How iPhone batteries work, and how Apple manages performance over time

Posted:
in iPhone edited January 2
Lithium-ion batteries such as those found in nearly every smartphone and tablet are not little, indestructable homogenous power reactors. AppleInsider delves briefly into the construction and chemistry of them, and discusses how an old battery works against not only device manufacturers, but the users themselves.


A battery is a physical and chemical process, and isn't eternal

First, some science 101.

Any battery stores electrical energy in the form of chemical energy, and can convert that energy into electricity. A battery requires an anode and cathode, separated by electrolyte that allows the flow of electrical charge between the cathode and anode.

When the battery is put under load, the anode releases electrons to the negative terminal and ions in the electrolyte through an oxidation reaction. The cathode accepts these electrons, completing the circuit for the flow of electrons.

This is the end of the story for non-rechargeable batteries. However, lithium-ion batteries can be recharged. So, when current is properly applied, the electron flow happens in reverse, recharging the battery.

This is not a linear reaction -- also called an "ideal battery." A battery's total capacity measured in milliamp-hours for mobile devices and overall lifetime including charge/discharge cycles are limited by engineering choices made by device manufacturers including charging circuitry and software, plus physical volume of reactants.

The physically larger the battery, the higher capacity, and the flatter the curve, and the less pronounced the voltage drop is when the battery charge is low. Given the current battery debacle, this is why the Plus-sized iPhones are less affected by slowdowns, and why the iPad doesn't seem to be at all -- physically larger batteries.

The better and more regulated the charging and discharging of the cell is, the longer the battery will live. The more temperature excursions the battery endures, like leaving it in a hot car, or outside in freezing temperatures, the shorter the life of the battery will be.

The output voltage of any given battery over a discharge cycle can be plotted over percent charge.




The reactants aren't eternal, though. In the case of a lithium-ion battery, "metal whiskers" can form in the cell, shorting out afflicted portions of the battery cell and cutting down on available power. Ultimately, the whisker formation, coupled with reactant depletion lead to a completely dead battery from under-voltage and non-reversible oxidation.

The slope of the output voltage graph varies

The voltage "curve" and slope varies a great deal based on a number of factors. Battery wear shrinks the horizontal axis. Obviously, device power demand decreases the amount of time it takes to progress along the curve on any given charge.

Damage to the cell from the environment, faulty charging gear providing more than allowable voltage, or other issues permanently increases the steepness of the voltage drop while the battery is being used. Operating temperature has a temporary increase in steepness as well, with low temperature having more of an impact on capacity than high.

A "dead" battery, really isn't

Just because your battery is "dead" doesn't mean that it doesn't have any stored power. This is evidenced by Apple popping up a "plug in" graphic when you try to start a phone with a drained battery.

In the case of a lithium-ion battery, the battery must have more than 2V capacity, or the electrode starts to oxidize. This happens fairly quickly and cannot be reversed by recharging. This is why a lithium-ion battery left idle for a period of time ultimately completely dies, or has next to no capacity.

In all likelihood, the reports of the Samsung Note 8 shutting down and not coming back to life after a drained battery are because the device is below the power that it can kick-start the charging circuitry, and not permanent damage to the battery.

Device manufacturers, and the curve

Without delving too deeply into battery manufacture, device manufacturers have to make choices based on the performance curve. Battery capacity and output voltage are two different things that have to be considered, to say nothing of safety factors.

Every device has a critical voltage, at which point it won't stay operational, may lose data, or crash entirely. The critical voltage is universal, varies from device to device, and is not just applicable to Apple.

Grossly simplified, manufacturers have to look at the battery performance curve, and pick a point that it is making engineering choices for. If they miss their guess, then as a device's battery chemically ages, it will shut down. Yes, this does happen in Android devices -- with the Nexus 6P from Sept 2015 now seemingly affected by the problem -- and no software fix to prevent the shutdown in sight.

If a lithium-ion battery is compromised by overcharge, overheating, damage, or simply age, the inner cells can "outgas" the contained flammable electrolyte mixture. An undamaged battery membrane will stretch and bulge to contain this material to some extent, but at some point the membrane will rupture explosively -- as practically demonstrated by the fires caused by the Galaxy Note 7 in 2016.

