iFixit dumps dust into Apple’s latest MacBook Pro keyboard

iFixit recently revealed that Apple's latest MacBook Pros include a new protective membrane under their key caps, and now the site has stress-tested the membrane it found to see whether the silicone prophylactic can keep dust out of Apple's delicate-as-its-name-implies butterfly key switch mechanism.

The teardown team first dumped particles of a “powdered paint additive” that is apparently coarser than sand into the keyboard and concluded that the membrane does actually herd the dust to the edges of the switch area, away from the vulnerable holes where the key cap clips into the switch.

Image: iFixit

The keycap-and-membrane assembly seems to form a fairly robust seal, although iFixit stress-tested the system with even more dust and found that with “aggressive typing,” particles can eventually make their way into the mechanism itself. The site didn't comment on whether the amount of dust it used to achieve this infiltration is representative of what a normal notebook keyboard might see in its lifetime.

The new key switch condom isn't perfect, though. After its initial dust dump, iFixit moved on to finer sand and found that a few grains of the beach builder can still infiltrate the mechanism and prevent the switch from operating.

For folks who might need to get inside the keyboard to blow out dust, there's good news: Apple is now using thinner key caps that are easier to remove than those on past butterfly-equipped MacBook Pros. Still, a total keyboard replacement will require technicians to replace the entire top case of the notebook itself. Hopefully this silicone membrane can stave off that repair for a long time.

Comments closed
    • FireGryphon
    • 1 year ago

    Basically the problem that butterfly keyboards introduced is still a problem, but you’re less likely to experience it. No thanks. I was *so* close to buying a MBP, but I’m not willing to gamble $700 per random repair.

      • DancinJack
      • 1 year ago

      Applecare(+), bruh

        • derFunkenstein
        • 1 year ago

        Even then that’s nearly $300 and only good for 3 years. I mean, I got it on mine because it’ll cover a drop or two, but I can see people balking at that specifically for coverage against a design flaw.

    • DancinJack
    • 1 year ago

    So, so stupid. I’d like to update my MBP but Apple isn’t making it easy.

      • chuckula
      • 1 year ago

      This is just part of the [url=https://en.wikipedia.org/wiki/Golden_Path_(Dune)<]Golden Path[/url<] to the ultimate 2020 goal.

    • uni-mitation
    • 1 year ago

    [quote<] The new key switch condom isn't perfect, though. After its initial dust dump, iFixit moved on to finer sand and found that a few grains of the beach builder can still infiltrate the mechanism and prevent the switch from operating. For folks who might need to get inside the keyboard to blow out dust, there's good news: Apple is now using thinner key caps that are easier to remove than those on past butterfly-equipped MacBook Pros. Still, a total keyboard replacement will require technicians to replace the entire top case of the notebook itself. Hopefully this silicone membrane can stave off that repair for a long time. [/quote<] See guys? The problems were not that big of a deal! Just snort the coke on a hooker's ass. Problem solved! uni-mitation

      • DancinJack
      • 1 year ago

      This is definitely an appropriate comment.

        • uni-mitation
        • 1 year ago

        So, do I get the good housekeeping seal of approval? That’s what I want to know!

        uni-mitation

    • chuckula
    • 1 year ago

    From the TR Twitter:
    [quote<]I don't understand this PR response, though. [b<][i<]Even for as dusty as it is[/i<][/b<], Cinebench provides meaningful deltas in performance for CPUs even out to 18C/36T and beyond. The idea that it's not accurate for recent x86 chips seems silly given that 2016 SKL cores still power MBPs today[/quote<] Apple clearly commissioned a Cinebench update as part of its keyboard protection initiative.

      • dragontamer5788
      • 1 year ago

      It seems like Cinebench has issues at 32C / 64T or so. But for anything less than 20C, it seems to be a good benchmark still.

      I think the overall technology-press needs to start thinking about a new multithreaded benchmark for the near future. But Cinebench is still good and workable. Its just the next-gen 28c (from Intel) or 32c (from AMD) madness is going to start seeing scaling issues.

        • Klimax
        • 1 year ago

        Iray (CPU mode) can fill that. Luxray or similar can work as well. (Just One needs to be careful with OpenCL drivers) Blender’s render can work too.

        Also cryptocurrency mining (there are several CPU only algos – Magicoin (includes 64-bit float computations and large precision computation using GMP/MPIR), Verium (each thrad uses several GB of RAM)

        • Sahrin
        • 1 year ago

        That and begin testing for multiapp-multithread workloads which is how much users are using machines now.

