computer shutting down due to bad heatsink, damaged cpu?

Hi all, we have a user who upgraded a machine to a new core I7 processor and upon booting up the machine was shutting itself off within a minute. It just sat at the BIOS screen, no OS on the machine yet. We took a look and the heatsink was missing two of the plastic connectors that fasten the oem Intel heatsink to the motherboard. Perfect excuse to upgrade that cheap oem heatsink, however, since the machine was shutting down is it likely the CPU overheated and was damaged?  

Unclear if the heatsink was damaged from the factory or if the user damaged it installing it. It was supposedly never able to stay on for more than a minute, usually shutting down after a few seconds.

Thanks.
JsmplyAsked:
Who is Participating?

[Product update] Infrastructure Analysis Tool is now available with Business Accounts.Learn More

x
I wear a lot of hats...

"The solutions and answers provided on Experts Exchange have been extremely helpful to me over the last few years. I wear a lot of hats - Developer, Database Administrator, Help Desk, etc., so I know a lot of things but not a lot about one thing. Experts Exchange gives me answers from people who do know a lot about one thing, in a easy to use platform." -Todd S.

dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
>>  since the machine was shutting down is it likely the CPU overheated and was damaged

The Intel CPUs are supposed to shut down if the heat is too great and this seems to have happened.

However what damage, if any, has been done can't be determined.  How many times did this happen?  How long was the machine switched off before the go button was pressed again?  All of these may play into if the CPU was damaged.

Test with a new heatsink, properly fastened, thermal compound properly applied.  If it works then yeah.
0

Experts Exchange Solution brought to you by

Your issues matter to us.

Facing a tech roadblock? Get the help and guidance you need from experienced professionals who care. Ask your question anytime, anywhere, with no hassle.

Start your 7-day free trial
JsmplyAuthor Commented:
Thats the plan. Going to pick up a good cooler master for them this week. We do have some spare core I3 oem heatsinks handy, are they interchangeable for a quick fix?  The I3 heat sink actually looks a bit bigger.

The user claims it was tried to boot a few times. They did claim to try to turn it right back on a few times before realizing there must be a problem. Also, when you say if it works then yeah, do you mean if it works no harm done?
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
>>  Also, when you say if it works then yeah, do you mean if it works no harm done?

Really mean if it starts, then good, yeah it starts.  Now cross fingers and let it keep starting and working properly for the next five years.

Can't comment on the heat sink bit as I'm more into old Athlon 462 boards.
0
Make Network Traffic Fast and Furious with SD-WAN

Software-defined WAN (SD-WAN) is a technology that determines the most effective way to route traffic to and from datacenter sites. Register for the webinar today to learn how your business can benefit from SD-WAN!

JsmplyAuthor Commented:
Actually thinking further, don't some systems shut down if the heat sink is not making proper contact?  Because it was pretty quick to shut down and was not that hot to the touch.
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
Not sure about that.  Far as I know the Intel has a thermocouple or similar device within it to monitor temperatures.

Dell's have a circuit based on the spring retaining clip to monitor their northbridge sinks to make sure they are fastened correctly.
0
JsmplyAuthor Commented:
Machins is a dell
0
sihtCommented:
The temps you need to worry about for the i7 9xx series are:

From Intel's Processor Spec Finder
 
All Core i7 9xx variants:

Vcore Max 1.375v
Tcase Max (CPU temperature) 68c
Tjunction (Core temperature) 73c
Tjunction Max (applies to overtemp protection such as Throttle and Shutdown) 100c

So long as the CPU is at or below Tcase Max and Tjunction everything is OK. If it hits Tjunction Max then is is over temp and will throttle/shutdown to protect the CPU.

If the i5 Cooler fits the board you can install it with some thermal compound and try to boot the machine, keeping a close eye on the temps in the BIOS. If it all looks OK in the BIOS let it boot to the OS and monitor the temps with a free tool such as RealTemp. If you really want to be sure it's OK you can stress test it with Prime95.

If your i5 cooler is keeping the CPU at or below Tcase Max and Tjunction then it's doing it's job and the machine is safe to use. Hopefully the overheating has not done any damage to the CPU and it will be stable.

