12 ms·

I don't see how you can get more cooling with a firmware change when the fans are already going at full speed. Does that mean the whole unit will now run hotte

I don't see how you can get more cooling with a firmware change when the fans are already going at full speed. Does that mean the whole unit will now run hotter before it starts to throttle? How is this going to impact some of those bad solder joints MacBooks are famous for?

coldtea 8y ago
>I don't see how you can get more cooling with a firmware change when the fans are already going at full speed It's not about what you do when the temperature if at its peaks and the fans are going at full speed, it's about how you avoid getting there. >How is this going to impact some of those bad solder joints MacBooks are famous for? When did MacBooks became famous for "bad solder joints"?
- stcredzero 8y ago
  Watch a bunch of Louis Rossman YouTube videos. He knows all about the seedy underbelly of Macbook hardware design mistakes. https://www.youtube.com/watch?v=1AcEt073Uds https://www.youtube.com/watch?v=1AcEt073Uds
  - dpkonofa 8y ago
    A few things that I take issue with here: 1) Louis Rossman has a vested interest in drumming up attention for his videos and has, on multiple occasions, used false statements and misleading practices to do so. Just like in the initial release video of the Touch Bar MBPs where he used an unshielded USB-A to USB-C cable and then complained that WiFi wasn't working. When used with a shielded cable, the issue went away and yet he never issued a correction or retraction. The issue was with the adapter he used, not the computer. Which leads directly to... 2) The issue stated in that video that required reballing was an issue with nVidia chips and affected all laptops that used that platform, both PC and Mac. To suggest that it was an engineering issue on Apple's part is disingenuous and only further reinforces the idea that he's not making these videos to be helpful but to drive traffic to his channel and repairs to his store. 3) Just look at how he used the Linus story to get the repair done at his shop. He has a pattern and, unlike what most people think, it's not altruistic. It's all to benefit him. I don't blame him but let's not pretend that he's doing this because he cares about users or the "right to repair" movement. He cares because he can mislabel and misconstrue the positions of the people against it to benefit his business. He'd love it if Apple was forced to provide him schematics and docs. He's in it for himself.
    - stcredzero 8y ago
      Sure, he's got his axe to grind. I mean, you can't watch Louis and not think he's got an axe to grind -- like maybe even literally in a makeshift workbench in his kitchen with a whetstone. But even if you watch his videos with the most cynical stance, it's still very eyebrow raising. Apple has had it's share of design 'doh! That's pretty much undeniable in gory detail -- far beyond the big headlines everyone knows about.
      - sjwright 8y ago
        The problem is his arguments are devoid of context or proportion. He is looking with a microscope—both metaphorically and literally—at tiny details on a circuit board which don't seem to be a problem for >99% of Apple's customers. Are these points of failure really any more common on Apple devices than their competitors? (There is a remarkable difference in mindset for the majority of MacBook and PC laptop owners. PC laptop owners will often just ditch their machine if it stops working. It's weird to say but I see it all the time.) Now I don't doubt that most of what Louis says is correct. But I also appreciate much of it from Apple's perspective. If everyone in the independent repair industry was at least as competent as Louis is today, Louis would have a point. But they're not, so he doesn't. Besides, Apple's intransigence creates the market which Louis takes advantage of; he shouldn't complain. Oh and I keep getting irritated every time Louis bangs on about those damn "refurbished" iPhone screen assemblies. He should know full well that the replacement glass used on these refurbished assemblies is cheap fragile non-Gorilla junk. I know multiple people who have had their iPhones independently repaired only to have the screen break again in a matter of days or weeks, under the most innocuous of circumstances. We are being scammed, and Louis defends the scammers. Otherwise it's a great channel.
        stcredzero 8y ago
        I would be very surprised if Louis specifically endorsed non Gorilla glass replacements. It's possible that he does. I just watch his repair videos. As far as that goes: don't buy them, and don't get repairs from someone who would sell you one! I know multiple people who have had their iPhones independently repaired only to have the screen break again in a matter of days or weeks, under the most innocuous of circumstances. I realize this will be unpopular, but my observation about people who walk around with perennially broken glass on their phones, is that some people can't keep nice things. If your lifestyle is that rough, either use a completely enclosing polycarbonate case, or get a more robust 2nd phone. Expecting a computer made largely of glass to stand up to roughhousing contradicts common sense.
- walrus01 8y ago
  there were a whole series of macbook pros from around 2010-2012 which had the nvidia discrete GPUs, BGA chips, come unsoldered from the motherboard on their own under normal usage. (edit: nvidia, not amd/ati radeon)
  - sixothree 8y ago
    Early 2008 MacBook Pro as well.
  - Fnoord 8y ago
