Oh, I got it. You added a SPDT switch and two control PIN for separating the TX and RX signal again. Good solution once the circuit could not be modified. The only thing needs to take care, the input 1dB compression point for this switch is 39 dBm @ 5V control signal or 35 dBm @3V control signals. Make sure use 5V as control signal for 6W PA.
If the circuit could be modified, we could just place the PA between the Lora IC and the switch like the HMC544A. Under this configuration, we still only need one antenna and just two control pins.
For the high-power RF switch, we may have more choices. I already started to measure the RF switches due to global chip shortage, will also update the results (Technical Information: rfmeasurement:testing_fixture_6_lead_qfn_spdt_rf_switch [Unit Engineering Wiki]).
Instead of using two outputs from the esp32, I used a sn74h04n to take a single gpio and give me the two signals required for the HMC544A. This way the esp32 only needs one gpio and the sn75h04n will also switch the HMC544A much faster than the esp32 is able to do it natively.
As the result, the firmware of device with PA could be totally compatible with current release. The RXTX/RFMOD PIN of the sx1276 or similar chips could still be used to control the new RF switch with an additional Inverter chip.
My plan was to build it in a way where in order to enable this setting, the user must also enter their ham radio license as a configuration requirement.
At least in the US, using output power >30dbm requires a ham license and the firmware should support proper behavior. 6 watts is about 38dbm.
I thought you may define a RF front end identification protocol. E.g. connecting different resistor voltage dividers to a given ADC PIN of the MCU. So that, the firmware could aware the rf front end’s TX power capability at runtime. Users will be prompted to confirm that they are eligible to use high-power transmitters. However, this mechanism is easy to crack because the users have the source code in their hands.
Yeah, we can use security chip + RSA signature + online verification. But it may not be worth paying these costs. After all, the purpose of this mechanism is just showing the attitude of developers, requiring users to have corresponding licenses to use higher transmit power. So, define a protocol based on resistor voltage divider and ADC could be one of the wise choice.
I am unable to edit the original post at the top, so adding a few new observations here:
Definitely buy a NanoVNA to test your antennas. I have got a few crappy antennas.
The SX1262 T-Beams on the 868 Mhz frequency seem to work better than any other chip and frequency. This is not my experience so far, but it’s the experience of all others on the forum and Discord.
As the result, the firmware of device with PA could be totally compatible with current release. The RXTX/RFMOD PIN of the sx1276 or similar chips could still be used to control the new RF switch with an additional Inverter chip.