[TowerTalk] Fw: Feeding single band HF yagis 500+ ft from the shack

[Lux, Jim]

On 5/18/22 11:02 AM, Jim Brown wrote:
> On 5/18/2022 10:33 AM, Lee Hiers wrote:
>> K1DG has a presentation he's done for Contest University.  Per his study
>> over time, he estimates that a 1dB improvement in your station 
>> results in a
>> 6% score increase in a contest.
>
> N6ZFO, who worked as a statistician in big pharma before retiring, did 
> a statistical analysis of Sweepstakes scores comparing high power to 
> low power (1,500 to 100W), and came up with a comparable number.
>
> My own observations were from 160M contests, when I was S&P up and 
> down the band without retuning my amp. The difference in power output 
> varied by 1 - 1.5 dB. I would call stations I'd miss on a second pass, 
> and failing to work them, retune to get back to full power and work them.
>
> Yes, as Jim Lux says, could be prop, but this happened often enough 
> for me to believe it was probably power.
>
> Copyability of a signal is largely signal to noise ratio. In my 
> professional life in pro audio, I did voice paging in office buildings 
> in the presence of background noise (largely HVAC systems, but also 
> office noises). When balancing those systems, a change of 1-2 dB often 
> made the desired difference. 


I suspect that it's SINR - interference is a big deal (e.g. breaking a 
pileup at the Rx end, which you may not be aware of at the Tx side) - 
and a 1 dB change could turn you from -0.5 dB under the interferer to 
0.5 dB over the interferer, and that makes a big difference in "Pdetect" 
- even if you ignore ionospheric fading. 0.5 dB is 12% louder


The Cocktail Party Effect (or selective attention) is what is operating 
here.  And it's well established that a "familiar voice" is worth quite 
a bit (or, I suspect, one's fist, if operating CW). Another key aspect 
is "trigger words" or specific phraseology - there's a reason why ATC 
talks with pilots using a standardized form, and, of course, classic CQ 
calls or contest exchanges are a similar thing.   From a straight level 
standpoint one can look at what's called the "Masking Threshold" - I'm 
sure someone has done it for speech, but most of what I see is for 
tones. And your ear is *really* good at separating closely spaced tones, 
some plots on Wikipedia (from Ehmer) seem to show that a few Hz is 
sufficient to reduce the masking level to 0dB for tones at 500 Hz.   
We're also really good at separating spatial sources, so binaural 
listening fed from two different antennas might also be very good.