The RFM9x datasheet seems to imply that 20dbm at 1% duty cycle is the maximum Tx power setting for RFM9x, yet the documentation in the RadioHead .h file suggests (with caveats) it may be possible to get as high as 23dbm:
/// We have made some actual power measurements against
/// programmed power for Anarduino MiniWirelessLoRa (which has RFM96W-433Mhz installed)
/// - MiniWirelessLoRa RFM96W-433Mhz, USB power
/// - 30cm RG316 soldered direct to RFM96W module ANT and GND
/// - SMA connector
/// - 12db attenuator
/// - SMA connector
/// - MiniKits AD8307 HF/VHF Power Head (calibrated against Rohde&Schwartz 806.2020 test set)
/// - Tektronix TDS220 scope to measure the Vout from power head
/// \code
/// Program power Measured Power
/// dBm dBm
/// 5 5
/// 7 7
/// 9 8
/// 11 11
/// 13 13
/// 15 15
/// 17 16
/// 19 18
/// 20 20
/// 21 21
/// 22 22
/// 23 23
/// \endcode
/// (Caution: we dont claim laboratory accuracy for these measurements)
/// You would not expect to get anywhere near these powers to air with a simple 1/4 wavelength wire antenna.
I'm leary to try it, so I thought I'd first inquire: is it possible to get any meaningful extra range from it at those higher settings?
Also, what sort of antenna would be a lot better than "a simple 1/4 wavelength wire antenna" at getting "these powers to air"?