FrSky apparently misled people somewhat by saying that with a redundancy hookup, control switches to the rx with the stronger signal. The way it is actually working approximates that so it's only somewhat misleading and simplified but signal strength doesn't determine which rx is in control. If you think about it, it really couldn't work that way, nor would you want it to. That behavior would require that the primary have access to the secondary RSSI and typically it does not. It would require some clumsy algorithm for determining which signal was better and for how long before switching. And what about 3 rx? Also, signal strength isn't a great indicator of link reliability, just better than nothing. Signal strength can be very high while interference is wiping out your control packets. Control packets can be coming through fine with low signal strength in a quiet RF environment.
Here is how it works. If the primary rx has a good packet from the tx it uses it and ignores the secondary input. If the primary detects a bad packet it ignores that and uses the frame from the secondary sbus. It's that simple, and that determination is made on each frame. Even if the primary has a strong signal it will still drop a few packets. When that happens it takes the frame from the secondary. If the signal to the primary deteriorates it will replace more and more frames until eventually all frames are from the secondary. There is never a point where it "switches control" to the secondary.
The reason I can make these statements with confidence is that they can be seen clearly in the Snoop logs to which Andrew referred. The full write-up is here:
https://www.rcgroups.com/forums/showpost.php?p=46575667&postcount=66
What was done was to contrive a setup in which the two receivers were listening to different signals from the same tx. A test channel was established such that the primary rx was getting a constant high value on that channel and the secondary rx was getting a low value. By logging the sbus out of the primary it could be seen which rx was the source of each frame. I'll attach graphs for two segments of the log file. In the first graph the primary rx has a good link and most frames are derived from that rx as can be seen by the graph being mainly high with only an occasional low value indicating the frame came from the secondary. In the other graph the opposite was true, the primary had a poor link and nearly all frames were coming from the secondary. All-in-all this seems like a superior methodology to the popular but wrong idea that is somehow switches control from one rx to the other based on signal strength.
It works this way with both ACCST and ACCESS.
--------------------------------------------------------------------------------------------------------
Nothing above has anything to do with telemetry which is completely unrelated electronically. I don't know too much about how the telemetry works with recent firmware but there have been some changes. Here are some things I think I know:
* It would be impossible for telemetry to indicate which receiver is in control in the above scenario since frames can come from either receiver determined every few milliseconds. The telemetry is much slower and couldn't keep up. Perhaps the best that could be done is which rx is providing most of the frames, or a percentage.
* It used to be necessary to shut off the telemetry on the secondary receiver to prevent the secondary telemetry from interfering. Newer ACCESS systems get around that and can provide telemetry from both (3) receivers.
* Mike Daily provided a hookup for "Trio" telemetry with a redundant rx hookup. This setup connects all s-port sensors and the receivers in parallel. Here:
https://www.rcgroups.com/forums/showpost.php?p=47398449&postcount=8918
* I've done some superficial testing of this hookup and the external sensor data seems to come through to the transmitter just fine with telemetry active on both rx. (I tested it with two R6 ACCESS receivers, a Variometer, and a Current sensor)
* In testing the above, I tried disconnecting one rx at a time. The external sensor data was received by the tx with only a momentary pause during transitions as a receiver was connected or disconnected. It's not clear how, but the receivers seem to work out which is in control of the s-port.
* I feel there is a lot I don't know about how the telemetry works with current ACCESS firmware but it seems to be evolving in parallel to Ethos.