Skylake CPUs may be damaged by pressure from some heatsinks

PCGamesHardware.de is reporting (Google translation here) that Intel Skylake CPUs may be damaged due to pressure from some third-party heatsinks. Some users have reported bent CPUs and even bent socket pins. The site says the cause for this problem may be the thinner substrate used in Skylake CPUs.

Courtesy of Modders Inc.

Several manufacturers have issued statements about the problem. Scythe says that coolers using its HPMS mounting system (which include the Mugen 4 and Mugen Max) may damage CPUs if the system suffers strong physical shock. The firm is offering customers a free set of replacement screws. EK Water Blocks' current-generation PreciseMount system should prevent damage, but the company does not recommend that modders use its older water blocks in Skylake systems.

Silverstone states that its current lineup of CPU coolers is designed to meet Intel's mechanical specifications, but is nevertheless checking to see if its older models are safe to use. The company claims it has yet to hear about any problems, though. A similar sentiment is echoed by Thermaltake, which says that most of its CPU coolers include a security system that prevents users from overtightening the screws. The story is pretty much the same with Thermalright, Arctic Cooling, EK Water Blocks, and Noctua, which all state their coolers are safe to use.

It's worth mentioning that nearly all the companies recommend that users remove heavy CPU heatsinks before transporting their systems in order to prevent damage due to shock and vibration.

Comments closed
    • south side sammy
    • 5 years ago

    something Intel has been needing for decades……… a “redesigned” heat sink and latching mechanism.

    • oldog
    • 5 years ago

    And… this is why I keep coming back to the Tech Report. This is very useful info!

    • albundy
    • 5 years ago

    heh, i’ll sign up for that class-action form!

    • MetricT
    • 5 years ago

    This would matter less if Intel would include a usable heatsink with their CPU’s. I couldn’t get either of my OEM Haswell heatsinks to actually hold onto the motherboard, they would both pop off after the slightest bump. So I *had* to use third-party heatsinks.

    Intel should make backplates with threaded mounts mandatory, and should include a heatsink capable of actually staying on and keeping the CPU from thermally throttling during a Prime95 run. Also, please use solder for your IHS. Doing all this at Intel’s scale couldn’t cost more than a dollar.

    • ozzuneoj
    • 5 years ago

    Strange timing. I tried to swap out the fan on an old P3 1Ghz EB CPU a few days ago and the system stopped posting. Turns out the heatsink rocking side to side a tiny bit cracked a fraction of a mm off one corner of the CPU die. Dead instantly. Replaced it with a p3 933 I had on hand but it was still sad, as that was the best CPU the board would support and I liked that…

    Who would have thought, the first working CPU I’d crack the die on would be 15 years old. I’m just glad it didn’t happen when they were $800…

      • drfish
      • 5 years ago

      2004, Athlon XP 2500+ (Barton), first and only CPU I ever broke (took a chip out of the corner installing the stock HSF). Poor little guy never even got to power up. :'(

        • USAFTW
        • 5 years ago

        I also remember a 9600 GT I had back in the day got chipped in the corner, but it kept working. Happened when I was under the impression that by good will and blowing enough times it would run cooler. Ahh, those push-pin heatsink mounts, still refusing to die…

    • JMccovery
    • 5 years ago

    So, we go from crushed dies to bent CPUs/socket pins?

    Reminds me of when I modified a Volcano 5 to work with my de-lidded K6-2 350… Poor chip.

    • TruthSerum
    • 5 years ago

    [url<]https://en.wikipedia.org/wiki/Strained_silicon[/url<] Isn't new technology wonderful?

      • NeelyCam
      • 5 years ago

      Maybe putting a third-party cooler on your Skylake makes the transitors faster, for some extra overclocking headroom?

        • TruthSerum
        • 5 years ago

        I just pipe my excess heat to dev/null. Transistors don’t know it’s not bacon.

          • Wirko
          • 5 years ago

          One day it’ll all come back to you.

