Hacker News new | comments | show | ask | jobs | submit login
Opinion: How Apple could have prevented the iPhone-slowdown controversy (marketwatch.com)
77 points by chmaynard 86 days ago | hide | past | web | favorite | 114 comments

> Knowing the battery size (which also limits the total peak power draw from the battery) and software system in play, Apple could have (and I would argue should have) designed a processor that would provide consistent performance throughout the life of the phone.

> Yes, this would result in lower peak performance for iPhone devices.

Aren't they essentially arguing that Apple should enable the throttling on day one instead of waiting until the battery degrades? How does that benefit anyone?

Yes, and I think this is a common strategy in the CPU manufacturing world. The companies developing and testing software for these devices probably don't realize that the performance of their software will start to degrade in 18 months.

I think this is a really important point.

Also, consider the effect it has on companies that have contracted that development out, and a year later, their app "sucks", without any code changes.

If they knew this from day one, or even day 501, then insisting on designs with non-user-replaceable batteries looks even more like planned obsolescence for financial gain.

If you design a chip with the whole lifecycle of the battery in mind, you can achieve better average performance for the life of the chip.

Pulling numbers out of the air, the argument is the chip was designed for maximum performance with a fresh battery, 100% performance. But later it's throttled to 80%. Or, Apple could have designed the chip for a steady 95% performance throughout the life of the phone.

I would prefer a phone that lost performance over time than one that was gimped at the start. I think the optics aside, most consumers would prefer that as well.

Depends on the customer. If you're planning on keeping it for five years, you'd like it to last the time. If you're planning on replacing it when the new one comes out next year, you'd rather it got maximal performance during that year. What it comes down to is which customer apple wants to cater to which is fairly obviously the latter.

Keeping a phone for five years isn't something I'd want considering the rate of performance improvement. A five year old iPhone would be an iPhone 5. The improvements in CPU/GPU since then are fairly astounding. Going from an A6 dual-core at 1.3GHz to an A11 six-core at 2.39GHz is just nuts. One is 32nm, the other 10nm.

Smartphones are the one area in computing that seem to still be having huge jumps in performance compared to desktop/laptop computers.

No. It's about building a chip that isnt going to crash when the voltage drops for a picosecond, allowing throttling to happen only as needed without fear of crash.

Essentially they designed a processor that cannot handle the normal variance of real life usage and so did a software fix that's very expensive to the user but very cheap for them. I am wondering is there a reason they can't be sued to oblivion for this?

Sued? Yes. To oblivion? No... not enough harm done, and Apple has good lawyers and deep pockets.

Are there processors out there that can handle voltage drops without crashing?

Pretty much all processors start off that way. The clock speed needs to drop to compensate, though.

Unless you actually put some thought into your hardware designs because you're making phones that have a life expectancy of greater than a mere 12 months that will be used around the world for years to come... in which case you can easily reference an electrical engineering textbook and design a circuit that can buffer enough power and supply sufficient voltage/current to meet your expected max sustained power draw with some very simple and cheap components.

But simple frequency scaling based off of _dynamically detected_ available power. There's no need to slow down a CPU based off the age of the battery so you can "guess" what the max speed you can safely run at without inducing a brownout is (and even less of a reason to do so in discrete steps instead of linearly scaling the performance, except if you're trying to do it discreetly) - all you have to do is monitor available power and scale the CPU clock rate accordingly in realtime. But if your circuit is poorly designed with no forethought as to what you'd do if the maximum rated power of the battery wasn't available, I suppose that's not so easy to retrofit.

Modern x86 chips have built in support for voltage drops, as part of margin recovery. As the chip changes activity profile increasing current draw (e.g. turning on all cores for multithreaded code) the voltage will drop until the regulators catch up. "Back in the day" you used margin to accommodate this. Today, circuitry detects these events on the fly and slows down the chip for a few microseconds waiting for the voltage to recover. This allows you to remove margin and operate closer to the limit.