    Discrete GPU on my former MBP 2010 would regularly crash. Would need to force it on the integrated, slower though less power hungry Intel GPU. Supposedly the whole Nvidia debacle is the reason why Apple hardware doesn't sport Nvidia anymore. Though you can get them to work sometimes with iGPU, its more safe to keep up with AMD. I'm not sure about FreeSync vs Gsync though, and I suppose Moonlight doesn't work on AMD.
- jwalton 8y ago
  > When did MacBooks became famous for "bad solder joints"? There has been more[1] than one[2] case where GPUs in MBPs stopped working because of bad solder joints. And ifixit has a guide about how to fix your mac by putting it in the oven[3] to reflow the solder joints. 1: https://www.macupgrades.co.uk/store/product_info.php?products_id=813 https://www.macupgrades.co.uk/store/product_info.php?product... 2: https://www.apple.com/support/macbookpro-videoissues/ https://www.apple.com/support/macbookpro-videoissues/ 3: https://ifixit.org/blog/6882/why-i-drilled-holes-in-my-macbook-pro-and-put-it-in-the-oven/ https://ifixit.org/blog/6882/why-i-drilled-holes-in-my-macbo...
  - slantyyz 8y ago
    I had one of those problematic 2011 MBPs, and reballing/reflowing/baking was only a band-aid. For many users, the issue was known to recur after implementing the fix. Even for the eventual repair order, Apple stuck in the same logic board with the same parts. I think the main difference was that those logic boards throttled the CPU to run cooler. That early 2011 Macbook Pro was the first laptop I ever had (by any manufacturer) that died. I was lucky I still had the MBP around by the time Apple issued the repair order. I was able to swap out my dead logic board and eventually unload the laptop.
    - linuxlizard 8y ago
      I had one of those problem units, too. After a lot of back and forth, Apple swapped the main board. I had no subsequent problems. Yay, Applecare.
      - slantyyz 8y ago
        My AppleCare had lapsed the time it died. The logic board was already replaced once while it was under Applecare, and maybe 3 months after the first logic board replacement, the GPU flickering issues started recurring. At that point, I didn't have the board replaced again under AppleCare because it was clear they were going to replace the logic board with one that for all intents and purposes, had the same design (?) defect.
  - 8y ago
- reaperducer 8y ago
  When did MacBooks became famous for "bad solder joints"? Have you noticed how low the bar is for "famous" these days? Whether it's some FaceTube "star" or tech "scandal" du jour, everything is massive these days and everyone on the planet is affected. Somehow, there are no small issues anymore.
  - bobwaycott 8y ago
    Hey, now. Don't go robbing the masses of their Two Minutes Hate.
    - dang 8y ago
      Please don't do this here.
      - bobwaycott 8y ago
        Wow. I'm pretty sure this is the first time in 8 years I've commented anything that's been flagged. I'm not entirely sure what "this" is. Is it saying "don't go robbing the masses"? Is it agreeing with the parent's comment by means of a literary reference that I thought was an approachable and analogous representation of the phenomenon the parent was describing? Seems whoever flagged this read it in a wildly different way than I intended it. Would it not be "this" if I had worded the same point differently--e.g., "You know, you're right, parent. These events totally remind me of 1984's Two Minutes Hate."? Genuinely curious what "this" is that I shouldn't do, dang, because I've spent 8 years intentionally avoiding doing it.
        dang 8y ago
        It's simple: the comment was unsubstantive and snarky. Bad combo, especially on flame-prone topics. We're looking for thoughtful comments here, as you know. The issue isn't just the quality of a comment in its own right—it's the quality of what it leads to. These things compound. Thanks for being a fine HN commenter and user!
  - always_good 8y ago
mikhailt 8y ago
The issue wasn't cooling; it was how the CPU incorrectly throttle itself that caused it to overheat again too fast. There's another thermal control unit involved and what people found was that the firmware for it was causing the CPU to drop down to 800mhz instantly when it detected the specific terminal event and when it reset that, it went back up into high-overheating load (such as the benchmark sustaining the high load) instead of staying at the right balance. In other words, it went into a loop of super low power load and then super-high load instead of the middle where the terminal balance is right for the load it is doing. Because of this low/high loop, the performance was not consistent and degraded as the overheating cause it to drop to 800mhz all the time. Where Apple likely messed up is that the firmware was not placing the CPU at the middle to avoid going into a loop, this allows the CPU to sustain a better CPU performance consistently and for an extended period of time.
wmf 8y ago
Note that bad solder joints only existed for 1-2 years and were solved long ago. AFAIK they were caused by NVidia (and MS) using untested lead-free solder balls without underfill.
masklinn 8y ago
The issue was not the CPU throttling itself, it was the CPU being configured to draw north of 100W and the VRM not being able to handle it, overheating, and leading to the system board hard throttling (down to 800MHz rather than slowly stepping down) the CPU until the VRM cooled down at which point the CPU would be un-throttled and the cycle would restart. Apparently the power configuration is the one Apple's always used but previous CPUs didn't come close to the current generation in ability to draw power, so they'd never reached the VRM's thermal limits. By lowering the CPU's maximum TDPup to 50~55W, they avoid the VRM over-heating and while the CPU does not turbo as high as with a 100W TDP allowance it can stay there just fine.

