The program works properly joining to The Things Network using OTAA and sending a dummy payload every 20 seconds… but in the interval between transmisions, the power consumption is about 150uA… I would expect something closer to 15 uA (that is what I get on the same hardware with a similar program developep using MbedOS).
I am using the RAK3272S breakout board, so there is nothing connected to the STM32WLE5, and the st-link is disconnected during power consumption measurements (all not used pins are configured as analog inputs and debug is disabled).
Has anybody got power consumptions in the order of 10uA?
Dear @kersing It’s only a testing program…
The final node (if I suceed in solving this issue) will measure sound noise and send the average every hour.
Silly me for working with TTN since 2015 and never noticing the rule that states you don’t have observe the fair access policy when you are just testing…
Have you checked the mbed initialisation code for controller pins not being used? The state unused pins are in can make quite a difference when sleeping in low power mode. Another thing to check, are both code bases using the same low power mode? And are all peripherals shut down?
When I start to deal with power I used to keep things simple at first, then adding features one by one.
Here below my mbed deep sleep test we made when created this PR for murata device.
you may notice there is the radio instance declared because we detected this one was consuming power even if not used (this may test bad init)
The is also compatible with LoRa-E5 and RAK3172 according selection of the correct target of course
Yes, with Mbed I’ve been able to reach below-10uA consumptions, but now I need DMA and that is not easy with Mbed, so I switched to STM32CubeIDE, and I agree it’s a pain to understand the code.
The hardware I’m using is the RAK3272 breakout board (just a RAK3172 with the SWD and power pins).