If you want more info on CPU temps and testing have a look here:

http://www.tomshardware.com/forum/256908-28-what-maximum-temperature-stock-core

HTH
Simon

0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
Dells may have something like this on the Northbridge http://www.experts-exchange.com/Hardware/Components/Fans_Cooling/Q_23480831.html

but not aware of anything on the main CPU.
0
_Commented:
>> spare core I3 oem heatsinks handy, are they interchangeable for a quick fix?

Not usually. The last I saw, the i7 (socket 1366) and the i3/i5 (socket 1156) use slightly different mount holes.
I have seen a couple of post around where somebody did it. But it's not a real good idea.
0
_Commented:
...but more importantly the mounting pins which affix the heatsink to the motherboard are spaced further apart. For socket 1366 motherboards the holes are spaced 80mm apart, for socket 775 motherboard that distance was 72mm. Socket 1156 motherboards take the middle ground with a spacing of 75mm.

http://forum.pcstats.com/showthread.php?69128-Intel-Socket-1366-Socket-1156-and-Socket-775-Heatsink-Compatibility-Issues
0
JsmplyAuthor Commented:
Coral, this board looks like it can be ordered with an i3, i5, or i7 processor (from Dell).  Seems to be a proprietary HSF, cracked open an i5 and an i7 (of this model PC) and both seem to be using the same Dell HSF.
0
_Commented:
Interesting. But then again, Dell does like to do their own thing.
Also it appears that some i7s are being made for the 1156
http://www.tomshardware.com/forum/263067-28-core-socket-1156

If you can get one of the i3 HSFs to fit tightly, it might be worth a try. Just be aware of the hole spacing, and don't force it.
0
JsmplyAuthor Commented:
Yep, did some searching online.  It appears to be a custom HSF.  Compared it with several OEM Intel heatsinks from i3, i5, and i7's and none match the Dell equivalent.  Seems like the Dell HSF will need to be re-used as the motherboard has a back-plate behind it setup for the Dell HSF.  It's either that or disassemble the whole thing.  We can re-use the Dell, just clean the Dell HSF and reapply AS5 sound like a plan?

Alsp, in regards to the original problem, that might have been the issue all along.  It looks like our user was trying to use the Intel HSF that doesn't fit on the Dell board just right.  The question is, was the machine turning off because it was overheating or because it didn't detect the HSF?  Would you think the i7 would get hot enough to overheat just by being at the bios setup for a minute or so?  The user said it was not overly hot to the touch (the ZIF bar, etc) after shutting it down to investigate.  

If we can get it to run, will CPU-Z or a similiar tool show the MAX temp the CPU has ever hit (similiar to SMART HDD data) to determine if it was truly overheated?  
0
nobusCommented:
imo  no software exists for showing that - unless you have a monitoring sqoft iinstalled with logging features.
check also the motherboard around the cpu to see if it is discolored from high temp; if so - better replace it too
0
sihtCommented:
It looks like our user was trying to use the Intel HSF that doesn't fit on the Dell board just right.

So they have removed the Dell HSF and replaced it with a stock Intel one? This person owes you lunch at the very least.

will CPU-Z or a similiar tool show the MAX temp the CPU has ever hit

If it was already installed on the machine there may have been a logging function, I'm not sure. It's unlikely.

We can re-use the Dell, just clean the Dell HSF and reapply AS5 sound like a plan?

Good plan. Keep an eye on the temps.

was the machine turning off because it was overheating or because it didn't detect the HSF?  Would you think the i7 would get hot enough to overheat just by being at the bios setup for a minute or so?

i7's can run hot, depends on the batch a bit. Processors generally heat up and cool down fairly quickly with cooling installed. One with no HSF will get hot quick. You're not going to know untill you try it with proper cooling. If it's already damaged you cannot hurt it any more by doing so.
0
JsmplyAuthor Commented:
Hi Siht,

Thanks.  Yes, generally lunch would be expected, but in this case it's being setup for charity (and the user was a volunteer) so hey, what are you going to do.  No worries, we will get them setup.

We will know tomorrow when we can try to fire it up with proper cooling.  If it works are we good to go, or could it work with a decreased life, etc?

Thx
0
sihtCommented:
If it works are we good to go, or could it work with a decreased life, etc?