It's a tradeoff. Building a chip with more margin costs perf and/or power too. So it's a game you can't win.

No. You can build circuit to accommodate small voltage fluctuations without crashing things, defeating the need to drop the clock speed 24/7. Lots of other chips do this. Apple dropping clock speed as phones age is a very convenient excuse for an anti-feature to sell new phones.

Seriously, isn't this discussion of the wrong topic altogether? CPU is not too strong, battery is too weak by design. Am I missing the point in this and other articles arguing about everything else but not about battery capacity? What if we had 10% more battery capacity and (oh horror) 0.1 mm thicker phone?

I recently swapped the battery on my iPhone 6. I can tell you that if the battery is too weak it is due to space constraints. Mabe that's what you mean by design, but given the form and space available, I don't think the battery could be stronger than it is.

Sorry I wasn't completely clear. When I say 'by design' I don't really mean 'planned obsolescence', I mean 'system design problem' because somebody wanted to shave another 0.01 mm of thickness. I would be surprised to see any other reason than stupid form over function that causes this.

You mean what they did in the recent round of iPhones?

Maybe. I hope so but we can't really tell if new batteries have security margin or not. It will show in 12+ months.

There is a lot of gymnastics this days, from people like John Gruber and Rene Ritchie to downplay what just happened. Truth is their 'explanations' are disturbing to watch and extremely unhelpful to Apple.

Plain and simple, battery is shaved so much that there is no surplus left (in the name of thinness), nothing to amortize normal battery degradation, so iPhone only works beautifully during initial review and couple of months after that.

I really love Apple and some of my daily living depends on it, but this crap has to stop already. Three years back until now there are many missteps in the name of fad and profit, couple of years more and it won't matter since I won't have any market in that ecosystem.

> Plain and simple, battery is shaved so much that there is no surplus left (in the name of thinness), nothing to amortize normal battery degradation, so iPhone only works beautifully during initial review and couple of months after that.

This is the crux of the issue. Apple is trying to "fix" a design flaw with software.

If it is only one hiccup I woundn't react half as much but we could see exactly the same 'form over function' flaw with Mac Pro recently. Mac Pro has no cooling margin so it often fries graphic cards as soon as you try using it for (surprise) pro stuff.

I really dislike gimmicks over reliability, and I count race to thinness there too, that can be seen with Apple in last couple of years.

> There is a lot of gymnastics this days,

> iPhone only works beautifully during initial review and couple of months after that.

I guess both sides are guilty.

> Plain and simple, battery is shaved so much that there is no surplus left

So what do we gain by a larger battery? Only that the performance envelope would be pushed even higher. You’re basically arguing for degrading the performance of the phone from day one rather than degrading it over time. Matter of preference I guess

I am arguing that you can't create the system where only 100% of speed can be achieved by having 100% non-degraded battery. iPhone must have 10-20% security margin in battery capacity so when battery degrades it doesn't require degrading performance. That can be achieved by 10% thicker phone, right?

That’s why I’ve been wanting an iPhone Pro since about the iPhone 5. A bit thicker, maybe even kind of blocky a la iPhone 4, so you can a decent battery in there. About the size of an iPhone 5 because real pros have shit to do, not watch movies. And since we’re pros, screw playing games on it, so you can put a slower, battery-sipping CPU in there. IP67 water resistance. I’d buy one tomorrow. What I describe is almost the SE, but the SE is just going to have the same problems delayed a year.

When we were doing the Apple Newton, one of my cow-orkers noted that the CPU had a divisor register that let you choose pretty much any clock frequency you could want, down to kilohertz.

His half-serious proposal to guarantee users "a 12 hour battery life" regardless of battery condition was poorly received by the rest of the team . . .

Is that different from what a prescaler is?

It was a "CPU throttle" register that controlled the ARM's clock, with four or eight bits of divisor from the base 20Mhz clock. It's been a long time and it's hard to remember h/w register layouts from several decades ago :-)

(Yes, we did the whole Newton on a 20Mhz ARM with less than 512K of RAM).

