search

mcci-catena / arduino-lmic

LoraWAN-MAC-in-C library, adapted to run under the Arduino environment
https://forum.mcci.io/c/device-software/arduino-lmic/
MIT License
650 stars 212 forks source link

Value of MAP_DIO2_LORA_NOP is wrong #217

Closed terrillmoore closed 5 years ago

terrillmoore commented 5 years ago

From @cstratton (thanks Chris!)

https://github.com/mcci-catena/arduino-lmic/blob/a86a87ec9c131818172a21c6e52c080ee4917181/src/lmic/radio.c#L267

this should be 0x0C, not 0xC0.

More context:

https://github.com/mcci-catena/arduino-lmic/blob/a86a87ec9c131818172a21c6e52c080ee4917181/src/lmic/radio.c#L261-L267

And per the Semtech datasheet: image

Using 0xC0 causes the value controlling DIO0 to be set to an unspecified value, here and here:

https://github.com/mcci-catena/arduino-lmic/blob/a86a87ec9c131818172a21c6e52c080ee4917181/src/lmic/radio.c#L538

https://github.com/mcci-catena/arduino-lmic/blob/a86a87ec9c131818172a21c6e52c080ee4917181/src/lmic/radio.c#L654

I can't figure out how this ever worked; the only explanation is that 0xC0 happens to cause the right value to be output on DIO0 (TXDONE if TXing, and RXDONE if RXing). I will fix this and call for testers (and do some testing myself).

terrillmoore commented 5 years ago

Also curious, the value for MAP_DIO2_LORA_NOP, if used in the proper location, is a sketchily documented value: image (SX1272 is similar.)

terrillmoore commented 5 years ago

I've pushed the fix on a branch https://github.com/mcci-catena/arduino-lmic/tree/issue217 -- I don't have hardware with me to test. I'll test tomorrow NY time, but I welcome feedback from others as to whether this breaks anything. I haven't looked at the Semtech reference code to see what they do. This might solve some long-standing mysteries, because it appears from the datasheet that DIO0 would never be activated (and thus we'd be dependent on the DIO1 signal, as we know we are from Feather work). @svelmurugan92, if you can test and report results, I'd be interested.

svelmurugan92 commented 5 years ago

Hi @terrillmoore,

Checked out this branch [https://github.com/mcci-catena/arduino-lmic/tree/issue217], build and verified in both SAMD (With Catena 4450) and STM32 (With Catena 4612) platforms. From the test results, Recent changes did not break anything. Tx and Rx works as expected.

Continue our development cycle with this branch changes. If any weird behavior, Will let you know

Thank you!

terrillmoore commented 5 years ago

Thanks @svelmurugan92. @cstratton also found https://github.com/matthijskooijman/arduino-lmic/issues/13, seems to be same issue. I'm going to start a pull request because it's clear that this is a proper change.