Hacker News Reader

Comment by m_mueller | Hacker News Reader

What drives me up the wall with all recent macbooks, besides the deteriorating keyboards, is how they need a charged battery to power on. Meaning, they have to constantly drain and charge the battery even with ac plugged in. Is is really that hard to provide what has been standard in laptops for years with an additional wall power circuit? Just planned breakage so you have to keep buying new ones? Apple in the passed stressed these kinds of details, now apparently you can't even tell whether it's charging anymore without looking at the screen.

View on HN

Edit: I point out Apple failing to write drivers FOR THEIR OWN LAPTOPS and I get downvotes? :/

Apple just doesn't seem to care about build quality beyond the cosmetics of their Mac lineup. A great example of this is how they have dual connections from the keyboard & trackpad to the mainboard (USB and SPI), while the UEFI and Linux have drivers to support the keyboard fully, macOS has no driver for the built-in USB connection.

What this means is when your coffee spills on your Mac, the keyboard won't work in macOS.

Louis Rossman did a video demonstrating Apple's lack of drivers for their own hardware: https://www.youtube.com/watch?v=geGmC1xI4zo

Because it seems fair enough as en engineering decision?

Apple also provide firmware updates far longer than most manufacturers (there was a recent-ish CCC? talk highlighting this).

> Because it seems fair enough as en engineering decision?

Huh? Supporting hardware with drivers is purely a software decision, its not as though software adds noteworthy weight to a device.

> Apple also provide firmware updates far longer than most manufacturers

Only for iPhones, on Mac they leave your iCloud email in in plaintext on the UEFI after the computer is wiped. For computers, expect 6 to 7 years of macOS updates at most, meanwhile Windows & Linux offer support to ancient i386 systems still.

> Supporting hardware with drivers is purely a software decision, its not as though software adds noteworthy weight to a device.

But writing drivers still takes engineering time and costs money. Why bother writing two different drivers for two different interfaces when they do the same thing?

> Why bother writing two different drivers for two different interfaces when they do the same thing?

Reliability and quality, the hardware is already there, why would you hamstring the product's reliability?

Because a failure that causes one interface to fail, but not the other seems really unlikely. Plus it wouldn't just be another driver to write, you also have to come up with some logic that determines if one has failed and use the other instead. This isn't linux where you can reasonably tell users "oh just blacklist this kernel driver and it'll use the other instead."

> a failure that causes one interface to fail, but not the other seems really unlikely.

Go follow that link in my original post. Its a fairly common water damage issue for Macbooks where the SPI bus will be damaged, but the USB connection will be fine.

> come up with some logic that determines if one has failed and use the other instead.

Once again, go follow that link. Apple already built the logic, just instead of switching to USB, it begs you to pair a bluetooth keyboard!

Ah, sorry, I wasn't somewhere I could watch the video. If that's the case, I take back what I said, it could probably make sense to include both drivers then.

No worries, I avoid linked videos like the plague myself. This one is decently short (for these types of repair videos) at 11 minutes, and is pretty easy to jump through if your pressed for time.

How much of your email are they storing in UEFI? And why?

The full email address is stored, there are a couple videos where Louis Rossman nabs the email address (occasionally with an SPI programmer) and contacts the original owner, finds out the laptop is stolen and returns it.

Oh the email address. That makes more sense. I thought perhaps they were storing iCloud email in the chip for your account.

This is not really true, the difference is much smaller and you will not boot windows 10 with the i386 feature set.

Since Windows 8 the NX bit is required, making it require relatively modern CPU's, mostly x86_64, allthough a few x86 has it as well.

So for example, you can run the 32bit version of win 10 on the 32bit Core Duo (it has nx), while OSX with its 64bit requirement needs a Core 2 Duo.

Newer OSX require a graphics card of DX10 level, Windows 10 require DX9 level.

> on Mac they leave your iCloud email in in plaintext on the UEFI after the computer is wiped.

You are probably referring to the "Find My Mac" token that's stored in the NVRAM. It's fairly easy to wipe it with one of these commands/procedures:

  - sudo nvram -d fmm-mobileme-token-FMM
  - sudo nvram -c # this one will wipe the entire NVRAM, use with care!
  - https://support.apple.com/en-us/HT204063

> For computers, expect 6 to 7 years of macOS updates at most

High Sierra can be installed on "Late 2009 iMacs", that right there is 8 years of security and feature updates. Those models will get security and point updates, at the very least, until the next version of macOS, for a total of 9 years (if not more, we'll see).

Last year's macOS Sierra won't install on my wife's 2010 Macbook Pro. Apple doesn't seem to always roll off support every year. I'm guessing some version adds new tech and they drop support for 3-4 years worth of hardware all at once (that hardware is usually at minimum 5 years old).

Thankfully I have no expectation my keyboard will work after spilling water on it.

> Apple just doesn't seem to care about build quality beyond the cosmetics of their Mac lineup. A great example of this is how they have dual connections from the keyboard & trackpad to the mainboard (USB and SPI), while the UEFI and Linux have drivers to support the keyboard fully, macOS has no driver for the built-in USB connection.

Well, it's complicated.

In the past mobile Macs used USB to connect the input devices like most other notebooks do. At a certain point Apple decided that it would be a good idea to switch to SPI for power consumption reasons.

Following that decision there were certain models which had wired both, e.g. the MacBook Pro 2015. While Apple switched already to use SPI there, Linux still used USB, as there was simply no driver support.

The Retina MacBook in 2015 was the first mobile Mac which didn't support USB for input devices anymore. The MacBook Pro followed in late 2016.

