I think iPhones don't run directly off external power even when plugged in -- they charge the battery, and run from that. For evidence, note that if your battery is completely dead, you need to provide power for a few minutes to get it some baseline amount of charge before you can turn the phone on.
It feels plausible that a battery with bad peak voltage could still be unreliable even while being actively charged, so it'd make sense if the throttling remained.
The battery is either charging or discharging on net.
I sometimes see colloquialisms like "The laptop isn't really running off AC power, the AC is charging the battery and the laptop is discharging it to run". I also see people say that it's bad to keep a phone or laptop on its charger all day, and I usually do that to no ill effect, so I stopped trusting common wisdom on batteries.
When you get down to the cell level, each cell has 2 terminals, and it's either charging or discharging, right? Unless the charge circuit is doing some green-blue trick to charge half the cells and discharge the others, nothing that runs off a battery can be running AC "through" the battery. They taught us in A+ class that UPS' run power "through" the battery but I think it's objectively wrong. Especially if it's a lead-acid battery which I _know_ only has two terminals.
You can draw a boundary around the battery and charging stuff and say, if 90 watts is going in and the computer is pulling 80 watts, the battery must be charging at 10 watts. If the computer is pulling 120 watts, the battery must be discharging at 30 watts. But there is not 90 watts of power flowing "through" the battery in either case. Adding a load just causes it to charge slower, or discharge instead of charging.
It's probably true that the power supply only converts from mains to whatever the battery charger wants, and there's no path from mains to the load without going through the battery's voltage, but as other posters said, the reason you can't run some phones with a dead battery is more related to peak load. I have definitely owned mobile hardware that worked _pretty well_ without a battery, but worked better with a battery to smooth out load during startup or CPU load spikes.
Figure 1 shows the architecture pretty much everything uses. It's done this way because it's simple and effective.
Basically ReactiveJelly has it right and there is KCL at a node that joins the DC output of the power adapter, the system and the battery. When the adapter is present, if the system draws more than the adapter can supply, the difference comes from the battery, discharging it. if the adapter can provide more than the system needs the difference goes to the battery, charging it.
The main issue is modern systems can change their load way faster than the adapter can respond. Therefore it's up to the battery to make up the difference during these rapid transients preventing system shutdown. As long as it doesn't happen too much (and the system might choose to throttle to guarantee this if is not true) the battery charges on average.
In the end the systems are a complicated thing because they are balancing: safety, battery lifetime, wanting to support many sizes of adapters, thermal considerations, charge time and system performance. And everyone has their own opinions on how to prioritize things.
I’v found this out when looking at what are the purpose of various chips on MacBook Air motherboard.
And suddenly things that didn’t make sense now makes sense: laptop with a dead battery is unbearably slow due to CPU throttling to 600 or 800Mhz. Unusable piece of trash until battery changed.
Most phones don't run directly from mains with an empty or removed battery.
The reason is that the power supply isn't guaranteed to be able to cope with the short peaks in power like from the various radios. It could lead to brownouts. And it's not a common enough usecase for manufacturers to bother supporting.
> And it's not a common enough usecase for manufacturers to bother supporting.
Woah now, I would like to disagree with that, phones running out of a charge is a pretty regular occurrence. In fact, I'd say the only reason people panic so often to keep a charge is because the fucking things won't turn back on immediately. shakes fist at sky It would be fucking GREAT if I could use it (turn it on) immediately upon plugging it into a power source. I have been pissed off about not being able to use my iPhone (or any device) while plugged in since "they" (electronics manufacturers) started doing this shit. Truly, I've seen grown ass people thrown into a panic because their phone is about to die and I promise 80% of that panic (for a plethora of reasons from silly to necessary) would be alleviated if they knew they could just plug the damn thing in and have it turn on immediately.
I have one phone that once got into a reboot loop: its battery was flat, so I’d provide it with a USB power source, it’d charge the battery for about five seconds, then it’d try to start up automatically (without me telling it to), but that would stop it charging the battery from which is all it seems to be willing to draw from at that stage of boot, so after another about five seconds it’d die of a flat battery again, then after another few seconds it’d recognise the power source and start the loop again. The entire cycle took about 17 seconds, like clockwork. Not sure quite why it went like that; historically it had been willing to charge from flat, and since then I think it is too, though I haven’t used it much for the last couple of years, and I can’t remember whether it normally powers on automatically in this case. I tried almost all combinations of four USB cables and five power sources (Surface Book, car with 12V → USB adapter, Surface Book AC adapter which has an additional USB outlet, Surface Dock, and a wall charger) before finding some combination that worked, not triggering the untimely and defective power-on.
I will say that I've used a smartphone for about a decade now, and I can count on one hand the number of times it has actually run out of charge. (With hitting <5% being about as infrequent.)
A battery management IC will be able to discern a battery's health.
Not saying you're wrong, but it could easily be the case that the iphone runs from wall power if it detects a healthy battery, but doesn't allow that if the battery is very dead.
It feels plausible that a battery with bad peak voltage could still be unreliable even while being actively charged, so it'd make sense if the throttling remained.