There's no easy way to tell. Enthusiasts who are into overclocking use benchmarking tools like Prime95 to test stability. Prime95 is designed to search for undiscovered prime numbers and involves a massive amount of calculations. The theory is that if a machine can run Prime95 for 24 hours without errors it can be considered truly stable.

http://www.overclockers.com/forums/showthread.php?t=335813

As it has gotten hot I would do some kind of stability testing even if it's just putting the machine through some heavy use testing before you deploy it. As for decreased life there's not really any way to tell. If it boots and passes some stability testing I would consider it a win and deploy the machine. If you're at all concerned that you are supplying the client a dud part then you will need to replace the CPU.
0
JsmplyAuthor Commented:
Thanks, will run Prime95 for 24 hours to be safe.  It's going to be deployed "in-house" so to speak, but will be used for important tasks so don't want to deal with a machine that's going to die out prematurely.  Thanks!
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
If you do use Prime95 then monitor the temperatures.  I recommend Speedfan http://www.almico.com/speedfan.php for that purpose.  I wouldn't do 24 hours though.

See Prime 95 http://files.extremeoverclocking.com/file.php?f=103 and http://www.mersenne.org/freesoft/ for Prime95.  15 minutes should be enough (but keep an eye on those temperatures).
0
sihtCommented:
Just be aware that things other than CPU can cause Prime95 to fail. The link in my last post is a pretty definitive how to on stress testing PC's.

Good Luck.
0
_Commented:
>> ...was the machine turning off because it was overheating or because it didn't detect the HSF?

Over heating.  It doesn't take very long for them to warm up.
I wouldn't worry about it too much. Intels thermal shutdown feature is there for just that reason. It's hard to 'cook' an Intel without really trying, these days.

Like suggested, I would test it, to make sure.
0
JsmplyAuthor Commented:
Thanks.  We assumed the Intel thermal shutdown feature would have saved it, however dbrunton in one of the early posts in this thread seemed concerned that the user started the machine right back up when it was shutting down.  Probably because it was not giving it time to cool down?  However, wouldn't the Intel thermal shutdown feature just not start up the chip then?  The user reported they tried to start the machine 3-4 times and each time it took less time for it to shut-off (presumably getting cooler and hotter pretty quick).
0
_Commented:
>> Probably because it was not giving it time to cool down?

Correct.
Although, starting it "3-4 times" is heading into the "really trying" area.    ; )
But since it was still trying to post on the last attempt, the odds are on your side.
0
sihtCommented:
Chips cool down fast too so it was probably cool enought to start but still hotter than ambient. Each time it started it was closer to the shut off temp as it had not had time to cool all the way down.

As coral47 says the thermal shutdown feature is there to stop the machine before it gets hot enough to damage the processor.

As for Prime95 24 hrs is if you really, really want to be sure. I usually test my machines overnight. Keep a close eye on it for the first 15 min to monitor the temps.
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
Probably still heat in the CPU that hadn't dispersed.  Each time it was being started up there was more heat there and hence the quicker shut down.

But again you won't know until you test.

Note my comment above about using Prime95 for 15 minutes only.  That should be more than enough time to see how it behaves or misbehaves.  You will only need 24 hour testing if this machine was going to be a gaming or server type machine.  For a normal user 15 minutes should be sufficient testing.
0
JsmplyAuthor Commented:
Actually thinking about it, is the processor overheating and intel's thermal protection kicking in or is the BIOS shutdown temperature being reached and that causing the shutdown?  I'd imagine the default BIOS settings on the Dell's are below whatever Intel's temp levels are.  Can't verify till the machine is up and running though.
0
nobusCommented:
if the bios has a separate temperature probe, it can be different; otherwise it uses the cpu in-built one
0
JsmplyAuthor Commented:
Nobus: Even if it used the CPU in-bult one, don't most Dell BIOS have a setting to shutdown BEFORE the cpu built in thermal shutdown point?

Either way, got the users problem corrected.  Cleaned off the heatsink and processor and applied the Dell heatsink with a fresh thermal compound application.  The machine idled at bios for about 25 minutes with no problem (although the Dell BIOS doesn't seem to show temperature information . . . but it didn't shut-off).

Going to try to install an OS now and run Prime95 on it.  Will report back.
0
JsmplyAuthor Commented:
Ran Prime95 for 30 minutes, machine ran it fine.  Max temp was 81C (spiked), usually under full load it stayed between 78C-80C.  
0
sihtCommented:
What processor is it exactly?
What were you using to measure temps?
Have a look here to be sure you are not over temp:

http://ark.intel.com/ProductCollection.aspx?familyID=28037
0
JsmplyAuthor Commented:
Intel® Core™ i7-870 Processor, used both RealTemp and Speedfan.  Didn't see a max temp listed on that link you sent, but according to online research the core i7 shuts down at 100C and with non-upgraded HSF it's relatively normal to hit 75-80C under load.  That's mostly from hardware forums, overclock forums, etc.