coldtea 8y ago

>I don't see how you can get more cooling with a firmware change when the fans are already going at full speed It's not about what you do when the temperature if at its peaks and the fans are going at full speed, it's about how you avoid getting there. >How is this going to impact some of those bad solder joints MacBooks are famous for? When did MacBooks became famous for "bad solder joints"?

stcredzero 8y ago
Watch a bunch of Louis Rossman YouTube videos. He knows all about the seedy underbelly of Macbook hardware design mistakes. https://www.youtube.com/watch?v=1AcEt073Uds https://www.youtube.com/watch?v=1AcEt073Uds
- dpkonofa 8y ago
  A few things that I take issue with here: 1) Louis Rossman has a vested interest in drumming up attention for his videos and has, on multiple occasions, used false statements and misleading practices to do so. Just like in the initial release video of the Touch Bar MBPs where he used an unshielded USB-A to USB-C cable and then complained that WiFi wasn't working. When used with a shielded cable, the issue went away and yet he never issued a correction or retraction. The issue was with the adapter he used, not the computer. Which leads directly to... 2) The issue stated in that video that required reballing was an issue with nVidia chips and affected all laptops that used that platform, both PC and Mac. To suggest that it was an engineering issue on Apple's part is disingenuous and only further reinforces the idea that he's not making these videos to be helpful but to drive traffic to his channel and repairs to his store. 3) Just look at how he used the Linus story to get the repair done at his shop. He has a pattern and, unlike what most people think, it's not altruistic. It's all to benefit him. I don't blame him but let's not pretend that he's doing this because he cares about users or the "right to repair" movement. He cares because he can mislabel and misconstrue the positions of the people against it to benefit his business. He'd love it if Apple was forced to provide him schematics and docs. He's in it for himself.
  - stcredzero 8y ago
    Sure, he's got his axe to grind. I mean, you can't watch Louis and not think he's got an axe to grind -- like maybe even literally in a makeshift workbench in his kitchen with a whetstone. But even if you watch his videos with the most cynical stance, it's still very eyebrow raising. Apple has had it's share of design 'doh! That's pretty much undeniable in gory detail -- far beyond the big headlines everyone knows about.
    - sjwright 8y ago
      The problem is his arguments are devoid of context or proportion. He is looking with a microscope—both metaphorically and literally—at tiny details on a circuit board which don't seem to be a problem for >99% of Apple's customers. Are these points of failure really any more common on Apple devices than their competitors? (There is a remarkable difference in mindset for the majority of MacBook and PC laptop owners. PC laptop owners will often just ditch their machine if it stops working. It's weird to say but I see it all the time.) Now I don't doubt that most of what Louis says is correct. But I also appreciate much of it from Apple's perspective. If everyone in the independent repair industry was at least as competent as Louis is today, Louis would have a point. But they're not, so he doesn't. Besides, Apple's intransigence creates the market which Louis takes advantage of; he shouldn't complain. Oh and I keep getting irritated every time Louis bangs on about those damn "refurbished" iPhone screen assemblies. He should know full well that the replacement glass used on these refurbished assemblies is cheap fragile non-Gorilla junk. I know multiple people who have had their iPhones independently repaired only to have the screen break again in a matter of days or weeks, under the most innocuous of circumstances. We are being scammed, and Louis defends the scammers. Otherwise it's a great channel.
      - stcredzero 8y ago
        I would be very surprised if Louis specifically endorsed non Gorilla glass replacements. It's possible that he does. I just watch his repair videos. As far as that goes: don't buy them, and don't get repairs from someone who would sell you one! I know multiple people who have had their iPhones independently repaired only to have the screen break again in a matter of days or weeks, under the most innocuous of circumstances. I realize this will be unpopular, but my observation about people who walk around with perennially broken glass on their phones, is that some people can't keep nice things. If your lifestyle is that rough, either use a completely enclosing polycarbonate case, or get a more robust 2nd phone. Expecting a computer made largely of glass to stand up to roughhousing contradicts common sense.