Under Siege 2?

So his argument is that the CPU should be throttled from day 1 so that it doesn't have to be throttled after a year of operation at a higher speed? How is that logical?

You could provide a smaller throttle to get the same effect over longer periods of time. Leveling the throttle over the average lifespan of the phone would provide a consistent performance experience. That’s how I read the argument at least.

How does he know the CPU isn't actually throttled already per his suggestion? :)

The simplest solution is to just tell the user that they need to replace the battery with a popup.

The fact that they didn't implement it and that doing so might benefit them financially obviously leads people to assume malicious intent.

Agreed. They can have a $79 in-warranty battery replacement program but if no one knows about it, it does little good to point to its existence later.

>How Apple could have prevented the iPhone-slowdown controversy

Easy, user replaceable battery. (For clarity by "user replaceable" I don't mean by the user paying $79 for Apple to do it; which is a rip off)

IMO if you design a phone (costing the better part of a grand) with a battery unsuitable for a years worth of use it's a design flaw, if the battery is not easily replaceable by the user then the phone is not fit for purpose.

Not to mention the junk this generates and resources this wastes by forcing users to buy a new phone, when a replacement battery would make it perfectly usable again. I hope some regulatory body would step in here, like the EU did for unified phone charger ports.

I (and many other buyers) don't want a user-replaceable battery. I think they look like crap, they make the phone look like crap, and in my history of owning iPhones since day one, I've never had to replace my iPhones battery. I did have one replaced through a warranty extension thing but that was because I was reselling the phone and wanted to give the buyer a perfectly new battery. Otherwise I would have kept the one inside.

Many phone owners these days put their device in a case.

How then can it then "look like crap" if aesthetically it's hidden under a layer of coloured rubber?

Not to mention you'd be hard pressed telling if a phone has a user replaceable battery just by looking at it. I should know because that's the first thing I ask because I won't buy a phone without a user replaceable battery.


I replacyed my old Samsung S4 Mini because the flash storage got super slow. It is known, you can test it and because android uses sqlite and there are apparently many small iops, the smart phone starts to become slow and lagy.

Who cares? I cared. Had to replace it because i can't / will not re solder flash chips. Who else cared? NO ONE...

My previous smartphone was also too old for warranty and had an issue with the mainboard. Also enough other people had this issue.

I don't mind when there are better lawys like 3 year warranty for specific specs like battry lifetime, broken pixel and iops but this is not a 'controversy'.

No one would have blinked if Apple had just built the throttle into Low Power mode and documented the X% cut to clock speed and left stock performance alone.

The problem is - when the battery is bad, the phone will shutdown even if the gauge says 60% or so, and that isn’t obviously “low power mode”. Source: saw it happening on my daugher’s iPhone 6, a year or so ago.

To be clear: I’m not apologizing Apple. They have “we’ll make decisions for the user” attitude in their DNK, but this was gone too far.

I wonder if that's what's happening with my Nexus 6P, which shuts down at 40ish%.

No, that's a widespread defect. I called Google Fi to complain about in October, and they decided to replace my out-of-warranty 6P with a Pixel XL (1st gen, but new).

Apparently they had run out of 6Ps and were trying to clear stock of 1st gen Pixels. They probably won't be quite so nice to you, sorry.

Yeah, I already called them and they said I could at most talk to Huawei to see if maybe they'll get it repaired :/

Why is throttling needed when the phone is connected to AC power?

Good question. Here's another that needs to be answered IMHO:

Do third party battery replacements return the processor to normal performance? Or is it just "official" Apple-performed battery replacements that do so?

I wouldn't be surprised if there were some sort of extra procedure needed to reset the processor throttling back to normal. Or, perhaps the throttling code checks for genuine Apple batteries somehow?

Apple needs to come clean with more details about exactly how this is implemented.