          • Redocbew
          • 1 year ago

          I still don’t understand why some people think that will show different results than the so called average stress test. Of course different workloads will have different behavior, but running multiple applications is not “a workload” that has it’s own characteristic behavior. It doesn’t really change anything, and it’s kind of impossible to run a test like this for a general case, no?

    • tipoo
    • 1 year ago

    Apparently the i9 can throttle down to 800MHz…Thanks for the credit card thick heat pipe, Jony. Yeesh.

    [url<]https://twitter.com/barefeats/status/1020167541329547265[/url<] [url<]https://9to5mac.com/2018/07/18/how-macbook-pro-throttles-with-final-cut-pro-x/[/url<] Edit: Here's a third source finding thermal throttling beyond last years chips or the i7, in the i9 [url<]https://www.youtube.com/watch?v=M1VWl8zZEsU[/url<] Double Edit: Now four sources and more attesting to 800MHz dips...There's giving the benefit of the doubt, but if it walks and talks like a duck, why are we doubting this phenomenon so much? [url<]https://old.reddit.com/r/apple/comments/90aqly/for_those_wondering_about_the_26_ghz_15_macbook/[/url<] (I see John Poole's update there, but it seems off base - the Barefeats twitter test wouldn't have been hitting encoder and going to idle, and Jeff B's test wouldn't have had the 800MHz dips go away in the freezer if that was the case) Last edit, maybe...Someone seemingly proved my theory below that the 800MHz dips may have been VRM throttling, because they didn't always occur at peak CPU temperature, and adjusting this wattage made them go away. Only after a few days of people trolling anyone suggesting the 800MHz dips were real... [url<]https://www.reddit.com/r/macbookpro/comments/91256u/optimal_cpu_tuning_settings_for_i9_mbp_to_stop/[/url<]

      • chuckula
      • 1 year ago

      That article you linked has this update:

      [quote<]John notes that at 800 MHz, the CPU is idling, awaiting further instructions. This 800 MHz dip that we see on the graphs is normal in the sense that 800 MHz is an idle frequency.[/quote<] So there is throttling, but no, the chips don't throttle down to 800 MHz under load. It's still definitely an issue with an obsession over thin/quiet preventing Apple from doing its job in providing an adequate cooling solution.

        • tipoo
        • 1 year ago

        I addressed that in my comment…Barefeats test wouldn’t have gone to the encoder and created a CPU idle period, and that too was forced down to 800MHz, and Bens original test wouldn’t have had those eliminated in the fridge.

          • chuckula
          • 1 year ago

          [quote<]Bens original test wouldn't have had those eliminated in the fridge.[/quote<] That's not necessarily true when executing a complex piece of software. The idle periods experiences in the throttling tests could be precisely caused by the earlier throttling producing an idle state later on in the process of running the software. If you look at the temperature graph carefully you'll see that there are prolonged periods of higher temperatures where the notebook did not throttle down to 800 MHz so if it automatically drops down to 800 MHz based just on temperature you would have seen it do so much more often and earlier in that run.

            • tipoo
            • 1 year ago

            Barefeats test was looped playback, so it would neither have idle periods nor move work to Quicksync –

            [url<]https://twitter.com/barefeats/status/1020167541329547265[/url<] I see lots of downvoting, I don't see any countering of this fact.

            • tipoo
            • 1 year ago

            >If you look at the temperature graph carefully you’ll see that there are prolonged periods of higher temperatures where the notebook did not throttle down to 800 MHz

            Throttling all the way down to minimum multiplier when the CPU isn’t overheating can be caused by VRM overheating. The platform sends a signal to tell the CPU to draw minimum power possible while the VRM cools down. In other words, the PL settings and temperature are telling the CPU that it’s OK to boost, but the laptop itself can’t handle it.

            • tipoo
            • 1 year ago

            Someone just proved my VRM theory too –
            [url<]https://www.reddit.com/r/macbookpro/comments/91256u/optimal_cpu_tuning_settings_for_i9_mbp_to_stop/[/url<]

        • barich
        • 1 year ago

        That doesn’t make sense.

        Why was it faster and with fewer drops to 800 MHz when only four cores were enabled?
        Why was it a lot faster and with no drops to 800 MHz in the freezer?
        What, in the middle of encoding, would cause the CPU to go idle? It can’t possibly be waiting on memory or storage. The CPU is by far the bottleneck.

        • abiprithtr
        • 1 year ago

        There is a sentence – right below that update.

        “[i<]Yet, this still has all the makings of a thermal issue. The export is faster when there are only four cores being used instead of six. And as shown, the export is obviously much faster when outside cooling is added.[/i<]" Means Apple is not able to design its laptops to remove heat quickly enough, given the hardware. Performance on tap is lower than what's on offer. This might be affecting more thin-and-light devices.