walrus01 8y ago
there were a whole series of macbook pros from around 2010-2012 which had the nvidia discrete GPUs, BGA chips, come unsoldered from the motherboard on their own under normal usage. (edit: nvidia, not amd/ati radeon)
- sixothree 8y ago
  Early 2008 MacBook Pro as well.
- Fnoord 8y ago
  Discrete GPU on my former MBP 2010 would regularly crash. Would need to force it on the integrated, slower though less power hungry Intel GPU. Supposedly the whole Nvidia debacle is the reason why Apple hardware doesn't sport Nvidia anymore. Though you can get them to work sometimes with iGPU, its more safe to keep up with AMD. I'm not sure about FreeSync vs Gsync though, and I suppose Moonlight doesn't work on AMD.
jwalton 8y ago
> When did MacBooks became famous for "bad solder joints"? There has been more[1] than one[2] case where GPUs in MBPs stopped working because of bad solder joints. And ifixit has a guide about how to fix your mac by putting it in the oven[3] to reflow the solder joints. 1: https://www.macupgrades.co.uk/store/product_info.php?products_id=813 https://www.macupgrades.co.uk/store/product_info.php?product... 2: https://www.apple.com/support/macbookpro-videoissues/ https://www.apple.com/support/macbookpro-videoissues/ 3: https://ifixit.org/blog/6882/why-i-drilled-holes-in-my-macbook-pro-and-put-it-in-the-oven/ https://ifixit.org/blog/6882/why-i-drilled-holes-in-my-macbo...
- slantyyz 8y ago
  I had one of those problematic 2011 MBPs, and reballing/reflowing/baking was only a band-aid. For many users, the issue was known to recur after implementing the fix. Even for the eventual repair order, Apple stuck in the same logic board with the same parts. I think the main difference was that those logic boards throttled the CPU to run cooler. That early 2011 Macbook Pro was the first laptop I ever had (by any manufacturer) that died. I was lucky I still had the MBP around by the time Apple issued the repair order. I was able to swap out my dead logic board and eventually unload the laptop.
  - linuxlizard 8y ago
    I had one of those problem units, too. After a lot of back and forth, Apple swapped the main board. I had no subsequent problems. Yay, Applecare.
    - slantyyz 8y ago
      My AppleCare had lapsed the time it died. The logic board was already replaced once while it was under Applecare, and maybe 3 months after the first logic board replacement, the GPU flickering issues started recurring. At that point, I didn't have the board replaced again under AppleCare because it was clear they were going to replace the logic board with one that for all intents and purposes, had the same design (?) defect.
- 8y ago
reaperducer 8y ago
When did MacBooks became famous for "bad solder joints"? Have you noticed how low the bar is for "famous" these days? Whether it's some FaceTube "star" or tech "scandal" du jour, everything is massive these days and everyone on the planet is affected. Somehow, there are no small issues anymore.
- bobwaycott 8y ago
  Hey, now. Don't go robbing the masses of their Two Minutes Hate.
  - dang 8y ago
    Please don't do this here.
    - bobwaycott 8y ago
      Wow. I'm pretty sure this is the first time in 8 years I've commented anything that's been flagged. I'm not entirely sure what "this" is. Is it saying "don't go robbing the masses"? Is it agreeing with the parent's comment by means of a literary reference that I thought was an approachable and analogous representation of the phenomenon the parent was describing? Seems whoever flagged this read it in a wildly different way than I intended it. Would it not be "this" if I had worded the same point differently--e.g., "You know, you're right, parent. These events totally remind me of 1984's Two Minutes Hate."? Genuinely curious what "this" is that I shouldn't do, dang, because I've spent 8 years intentionally avoiding doing it.
      - dang 8y ago
        It's simple: the comment was unsubstantive and snarky. Bad combo, especially on flame-prone topics. We're looking for thoughtful comments here, as you know. The issue isn't just the quality of a comment in its own right—it's the quality of what it leads to. These things compound. Thanks for being a fine HN commenter and user!
- always_good 8y ago