30 minutes of blended tests on Prime95 stayed around 78-80C
0
JsmplyAuthor Commented:
*30 minutes of the blended test, didn't mean to insinuate the test was stopped, that was non-stop for 30 minutes.
0
_Commented:
Sounds good.
I would still kind of keep an eye on its temp for awhile. For an overclock or stress-test, those temps aren't bad.
But for a standard rig, when they want to run above 70C all the time, I get nervous.
0
sihtCommented:
http://ark.intel.com/Product.aspx?id=41315

Tcase = 72.7c

You're right though, 75 - 80c is not unusual. I got my i7-950 to run a lot cooler at stock by lowering the cpu Vcore voltage in the BIOS.

Sounds like the CPU is OK though,that's a win!
0
JsmplyAuthor Commented:
Just let the Prime95 run for another 42 minutes on the test they claim generates the most heat.  It still ran between 77C-80C (never saw it go above 80C this time, last time it peaked at 81C for just a second).

After stopping the test, the Cores (1-4) dropped to 40 C almost immediately and a minute or two later they were down to 30C which seems to be "idle" temperature with just Windows loaded (this is a fresh OS load so really no background software . . . loaded the OS and went straight to prime95 and speedfan).

Is it safe to assume the CPU survived and is stable for production?  Should we leave it running 24 hours to be safe?

0
JsmplyAuthor Commented:
(It meaning Prime95)
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
I wouldn't bother.  

It held good for 42 minutes under maximum load.  In real life it won't be exposed to the type of stress you've just given it.
0
_Commented:
Dealers choice, but I wouldn't bother with it. Let it rest overnight, and make sure it boot from cold tomorrow.
Then "load it and launch it".   : )
0
_Commented:
I hate it when he does that.   ; p
0
sihtCommented:
Sounds like it's probably fine. Next job!
0
JsmplyAuthor Commented:
Thanks for sticking with it.  Going to close the question.  If either of you care to reply, because its a Dell mobo it appears that mounting an aftermarket heatsink would be a pain (Mobo removal, etc).  Would a more powerful case fan help with CPU temps at all?  If not, it seems 78-80C for the Core i7 under Prime95 is not "bad" but thought it might help.
0
JsmplyAuthor Commented:
Thx all!
0
_Commented:
Thank you much.  : )

>> ...more powerful case fan help

Maybe. But I would wait and see what it averages at with a normal load, first.
0
sihtCommented:
See how it goes. Most office desktops hardly do any work, the CPU's are just idling most of the time. If it gets too high too often you could think about it but you probably won't need to.
0
JsmplyAuthor Commented:
Thanks.  Was more curious if the case fan would have much effect on the CPU or if that's really only effected by the HSF or CPU fan.  For what it's worth, the machine will be a file server at times and run some VMs, to it will get some excercise.
0
sihtCommented:
OK, I wouldn't bother unless you are having heat issues. If you are then a case fan will not have anywhere near as much effect as a better HSF will.

As others have said Prime95 works the CPU as hard as it can be worked, even gaming won't work it as hard, so if your temps are acceptable under Prime95 then they will probably be fine with anything else you can throw at it.
0
dbruntonQuid, Me Anxius Sum?  Illegitimi non carborundum.Commented:
Check the case.  Sometimes they have provision for an inlet fan and an exit fan.

If no one cares about the noise that fans make (the computer might be in a server room) then a more powerful inlet and exit fan could help keep temperatures down.  Rerouting wires can help as well.  Converting PATA cables to SATA with suitable adapters can also help with reducing cable clutter.  The last is something I'm doing on my home machines.
0
_Commented:
It definitely has some effect. How much, varies.
I still wouldn't worry unless it just stays in the high 60C+.
Might be worth the effort to knock a couple of degrees off the top then.
0
It's more than this solution.Get answers and train to solve all your tech problems - anytime, anywhere.Try it for free Edge Out The Competitionfor your dream job with proven skills and certifications.Get started today Stand Outas the employee with proven skills.Start learning today for free Move Your Career Forwardwith certification training in the latest technologies.Start your trial today
Components

From novice to tech pro — start learning today.