Comment by 0x_rs
3 years ago
Most (if not all) current phones available in the market cannot even boot without a battery connected, is that correct? I'd feel a lot more comfortable without having to worry about it and risking forgetting about it for any unspecified amount of time just to find a pillow protruding from the back case. I really find it a shame that we have relatively powerful, energy efficient, mass produced devices in our pockets we eventually dispose of when they could have a second life serving another purpose. postmarketOS is a really interesting project, and if most phone manufacturers weren't actively user-hostile I could've used some old phone instead of my Pi(s) for whatever project I'm interested into, but I however think it's fruitless until strict regulations allow users to do whatever they wish with their handheld PCs. Due to this, among other things, I find all the talk about recycling and whatnot hypocritical from big companies, that should be the last stage of something's life, not dictated by obsolete security patches or the likes.
On the battery issue. A place I used to work had iPads stuck on meeting room doors with a webapp for booking and displaying who'd booked meeting rooms. They were puffing batteries in 9-12 months of being plugged in 24x7.
I got a few cheap power point timers, and set them to only charge for 1 hour twice a day, and we didn't have a battery failure in the 3 years after that while I was still there.
(Totally agree with the rest of your comment too...)
Was this recently?
I forget when, but in last ~3-5yrs, I've noticed iOS seems to handle this. If you leave a device plugged in for a long period you eventually getting a notification saying it's not gonna charge as much, and it caps the max battery to 80%.
Just wondering if your story is before or after that feature.
edit: I googled, it was iOS13, so 2019 - https://support.apple.com/en-ca/HT210512
Way before. 2014-ish.
Phones can't run without the battery because their peak power usage can exceed what the phone chargers provide. So the battery is used to cover these micro power spikes.
Throttle the CPU if you find that the phone or charger get warm (too much power supplied long-term) or if the phone crashes (sudden power demand -> low voltage and supply can't react).
Even with 80w-120w chargers these days?
You have several bottlenecks in the chain. The phones won't draw more than what their internal power distribution system was laid out for. If it requires a battery internally to run stable you can't just fix that by theoretically supplying more at another point. Batteries are very convenient because they can act as a buffer that is able to supply a high peak current. Modern CPUs and GPUs need a high peak current if you don't take countermeasures. Supplying that peak current from an external power supply would be quite expensive.
To rephrase your question:
Could we build a new phone that runs off a 80..120 W charger without battery? Absolutely yes.
Will it help older phones that were not designed to use such a charger? No, most likely that charger will change nothing.
Those aren't the problem. It's that you can't control what the user will do/use (a sort of unbounded problem). The 0.5w or the ultra ultra cheap chargers that let AC signal through or maybe it's not giving true/clean 5/10/20v.
Now you have to have an extra decision circuit that judges whether or not the wall is "good enough to boot the phone" and then ask if it's good enough to run it as well. Last thing you want is wall power to dip randomly and you can't switch fast enough and simply shutting off.
I understand why you would pick the battery as the requirement. You know the exact behavior of when the phone is safely bootable vs not w/ a predictable DC signal.
1 reply →
I wonder if you could make a virtual battery, that was really just a dummy load that always said the battery was at 100%, so a charger could be used.
6 replies →
I think even those could be a problem. Wired electronics contain capacitors to handle the spikes while phones use the battery as a cap.
That should be sufficient. The most peaky loads are RF transmit, can be 10-20W instantaneous. If you add that with GPU operations and backlight, camera, etc. But I don't think it'd easily go over a few tens of watts.
I use a Magisk module called ACC (https://github.com/VR-25/acc) to set the charge/discharge range to be between 60-75% for my permanently connected old phone. I'm using it to show a Grafana dash of my computer's resources.
Wait, I assumed manufacturers (or ROM makers / AOSP) did this by default by not charging it to 100%?
Then again, why would they? The faster the battery goes the sooner the end user will have to either replace the battery (which is almost impossible with phones these days) or ditch the old phone and get a new one.
There's also https://f-droid.org/en/packages/com.slash.batterychargelimit... which just needs root.
This actually happened with me :) I keep the phone wherever and don't look at it a lot; by chance I saw that the battery was bloated. The phone/server has been running fine since I bought the replacement, around 1.5-2 years ago.
Uptime has been really good, only downtime I had was when my wife disconnected the phone a couple of times and forgot to plug it back. Even when I moved a couple of months ago, the battery held, and downtime was basically just the time between disconnecting old wifi-reconnecting new wifi in new place.