Bangle2 died on low power - did not shut itself down?

[espruino-discuss2]

Posted at 2022-12-15 by Kosi2801

Firstly thanks for the effort in the previous months to reduce the power consumption of the watch. When I bought it in February it ran a full week on a single charge and I expected that to go down over time when I added more apps/watchfaces.
Instead in the previous run it managed to not bother me for two weeks straight 🙀

Maybe that's a reason that yesterday I forgot to charge the watch and unfortunately it ran flat. What surprised me is that it apparently did not shut itself down in time and I also could not turn it off anymore when I realized I was on 0% (but still living).

When I tried to get into the Launcher menu the screen just garbaged up and I couldn't do anything besides reloading which then corrupted the screen a bit differently again.
My assumption is that the power was so low that either the CPU or the screen drawing did not have enough power anymore to redraw the full screen and in that state I could not reach the "Turn off" menu anymore.

I would have assumed that the watch turns itself off before it reached such a low state but either it wasn't there yet (but already pretty much unusable).
Or does this self-shutoff maybe not exist at all (or is broken?)?
If the power runs out (or goes super low) while the watch is still alive, is there a risk that it corrupts its settings or other storage?

Is there another way to shut down the clock without the settings-menu? I understand that different timings of pressing the button are just different stages of "resets" but I didn't want to reset just shut down the watch safely.

---

Attachments:

• edited_Screenshot_2022-12-15-08-05-57-58_b3878e0b713ed77b76a4f3099169c9cb.jpg

[espruino-discuss2]

Posted at 2022-12-15 by @gfwilliams

That's an odd one - what did the corrupted screen actually look like? Is it possible the watch display just got really cold in this weather?

Usually there's a little more power left after it reaches 0% (and the watch will just work normally reading 0% until it turns off) but the hardware should automatically shut off when the battery reaches ~3v, which is well within spec for the components.

To get low enough for real problems, I'd expect from that Gadgetbridge graph that the charge display would have stayed on 0% for some time

---

Posted at 2022-12-15 by Kosi2801

Unfortunately I did not make a picture of the screen. It was definitely not connected to temperature as I was inside the whole time and the temperature-value (displayed as part of my LCARS watchface) always showed 30-33°.

But I can describe my actions and what I observed:

• start at normal LCARS watchface, locked. no visible distortion.
• press button. lock widged disappears
• press button. "loading" box appears for a bit then the upper ~1/4 of the screen looks like the Settings/Launcher menu but with small quirks here and there. Below that a part of the "loading" box remains a bit distorted and on the remaining screen there are partly color spots/leftovers from the LCARS watch and partly stripes/dots that look like from the Settings/Launcher menu.
• press button for 2s. In the middle of the screen some movement and 1 or both vertical lines for the "loading" box appear, but only part of the content.
• no matter what clicking or touching the screen I did it stayed corrupted.
• after ~2-3mins the screen recovered and showed the LCARS watchface again properly.

---

Posted at 2022-12-15 by @gfwilliams

Thanks, wow - that is strange...

after ~2-3mins the screen recovered and showed the LCARS watchface again properly.

Was that on charge, or just normally?

It sounds a lot like the communications to the screen were a bit dodgy. Hopefully it won't happen again, but please keep us updated if it does!

---

Posted at 2022-12-15 by Kosi2801

I did not constantly monitor the screen but just noted that after 2-3mins it was ok again. I did not do anything on the watch in that few minutes (no screen touch or button press).

My assumption is that because of the low power something did not correctly work/communicate anymore, probably as you said "just" the communication to the screen.
And when the backlight was off and the battery could recover for a short time again, maybe it just raised the power enough to pull off a full redraw again.
Yes, this has never happened to me before but usually I also don't allow the battery charge to go below 10%. Just this time I completely forgot.