How we got here

Mobile processors don't consume a set amount of power. They draw more volts depending on how hard they are being worked.

Benchmarking applications are designed to run a processor for as long as it can, as hard as it can to complete a calculation, or do a series of them in an set period of time. This will invariably engage any voltage-related down-clock routine as the processor is slammed by the testing -- and it is how Apple's specific method of fixing the random shut-down problem in a voltage-constrained situation was discovered first on Reddit, and then by Geekbench curator John Poole.




It isn't clear how many users actually experienced or noticed the slow-downs, but the random shut-downs before iOS 10.2.1 were certainly noticeable. According to Apple, the slow-downs aren't permanent, and are only invoked with the situation demands it. A battery less chemically depleted will down-clock less often, and a user not pushing the device hard may never see it, even with a near-defunct battery.

There are two different situations surrounding Apple's batteries in the iPhone 6s, and they are unrelated. The first is the battery replacement program that Apple executed for a batch of batteries installed in select iPhone 6s units. In that case, the battery had known manufacturing issues, and Apple replaced them free of charge to owners of the afflicted serial number range.

The second is the shutdown fix it implemented in Feb. 2017 with the release of iOS 10.2.1. What it did was temporarily down-clock the CPU in the phone, to keep the device's voltage demand less than the critical voltage where the phone would shut down seemingly at random. The iOS 10.2.1 is often conflated with the recalled batch of batteries for the iPhone 6s, when in reality, they have nothing in common.

To fix the problem, Apple was less transparent than it should have been, perhaps, in its very brief release notes saying that it had fixed the random shutdown problem. The aforementioned Nexus 6P that looks to be having the same low-voltage problem probably won't get a patch preventing the random crashes.

But, no matter how much money Apple makes, it can't escape the laws of physics, nor the reality of chemistry.
applesnorangesthedbaMisterKitClarityToSeebadmonkblubick
«13

Comments

  • Reply 1 of 52
    Very informative. Thanks 😉
    racerhomie3ClarityToSeeGeorgeBMacblubickjony0
  • Reply 2 of 52
    dws-2dws-2 Posts: 183member
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    I'm wondering if the "critical voltage" is higher on the iPhones 6  and later (vs. the 5s and earlier) because that's when the issue started. Maybe newer processors, being more powerful, need a higher critical voltage. Also, iPhone batteries tend to be smaller (iPhones being more power efficient) than Android batteries, and bigger batteries are less affected, so I'm wondering if this is less of a problem in the Android world. 
    ClarityToSeeradarthekatGeorgeBMacblubickjony0
  • Reply 3 of 52
    Important info for anyone worried or confused by the battery controversy. The average mobile phone customer (iPhone or Android) probably knows that the battery capacity will decrease over time, but that's about it. Apple wasn't doing anything underhanded with the 10.2.1 update.
    baconstangGeorgeBMac
  • Reply 4 of 52
    rogifan_newrogifan_new Posts: 3,017member
    dws-2 said:
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    I'm wondering if the "critical voltage" is higher on the iPhones 6  and later (vs. the 5s and earlier) because that's when the issue started. Maybe newer processors, being more powerful, need a higher critical voltage. Also, iPhone batteries tend to be smaller (iPhones being more power efficient) than Android batteries, and bigger batteries are less affected, so I'm wondering if this is less of a problem in the Android world. 
    Check out Rene Ritchie’s Vector podcast. Just before Christmas he had a round table discussion with a John Poole from Geekbench and a few others to discuss what’s happening and what it means. It was a balanced podcast with facts not outrage or hyperbole.
    dws-2
  • Reply 5 of 52
    thedbathedba Posts: 443member
    dws-2 said:
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    I'm wondering if the "critical voltage" is higher on the iPhones 6  and later (vs. the 5s and earlier) because that's when the issue started. Maybe newer processors, being more powerful, need a higher critical voltage. Also, iPhone batteries tend to be smaller (iPhones being more power efficient) than Android batteries, and bigger batteries are less affected, so I'm wondering if this is less of a problem in the Android world. 
    Check out Rene Ritchie’s Vector podcast. Just before Christmas he had a round table discussion with a John Poole from Geekbench and a few others to discuss what’s happening and what it means. It was a balanced podcast with facts not outrage or hyperbole.
    I listened to that podcast and also just heard John Gruber's episode 210 from Saturday Dec. 30. They're all very informative but in this day and age of binary journalism, right/wrong, left/right, good/bad etc. there is no place for nuances, unfortunately. 
    Many have made their minds up and are using this to confirm their personal bias. 
    racerhomie3chiaGeorgeBMacdws-2jony0
  • Reply 6 of 52
    dewmedewme Posts: 1,506member
    Very well done Mike, informative, logical, and clearly presented.

