I have been reading some of the posts in this thread with a growing sense of incredulity. The claims made by Prog are either ill-informed or seem designed to deliberately mislead. I am sure that the AirSpy receiver is a perfectly fine receiver, but the claims of 85-90 dB of SFDR for the receiver are absurd and misleading.
If we consider dynamic range, I am sure that even Prog would agree that the whole reason for wanting to maximise dynamic range is to enhance the ability of a receiver to successfully receive a weak wanted signal in the presence of one or more stronger unwanted interfering signals. If we didn’t have interferers, we wouldn’t really care too much about dynamic range. Many factors come into play to determine a receiver’s dynamic range and the performance of the ADC is JUST ONE OF THEM. If the receiver uses a tuner that pre-selects a segment of spectrum by the use of a mixer, synthesizer and analog filter, then the noise performance, intermodulation performance and phase noise of this tuner will all contribute to a loss of dynamic range.
Decimation trades off quantisation noise against bandwidth within the ADC, so there is no such thing as a free lunch. It is possible to get lower ADC noise via decimation, but only at the expense of reducing the receiver bandwidth. If the wanted signal is very weak, such that the C/N ratio into the ADC is dominated by KTB (noise from the source) and the LNA NF, then no amount of decimation will improve this. Unless the SDR platform is simply an ADC connected directly to the antenna, decimation will NOT boost the receiver dynamic range UNLESS it is already limited by the ADC alone. Correctly setting the gain can optimise the trade-off of noise vs intermodulation, but no amount of decimation, or gain control can do anything about phase noise and reciprocal mixing.
Unless the tuner uses $50k Rhode & Schwartz or Agilent signal generator for the LO, then reciprocal mixing is likely to be one of the main factors constraining the receiver dynamic range. A typical modern monolithic silicon tuner is most likely to use a low power fractional-N synthesizer with a 3rd order sigma-delta modulator, and is unlikely to have phase noise of better than -110 dBc/Hz at offsets between 10 KHz and 100 KHz (in fact, it is most likely to be worse than this). Now consider two signals each of 10 KHz bandwidth and separated by say 50 KHz. If the wanted signal is say 60 dB below the interfering signal at a 50 KHz offset, the phase noise and reciprocal mixing will limit the C/N of the wanted signal at the output of the mixer to 10 dB. If this represents the minimum tolerable CNR for acceptable demodulation (as an example) then the dynamic range of the receiver cannot be greater than 60 dB. No amount of decimation or ‘smoke and mirrors’ will improve this no matter what Prog claims.
When accounting for all other factors (noise, and non-linearities) , it is very hard to produce a receiver with an instantaneous dynamic range of more than 60 dB. Having more ADC dynamic range certainly helps, but for ADC SFDRs above 65-70 dB, the effect becomes marginal.
In a ‘real world’ receiver, the only real advantage of a narrow band ADC with >70 dB of SFDR is that it helps obviate the need for a well engineered AGC loop because the receiver can operate at lower gain levels and be more able to tolerate signal level fluctuations without crashing into the ADC noise floor.
As I have said, decimation to boost ADC SFDR at the expense of bandwidth is a useful tool but adding a narrow band analog filter and gain in front of the ADC gives effectively the same result. However, unlike what Prog claims, it does not magically boost the receiver dynamic range by 25+ dB
