RFM95W output power

[LogicalDave]

I'm using a HopeRF RFM95W RF module in some prototypes and the maximum transmit power I've been able to get is roughly +18.18dBm at 915MHz (see attached photo showing a CW transmission with the radio configured for +20dBm).  This may be reasonable since the maximum transmit power of the transceiver's PA is +20dBm and there is some IL associated with the module's RF switch and harmonic filter.  However I may also be messing things up with the DC power filtering (or something else).

The registers associated with power level are configured as follows: RegPaConfig=0x8F, RegOcp=0x2F, RegPaDac=0x87

Has anyone been able to get more RF power out of these modules?

[szoftveres]

there is some IL associated with the module's RF switch and harmonic filter

On top of that, I''m assuming there's also an RF cable between the spectrum analyzer and this module. The PA max. output is a function of Vdd and process tolerances, and the Siglent SA isn't an ethalon-grade instrument either.
1.8dB difference (or loss) doesn't seem unreasonable at all to me.

[LogicalDave]

It certainly could be an accumulation of loss and error. 

The SA (Siglent SSA3021X) is newish and in calibration; the total level measurement uncertainty spec is less than 0.7dB.  I have better SAs, but I have tested this one with a calibrated, leveled signal generator and it is *much* better than 0.7dB level accuracy that at 1GHz.

The vendor claims +20dBm out of the module; I seriously doubt that because that's the rated transmit power from the radio (before switch and filtering). I would have believed up to 1dB of total IL for the switch+filtering that are built into the module plus the cable, but 2dB seemed excessive which is why I asked what others are seeing.