[1000mp] Oooops, re diversity

[Tom Rauch]

In haste to send an e-mail to Anders, I clicked on what normally 
reply to original poster in my program and out it goes to the 
reflector. Worse yet, I did that twice.

All that aside, I don't want to sound like I am being critical of 
Anders but rather was trying to get Anders or someone to start a 
dialog that will result in including technical information on how 
diversity works. Diversity is finally getting some attention, the new 
Ten Tec rig seems to have more focus on true diversity than anything 
we have had available so far.

My thinking was that if we know the goals to move us towards optimum, 
and the reasons some systems don't work as well as others, then 
anyone experimenting can make better choices and understand the 
reason for results.

In quick summary here is what I hoped we would always include in any 
technical article :

1.) The receivers should be as similar as possible in bandwidth, and 
as narrow as possible for the mode. 

The reason is noise power is directly proportional to receiver 
bandwidth, while signal level is constant as long as the signal is 
narrower than the bandwidth. If you have a 100Hz filter bandwidth and 
a 100Hz signal bandwidth, and switch to a wider 1kHz bandwidth, you 
increase noise ten dB.

(This is most of the reason why people think PSK is significantly 
better than other modes like CW. The digital system uses a ~70 Hz 
filter in the computer but and the "ear" listening hears the signals 
through a 2.1kHz SSB bandwidth of the receiver. Switch to a 100Hz 
filter and copy a slow CW signal at slow typing speeds, and that 
"apparent" advantage evaporates.)     

If you mix two receivers and one has significantly wider bandwidth, 
you are adding needless noise to the "system".

  2.) Your brain will work better on signal near noise if the noise 
is very different sounding on the two antennas, and the signal is the 
same.

That requires the antenna "hear" different phase noise but the same 
or at least a slowly changing phase of signals. The effect is like 
the noise sounds hollow or moving, while the signal sounds the same. 
This effect is unique to wide spacings, and lets you dig deeper into 
noise (good CW ops can actually copy signals at or below noise floor  
level).

I'm not sure how this would work on SSB, maybe someone else has done 
it.

3.) The antennas, while having similar beamwidths (so they have 
similar S/N ratios), have to respond differently to fading or you 
lose QSB diversity. 

You can do that better by putting the antennas far apart, but is also 
does work with antennas close together if each antenna has 
defined nulls in the desired direction that are at very different 
angles. This is what Anders does.

Like with antennas, we can never build a perfect system. A perfect HF 
diversity system would have smaller optimized diversity cells 
with phase locked identical good receivers spread over hundreds of 
miles, while a useless one would have a wide receiver on a poor 
antenna augmenting a better receiver on a better antenna with the 
same pattern and with the antennas only a fraction of a wavelength 
apart.

All of us would be somewhere between these two extremes, but it is 
helpful to always know what optimum and zero are so we can 
move towards optimum.

Ten Tec recently seems to have made an effort to REALLY allow us to 
have the better equipment without homebrewing, and hopefully Yaesu 
and others will enter a new race so our equipment gets better...not 
prettier or just more complicated with less performance.




73, Tom W8JI
W8JI@contesting.com