Investigating Reports Of Intel Skylake CPUs Damaged By CPU Coolers

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.


If your friend transports his build and he's too lazy to remove the heatsink, well... It's his own damn fault.
 
I see some people defending intel here, but this is not the first time they've cut corners that have affected performance and reliability. It wasn't that long ago that people experienced issues because intel cut corners on the thermal material.

Stop defending intel and let them know that the lowered production costs aren't worth the damage to their reputation.
 
Wonder how the boutique computer vendors are going to handle this so far as shipping the completed unit. They surely cant require the customer to install their own cooler.
 
Noticed that the "bend" part was on the upper side of the socket and CPU so it means that the weight is not the problem but the over tightening of screws.
 
Interesting. Why did they make the material so thin? Is that partially why skylake gets incredibly low idle temps?

The substrate was designed to be thinner due to the general market trend of thinner mobile systems. Since consumers and advertising stresses the thinness of new mobile devices like Apple's Air products or Microsoft's Surface tablets, the ability to make a smaller CPU package (i.e. die on substrate) is seen as a desirable product goal.

While this wouldn't necessarily affect socketed CPUs directly, it is reasonable for Intel (or others) to use the same substrate specifications for their soldered CPUs found in embedded and mobile devices, as for their socketed desktop and server CPUs.
 
I have gotten the cooler master hyper 212 EVO myself getting ready to build my gaming machine with I 6600k. It is heavy I haven't put it together yet. I had a shipment from newegg and the chip box was partially crushed with motherboard,,memory chip, fans and 2 hard drives it wasn't packed correctly and the ups shipped the box and dented on two corners.
 
Well I have a skylake 6600k and evo cooler master, I never try to make so tight just the enough, well anyway too bad for pepople that have that problem.
 


sounds good
 
got a response from CM...
" Yes the V6 GT is compatible with Skylake and actually uses the same installation kit as the V8. Overall, though just don't over tighten the screws for the cooler and everything will be fine. "
words of the day for me
[OVER TIGHTEN]
 


so no torque rating is been used to tightening the screws with 2 fingers on small screwdriver 1/4 turn more after it stop?
 
idk, when I installed mine, when it stopped I didnt bother...
also my heatsink still turns when you put enough force to twist it.
I'm so damn scared of checking it now even though I know for myself it was not that tight.
 


sound scary enough so checking what other people say
 
The silicon substrate doesnt have to be thinner . Intel are just cutting it thinner o shave costs .
If the socket and heat spreader were solid then the substrate could not bend . The Heat sink is clamped to the board over the socket remember , and not to the actual cpu its self .
 
I think more of this has to do with installation and clamping force. Someone feel free to point out any errors but when you mate two flat components (cpu ihs and cooler base) to one another, they act as one. The cooler base is flat/solid/rigid so won't press into the cpu's ihs anywhere.

It could potentially act as a hinge if any of the four corners aren't as tight as the others (tight up to rated force, non crushing). The substrate shouldn't flex if everything is installed properly, considering the back plate of larger air coolers pressing from behind, the base of the cooler pressing from the front, effectively sandwiching the cpu and its substrate between two solid surfaces (including motherboard pcb and socket area).

Place 3 wooden boards together and clamp them snug then try to bend/flex the center board. Due to the 2 outer components being rigid it doesn't happen. What 'can' happen however is the center board could be smashed if the two outer components are clamped too tightly.

If I were to place a literal piece of tissue paper (ridiculously thin/flimsy) between two 1/8" metal plates and clamped them, the tissue paper isn't going anywhere - it will remain flatly pressed between the two plates assuming the plates were tightened evenly.

I would agree that heavy/bulky air coolers can place stress in shipping but this has typically been a problem with the pcb of the motherboard flexing, not bending/breaking the cpu. The force from a larger air cooler with backplate is transferred to the motherboard pcb, not the smaller cpu substrate clamped between them. This is why it sounds more like mounting force or uneven mounting pressures (2 screws tighter than the others).
 
 
I see what you're saying. On the other hand, the pins are offering considerable resistance if being compressed under the rated 50lbs of static load. 50lbs is a lot of weight, even large air coolers are only around 2-3lbs. Naturally that will change with dynamic loads and clamping force will increase the pressure but still, 50lbs is 50lbs of force.

Having a given rating of 50lbs for the cpu/substrate would be pointless if the socket/pins and any void space was so weak it could only handle 10lbs. Then again it would point to a weakness in the motherboard/socket rather than the cpu itself. Cpu cooler bases are often the same size or larger than cpu ihs plates so there shouldn't be a concentrated force in the center of the chip but rather evenly and around the supported edges. The backplate will only strengthen the whole setup.
 


From the pics it looks like the cpu substrate is only supported properly in the corners [which are being bent by the pressure] . That suggests the pins have a minimal effect . Though I have never had a really good I think they are spring mounted to push up as the socket is closed , That "spring"is obviously only just enough to guarantee contact
 

The size of the CPU cooler base and socket are irrelevant: the whole purpose of mounting force on the heatsink is to press the heatsink's contact area against the IHS and the CPU core. The IHS is hollow, held slightly off the substrate by adhesive, and the bulk of mounting force gets applied to the CPU die through the IHS, over the hollow area of the socket where the CPU's on-substrate bypass capacitors are.

You also have some heatsinks with slightly convex contact area which would further focus contact force directly in the middle of the IHS.

 
The substrate IS thinner due to less layers needed for skylake, but its pretty flexible. It would take a hell of a hit to crack it. No way simple mounting force would do the job. The IHS is actually thicker to keep mounting height the same so just the very edges are weaker, but overall this chip is stronger than broadwell. I will say that with the IHS removed (delidded) you can put enough pressure to bow the chip and make the outter pins lose contact.. That shows up as memory errors or just not recognizing memory, but I've never seen even that with a virgin chip, and I mount as high as 70lbs.
 
After seeing how a 600g -1000g+ cooler could bend the motherboard, I have never purchased any of those coolers. To mount these cooler to the cpu socket, they all require a lot of worrisome pressure and stress. It's not unthinkable that some of these heavy coolers can damage the cpus.
 

Yeah, he knows jack shit about computers, and the only way it was built was with me doing it. (Not complaining, it had a 980 Ti) I'll have to warn him though.
 

50lbs is the specified maximum. There will be a margin on top of that for manufacturing variance, another margin for momentary dynamic loads, and a safety margin on top of that.

If you take your PC and drop it from one foot high, as it would likely do a few times during shipping, a CPU with a large and heavy tower HSF will likely see peaks exceeding 200lbF.
 
Status
Not open for further replies.