials for me at the job. So: I've got some time to work with this problem machine.

The first thing I did was to grab an older machine from my render farm and use it for my sacrificial lamb.

I wanted first to see if either of the GPUs I had used in the problem machine were ruined. In order to do this I had to swap out the s. lamb machine's PSU because it doesn't have the extra power connections for the graphics cards (geforce GTX460 and GTX 670 4GB). So I grabbed the original Corsair AX750 from the problem machine and plugged everything in.

The good news so far is that the machine boots and I have display with both cards. This would indicate that the cards are undamaged wouldn't it?

Also really, really interesting: I noticed that the fan isn't spinning in the AX750 PSU. Does this mean what I think it does? What do you all recommend at this point?

As always, thank you for all your advice and help on this wonderful forum.
m
0
l
a c 78 à CPUs
September 27, 2012 6:01:23 PM

I have never checked mine, but I read that some high-end PSU's will turn the fan off when the load/temps are good. So it may be normal. You could try applying a load and see what happens with the PSU fan.

And so far it would seem that the components you tested in the working rig are, at the very least, not immediately causing the symptoms you see in the bad unit.

I would continue to troble shoot using only one suspect item at a time.
m
0
l
September 27, 2012 7:59:14 PM

I am told that the AX750 fan only speeds up under certain load. So I'm thinking it's ok. I will continue with more tests. Thanks!
m
0
l
October 4, 2012 4:27:18 PM

I have sent back the CPU to Intel for warranty replacement. I have replaced the mobo. I will buy some different ram, probably GSKILL Ripjaws, per recommendation on another forum. Wish me luck!
m
0
l
October 10, 2012 6:28:01 PM

okay I have a new mobo, a new CPU and some higher quality ram on the way, so I'm starting fresh!

Here's a (possibly stupid) question: can a bad SSD assasinate other components? I'm using the original because it passed all the tests the manufacturer told me to run.

And I'm a bit gun-shy about assembling this new build. Should I do it out of the case first? That might be difficult with the Corsair H100 for cooling...but I'm interested in any and all advice that will prevent me from having to eat another mobo.

Thanks very much for any advice.
m
0
l
!