ArduinoCore-stm32l0
ArduinoCore-stm32l0 copied to clipboard
GrumpyOldPizza
Join for AU915 not respecting subband setting ?
Is it possible to use this core with Arduino-LMIC rather than the LoRaWAN library?
OK. My device is getting stuck on joinOTAA for 2-3 minutes. I have tested with another node running Arduino-LMIC (on different hardware) and the join is instant. Both are next to the gateway and I can see good RSSI levels once the join has happened.
Just to clarify. The join takes time, but happens.
What region are you using ?
EU868 for example needs to respect a duty cycle requirement. So if the first join fails, it needs to wait some time.
What gateway are you using ?
Yes, join happens after time. I've been trying (so far unsuccessfully) to find other code to test with to rule out hardware issues (it's a custom board using the Murata Type ABZ module). Region is AU915. Gateway is pi/RAK831 running the multi protocol forwarder. No issues with other node running lmic which activates instantly and consistently.
On Mon, 25 Mar. 2019, 10:49 pm Thomas Roell, [email protected] wrote:
Just to clarify. The join takes time, but happens.
What region are you using ?
EU868 for example needs to respect a duty cycle requirement. So if the first join fails, it needs to wait some time.
What gateway are you using ?
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/GrumpyOldPizza/ArduinoCore-stm32l0/issues/76#issuecomment-476163054, or mute the thread https://github.com/notifications/unsubscribe-auth/AAhtKLlNov669qKPZIdERWq8EDYv-sZ4ks5vaLfIgaJpZM4cF84L .
Interesting.
What subband are you using ? Channel frequencies ?
AU915 is like US915, so speced for 64+8 channels. A typical gateway can only do 8+1 channels. So you need to do a LoRaWAN.setSubBand(subband). 1-8 selects the band, 0 reverts to a 64+8 setup.
I suspect that LMIC either by chance finds the proper subband (or you configured it) right away, or it does not respect the required backoff procedure.
I have just been doing some more testing. I can see an initial burst of current when the device first resets (which looks like a tx attempt), then it just looks like it sits idles for 2-3 mins, then another burst where it activates. It then uses SF9 for subsequent transmissions (the other node uses SF7).
I managed to get https://github.com/ARMmbed/mbed-os-example-lorawan running and joins are much quicker 4-5 seconds indicating the hardware is probably OK and there is something going on with the library.
Update: Following https://github.com/ARMmbed/mbed-os/blob/master/features/lorawan/FSB_Usage.txt and limiting to subband 2, registration is now virtually instant so inline with the other node running arduino-lmic