EFHW Experiments continued

I now experimented with different combinations of Coupler link turns and counterpoise lengths but keeping the antenna wire constant at 66ft 6 inches. I am now starting to get a feel for the variables.

I shortened the length of the counterpoise to about 6 feet. After slightly adjusting the resonant frequency for 7020Khz by adjusting the turns spacing on the toroid I was then able to improve the SWR down to 1:1 by reducing the turns on the primary link winding down to two turns. However the bandwidth was reduced significantly (I did not measure it) and the circuit was very difficult to keep stable as far as resonant frequency was concerned and as far as SWR was concerned. Small changes to the inductance resulted in big changes to the resonant frequency and SWR. I then went the other way and increased the turns to 4. With this setting I had to increase the length of the counterpoise to over 25 ft in order to achieve an SWR of 1:1. However the circuit became much more stable and the bandwidth expanded to easily cover the whole 40meter band.

So what is happening? Which setting is optimum? Does it really matter?  Looking at the AA5TB Figure 3 graph I note that as the counterpoise is lengthened that the end fed impedance would rise towards 5000 ohms. This would mean an increase in turns ratio would be needed to get to 50 ohms. So for 18 turns on the secondary I would need 1.8 turns. If I shorten the counterpoise then the end fed impedance would drop to say 1800 ohms. So for 18 turns on the secondary I would need 6:1 ratio which would mean 3 turns.

This result does not appear to make sense. Will think on this further.