Due to the length of the wire leading to the AC adapter, it probably can't respond quickly enough to the voltage drop. The battery is much closer. A solution would be to use a big enough capacitor on the phone end of the wire, but phones are very space constrained, and since the battery is non-removable, it can be used as sort of a "capacitor" for these instantaneous power demands, so there's no need for the extra cost and space of beefier input capacitors.

I know what you’re saying and see where you are coming from, but that’s not how it would work in reality. Presuming sufficient voltage at the the cable (ie no extreme loss due to internal cabling resistance) the length of the cable should not matter.

The real limiting factor is that a basic USB plug charger puts out only 2A while the iPhone’s peak draw at the same 5V can be well north of 2.5A if you’re using the GPU to its fullest.

However it’s not just the power cord that’s supplying current when you’re plugged in - if it were designed correctly there’s no reason those 2A couldn’t supplement whatever your battery is already putting out and only charge it with what’s left over. In fact, that’s how iPhones can stay at 90% while plugged in to a car charger, running GPS, playing audio at the loudest, and streaming some data in the background while some android models would discharge even while plugged in... but both would be dead a lot sooner if they weren’t plugged in.

The real problem is just shoddy engineering. Hardware isn’t much different from software: what all this boils down to is the equivalent of someone writing code that went into production across millions of devices around the world that assumed input x would always be greater than y and didn’t bother with any exception handling or bounds checking. And when that voltage/current combination coming in is insufficient.. an unhandled brownout exception occurs.

There’s also the USB-C adapter which can fast-charge iPhone/iPad quickly via Apple-specific charging protocol, which clearly provides much more power than than the battery does.

And if USB-C doesn't provide enough amps for you, there's always jumper cables.


I have to admit I’m a bit disappointed. I was hoping to see a real, full-size car battery.

I actually have three absorbent glass mat batteries hooked up to a Noco Genius battery tender and a 1000W 12V inverter to provide battery backup for my sump pumps (long story). Losslessly converting that to 5V shouldn’t be too hard.

Oh absolutely. I was only playing devil’s advocate - in practice even 2A is enough juice and my iPhone 6S would pass out at 60% _within a year of buying it_ when I was doing nothing more than reading Hacker News - I can’t imagine that takes up much power at all!

This is another very good question that didn't get proper answer yet.

I have noticed significant slowdown on my 2012 Mac Book Pro, after installing High Sierra. Up to the point of not being to start any app for several seconds, which wasn't the case with Sierra or before. So in this new light it may as well be throttled down too. Anybody else noticed this?

I'm running High Sierra on a 2012 MacBook Air and haven't noticed any slowdown. Heats up if I run a lot of things at once, but individual apps seem fine.

What if the flaw was earlier in the design? The iPhone has a battery that's 2/3 the size of other phone batteries. Would there even be a problem with a 1000mAh bigger battery?

Apple could have just implemented replaceable batteries and all this situation could have been avoided.

Edit: When I mean replaceable, I mean replaceable batteries without the user having to go an Apple store. Like in the old days, taking the cover off and putting the new battery back in.

Apple batteries are replaceable. $79 and an hour at the Apple store will do the job. If you want to save money you can get it done at a third-party shop for less than $50.

Where Apple made a mistake was in not communicating to its users that this is something they need to do on an annual basis, so people could budget it in to their cost.

The hour at the Apple Store is the deal breaker. My time is too valuable for that.

You don't have to spend the hour at the Apple Store.

You drop the phone off, leave, and come back at literally any time more than 1 hour later. Go to your important meeting. Do some shopping. Have that doctor's appointment you've been putting off. Have a nice long lunch. Whatever. You don't actually lose that hour.

I mean, obviously.

I love Apple products. Every major computing device I use is made by them. I am lucky to be closer than the vast majority of people to an Apple store and even for me this is a huge pain in the ass. I think it is incredibly unreasonable to just hand wave this away like its no big deal. It absolutely is and to pretend otherwise is naive at best.

