As discussed elsewhere, an omnidirectional CP antenna is nice for a robot to use. A popular configuration is the Lindenblad, which is basically a circular array of tilted dipoles. Since I already needed a plastic bucket, the thought of building the antenna by putting copper foil tape on the surface of the bucket is appealing. I wanted to work with my existing ATV transmitter, which can work at 434 or 426 MHz, conveniently centered around 430 MHz (the middle of the 70 cm amateur band). A bit of fooling with NEC2 resulted in the following model file:
CM bucket lindenblad test
CM 4 elements on a circular form, tilted 30 degrees
CM form is 13 cm radius (10 inch diam)
CM Free space
CM Pattern calculated at 430 MHz
CE Evaluated at 410 through 450 MHz
GH, 100, 41,.47, .148, .13,0,.13,0,0.005
GM, 100,3, 0,0,90,0,0,0,
The angle of the elements is 30 degrees relative to the horizontal, and the height of the "winding" is 14.8 cm. An immense amount of cut and try got the length adjusted so that the reactive component disappeared at 430 MHz, leaving a resistive component of 91 Ohms. The VSWR (assuming a 91 Ohm feed for each element) is below 1.1:1 from 420 to 439 MHz, so the antenna is fairly broadband. The pattern shows 1.4 dBi of gain at the horizon, dropping to 0 dBi at about 40 degrees away. The radiation is RHCP, with the axial ratio being 0.80 at the horizon and 0.86 at 40 degrees elevation, favoring the horizontal polarization for both.
Over a ground plane, however, the pattern looks a lot worse. The whole idea of a CP antenna is to reduce the horrible multipath you get with a linear polarized antenna. At least with this antenna, when one polarization has a null, the other polarization has a peak, so a 45 degree angled antenna might work fairly well. With a lossy earth (the real case), the nulls won't be as deep.
Pattern GIF(10K) Pattern PNG(10K)
The question will be: how to feed this antenna from a nominal 50 ohm source? A quarter wave transformer using 75 ohm coax would transform the 91 ohms to 62 ohms. However, we have to feed all 4 elements. Using a pair of pieces of 75 ohm coax in parallel improvises 150 ohm transmission line. This transforms the 91 to 247 ohms, which, if we feed 4 in parallel will be about 62 ohms, a match of 1.22:1, which is probably ok. Usual foamed dielectric cable TV coax has a velocity factor of 75%, so a 1/4 is going to be around 70/4 cm or 17.5 cm (6.9 inches).
Another approach would be to make the dipole elements folded, which effectively transforms their impedance by a factor of 4, for a feed point impedance of 360 ohms. Quarter wave lines of 300 ohm twinlead would transform this to 247 ohms, which paralleled gives you the same 62 ohm feed impedance.
Or, how about this, connect a half wavelength long TL between each feed. The TL can be any impedance. If you connect the 4 elements all in parallel, the impedance will be 1/4 of the feed point impedance: 23 ohms, give or take. One can do the folding trick to get the impedance up to 91.