    Knowing that Apple is presented with the same engineering, science, fabrication, manufacturing, and materials challenges that all other smartphone makers are faced with yet is still able to produce the best-in-class product in most every product category that they compete in says a lot of positive things about Apple and the quality of people they employ and partner with. We enthusiasts sometimes dwell way too heavily on the negative fringe cases and forget how well Apple excels in the fundamentals of engineering and business. Fortunately, the market knows and rewards them accordingly.  
    racerhomie3chiabadmonkGeorgeBMacblubick
  • Reply 7 of 52
    78Bandit78Bandit Posts: 184member
    The biggest question to me is why didn't Apple have a clue this was going to happen as fast as it did?

    Apple engineers knew enough about the A8 chip to claim 25% CPU and 50% GPU performance increases over the A7 while needing half the battery power.  I can't imagine they ever thought the phones would start unexpectedly shutting down after only a year or two.  It was only after the unexpected shutdowns continued to become more prevalent and they introduced specific diagnostic software in iOS 10.2 to try and figure out what was going on that they identified the voltage problem.

    This significant issue caught them completely off guard which would normally be unfathomable for Apple.  Something unanticipated happened with the hardware.  Possibly the actual real-world power draw was higher than the engineers expected, the A8 had a higher critical voltage requirement, or the batteries deteriorated faster than previous ones.

    This article is full of great information as to what is happening, but I don't think we'll ever figure out exactly why the phone didn't perform as the engineers expected.
    ivanhdws-2ClarityToSeelarrya
  • Reply 8 of 52
    So this is Apple's secret sauce: instead of replaceable batteries, Apple slows down the custom CPU over time to make devices appear to be usable for longer period of time. This technique probably was in the very first iPhone, maybe even iPods days. Samsung probably knew about it because it was building CPU for the early iPhones. This is probably why my Nexus 7 had to be plugged in and can't seem to keep its charge after a couple of years. This is why my iPad 3 is still usable after 6 years. It is all because the battery is old.
    ivanh
  • Reply 9 of 52
    Mike WuertheleMike Wuerthele Posts: 3,129administrator
    78Bandit said:
    The biggest question to me is why didn't Apple have a clue this was going to happen as fast as it did?

