@a-lurker said:
So using your figure of 2.6 dB, which looks like an appropriate figure but could be higher; then this suggests, It's really a waste of time using the LNA+PA PCBs unless they are used at both ends.
I largely agree, some refinements:
If the sensor is mainly sending data and the gateway only acking, it is probably wise to have the PA radio at the sensor as I expect the data messages to be longer than the Acks
2.6 dB is not nothing, the range is extended by 34% (new_range = old_range * 1.34) assuming free space
If the receiving side sees interference (ACI), a 20 dB louder received signal is for sure a better thing
Any improvement people may witness ... is probably just because the LNA+PA PCB uses a far better antenna
Yes, good antenna's can for sure help but be sure you know what you are doing. If you use a dipole antenna and point it to the other node it will not help you and likely even kill your connection.
The antenna efficiency of the on-board antenna's can be pretty good, assuming things are matched correctly. If the external antenna is also matched correctly, it will have about the same efficiency. The same efficiency means that same total power is transmitted.
The remaining difference then is that but the on-board antenna will have less directivity so the energy is spread out more evenly where for instance a dipole will focus most energy perpendicular to the dipole. A pretty good trick, that is how we can still receive signals for the Voyager 1 that is almost out of the solar system using a 20m dish antenna. But ... do point this antenna 1 degree wrong and you will not see the signal.