    • jihadjoe
    • 5 years ago

    They should do like the auto industry and release torque specs for those heatsinks. A lot of damage is done because beginners just tighten stuff as far as it will go, and of course not having any guidelines to follow just makes it worse.

    “Apply heatsink, tighten bolts to 6 ft-lbs.”

      • TruthSerum
      • 5 years ago

      How many tiny inch-minute torquing phillips #2 head screwdrivers do you have?

        • bthylafh
        • 5 years ago

        [url<]https://en.wikipedia.org/wiki/Torque_screwdriver[/url<]

      • bthylafh
      • 5 years ago

      *twitch*

      Newton-meters. Might as well use proper units if you’re going to start something new.

        • NeelyCam
        • 5 years ago

        Don’t you know that american units of measure are superior?

        I find it silly when elitist Europeans use kW for cars instead of HP. It’s not a friggin’ solar panel!

          • bthylafh
          • 5 years ago

          Foot-pounds. You heathens probably aren’t even circumcised.

            • TruthSerum
            • 5 years ago

            Circumcised with a torque wrench? That’s hard to do.

            • bthylafh
            • 5 years ago

            With that attitude.

            • jihadjoe
            • 5 years ago

            Metric debate = R&P for geeks lol

          • bhtooefr
          • 5 years ago

          Fun fact: There’s several different currently used definitions of horsepower.

          Mechanical/imperial horsepower: 550 ft-lbs per second, or 745.69987158227 W (this is what US car power figures are given in)
          Metric horsepower: 75 kgf-m per second, or 735.49875 W (this is actually, rather often, what European car power figures are given in)
          Electrical horsepower: 746 W exactly (this appears to be along the “look, we’ve [i<]actually[/i<] metricated, by defining our customary units in terms of metric units!" lines) Boiler horsepower: 33,475 BTU/hour, or 9812.5 W (apparently this was because early boilers needed to burn that much for a steam engine to produce one mechanical horsepower) There's only one, as far as I know, definition of kilowatt in use. (And, I find it annoying when European cars use metric horsepower, but state it as "hp", not PS or CV or whatever.)

    • Srsly_Bro
    • 5 years ago

    I think the main issue is people think that if the HSF is tightened to the highest degree it’ll perform better with each turn of the screw/knob.

      • TruthSerum
      • 5 years ago

      Technically true! Until you reach “fastener equilibrium” as I call it.

      • willmore
      • 5 years ago

      Well, I don’t know about your heatsink, but *mine* goes to 11. That’s one more. So, it’s better.

    • Bensam123
    • 5 years ago

    Seriously, no Intel Bendgate yet? This was a freebie… XD

    This is a pretty substantial deal. Processors don’t have a lot of tolerance when it comes to functioning after physical abuse…

      • xeridea
      • 5 years ago

      It would be Crushgate

      • rika13
      • 5 years ago

      You mean Skygate.

        • Bensam123
        • 5 years ago

        lol…

        • The Dark One
        • 5 years ago

        No that’s too obvious…. Uhhh, let’s say “Bendlake.”

          • willmore
          • 5 years ago

          No, it’s “SkyFail”.

    • Chrispy_
    • 5 years ago

    I’ve seen motherboards fail because third-party heatsinks apply so much pressure that even the backplate is not enough to stop the whole setup bending

    Skylake being less strong means that rather than killing your $150 board you’re now killing your $250 chip instead. Same old problem as always, just a different victim this time (though, arguably, your wallet is always the real victim).

      • Klimax
      • 5 years ago

      And then Skylake-E comes… (Also original victim would be 300-400USD board.)

    • f0d
    • 5 years ago

    another cost cutting maneuver from intel 🙁

    i dont mind them trying to save a few bucks but when it affects their products like this i just shake my head

    when they started putting subpar thermal goop under the headspreaders (instead of soldering like they do with xeon and sandy bridge and previous cpus) i was hoping they might actually reverse their decision eventually but oh no they come out with this crappy idea of a thinner substrate

    you make the best cpu’s intel but i just dont like how you are cutting costs like this

      • chuckula
      • 5 years ago

      [quote<]another cost cutting maneuver from intel :([/quote<] It's not cost cutting. In fact, it probably costs intel more money. The thinner PCB does have beneficial electrical characteristics for Skylake, however.