    This article is full of great information as to what is happening, but I don't think we'll ever figure out exactly why the phone didn't perform as the engineers expected.
    I think you're right -- I also suspect that a great deal of this tale won't be told. It may be something as arcane as just very very slightly off chemistry in the battery, or like you said, a slightly different critical voltage.
    edited January 2
  • Reply 10 of 52
    Important info for anyone worried or confused by the battery controversy. The average mobile phone customer (iPhone or Android) probably knows that the battery capacity will decrease over time, but that's about it. Apple wasn't doing anything underhanded with the 10.2.1 update.
    I would think (don't know) that the majority of Apple customers appreciate Lithium-ion batteries degrade over time.  Depends what you mean exactly by "underhanded" however my interpretation is that when Apple decided to slow down a device via IOS because the battery in that device is not up to a performance threshold AND not be transparent about it, then that is being underhanded.  Apple admitted as much in the press release that came out a few days ago.
    ClarityToSeebonobob
  • Reply 11 of 52
    ivanhivanh Posts: 179member
    Sorry Mike, I know that your intention in writing this article is to help Apple to contain damages. But,
    1. Apple slows down my iPhone 6 Plus (it) from 1400MHz to 936MHz randomly, while no app was running (besides CPU DasherX with which used to show the CPU speed). It was in room temperature, i.e. not in “extremely cold” temperature, but it was getting hot by itself that trigger me to test it.
    2. The iPhone 6 Plus is only 10 months old and the battery condition is 94% as revealed by coconutBattery app when connected to a Mac.
    3. Throttling happens too when the iPhone 6 Plus is attached to a 2.1A Apple original charging source.
    4. All my family members and many friends with iPhone 6 Plus experience speed throttling more severe than iPhone 6s.
    So, all of the battery myth told by Apple and you, fail to give an explanation why my iPhone 6 Plus has been throttled!
    ClarityToSeeHabi_tweet
  • Reply 12 of 52
    dws-2 said:
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    I'm wondering if the "critical voltage" is higher on the iPhones 6  and later (vs. the 5s and earlier) because that's when the issue started. Maybe newer processors, being more powerful, need a higher critical voltage. Also, iPhone batteries tend to be smaller (iPhones being more power efficient) than Android batteries, and bigger batteries are less affected, so I'm wondering if this is less of a problem in the Android world. 
    The problem far more widespread in the Android world .Android has built in CPU throttling.Even besides that ,even the Latest LG & Samsung phones ,with heavy use , has had random reboots.
    This is the problem Apple has been trying to solve.
  • Reply 13 of 52
    ivanh said:
    Sorry Mike, I know that your intention in writing this article is to help Apple to contain damages. But,
    1. Apple slows down my iPhone 6 Plus (it) from 1400MHz to 936MHz randomly, while no app was running (besides CPU DasherX with which used to show the CPU speed). It was in room temperature, i.e. not in “extremely cold” temperature, but it was getting hot by itself that trigger me to test it.
    2. The iPhone 6 Plus is only 10 months old and the battery condition is 94% as revealed by coconutBattery app when connected to a Mac.
    3. Throttling happens too when the iPhone 6 Plus is attached to a 2.1A Apple original charging source.
    4. All my family members and many friends with iPhone 6 Plus experience speed throttling more severe than iPhone 6s.
    So, all of the battery myth told by Apple and you, fail to give an explanation why my iPhone 6 Plus has been throttled!
    Being slow isn’t the same as throttling peak power draw. 

    Dunno DasherX but it sounds like an app. 

    Accusing Mike of working for Apple PR reveals you to be cluelessly paranoid. 
    racerhomie3Mike WuerthelechiamacxpressbadmonkGeorgeBMac
  • Reply 14 of 52
    Mike WuertheleMike Wuerthele Posts: 3,129administrator
    ivanh said:
    Sorry Mike, I know that your intention in writing this article is to help Apple to contain damages. But,
    1. Apple slows down my iPhone 6 Plus (it) from 1400MHz to 936MHz randomly, while no app was running (besides CPU DasherX with which used to show the CPU speed). It was in room temperature, i.e. not in “extremely cold” temperature, but it was getting hot by itself that trigger me to test it.
    2. The iPhone 6 Plus is only 10 months old and the battery condition is 94% as revealed by coconutBattery app when connected to a Mac.
    3. Throttling happens too when the iPhone 6 Plus is attached to a 2.1A Apple original charging source.
    4. All my family members and many friends with iPhone 6 Plus experience speed throttling more severe than iPhone 6s.
    So, all of the battery myth told by Apple and you, fail to give an explanation why my iPhone 6 Plus has been throttled!
    I don't care about "Apple damages." We have no connection to them.

    You're confusing normal operation with throttling a CPU because of a bad battery.

    1) Yes, that's normal. The processor doesn't run at 100% all the time. It runs at full speed only when it has to to (guess why) preserve battery life.
    2) Okay, see point 1
    3) Okay, see point 1. And, connected to a power source changes nothing as far as CPU speed goes.
    4) Okay, see point 1

