Because the TV transmitter is hanging from a balloon, the old setup used a cloverleaf antenna. This gives a radiation pattern in the form of a half sphere, pointing down. I am told that the advantage is that the radials are in a 45 degree angle, and the field has circular polarization in the vertical plane. Because the field is "both horizontally and vertically polarized" (note the quotes), the receiver antenna can be in any position, and there is less fading on the signal when the transmitter antenna tilts or turns in the wind.
For the new clover leaf antenna I used Oscar Lang's clover leaf calculator to calculate the length and position of the elements. My first attempt resulted in an antenna which resonated about 60 MHz higher than calculated. At these frequencies (1 GHz and up), dimensions and material can make a lot of difference.
After changing the size and experimenting with the construction I came to an antenna which resonates exactly at 1252 MHz, and has almost -30dB return loss:
The final version of the antenna has a PCB mounting point for the radials. This adds capacity and throws the calculation of the resonant frequency off, but it was worth the struggle:
The end result is of course covered in epoxy resin to make it waterproof. The radials are made of regular copper wire. The wire was twisted with a hand drill compress the molecular structure, which makes the wire more rigid. Thanks for the tip, Remco!
Of course, building an antenna like this would have been impossible without Mischa loaning me a spectrum analyzer with a tracking generator covering these frequencies. Playing with that tool was a lot of fun and I've definitely added that to my wish list :-)