mikhailt 8y ago

The issue wasn't cooling; it was how the CPU incorrectly throttle itself that caused it to overheat again too fast. There's another thermal control unit involved and what people found was that the firmware for it was causing the CPU to drop down to 800mhz instantly when it detected the specific terminal event and when it reset that, it went back up into high-overheating load (such as the benchmark sustaining the high load) instead of staying at the right balance. In other words, it went into a loop of super low power load and then super-high load instead of the middle where the terminal balance is right for the load it is doing. Because of this low/high loop, the performance was not consistent and degraded as the overheating cause it to drop to 800mhz all the time. Where Apple likely messed up is that the firmware was not placing the CPU at the middle to avoid going into a loop, this allows the CPU to sustain a better CPU performance consistently and for an extended period of time.

wmf 8y ago

Note that bad solder joints only existed for 1-2 years and were solved long ago. AFAIK they were caused by NVidia (and MS) using untested lead-free solder balls without underfill.

masklinn 8y ago

The issue was not the CPU throttling itself, it was the CPU being configured to draw north of 100W and the VRM not being able to handle it, overheating, and leading to the system board hard throttling (down to 800MHz rather than slowly stepping down) the CPU until the VRM cooled down at which point the CPU would be un-throttled and the cycle would restart. Apparently the power configuration is the one Apple's always used but previous CPUs didn't come close to the current generation in ability to draw power, so they'd never reached the VRM's thermal limits. By lowering the CPU's maximum TDPup to 50~55W, they avoid the VRM over-heating and while the CPU does not turbo as high as with a 100W TDP allowance it can stay there just fine.