828668 - prototype remote power solution for phones

Status: RESOLVED DUPLICATE of bug 917378

(Testing :: General, defect)

Description

[(not currently active) Ted Mielczarek]

We do some automated testing on real phones (Eideticker, Autophone), and everyone who does it has experienced issues with reliability resulting in phones needing to be manually power-cycled.

I did some investigation, and it looks like we can probably automate this. There are two potential routes we can take:
1) Automated button pusher to simply press the power button on phones. This might get us most of the way there, as the failure modes experienced are often "phone doesn't come back up after a software reboot", and powering the phone on via the power button works. bc has seen instances of phones needing their batteries removed and replaced to get out of a particularly hard lockup, so this wouldn't fix that.

2) Hard-wired power in place of battery + automated button pusher. If we take the battery out of the phone and replace it with hardwired DC power, it becomes simple to automate "taking the battery out" by simply cutting the DC power. We'd still need the automated button pusher component because phones will not power on automatically when you restore battery power. The question here is whether all the phones we care about will accept direct DC power without caring whether a real battery is attached. Most smartphone batteries nowadays have 3 or 4 pins, where the extra pins are often a thermistor to measure battery temperature and a data pin with an I2C bus.

As a variant on 2) wlach suggested that we might be able to use the existing phone battery and wire it to the phone terminals but with a relay in the middle so that we can cut the power as desired, but with data pins untouched. This might be the sweet spot here, where the phone will still see a real battery but we can kill the power easily.

I'm going to do some experimentation here and see what works and what we might want to do. If this looks feasible, we will probably find someone else to do the actual work.

Comment 1

[(inactive) Jim Chen [:jchen] [:darchons]]

I can help with this if needed! Putting the EE degree to use :)

Comment 2

[Jon Hylands [:jhylands]]

Attachment: Battery Harness

Note that I am already deep into building a solution for this. I have a 3D printed harness that currently measures current draw of the battery, and in the next revision (which will be started next week) will also provide a remote-disconnect feature for battery power. We (performance and automation teams) are investigating having a bootloader change applied so that reconnection of the battery will boot the phone, which will remove the need for a remote button-pusher.

Comment 3

[Jon Hylands [:jhylands]]

Wow, I just looked at the dates of the first two messages, and got the month/day backwards, so I thought this was a recent post, and not something from nine months ago...

Comment 4

[cmtalbert]

(In reply to Jon Hylands [:jhylands] from comment #3)
> Wow, I just looked at the dates of the first two messages, and got the
> month/day backwards, so I thought this was a recent post, and not something
> from nine months ago...

Jon, I think we should dupe this bug against the relevant one for the work you're doing.