We have a nrf9160 chip on a custom board with zephyr as the operation system. Modem firmware is 1.3.7. The board is battery powered, but can for development purposes also be power externally. The device sporadically show the following issue, both with battery or external power supply:
The device start and try to establish an LTE connection to send an udp packet. With some debug statements in the LTE event handler I see that there are some LTE events coming. After the RCC update event it seem the board is reseted and reboots. This take around 2min and is repeated forever.
The issue occur with our own firmware as well with some of the nrf samples (I tested some, not all), e. g. nrf/samples/cellular/http_update/application_update. With the nrf samples I see also the output "power-on reset" even with an external power supply which is able to provide several amperes. Thus it is hard to imagine there is some power problem.
The issue is sporadic with both the hardware and the software unchanged. We did not see it for weeks, but yesterday it was visible again all the day. I tried to find a reason for this issue and I exchanged the software, the SIM card and tried several external power supplies. I have a working device and I also took the firmware from that device and put it on the rebooting device. The issue was unchanged. It continuously rebooted when establishing the LTE connection. I let the device run with the nrf sample until this morning. The output showed constant reboots every 2min. Then I re-flashed our firmware on the device and the issue disappeared - without changes in the hardware, power supply or modem firmware. Thus I would exclude any reason concerning the hardware or the power supply. It seems there is some reason for this issue in the modem or the modem firmware. Unfortunately I'm not able to reproduce this issue. It is only visible from time to time, but when it is visible it seems to be stable.
Do you know anything about this issue or something similar? All I could find up to now was some power supply issue or an application bug, but I think I can exclude these reasons here.
Note: we have the nrf9160 in production use and plan to use this chip in several thousand devices. If we cannot solve this issue sufficiently we are forced to redesign the hardware again and use another chip.
Best regards, Torsten