Then you can afford iPhone upgrade annually.

A swappable battery requires physical containment on both the battery and the socket that adds weight and eats up volume that could otherwise be used for more battery.

I am sure the Apple geniuses could have invented a clip-on battery that takes almost no extra space at all. I bet they wanted to make these phones obsolete after 2 years to force us to buy more.

If you're so sure this is possible, design a better phone and get rich.

Then it must be better for Apple to cripple old phones. /s

They are replaceable... for $79. Which is probably close to what they would have charged for additional batteries if they had gone with a hot swap design.

Actually, I bet they’d sell a $79 normal and $129 extended ‘pro’ version.

Average Joe wouldn't know it was a battery issue making their phone shutdown. Oh, and batteries ARE replaceable.

Easily replaceable. Like it was before smartphones. Your device has either a shit design or or are deliberately trying to rip people off if you need to go to a specialist to replace a damn battery.

Also my laptop from 6 years ago can show the battery health in percent, such technology is not magic.

There have been plenty of smartphones with replaceable batteries. It's just in the last couple of years that everyone has given up on them. I'm still using a Galaxy Note 4 with a replaceable battery. I haven't replaced it yet, but just a few weeks ago I had to pull the battery to get the phone out of a reboot loop. (That was the first time I had seen that with this phone.) If I hadn't been able to remove the battery, I would have have had to wait for the phone to drain it completely and shut off.

I prefer built-in (integrated) batteries.

If they sold a version that had replacable battery and the current one I'm not even sure I'd get the replacable one if it was worse sealed. I replace for $39 after warranty is expired which is pretty much nothing.

I don’t know if I’m just too old having lived through the old nickel-cadmium “memory effect” days, it everybody I know - techie or not - already knows and accepts that batteries “get old” and then “die faster” and “don’t charge right” as the device ages.

Lithium battery tech today is an incredible step forward, but simply acknowledging the still existent limitations of the battery and it’s lifespan seems like a much simpler and saner solution than building a house of cards trying to hide that reality.

Battery dying in 2 hours after a full charge? Time to get it replaced. If your device is still under warranty, it’s free. Otherwise it’s 80 bucks and you’re golden. What’s the big deal?

Instead we have this nonsense approach to pretend that batteries live forever at such a cost.

And then you have the people that pop up out of the woodworks in each and every single iPhone battery life thread to say "it's not that the battery dies faster, it's that the phone suddenly shuts off at 60% [or whatever]," which is an absolute load of BS. The iPhone 6S was the first device to have that problem (I went to the Apple store within a year of its release with my iPhone 6S and spent over 3 hours explaining to the tech what was going on despite what his fancy charts and on-chip reporting software saw until he relented and gave me a replacement phone well before Apple ever acknowledged the issue and initiated the limited recall) - and it's the equivalent of a bug in the hardware design.

As a battery ages, the "definition" of 100% slides (as the charge cycles go up the charge capacity goes down - although it's not that simple since li-ion batteries also retain some form of the old cadmium cell memory effect that doesn't scale in a directly linear fashion with the charge cycles, but not to an appreciable extent that would have a bearing on this particular matter), but there's no reason that a device should die at 60% of its _remaining_ battery life. Except that the iPhone 6S' hardware was designed in such a way that it demanded current that could only be supplied at a voltage in excess of what the reduced-capacity battery could supply, leading to the problem at hand.

This is just me guessing, but I'm fairly convinced of the truth here: Apple didn't implement CPU scaling to improve the end user experience nor to make money off of "planned obsolescence" but rather to avoid having to replace phones/batteries within and without the warranty period (as a recall was warranted due to the faulty design) and decided to implement a software _workaround_ (and decidedly _not_ a _fix_) to get all but the most susceptible phones to avoid the scenario that would trigger the brownout.

There is no such thing as "100%" or "60%" or any other battery meter number. They are all guesstimates based on the battery's history and cycle count.

