antenna-theory.com

[Rykk]

I'm currently helping a start-up company that makes Quadrifilar Helical antennae and have encountered a possible problem with their very low budget testing process - or, most likely, my marginal understanding of these antennas.

We're currently testing two 965MHz RHCP turnstile antennas for use in testing the quadrifilars. They are set up facing each other with one acting as the transmitter and the other as the receiver. The antennas are connected to the 2 ports of a Vector Network Analyzer. They are made of two y shaped copper tape dipoles in a sort of "x" pattern laid onto a plastic (lexan?) sheet with a reflector below. The lengths of one of the two "v" elements of each dipole are shorter to supply the polarization. They are being fed thru RG316 coax to an SMA "rocket launcher" connector. One set of 2 elements connects to the conductor of the SMA connector and the other to the ground. The antenna is not being fed in quadrature. The distance apart is 1 meter and two meters. Elevation patterns are tested by turning the receive antenna on a PVC mast. The test area has no anechoic material as yet and is merely just a carpeted room with a couple of tables and a bench. Very crude, I know.

Here's the source of my puzzlement: When the antennas are facing each other in the same orientation, mirroring each other, I get a forward S21 response curve on the VNA of a certain amplitude. But - if I rotate one of the antennas 90 degrees such that now they aren't oriented long elements to long and short ones to short ones but rather are now long to short, I get a curve a good 8-10dB higher in amplitude. When I rotate the one to 180 degrees - or upside down in relation to the transmit antenna - I get a curve very close to the initial one in amplitude. Go another 90 degrees to 270 and I, again, get a curve much higher in amplitude.

So, do you reckon that these antennas have a problem, or is this merely an expected result due to the phase relationship of the radiated wavefront of the transmitter illuminating the receive antenna which is physically 90 degrees out of "phase" with the transmitter? I have noted this same behavior to a somewhat lesser in amplitude difference when rotating a quadrifilar antenna that's receiving it's signal from one of their turnstile antennas, as well.
Thanks,
Rick

[bigSteve]

What you are describing is exactly what I would expect from two linearly polarized antennas. I suspect your RHCP antennas aren't really circularly polarized

[Rykk]

I'm thinking you may be right. The engineer I'm working for is a brilliant guy but I'm thinking the design of these is all wrong. I know that, with a quadrifilar helical antenna, polarization is achieved by shortening one of the filars on each of the two "dipoles". All of the examples of CP turnstile antennas I've found on the 'net have one of the dipoles fed 90 degrees lagging by using a 1/4 wavelength length of coax. I'll have to bring this up to him.Thanks,
Rick

[Rykk]

Ok, I think I've answered my question:

1) A Turnstile type antenna must be FED in quadrature for it to be circularly polarized. My boss's are not.

2) When measuring the response for a CP QHA and using a spinning dipole as the transmitter, the gain (Axial Ratio?) will increase by a good 16Db or so in the 90 and 270 degree positions. If the turnstiles were RHCP, the difference throughout 360 degrees would be - nominally - less than 1Db.

Therefore, these two turnstiles are NOT circularly polarized and - being linearly polarized - would exhibit the behavior I've seen in the lab when using one as Xmit and the other as Rcv and rotating one of them.
What do y'all think - Am I correct in my conclusion?
Thanks,
Rick