That left Linux without a working driver for the input devices. A year ago somebody started to reverse engineer the SPI protocol to get a working driver: https://github.com/cb22/macbook12-spi-driver/

The release of the SPI-only MacBook Pros end of last year sparked the further development of this driver, so it's slowly reaching a state were everything is just working (maybe except for advanced Touchpad features).

That driver is still not included in any major Linux distribution, as it's still pretty experimental and also required patching the Linux kernel for quite some time, because Apple does some ACPI-related stuff different than the PC-world. Thankfully support for Apples way will land in the upcoming Linux 4.14 (https://github.com/torvalds/linux/commit/53ac64aac9af8cd0e54...) making some code in the driver redundant (https://github.com/cb22/macbook12-spi-driver/pull/46).

So I would say that whole topic isn't about not to care about build quality, but about to care about power consumption. Still having USB on some generations of Macs in addition to SPI was probably simply to ease the migration by unbundling the hardware changes from the software changes.

It's more about voltage than power, so you are wrong that they necessarily drain the battery when plugged in . Anyways, this isn't new at all, it was the same with PowerBook Duo.

Why does every other laptop manufacturer include circuits to handle this use case, yet Apple has failed to do so for 24 years?

Same deal with the lack of keyboard texturing on Macs, providing a flat surface for bacteria, oils and dirt to collect on your laptop's keyboard. Other manufacturers commonly add an anti-microbial texture to their keyboards, which also wards off most of the grease you see on Mac keyboards.

Point of reference here - I plugged in an old 2008 Macbook Pro with no battery (it had expanded, so I took it back to Apple for recycling), and it booted up fine from a Magsafe 1 charger. So I don't know if that's the case.

It was the case for me and @m_mueller apparently, my old MBP had one of those inflated batteries from heat likely, the only other device I've seen a battery puff up like that in is an old Blackberry. No bueno!

Never seen significant puffage in the used/broken Li-Ion battery packs I've rebuilt either, so presumably that is moderately unique to Apple?

It is unique to the laws of physics in this universe. If you look around, you'll find video of exploding flashlights, and it wasn't Apple that recently had airlines forbidding you to bring their devices onto planes...

So, that's drivers, batteries, and keyboard-coatings; what's next?

Saving your email you used for iCloud to the UEFI in plain text and providing no way to wipe it? Could take a sledgehammer to it...

Never said that battery failure was something that only occurred with Apple, just the way they fail (puffing up) was something that has happened to myself and two of my friends, meanwhile these 5+ year old battery packs from HPs and Dells that I've been pulling apart had a cell or two fail to retain a charge, but they don't significantly expand in size usually.

> providing no way to wipe it

sudo nvram -d fmm-mobileme-token-FMM

Same here. That battery was pretty scary.

Personally, I don't remember owning a battery removable Mac notebook between the PowerBook G4 to the MacBook Pro that couldn't function without a battery while plugged in.

The Duo was the first, and not every powerbook after it copied this design. I think they might have pulled it out again for the G1 Air, but I'm not sure.

I have a summer 2009 MBP whose battery failed in... 2013, I think. I removed the battery and have been running it off just the Magsafe adapter ever since (just as a low-power home server).

From time to time, Apple has cheated to make their laptops thinner by skipping a capacitor or two. This started with the PowerBook duo (the first ultralight, actually, Apple basically defined the modern laptop form back then), and they pull out this trick whenever they need it to fit in their design envelope (literally and figuratively). I'm sure other laptop manufacturers do also, you probably just never encountered any other ones.

Clean your keyboard more regularly.

> Clean your keyboard more regularly.

Just touching an Apple keyboard gives it a slight grease, despite regular cleaning. You don't get that on other keyboards, as they are textured slightly to prevent it.

This isn't true; I've actually done some anecdotal AB testing on this. My wife cleans her Macbook keyboards regularly and it's always non-greasy to my touch, they always feel new. I on the other hand, am less diligent, and I can tell.

Compare it to a Corsair keyboard or a Thinkpad keyboard, you can get that new feeling much longer in between cleaning. Additionally, your wife may have drier skin, meaning less or no grease.

Except we both use the same computers, so I expect it goes beyond user just to habit.

Also, the texture doesn't make the keyboard cleaner, it just makes it feel less greasy when dirty. That isn't considered a good thing to many people who are more likely to fit into Apple user demographics.

The same trick was used in many Nokia phones. Probably the most extreme instance of this is one HW revision of 5110, where the final RF PA stage consists of single NPN transistor in common emitter circuit connected directly to battery terminals (other HW variants have HP power amplifier IC in the same place, but still, powered straight from the battery)

> Meaning, they have to constantly drain and charge the battery even with ac plugged in.

I'm nearly certain this is wrong

partially, charger is too weak to power whole laptop, so the whole thing requires battery as a buffer.

Similarly X220/230 will throttle to half the speed (~1GHz) when using 65W charger (instead of recommended 90W) and no/empty battery.

Seriously? I did not know this. The battery in my old MSI died recently (no longer charges) but I can still use it plugged into a wall. I thought "Well that's good. At least it's not like a shitty cellphone where it has to build up a minimum," but .. guess that's not true now for MacBooks?

I don't quite understand what you're saying. Are you asserting that modern Mac computers will not turn on without a battery installed (which is not true in my experience), or are you asserting that trickle power is a bad thing (also not true in my experience)? Maybe I'm missing your point altogether.

Edit: did some research, I will admit it is possible that newer MacBook pros with an embedded battery may not power on if it is not connected (not sure about this). But I know my 2010 15-inch MacBook Pro does power on and run without a battery connceted (I use it as a media server), and I'm pretty sure most MacBook airs would run without a battery connected as well.

Doesn't seem to impact battery life much, so they handle that properly at least - I've got old Macbooks that were on AC all the time, marginal degradation of the batteries.

Reply on news.ycombinator.com