Li-ion batteries don't have a simple voltage-charge relationship like lead-acid batteries do so you can't actually know the real charge state of the battery (and even those have a number of ways the battery chemistry can get messed up).

I'll also note that your specific battery history matters a lot. If you leave your phone on a hot car dash in the summer even once it will dramatically alter that battery's performance. It isn't as simple as saying it cuts the lifetime... the battery's ability to deliver peak current is affected.

(I’m sorry I wasn’t clearer, but I am not in any way insinuating that there isn’t a problem where iPhones shut down at > 10% as they age - the “BS” referred to are the claims that it’s an intrinsic problem with batteries that couldn’t be dealt with otherwise and not a symptom of bad design on Apple’s behalf that failed to properly handle limited power scenarios. I’ve posted a more detailed comment about what the right thing to do in this situation is earlier today, you can find it in my comment history. Additionally, there is no doubt that Apple has had this problem in their designs for some time and across a range of products but the iPhone 6S was by far the worst culprit.)

Same experience: 6S started shutting down at about 20% after less than a year. I recall searching for this problem and finding that Apple acknowledged that "it's a bug" and that they were working on a fix. The fix came out, but they didn't explain in what way: since then I have an underperforming 6S (50% of average geekbench score) and nobody told me why untill now. Totally agree with your guess of the reason: it was cheaper to do that in software than to replace a lot of batteries, also avoiding the PR nightmare.

> And then you have the people that pop up out of the woodworks in each and every single iPhone battery life thread to say "it's not that the battery dies faster, it's that the phone suddenly shuts off at 60% [or whatever]," which is an absolute load of BS. The iPhone 6S was the first device to have that problem ...

Nope! I had an iPhone 3GS which would die at 30-40% if I turned on GPS.

I'll add my voice as well.

I had a 5s that shutdown randomly at around 30-40%.

On a hunch I replaced the battery and the problem was solved. I can't say if performance improved or not since I gifted it to someone.

>> The iPhone 6S was the first device to have that problem

My 4S and my 2009 MacBook Pro would both randomly shut off under 40% or so after they were 2 years old.

The fun part is, everybody easily accepts that battery degrades. But thinks hardware shouldn't.

I have a 6 year old battery in my 2011 Macbook Air and it seems to handle the peak load of my quad core i7 just fine.

Phones and laptops are different beasts. I'm going to make up some numbers here to simplify, but the idea is the important part.

Note: I wrote this whole comment with C inverted the first time. Too little sleep. It is now fixed thanks to user revelation.

First: Battery people like to talk about discharge and charge rates in a unit called "C". If you drain a battery at 1/10C, you will totally use it up in 10 hours. If you drain it at 1/2C you will totally use it up in 2 hours.

Laptop: 5000 milliamp hours at 10 volts for 50 watt hours of energy. This will last for 10 hours of use if you average 5 watts. You will be discharging at 0.1C on average. Your CPU might be a 15 watt CPU, so you can burst to maybe 20 watts total, or 0.4C.

Phone: 1500 milliamp hours at 4 volts for 6 watt hours of energy. This will last for 6 hours of active use, so that is 1 watt. You are discharging at 0.16C on average. Let's guess the CPU can surge to 5 watts. That takes you to 6 watts total use and a discharge rate of 1C.

So that is a 2.5x safety margin that the laptop has for degraded batteries over the phone.

Summary: Phones operate much closer to the current limits of batteries compared to older laptops. Accumulated wear or damage to the battery will affect them sooner.

I think you got it backwards, "C" is for capacity (useful since (dis)charge current obviously scales with capacity, ceterus paribus) and a 2C discharge rate is twice the current of a 1C rate.

You also need to be careful in your calculations. Battery discharge rates are given in Amps (or C multiples) not Watts because over a cycle, cell voltage will naturally drop. Discharging at 5 Watts puts more strain on an empty than a full cell since the first requires higher current to compensate for the lower voltage. Amps reflect the "demand" better.

