The obvious kite antenna is a random longwire, however this has some major disadvantages. Firstly, if the end is current driven then high RF currents are present next to the operator, an efficient ground system is required and most importantly, the radiating section is near the ground, not 200 feet in the air! As the wire gets longer, the antenna becomes directional - along the length of the wire, this is useless for both DX and NVIS contacts.
The end fed half wave is simply a dipole fed at one end, a voltage node. As the current at this point is very low it does not require a substantial ground. This antenna was employed by the Zepplin airships, giving rise to what has become known as a 'Zepp' antenna. In the Zepp, a quarter wavelength of ladder line is used as a transformer, being driven at one end by the transmitter (the current node) and in turn driving the antenna at the other (the voltage node). In this configuration the earth connection is just about redundant and we have a good approximation to a single ended balanced antenna!
The diagram below shows my standard set up.
The screw in 'dog stake' (available from pet shops and kite stores - for big kites use a stainless steel one) is used as both the kite anchor and ground connection for the radio and generator. The radiator element is connected to the kite line using a thin elastic cord and is held to the main kite line using plastic curtain rings. Alternatively, set up below the kite and have a true vertical. The radiator is driven by a length of ladder line. Big kites can lift quality commercial ladder line or ribbon feeder; home made light-weight feeder is better for weaker kites or low wind days. The twin feeder is connected to the ATU balun and the ground connection of the entire system taken to the dog stake. Check that there is a dc ground path from both sides of the ladder line (or long wire) , be especially careful if using a Z-match tuner rather than a T-match and balun as the antenna may be floating - when the static flashes over the radio front end is dead.
A word of warning when running QRO; corona at the tip which can cut through the kite line if the two are touching. To avoid this, leave a few inches of wire dangling or cover the tip in a blob of hot melt glue.
A prime requirement is that the kite is stable and generally the higher it is flown the more stable it becomes, as it is clear air. In the UK, the highest a kite can legally be flown without CAA permission, and in unrestricted air space, is 60m (just under 200 feet). Therefore I always fly my antenna kites to the maximum height permitted, even if the antenna wire is only to be raised to a fraction of this.
A lightweight setup
The use of ladder line or twin feed permits the radiating element to be held well above the ground, like getting a vertical dipole on a big tower. However, handling ladder line in the field can be awkward and the setup is somewhat bulky. For lightweight QRP work it is far more convenient to feed a half wave wire directly from an LC matching network, like the one shown below which will tune 40m and 30m.
It is built on a small peice of circuit board that has been tinned with solder (to cover up some over etching!).
It comprises a single 15 turn coil wound on a T80-2 toroid. One end of the coil (to the right) is connected to ground (the tag of the BNC from the rig) and the RF from the rig is fed into the coil between the second and third turn (a turn is counted as a pass through the centre of the toroid). A small 500pf compression type variable capacitor is connected across the whole coil (turn 0 to turn 15) and the antenna connection is made at the top end of the coil (turn 15). The copper beneath the capacitor has been etched away.
There is a fantastic web site about end fed antenna and matching units by Steve AA5TB and if you contemplate using end fed antennas than a visit should be mandatory! He describes a simple and accurate tuning procedure. A 4k7 resistor is connected between the antenna output and the ground connection (this simulates the high impedance of a purely resistive end fed antenna). The capacitor is then tuned until a minimum SWR is measured at the operating frequency. If you are lucky to have an antenna analyser then set it to the operating frequency and tune the capacitor until X=0. If the SWR is not low enough, or the resistance R is a long way off 50 ohms, then one can also play around with the number of turns and feedpoint. This was what I did until I got R=52, X=0 for both 7.030MHz and (with retuning) 10.120MHz, resulting in the circuit above. The matching network is then set up for that particular frequency. Next a wire of approximately the correct length (half wavelength) is connected and gradually 'pruned' until the SWR is acceptable (or R=50, X=0). It is important to make this last adjustment by changing the length of the wire and not by adjusting the matching newtork as you want to make the wire look like the resistor, i.e. a purely resistive. Obviously, my matching box only covers one band and to use another I have to go through the tuning procedure again, Steve shows a matching unit using a polyvarcon capacitor that he has calibrated for different bands and this is a good approach if you are using a multiband rig (like an FT-8178) but you still need to change the wire.
Once adjusted, I have found that the antenna is very robust and does not need further adjustment, whether used vertically from a kite or relatively low and horizontal when tied to a tree. It is ideal for kite use as there is a dc path directly from the antenna, through the coil, to the outer part of the BNC (and onwards to the rig) so a connection to ground should be made via this path to prevent static charge build-up.
Note that even at QRP powers the feedpoint voltage is very high (hundreds of volts).