terrillmoore
commented
4 years ago [I really need to take the debug prints out of the LMIC -- they are a snare for those new to the code. The debug prints are only for debugging the LMIC, not for debugging your system. You have a system problem. LMIC prints are not useful for finding problems, in fact they make other problems crop up..]
Almost certainly because of your debug prints, you are not receiving downlinks. Debug prints while the LMIC is operating change the timing and the LMIC is not designed to detect that or recover from it.
Please use the compliance sketch in the examples, possibly modified if you like to up how often the dummy app transmits, and try that. It uses a logging technique, and prints when time is not critical to the LMIC.
Thanks, --Terry
dolfandringa
Describe your question or issue I have a custom STM32F103 board that I use with an RFM95 chip and arduino on the US915 plan. The gateway is 2-3m away from the node and is an orange pi zero with a rakk2245 concentrator running chirpstack gatewa-bridge and the lora packet forwarder that rakk supplies with the 2245.
I have turned on ADR on the node and use OTAA. When the node sends data, it fails to send though on the higher data rates. The gateway sends it LinkADRReq messages to use data rate 4 (500khz/sf8), but then when the node sends the next message, it fails to send on dr4 and steps down the datarate gradually until it succeeds in sending. After a successful send, the gateway often tells it to switch back to DR4 again. And the poll messages that get sent by the node as a response to the LinkADRReq messages, do arrive fine and are sent on DR4. So I wonder why it is failing to send the new messages on DR4? Could there be a software (timing or something) issue that causes this? On my RFExplorer, I only see high peaks for some of the send attempts, but not on all. And even the ones that do cause high peaks, at least don't registered as being sent successfully straight away (more send attempts follow). In the sample output below there are peaks visible in RFExplorer at 16:33:11 and 16:33:24 but nothing really before or after that.
Below I have copied the output of my node that logs changes in frequency, opmode, channel, datarate, etc and TXSTART/TXCOMPLETE events that get triggered. The POLL messages are responses to MAC commands from the gateway.
The two screenshots from the peaks in RFexplorer are attached, and I also added the saved RFExplorer output if someone happens to have RFTouch and is interested.
Environment
This information is all important; it's hard to help without a complete set of answers.