Your summary is still correct though, there is more margin for degradation in a laptop battery and space for decoupling.

Yes I did write it all backwards. I'm going to go back and fix the post, so your comment is not going to look insane.

Most laptops with dedicated GPUs can drain their battery quite a bit faster than a phone can, but they don't exhibit instability on batteries nearly this young.

Without making the example too complex, it is the instantaneous peaks which get you in trouble. You could have your average power consumption approach 2C (totally discharged in two hours) and be just fine, that is the specified maximum for discharging something like a standard 18650 lithium cell. It is the surge that is going to cause an unacceptable voltage drop.

So, tiny capacitor?

And I don't mean discharged in two hours, I mean discharged in 40 minutes when you keep it at max usage. Faster than a phone.

If phones had active cooling and could thermally sustain a maximum CPU+GPU load for 40 minutes, they would discharge that fast or faster.

Surprised you have a quadcore macbook air, since they were never made.

You are right, it is dual core. I forget the mobile variants can be dual core. *4 virtual cores :p

Different devices have a longer consumer replacement cycle and less need for out-of-the-box energy density. For the same level of capacity degradation (80%) MacBooks are designed to 1000 cycles whereas iPhones batteries are designed for 500 cycles [1]

[1] https://www.apple.com/batteries/service-and-recycling/

Batteries degrade with cycles and heat. Maybe your battery has few of both? Or a new battery?

Because it doesn't degrade, at least not in ways that are meaningful to the customer.

The lede of the story and pretty much every one I've read is false.

>For years consumers have complained and theorized that Apple would purposely degrade the performance of older iPhones, pushing them to upgrade faster than they normally would.

>This week the theory of slowing iPhones was validated by external sources and eventually corroborated by Apple.

The conspiracy theory we're talking about is that Apple slows your old iPhone to make you upgrade to new ones and it goes back all the way to the first iPhones. It's false on its face because this fix wasn't added until iOS 10.2.1 [1] and it wasn't what Apple "corroborated" at all. And for the "Apple should have been transparent" crowd, here's what it said in the release notes:

>iOS 10.2.1 includes bug fixes and improves the security of your iPhone or iPad.

>It also improves power management during peak workloads to avoid unexpected shutdowns on iPhone.

>For information on the security content of Apple software updates, please visit this website: https://support.apple.com/HT201222

There's some extremely bad writing on the internet about this story meant to capitalize on the general controversy. Honestly, Apple should find the most egregious one and sue for defamation, and relish the opportunity to testify under oath if it goes to trial.

[1] https://support.apple.com/kb/DL1893?locale=en_US

I agree with most of what you wrote, and I do think it was the right call for them to implement this feature for iPhones with batteries having trouble, and I agree most if not all the articles about this have been poorly written and badly misinformed. However...

> for the "Apple should have been transparent" crowd, here's what it said in the release notes

I can't agree that that's sufficient. That little one liner in the release notes just isn't enough.

On macOS, they did more. The battery menu extra will actually display a warning when clicked if it detects that your battery is having trouble. It shows as an extra row at the top of the drop-down menu with a little little caution triangle icon and the string "Service Battery". When you click this, it takes you to a help article advising you to "take your computer in for service". It's not perfect, but it's better than not telling the user anything.

Apple could and should have done something similar on iOS, especially given that there's a whole section in Settings for just the battery. They could have displayed a similar alert icon with similar help text, for starters. A full-screen alert advising the user to get their battery serviced whenever this condition was detected would have been even better. Burying it as a single sentence in the release notes wasn't all that helpful or informative for customers.

They did in the 10.2.1 update [1]. As well as interviews with the tech press about it [2].

[1] https://support.apple.com/en-us/HT207453

[2] https://techcrunch.com/2017/02/23/apple-says-ios-10-2-1-has-...

My iPhone 6 is benchmarking at 25% lower than the average when fully charged and this warning is not being displayed. So clearly shenanigans are at play.

Are you testing on a fresh install of iOS with no background apps?