        • f0d
        • 5 years ago

        dont really care honestly

        its a stupid thing to do just like how they put low quality goop under the heatspreader instead of soldering like in sandy bridge

        and for downthumbing me im gonna up thumb you
        take that you malevolent man you 😀

          • chuckula
          • 5 years ago

          “its a stupid thing to do”

          Improving the electrical characteristics of your chip is not stupid.

          The fact that a relatively small number of after-market overweight heatsinks might damage some sensitive electronics when screws are overtorqued is unfortunate, but it isn’t Intel’s fault either.

            • kuttan
            • 5 years ago

            “Improving the electrical characteristics”

            Where did you get the definition ?? Its another clear cost cutting measure from Intel and this time it affects Chips durability as well. [b<] Even if thin PCB can improve electrical characteristics of the chip still chips durability is supreme. [/b<] Its pathetic you blindly defends Intel.

            • chuckula
            • 5 years ago

            Yeah, so when TR reviewed the chip how come they were able to run benchmarks?

            I mean, according to you these chips spontaneously disintegrate when exposed to air, so obviously it was completely impossible for TR to get a working sample installed in a motherboard.

            How come there haven’t been tens of millions of returned chips?

            Oh wait, it’s because in the real-world this isn’t a real-problem and you’re a two-bit loser who probably doesn’t know which end of a soldering iron gets hot, but the Internet has given you delusions of god-hood.

            • f0d
            • 5 years ago

            wow way to go overboard in no way was anything like
            [quote<]I mean, according to you these chips spontaneously disintegrate when exposed to air, so obviously it was completely impossible for TR to get a working sample installed in a motherboard. [/quote<] ever said there is no need to blindly follow any company we just think its **maybe** an issue because there were zero bending substrate issues previously with intel cpus but there is now they are thinner i personally worry because some of the massive heatsinks i install for people have really high pressure mounting systems i would also like to know what improvements to the "electrical characteristics" were needed for the substrate to be thinner that the cpu wouldnt function if it was thicker like previous cpu's were where is the article explaining why this was needed or the chip wouldnt function? [quote<] you're a two-bit loser who probably doesn't know which end of a soldering iron gets hot, but the Internet has given you delusions of god-hood[/quote<] classy abusing someone ALWAYS gets your point across successfully - clearly you win congrats.!

            • kuttan
            • 5 years ago

            You provided Zero facts to counter what I said. Your little ego get wounded is what I can see here. [b<] Gold Intel TROLL ?? [/b<]

            • DoomGuy64
            • 5 years ago

            Yup. I wonder how he’d react if this was AMD having the problem.

            • chuckula
            • 5 years ago

            Well, when AMD intentionally released drivers that caused GPUs… that are more expensive than any Skylake chip.. to overheat and in some well-documented cases permanently fry…. I didn’t say anything to be nice to AMD.

            And that’s official software officially released by AMD, these are an extremely limited number of third-party products that are being installed by teenagers who can’t spell torque, much less understand what it means.

            However, I can chance that stance to address losers like you.

            P.S. –> Mark Papermaster has officially stated that Zen will be “competitive” with those Skylake parts when Zen finally launches in 2017. So even AMD’s chief disinformation officer is spinning Zen downward now.

            • DoomGuy64
            • 5 years ago

            Guess that answers my question. This is exactly how you’d react if it was AMD. Intel gets a pass, and/or damage control.

            You know, it doesn’t bother me that either company will occasionally screw something up. What bothers me is the double standard. Not only is it immoral, but the FUD appears untrustworthy and gets ignored after so long.

