Increasing the value of DRV_USART_BAUD_RATE_IDX0 to, say, 921600 helps making the logs a bit more readable, though still not perfect. These problems might also be caused by my UART-to-USB cable. And of course, the same setting must be used when capturing the output, like in LOG="ttn-gateway-date +’%Y%m%d-%H%M%S’.log"; pio serialports monitor --raw -b 921600 | while read l; do echo "[date +’%F %T’] $l" | tee -a $LOG; done, or when using a Raspberry Pi to capture the output.
This allows for changing the log level to SYS_ERROR_DEBUG in:
…and for enabling some disabled logging in the original code, and adding many more calls to SYS_DEBUG(SYS_ERROR_DEBUG, ...).
I added the following at the end of sendCommand in src/app_lora.c to know when things go wrong (not sure why the code uses status *= ... which will still try all next configuration steps if a previous one has already failed):
SYS_DEBUG(SYS_ERROR_WARNING, "LORA: sendCommand %s\r\n", gotresponse ? "OK" : "ERROR");
…yielding:
CNFG: Configuring LoRa module
LORA: Changing state from 2 to 4
LORA: Starting reconfiguration
LORA: send_cmd: 0x23 0x31 0x1 0x0 0x0 0x55 0xd
LORA: recv_rpl: 0x23 0x31 0x1 0x0 0x0 0x55 0xd
LORA: sendCommand OK
LORA: send_cmd: 0x23 0x3a 0x1 0x0 0x0 0x5e 0xd
LORA: recv_rpl: 0x23 0x3a 0x10 0x0 0x1 0x1 0x4c 0x47 0x38 0x35 0x30 0x31 0x36 0x30 0x31 0x37 0x38 0x32 0x4 0x1 0xd
LORA: sendCommand OK
LORA: version: 01
LORA: configureRXChain(0, ...)
RF: 0,1,867500000
LORA: send_cmd: 0x23 0x34 0x6 0x0 0x0 0x1 0xe0 0xff 0xb4 0x33 0x24 0xd
LORA: recv_rpl: 0xd
LORA: sendCommand ERROR
LORA: configureRXChain(1, ...)
RF: 1,1,868500000
LORA: send_cmd: 0x23 0x34 0x6 0x0 0x1 0x1 0x20 0x42 0xc4 0x33 0xb8 0xd
LORA: recv_rpl: 0x23 0x34 0x1 0x0 0x0 0x58 0xd
LORA: sendCommand OK
LORA: configureIFChainX(0, ...)
Above, the call to configureRXChain(1, ...) actually consistently succeeds. Or maybe the if(appData.rx_uart_buffer[1] == command) in the call to configureRXChain(1, ...) is seeing the response of the earlier call to configureRXChain(0, ...), as it somehow gets in too late? The if(TIMEOUT(4)) is not being hit. After the above lines the log is a mess again, but I see both OKs and ERRORs.
I need to run now; just posting my early findings in case it helps someone right now; will try to trigger the default configuration later, to ensure the config that is fetched from the internet is okay. To be continued…