PLL, measuring spurs etc:
I tested fractional mode on the ebay ADF4350 evaluation board.
I wanted dBc (relative to main carrier), did some coding on my xpsa spectrum
could well be wrong, but anyways,
I only see the phase discriminator 100 kHz reference generated sidebands appear at
and that is because of the wide loop bandwidth.
The noise floor is set by the sdr DVB-T tuner sticks, this R820 about a few
dB better than the E4000 (-44dB).
So what is below the noise floor could be anything.
To me 39 dB down(in power) is OK,
but I still have to find the components in the loop filter in this multilayer
where everything has other numbers, and possibly even other values,
than in the AD reference board.
So that will take some Ohm measurements on that board to find the right SMDs.
The requirement here is -30 dBm or -50 dBc highest value applies (regulations),
but that goes for out of band.
This 100 kHz is way in band (1.5 MHz bandwidth), so it all depends on if it has
an effect on the QAM constellation, next thing to be tried before 2016.
This shoot not bee compared to lasers trying fuzzion, or to time with nuts.
Hey, but it is interesting to play with this.