Quadrifilar helix antenna (QHA ) return loss -7db unable to achieve -10db
A little jpg image of the general structure might help.
There are many forms of QFH, including types that are fed from the top, open top and bottom fed, sometimes the phase shift between the opposite arms is achieved by making one pair a loop, slightly smaller, within another larger loop that is fed.
One way was to use one of the helix as a coaxial feed from the bottom, to emerge at the top to drive opposite arms. This is the so-called "infinite balun" method.
Another way was to use a metallic tubular central support, which does not affect the radiation pattern, but provides a tube for a stepped quarter-wave coaxial matching section which has a 50-Ohms input.
The matching part of the tube, made of another tub set inside it, is contrived to have an impedance the geometric mean of the driving point impedance of the QFH at the top, and 50 Ohms. ie. SQRT(50*Zqfh).
The length of the tube is made quarter-wave for the frequency of interest.
The centre-conductor up the tube is supported by disc cylinders of PTFE.
thanks , this is my design of QHA . i obtain desire frequency , the problem is my return loss are at -7db the most , which did not achieve less than -10db . Any suggestion of vary parameter to improve return loss ?
Thanks for the picture.
The antenna you show has four part-helix legs around a square, achieving 180° from each start by going between diagonals.
This is a very loose approximation to a helix, and my first thought is that by starting in the centre of a side, rather than a diagonal, using 3 segments to make it to the opposite side, it becomes more helix-like. The feature of a QFH is the wide range of off-axis directions that still have a good circularly polarized quality. The way this happens requires the complex summation of fields from the continuously changing directions in a true helix elements on a cylidrical shape.
The classic quadrifilar helix is a combination of linear elements and loops. If the helix angle was 90°, you would have linear, and if it were 0°, it becomes a loop.
If the elements at the bottom were all joined together, and the elements fed at the top with opposite legs fed 180° out of phase, say using a rat-race hybrid, and the remaining two also fed 180° out of phase relative to each other, but 90° out of phase relative to the other pair, then the radiation pattern would be the classic "ball" shape, with no need for a ground plane (suitable for spacecraft).
By altering the aspect ratio, (how long compared to how wide), one can get the pumpkin-shaped pattern where the gain straight up to the Zenith is less than the gain toward the horizon, as suitable for satellite ground station kit.
The driving point impedance at the element feeds cannot be directly connected to standard 180° Hybrid combiners. Each has to be via a matching section - a piece of quarter-wave long transmission line of impedance the geometric mean of the combiner output, and the helix element.
If fed from the bottom, with open elements at the top, there is the problem that the impedances can get very high unless the lengths are right. There are types that sometimes connect the opposite legs at the top with capacitors that provide loop tuning. To feed any of these structures, you need the signal to first be split in a hybrid network that provides 4 signals, each delayed 90°, 180°, 270°, 360°. Then each requires a matching section to be able to drive the elements, whatever their impedance turns out to be. The match sections need to have equal delays. Most of this can be implemented in microstrip as part of the antenna itself.
In the past, I have simulated hundreds, and built only four QFH. The best was a top-fed version using a cavity resonator filter and match section built into a metal central support shaft which had ferrite chokes on the outside. The fact it was metal did not affect the pattern. QFH is such a popular antenna that you will find a huge wealth of construction and theoretical information.
The first step is to discover the impedance at the driving points of the elements.
If the impedances are not too extreme, here is some possibility you can figure a non-standard driving network that can connect directly to the elements, so that you need only one matching section to allow 50 Ohms feed.
There is some possibility you can mess with the structure dimensions, or add coupling tricks so that the elements present a convenient impedance to feed. This may be elegant, but it is very hard to do.
They are better built circular, but if you must use a square mandrel, I would at least try for more of a helix by going between the centres of the sides of the square. Also - I think S11= -10dB is not so good. A 14dB RLR (VSWR=1.5) would be better, and 21dB RLR (VSWR near 1.2) is not impossible.