And here are the results.
The first picture is with power level 3 when above 300MHz, second picture is with powerlevel is 0 for all frequencies.
The noise is much higher with low power and that can be improved by running from a well regulated 5V supply.
The scale of the logmag S11 is different so do not compare the angle but only if the cable loss is a straight line (which has not improved.....)
So, although running at low power has a big impact as the uncalibrated jump at 300MHz is now almost gone, but this was mostly hidden by the calibration, There is maybe some impact above 600MHz. A bit less wild maybe.....
--
Erik, PD0EK