[hpsdr] 12bit vs 16bit ADC

[Helmut Oeller]

Hi,
Hans, HB9CBU, developer of the ADT-200A, the first direct-sampling TRX for
hams, showed in his 2011 Friedrichshafen presentation the relation of
nominal bits vs. ENOB, based on AD types:
http://www.adat.ch/pub/Vortrag_Friedrichshafen_250611.pdf  (see page 11).
ENOB difference between the 12/16 bit versions is 1.4 bit, which is equal
8.4 dB dynamic difference with reference to the entire 1.Nyquist zone.
Everyone has to decide, what is sufficient for his application.
The efficiency of diversity is more complex and depends on lots of external
factors. VHF and higher shows completely other behavior than HF.
Polarization, elevation and other characteristics of the used antennas and
the steadily changing reflection effects at the ionosphere are significant.
Not the amplitude but the phase of the arriving signal is the critical
parameter. Tests with two 16 bit coherent frontends exhibited reasonable SNR
improvement of 15 dB and more. A 'noise-valley' up to 30 Kc broad looks very
impressive at the panadapter. Similar tests with a 14 bit Red Pitaya were
not so successful.  The performance of the sampling oscillator seems
critical. When we 'll have Hermes II available, we can continue with 14 bit
ADCs.

73, Helmut, DC6NY