Total cancellation in corporate feed network for patch antenna

Question

I'm trying to build an array of patch antennas in order to receive a GPS signal. The idea is to build 4 patches and connect them with a corporate feed network. An example of corporate feed is this one.

In my design the 4 patches are oriented in 4 different direction (eg. the first one looking at EAST, the second WEST, the third NORTH and the last one looking at SOUTH).

Now I have two question about this design:

1. The output are not matched (since I'm using T-junctions) so some of the power received will come back at the antenna and will be re-radiated. Is this a problem?
2. Can I have total cancellation if, for example, two patched receive two identical signal with 180 degree of phase shift between them?

Unfortunately I can't use other antenna design such as dipole antennas or helix. The problem is that this antenna will fly on a home-made rocket: the rocket will spin fast after lift off so I decided to design this array in order to achieve the best omnidirectional antenna possible while preserving the aerodynamics of the rocket itself.

The rocket is fully metallic, so, no antennas inside.

Answer 1

I recommend eliminating the array altogether and instead put a single patch antenna inside the rocket body facing upward.

The frequencies of GPS signals will freely penetrate a non-metalic body, nose cone, and parachute. Place any other electronics below the patch antenna so as to not obscure its view of the satellite constellation.

A patch antenna in this orientation will be largely insensitive to rotation of the rocket body and the patch will be positioned with its maximum gain facing the constellation.

The mission weight savings will be an added benefit.

Edit:

The OP added the condition that the entire rocket is of metal construction. The Haigh-Farr WrapAround Antenna is specifically designed for sounding rocket purposes. It is generally machined into the metal body of the rocket. In addition to GPS reception, the device can incorporate telemetry antennas.

Other sounding rocket designs have encapsulated a helix antenna in a synthetic tip of the metal nose cone. This has proved suitable for the trajectory through apogee. Switchable patch antennas are then used to track the decent trajectory.

Answer 2

On 1) let me take you through a though experiment I created, struggling with the same problem with a ring of antennas fed to make an Omni. You can back it up by calculation later.

Antennas are reciprocal devices. Gain in transmit is the same as gain in receive. If you transmit into the bottom port, the corporate feed / matching network works perfectly, power splits evenly and each patch radiates the same power, in phase (assuming no coupling between antennas). No need for proper isolated Wilkinson splitters with resistors; the resistors would not dissipate any power.

So why worry about power reflection in receive? Just take it on trust that when receiving, the gain will be patch gain - 6 dB (assuming the patterns don't overlap, etc)

The picture when receiving is more complicated - what happens to a signal arriving on only one patch, as the network is far from matched this way around. You can work out the impedances if you're careful. The answer is that some of the power is reflected back to the receiving antenna, some is re-radiated from the other antennas in the array, and only 1/4 of the power comes out of the feed at height bottom.

Can you do better? No, not without a switch. Using isolated Wilkinson splitters, you know you'll suffer 3 dB each time, the power being lost in the resistors. There's also no matching network that can efficiently combine power from any of four patches, to a single output, without loss.

If you trust the reciprocity, you don't have to do the analysis.

In summary, some power will be reflected, but it's not a problem, and there's no way to reduce the reflection/absorption losses in the network.

Answer 3

The output are not matched (since I'm using T-junctions) so some of the power received will come back at the antenna and will be re-radiated. Is this a problem?

A problem for you? It will reduce your received power, of course.

A problem for anyone else? It is no more problematic than any random bit of RF-reflective metal in the environment — better, in fact, since your receiver and line loss will attenuate the signal.

Can I have total cancellation if, for example, two patches receive two identical signal with 180 degree of phase shift between them?

Yes. Every antenna design has nulls somewhere, for some combinations of transmitter direction and polarization; this is a fundamental property of receiving coherent radiation. The only complete solution is to use two or three antennas (possibly built into a single system that has multiple feed points) each with independent receivers.

Antenna design for a single antenna is a matter of arranging so that the nulls land in a place you don't care about as much.