Those are hard questions to be answered. We should take into account at least
- physical network layout (which nodes can TX and/or RX to each other)
- all network parameters
- overall load on the network
- duty cycle restrictions
-
In theory, if I’m not mistaken, the first step could be counted by: [GPS fix time ms] + [random TX delay ms] + [packet TX length ms] + ( ( [random TX delay ms] + [duty cycle wait timer ms] + [packet TX length ms] ) * [number of nodes/hops and needed retransmissions] )
-
The same as the first one.
-
a) Initial message: [random TX delay ms] + [packet TX length ms] + ( ( [random TX delay ms] + [duty cycle wait timer ms] + [packet TX length ms] ) * [number of nodes/hops and needed retransmissions] )
-
b) Acknowledgements: [random TX delay ms] + [packet TX length ms] + ( ( [random TX delay ms] + [duty cycle wait timer ms] + [packet TX length ms] ) * [number of nodes/hops and needed retransmissions] )
All message types are treated the same way. In addition, all messages are, more or less, flooded at the moment. Therefore, currently most important factors are network parameters (SP, Spreading Factor), and the amount of hops/retransmissions needed. Duty cycle limits are not enforced yet.
There might be flaws in those formulas, and hopefully someone corrects me if needed.
More about airtimes can be found here: Antenna improved range