RFLink - effectiveness of response processing for reliable interactions

[johngouk]

This post is a summary of my experience in setting up reliable commands to RFLink, using my setup. It ends well, with no problems, so only read if you're interested!

I'm using RFLink with an Arduino Mega, supplied by Nodo, connected to a RPi 3 running Node-RED. Frankly, I love it, it all worked right out of the box and I even managed to solder it all together without frying anything. Fantastic!

My only question is - do other RFLink users process the returned responses to commands? To what extent do packaged RFLink solutions use the returned responses?

My setup was sending a burst of 6 commands, generated by timers, and ignoring the responses, since my initial experiments indicated that worked ok. All commands are serialised through a single Serial node, and incoming data likewise. However, I noticed an occasional "20;xx;CMD UNKNOWN" coming back from what were, from a Node-RED view, perfectly formed commands. Experience with other Serial devices led me to believe that data might be being lost due the processing of received commands while others are being received - it was a theory anyway!

I implemented a Node-RED Function node, to which all Incoming and Outbound messages are directed. This node basically only transmits a command if there is no unacknowledged command outstanding, buffering Outbound messages for sending if received while transmitting. Incoming responses are used to indicate transmission completion and the commencement of another transmission if there's any messages in the buffer. Read more and code at http://arduinoarsingabout.blogspot.co.u ... olved.html.

This appears to have cleared up the errors - I now have 0 negative responses over a 12 hour period already, as displayed by the Status of my new node. The lights work as required, so all good. Lovely. Thanks for listening - Happy New Year!