      • LocalCitizen
      • 1 year ago

      will the freezer test be part of all laptop reviews from now on?

      • tay
      • 1 year ago

      It’s idling at 800 MHz. I’d have to see the temp and load to go with the frequency to believe 800 was the throttle speed. This is just people going overboard and testing beyond their abilities thus making mistakes

        • tipoo
        • 1 year ago

        That’s in my comment, am I taking crazy pills or did no one make it past line 1 lol.

        Barefeats test would not have gone to the encoder and gone idle – that was also throttling at 800MHz. And if John Pooles encoder theory was right, the original test wouldn’t have had those 800MHz dips go away in the fridge.

          • tay
          • 1 year ago

          Yeah i read the 9to5mac tests after replying and it doesn’t make sense. Is the GPU throttling and causing the CPU to go into idle as there is no work to do?
          Going with AMD has been a dumpster fire for Apple IMO.

            • tipoo
            • 1 year ago

            Barefeats test was a video playback loop, no idle periods, and it still dove to 800MHz after two minutes

            [url<]https://twitter.com/barefeats/status/1020167541329547265[/url<]

            • tay
            • 1 year ago

            But can that be almost entirely GPU related? Let’s wait for someone more reputable than “bare feats”

            • tipoo
            • 1 year ago

            Who would that be (I mean, TR, obvs, but Barefeats is legit)? Barefeats has been around for years and does good testing, surely many TR’ers can attest to them.

            [url<]https://barefeats.com/[/url<] There's yet more people chiming in for the 800MHz dips, after Jeff Benjamin and Barefeats...If it walks like a duck... [url<]https://old.reddit.com/r/apple/comments/90aqly/for_those_wondering_about_the_26_ghz_15_macbook/[/url<]

      • tanker27
      • 1 year ago

      How is this even a concern? Since the introduction of the i series Intel chip (and probably before that, but i’m old and can’t remember) all CPUs throttle during idle unless you a) overclock and/or b) running high performance mode.

      /facepalm

        • tipoo
        • 1 year ago

        Err, this isn’t about idling, which isn’t what anyone means by throttling. It’s not during idle…The first of my links was a looped playback test. I mentioned John Pooles Idling theory, but it doesn’t appear correct.

        [url<]https://twitter.com/barefeats/status/1020167541329547265[/url<]

          • tanker27
          • 1 year ago

          Well throttling during load is entirely different from throttling during idle. If the graphs are true then throttling during load would be very concerning. If true I’m not sure who’d be at fault but it seems like an easy fix.

            • Spunjji
            • 1 year ago

            Apple is at fault, because both the cooling solution and power delivery hardware aren’t up to sustaining the chip’s standard power states – let alone the higher turbo states on the top-end i9 processor that they will happily charge you more for.

            • tipoo
            • 1 year ago

            They both are. Notebookcheck noted this Coffee Lake gen can draw more power than the previous Kaby Lake R chips, and that higher boost at the start appears to create a thermal runaway situation where it dives harder after. Apples fault is shipping it with the same thermal design as before – didn’t they test it, and if they did, did they find 800MHz dips acceptable? Dropping to base speeds is nothing new for macbooks, but this throttling past it is severe.

      • tipoo
      • 1 year ago

      Looks like the VRM power delivery theory was confirmed, now with part pictures

      [url<]https://www.macrumors.com/2018/07/24/throttling-i9-macbook-pro-power-delivery-chip/[/url<]

        • Phartindust
        • 1 year ago

        Hmm Apple is claiming that it’s a firmware issue, and they have release a patch to address it. guess we’ll see of it works.

        [url<]https://hothardware.com/news/apple-macbook-pro-intel-coffee-lake-throttling-macos-flaw[/url<]

    • TheRazorsEdge
    • 1 year ago

    Apple in a nutshell.

    1. Make a bad design decision in the name of aesthetics.

    2. Over-engineer the design until it works.

    3. Eventually it will make its way into competing products.

    It’s a decent system as long as you’re not buying the first gen of Apple tech.

      • Chrispy_
      • 1 year ago

      It looks like they’re still working on step 2. The condom helps, but it doesn’t completely remove the problem and once your keyboard starts to stick you need to pretty much ruin the laptop to get at the keyboard.

      Oh, also, the second time you replace a mac keyboard you have to toss the whole upper deck, because the first replacement keyboard required you to destroy the rivets and attach the new keyboard with self-tapping screws into the alloy shell itself.

Pin It on Pinterest

Share This