Yes and yes. My experience is shared with a lot of other random commenters. Clearly the threshold for this battery warning is lower that the threshold for throttling.

The lede of the story and pretty much every one I've read is false.

Meta comment: Why is it that we routinely debunk shoddy journalism here on Hacker News but struggle to see why people call the vanguard of mainstream media "fake news"?

I see articles from all venues get eviscerated here on Hacker News. Can we be honest with ourselves and just say that if they're this bad about complex technology subjects, perhaps they might be equally as bad on complex geopolitical or economic issues?

Because we shouldn't think about the media categorically one way or the other, as if it's either all fake or everything in the media is true. The media gets things wrong, and we should correct their mistakes.

However, the political aspects of discrediting entire news outlets or the entire news community because of critical or negative coverage of certain politicians or political parties is an attitude that promotes mistrust in our society and creates an environment where nefarious political actors can successful avoid scrutiny.

> Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well. In Murray's case, physics. In mine, show business. You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. I call these the "wet streets cause rain" stories. Paper's full of them.

> In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.

― Michael Crichton

What happens if you charge your throttled iPhone. Does the performance improve while charging vs running on battery ?

Rather than throttle, Apple has done something similar since the A10/A11 processors with their additional low-power cores. Presumably this should reduce the short-term wear on the battery and extend the life of the cells.

then they should have charged less for their products. or created a more environmentally friendly design. my guess is that apple's PR team made this article happen


If you have tell someone they're wrong about there being a controversy, there's a controversy.

Just add a setting. Prefer performance, or prefer battery life.

Is that so hard?

What arguments are there against this other than planned obsolescence? Educate me. Downvotes aren't arguments.

Performance isn't lowered to extend battery life, it's lowered because the degradation in old batteries makes it impossible to reliably run the processor at max clock; if you flip a switch to force high clocks despite an old battery, your phone will randomly shut down sometimes.

So they couldn't really run the CPU full speed for years. They sure could have, say, highlighted to users that a battery replacement can make an old phone run better for longer, but of course they'd much rather users replace their phones. Or, as another comment here suggested, they could have left more margin--overspecced the battery so it could run the CPU at higher speeds even once it was somewhat degraded, rather than putting so much of the efficiency gain over iPhone generations into thinness and lightness. Might also help battery life for heavy users.

Performance isn't lowered to extend battery life, it's lowered because the degradation in old batteries makes it impossible to reliably run the processor at max clock; if you flip a switch to force high clocks despite an old battery, your phone will randomly shut down sometimes.

That may be what they're telling people, but there is absolutely no engineering basis for it. Battery voltages decrease gradually under all normal operating conditions, and switched-mode converters can extract useful energy from them throughout the entire discharge regime.

The main constraint on battery voltage isn't some mysterious "random shutdown" effect, but the health of the battery itself. If you allow the battery voltage to fall too low -- which will, again, happen gradually and not instantaneously -- its service life will be greatly impaired. This is not something that happens without warning. It's a perfectly well understood effect. Once the phone's CPU sees this happening, it will have plenty of time to post a "Shutting down in xx seconds" message to the user, go into emergency low-power mode, or otherwise take action to protect the battery.

Even those that skeptical of what Apple did here seem to think there's some real limit on safe power draw involved. If that can be shown false it could probably be a popular blog post here at least.

From Apple's statement: "Lithium-ion batteries become less capable of supplying peak current demands when in cold conditions, have a low battery charge, or as they age over time, which can result in the device unexpectedly shutting down to protect its electronic components."

Apple seems to be saying that throttling might occur even if the battery is perfectly fine. This points to a fundamental design flaw across many different models and generations, which is the main conclusion of this article. One could even argue that Apple is using a "cheat device" that produces different performance during benchmarks vs. actual real-world use. Sound familiar? This is exactly what VW did. Is it illegal? We will find out as the lawsuits proceed.

Applications are open for YC Summer 2018

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact