[TowerTalk] Combining antennas

[Jim Lux]

At 02:02 PM 2/3/2004 -0500, Pete Smith wrote:
>At 10:51 AM 2/3/04 -0800, Jim Lux wrote:
>...
>>So, the SNR will be 0.5Sa/.707(Na+Nb).  If Na=Nb, this is .5/1.414 or 
>>about 4.5 dB worse than it was before.
>>
>>If the noise is correlated, then it will either be as bad as 2x or 
>>completely nulled out, depending on the phasing.
>>
>>If you have a good hybrid junction, you can get more of the power to flow 
>>to the radio, rather than the other antenna (assuming everything is 
>>"well-matched").  Assuming an ideal hybrid, then, the SNR at the radio will be:
>>
>>Sa / (.707 *(Na + Nb) )  for uncorrelated noise (a 1.5 dB hit) , or, 
>>Sa/(Na+Nb) for correlated noise (a 3 dB hit)
>>
>>In more practical terms, what you're doing is creating an antenna with 
>>more "capture area" with two lobes.
>
>If I understand correctly, in domestic contests where signals most often 
>are a LOT above the ambient noise level, the receiving penalty will be 
>acceptable.  Once in a while, for example when working a weak back-scatter 
>signal, it will be important to find the single antenna that gives the 
>best copy.  Happily, this is also probably the situation where you would 
>concentrate all your transmit power to the best antenna too.  But for 
>maximizing signal over the widest possible geographic spread, being able 
>to feed two or three antennas at once will be a good thing, as Martha 
>would say.
>
>Are these fair conclusions?


I think so.. What you've basically got is an antenna with "adjustable" 
beamwidth, which can be very handy.  Add adjustable phasing, and you could 
have a "interference nulling" system, which is even handier.

I have to say that all that I said above is nice and theoretical, and I've 
never actually tried it.  I suspect that more than one person on the list 
has tried this and can tell you if theory and reality meet.