I have failed to set up the failsafe which seems to be linked to my particular setup. I have a Taranis Tx, FrSky X8R receiver feeding a Pixhawk via the Sbus or serial RC connection. I have set the failsafe on the TX to no pulses, rebound and set the failsafe on the Rx for no pulses as well. The S bus output from the Rx viewed on a 'scope clearly alters in some way on loss of TX and the PixHawk reports loss of Tx. As far as I can see the Rx individual channels go into Hold as shown on the Mission Planner Failsafe setup screen. The individual channel outputs on the Rx operate correctly and go to no pulses.
That behavior had me fooled a while back.
Probably not explaining this very well but AFAIK if you select “no pulses” as your FS action, then the X8R will include a special FS code in the sbus stream to trigger it. The RC level inputs are not relevant.
What I have found is that although the PixHawk (with 3.6.7) recognises the radio failure the RC channels go into Hold and don’t seem to trigger the failsafe. As I read the manual the Throttle channel 3 must reduce below the FS_pwm setting tio trigger a failsafe. But, it does not change so no failsafe is triggered.
It doesn’t work that way with no pulses. It sets a SBus failsafe flag and the channel 3 setting has no effect. The channel 3 setting is only used if low throttle signal is being used to trigger failsafe instead of no pulses.
It is sometime since I fiddled around with this but remember observing on the scope, as you commented, a change in the rc pulse lengths but not “no pulses” when the tx was switched off. That was when I read about the special sbus failsafe code that is sent by the X8R to trigger the failsafe.
I am using Matek Wings on two different machines and although I have never had a real signal loss, simulated testing in the field has always worked as expected ie the aircraft return home in one piece.