[TowerTalk] 1 or 2 dB

[Lux, Jim]

On 5/19/22 11:38 AM, Jim Brown wrote:
> On 5/19/2022 6:23 AM, Lux, Jim wrote:
>> I'm not so sure that it's out of reach. yes, trying to implement it 
>> with gear from 1980 would be challenging. But with more modern 
>> equipment, where the "radio" is a black box controlled by a "front 
>> panel" or "computer" it gets easier.
>
> The Elecraft K3 with second RX that is the same as the main RX, and 
> which can be synced with the main, allows diversity reception, and 
> I've been using it since 2008.
>
> Diversity requires an antenna for each RX, spaced as widely as 
> practical from each other. It was invented in the earliest days of 
> radio to counter the effect of selective fading, which is the the 
> cancellation of two or more arrivals of the wavefront from the same TX 
> that have followed different paths, arriving at different times. The 
> time differences cause the arrivals to have a variable phase 
> relationship with each other, combining algebraically to cancel or 
> add, depending on the resulting phase relationships. Diversity works 
> best when the antennas have the greatest spacing, so that when 
> cancellation is occurring at one antenna, it is less likely to do so, 
> or even to increase, at the other.
>
>> And the diversity combining - doing it in analog is hard, but in the 
>> digital domain it's much easier, and for the most part it can be done 
>> at audio (or post down conversion to baseband or low IF).
>
> As diversity has been practiced since the beginning, combination is 
> done in the brain of the operator, with audio from the two receivers 
> in opposing ears. That's how it's done in the K3. The result is a sort 
> of spatiality to the sound, a bit like the true stereo image produced 
> by a spaced pair of microphones dedicated to left and right loudspeakers.
>
> Combining the outputs of the two receivers to a single (mono) channel 
> is problematic, because the phase relationships at audio have a good 
> chance of cancelling. 


For SSB, yes - a simple summing won't work.  But it's widely used in 
other systems where there's some processing or where the baseband phase 
is reliable  - For instance, on AM or FM, the instantaneous audio phase 
will match, so you coherently combine - typically modern diversity 
receive does some sort of weighting on the basis of SNR - the stronger 
signal gets a heavier weight, and when there's fading, it smoothly changes.

I will say that there are *bad* implementations - I had a car radio that 
did diversity on FM, but the two paths were noticeably different time 
delay (as in milliseconds) so you could hear an apparent "echo" as it 
switched from one to the other.