            As for the AMD driver thing, which isn’t a CPU issue, I can see why you wouldn’t want to say anything. AMD officially allows their chips to hit 95c, which causes the GPU to throttle. Most people would have noticed that, and it shouldn’t have caused mass failures. The card(s) that did fail were apparently not built to spec, and the user(s) either didn’t notice, or ignored the throttling. The failure(s) were also not well documented, as AFAIK it was [i<]one guy[/i<] on Reddit. No pics, we're just taking his word for it, which I'm still willing to give him the benefit of a doubt. It happens, but I would have noticed it. I've also hit 95c on my GPU running benchmarks. It didn't kill the card, but it does cause throttling and lower framerates. MSI uses horrible thermal paste which must have stopped working after a few months, so I solved it by applying some Antec formula 7. Much better temps now.

            • TruthSerum
            • 5 years ago

            “it’s because in the real-world this isn’t a real-problem and you’re a two-bit loser”

            It’s a real problem. “how big” is a relative description and all losers are two bit, so yeah.

          • Chrispy_
          • 5 years ago

          ermahgerd 100 edits! You win the round-number-of-post-edits prize. Did you script that?

            • Redocbew
            • 5 years ago

            I’m going to say that was a re-transmission from a phone or tablet or something that got stuck somewhere in the middle.

            • bjm
            • 5 years ago

            No, he clicked “Submit reply” and his heatsink fell off his CPU at the same time. Yes, yes.. that’s what happened.

            • f0d
            • 5 years ago

            nah i had a few jd’s and coke and was bored during a tv ad break and wanted to screw around and do something stupid

    • Krogoth
    • 5 years ago

    I blame retention mechanism for some oversized HSF units that out there *cough* Thermaltake *cough*.

    At least, we no longer have to worry unable cracking dies unless you only of those extreme overclockers who delidded their chip.

    • chuckula
    • 5 years ago

    David Bowie and Freddy Mercury have the answer: [url<]https://www.youtube.com/watch?v=Uuzmub5PXNQ[/url<]

      • K-L-Waster
      • 5 years ago

      Downvoting due to earworm…

        • morphine
        • 5 years ago

        Oh, you liked that riff? Let me [url=https://www.youtube.com/watch?v=rog8ou-ZepE<]ruin it for you.[/url<]

          • thanatos355
          • 5 years ago

          Oh, you like ruining [url=http://www.dailymotion.com/video/xcru59_jedward-under-pressure_music<]things?[/url<]

            • Bensam123
            • 5 years ago

            And it all goes back to the source…

            [url<]https://www.youtube.com/watch?v=GUl9_5kK9ts[/url<] [url<]https://www.youtube.com/watch?v=au3-hk-pXsM[/url<] [url<]https://www.youtube.com/watch?v=ykwqXuMPsoc[/url<] [url<]https://www.youtube.com/watch?v=EIyixC9NsLI[/url<] I've never heard anything more addicting. Pretty sure he was on LSD while making these. Pretty sure most of those have 10 hour loops as well.

        • nanoflower
        • 5 years ago

        That’s not an ear worm. This is an earworm. [url<]https://www.youtube.com/watch?v=gvJrCc8li6s[/url<]

          • TruthSerum
          • 5 years ago

          I’m content to NEVER, EVER play that, ever. EVER.

          • NeelyCam
          • 5 years ago

          OMG. Just got the Presequel for PS3 – can’t wait for that holiday entertainment.

          • Redocbew
          • 5 years ago

          I have no idea what that was, but I never want to hear it again.

          • biffzinker
          • 5 years ago

          Try this link instead: [url=https://www.youtube.com/watch?v=8vpi2ZH6WPA<]Borderlands: The Pre-Sequel Song Mainframe[/url<]

            • biffzinker
            • 5 years ago

            Last one: [url=https://www.youtube.com/watch?v=Keo9k1PESec&list=RD8vpi2ZH6WPA&index=4<]It's Clappie[/url<]

            • nanoflower
            • 5 years ago

            Thought about going with that one but I figured actually having the ear worms on the screen might get the point across more than just a good song. 😉

Pin It on Pinterest

Share This