Worked fine for ~1 month, then stopped transmitting for no obvious reason. Physically intact, opened the case, disconnected/reconnected the battery - no signs of life. The charging circuit works because the red LED lights up when connected through the USB port and eventually turns "off".
No other signs of life. Any ideas of what might have gone wrong? I am mostly concerned not so much with reviving this specific unit as I am with figuring out the root case to prevent this from happening in the future (unless it is just a random failure - AKA "bad luck").
Thank you for any hints...
Is it able to change modes? Unplug the battery and hold down the MODE button while plugging the battery in. If the LED goes into blinking magenta (red and blue at the same time) and eventually blinking yellow, and stays in blinking yellow when released that would be good to know.
Thank you! It did actually wake up when I followed the procedure you described. After white LED it went into blinking green and connected. Maybe I released the button too early, but it did come back to life as if nothing happened... A mystery. I have a watchdog enabled, so it should have rest if it hung somehow.
Any ideas as to what could cause the unit to get stuck with the white LED "on" and then to not even respond to disconnecting / reconnecting the battery? Just trying to figure out how to avoid this in the future.
It's not clear what happened. My only guess, and it is just a guess, is that you ran into the nRF52 power down leakage current issue. If the MCU has the power removed from the 3V3 rail, but there is still power on the GPIO, the device will fail to reset. Also, it won't respond to the reset button until power is completely removed for long enough for the MCU to discharge, which takes a while because it has a low leakage current.
This occurred on the Tracker One and it would occur if there were peripherals attached to the M8 port with external power that were connected to the I2C, Serial, or GPIO. If the data lines did not de-power when the 3V3 rail was powered down, the device would enter the unable to reset state. We solved this using an analog switch to disconnect the ports and GPIO on 3V3 power-down.
The Monitor One does not have this circuit because there are too many GPIO lines on the expansion interface, but it is generally less likely to occur, at least the with the I/O card. If you've implemented a custom card with external power that has access to GPIO this is a strong possibility.
OK, thank you. So it is a latchup issue of sorts. The internal cap long discharge time is something I thought about after the fact, when the unit did come back to life after the MODE button trick, but sadly not before, so cannot try and replicate.
Yes, it has a full custom card with LORA radio, display capacitive buttons high precision A/D converter and a bunch of other stuff. The only external power source is coming in through the Monitor One charging circuit.