    Your phone is operating as normal. Mobile processors aren't binary in speed state - they aren't not running or at full throttle with nothing in between. They have an entire range of speed settings, as you can see in the Geekbench graph peaks, measuring the throttling with 10.2.1.
    edited January 2 chiaStrangeDaysClarityToSeeGeorgeBMacdws-2
  • Reply 15 of 52
    ivanh said:
    Sorry Mike, I know that your intention in writing this article is to help Apple to contain damages. But,
    1. Apple slows down my iPhone 6 Plus (it) from 1400MHz to 936MHz randomly, while no app was running (besides CPU DasherX with which used to show the CPU speed). It was in room temperature, i.e. not in “extremely cold” temperature, but it was getting hot by itself that trigger me to test it.
    2. The iPhone 6 Plus is only 10 months old and the battery condition is 94% as revealed by coconutBattery app when connected to a Mac.
    3. Throttling happens too when the iPhone 6 Plus is attached to a 2.1A Apple original charging source.
    4. All my family members and many friends with iPhone 6 Plus experience speed throttling more severe than iPhone 6s.
    So, all of the battery myth told by Apple and you, fail to give an explanation why my iPhone 6 Plus has been throttled!
    Calm down dude.Coconut Battery & 3rd party apps, don’t have the same level of access as Apple does .They will release an update with capacity information soon.
    But I believe,you need a restore from backup.Upgrading without restore does seem to slow down some phones.
    A wipe from Settings & a restore from iTunes or iCloud should fix most issues.
    edited January 2
  • Reply 16 of 52
    ivanh said:
    Sorry Mike, I know that your intention in writing this article is to help Apple to contain damages. But,
    1. Apple slows down my iPhone 6 Plus (it) from 1400MHz to 936MHz randomly, while no app was running (besides CPU DasherX with which used to show the CPU speed). It was in room temperature, i.e. not in “extremely cold” temperature, but it was getting hot by itself that trigger me to test it.
    The CPU doesn't always run at the same clock speed. The whole point of apps like CPU Dasher X is to monitor changing aspects of the system. If you go onto the App Store right now, you'll see that one of the added features for that app was changing the frequency listing to include both Max Frequency and Current Frequency...so it's normal for it to change. 
    chiaStrangeDaysClarityToSee
  • Reply 17 of 52

    78Bandit said:
    Apple engineers knew enough about the A8 chip to claim 25% CPU and 50% GPU performance increases over the A7 while needing half the battery power.  I can't imagine they ever thought the phones would start unexpectedly shutting down after only a year or two.  It was only after the unexpected shutdowns continued to become more prevalent and they introduced specific diagnostic software in iOS 10.2 to try and figure out what was going on that they identified the voltage problem.
    I think you're overstating it. Apple has said that the voltage issue that could result in shutdown (outside of a defective battery) is either an old battery (capacity below 80%), battery with low charge (20% or lower seems correct according to that chart in the article), or a cold battery. I don't think those are really "unexpected" per se when it comes to Apple engineering. I'm sure they already knew those scenarios were possible with lithium ion batteries. There's always a possibility that the 6/6s SoCs are a bit more susceptible to those types of issues though...last two A series designs prior to the introduction of the power efficient cores used in the Fusion/Bionic.
    StrangeDays
  • Reply 18 of 52
    ivanhivanh Posts: 179member
    Very informative. Thanks 😉
    It's misleading and de tour the attention that other iPhones and other reasons of throttling down the performance.  
    ClarityToSee
  • Reply 19 of 52
    ivanhivanh Posts: 179member
    dws-2 said:
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    "The best article", in what?  I guess it's the best article to contain the damages to Apple. It's the best article to divert your attention that other iPhones, e.g. iPhone 6 Plus, have also been throttled.  It's the best article to sell the possible huge battery inventory left in stock and waiting to expire.  Perhaps there are too plenty of iPhone 6s replacement batteries and not much of iPhone 6 Plus. 
  • Reply 20 of 52
    Mike WuertheleMike Wuerthele Posts: 3,129administrator
    ivanh said:
    dws-2 said:
    This is the best article I've read so far. It's nice to get more information. Keep up the good work.

    "The best article", in what?  I guess it's the best article to contain the damages to Apple. It's the best article to divert your attention that other iPhones, e.g. iPhone 6 Plus, have also been throttled.  It's the best article to sell the possible huge battery inventory left in stock and waiting to expire.  Perhaps there are too plenty of iPhone 6s replacement batteries and not much of iPhone 6 Plus. 
    Did you even read the responses to your comment about the nature of mobile processors and how they work in real life? Your unique take on a conspiracy theory doesn't work.
    edited January 2 StrangeDaysjwdawsoClarityToSeeGeorgeBMac
Sign In or Register to comment.