Antenna phase center 1

[biff44]

If you had a small 6 GHz antenna, and had to measure its phase center very accurately, how would you do it. I am especially interested in if the phase center shifts slightly at different angles.

 

[logbook]

If you use the 6GHz carrier at the transmit, but also split it into the receiver LO, you get a nice phase sensitive receiver. Either move the receive antenna for a minimum or use a phase shift in the LO path to do the same thing. Now, when you move the transmit antenna, you will be able to see the movement of the phase centre very accurately.

 

[biff44]

I can measure the free space phase shift as it is moved. What I want to know is HOW to move it and WHAT do I look for. In other words, do I move it a certain way, like in a circle of 4" diameter, and watch where the received phasor shows up as an ellipse of something and call the center the phase center?

 

[logbook]

This is my interpretation and extrapolation of the relevant information from Kraus, Antennas, 3rd ed.

The phase centre of the antenna is found by finding a equi-phase front in the far field. You would move the receiving antenna in a circle centred on the expected (geometric) phase centre of the antenna (your first guess). You would then change both the radius of curvature (slightly) and the centre of curvature to get a better equi-phase front. This sounds like a slow (but highly automated) process. Note that the receiving antenna is moving in an X-Y plane as well as in an azimuth angle just for doing measurements in one plane.

I suppose that at each X-Y position you should also change the azimuth angle slightly to maximise the received signal.

I have never done anything like this, so this is just a theoretical discussion on my part.

 

[Higgler]

Sounds right Logbook.
I've measured a few of them. Do you have an antenna chamber biff44? Typically you rotate the antenna, measure phase, then slide it back or fwd and remeasure phase. Then note where your antenna is compared to the center of rotation. Wide beam antennas are tricky to make a good setup, since things bounce off the absorber around them, narrow beam antennas are easier.

If your antenna isn't a monopole or bicone, then your phase will start moving/shifting at wider angles, say 60 degrees off boresight or more.
You'll typically measure an H plane phase that's longer at wider angles, and an E plane phase that's shorter.
Picture an X band horn antenna, at boresight the energy radiates from say 5 inches inside the antenna. As you move in the Eplane, the radiation is from the front edge as the electrons turn the corner, and hence you have moved 5" closer or shorter. In the H plane, the radiation appears to be from the far side of the horn since it's a flat plate with plain view to the far angle (the near wall outer edge will also radiate a little, but much less than the far wall flat inner surface) you then measure slightly longer phase delay in the H plane at far angles.

Phase center is a sum of all the energy from the antenna. That data is then combined to a single point.
The phase center should be spec'd as a single point, with +/- phase error, over a range of angles.

Are you building a phased array? If it's an array, the elements near the one you're testing affect your phase center results.

kch

 

[logbook]

On further thought my previous suggestion was over-complicated.

Leave the receive antenna stationary. Pivot the transmit antenna. You then only need a simple pivot rather than an X-Y-azimuth mount! Move (slide) the pivot point on the transmit antenna until you get a good/acceptable equi-phase front. To fully automate this you need a slide and a pivot on the transmit antenna, but there is no question of having to move radius of curvature as well of centre of curvature as I had suggested previously.

I think you could even do this test manually since it is so simple, provided you had a stick to do the rotation from a distance. You wouldn’t want an “ugly bag of water” near either antenna!

 

[biff44]

Hey thanks guys. That makes some sense. I do not have an anechoic chamber--I will be doing the test outside. I will try to keep reflecting objects far away. I do have a nice plastic pipe/rotating motor I can mount the transmit antenna on. I can just move the transmit antenna to overlap or move it further away from the rotation centerline, and whenever I can rotate it and get an equal phase at the receive antenna, then the point that is over the center of rotation is called the "phase center" of the transmit antenna. That I can do!

 

[logbook]

Of course if we were clever, we should be able to predict roughly how far to move the transmit antenna pivot point once we take one incorrect set of phase measurements. This would speed up the measurement process. Even knowing which way to move the pivot point would be a start. This should be pretty obvious from a bit of experimentation, if the theory is too hard (which it shouldn't be).

 

[VEBill]

'Phase center' seems to imply a point. Point assumes a circle. Do all antennas have a perfectly circular phase front, and thus a phase center which is a tidy point?

For 'a small 6 GHz antenna' this might not be a real concern.

 

[Higgler]

Antennas don't have a perfect circular phase front, but some are much better than others. Small antennas (monopoles, patches) on a ground plane are the best.

phase vs angle consistency; the primary criteria for use in amplitude phased arrays or phase interferometer arrays.

phase versus angle and phase stability versus frequency, primarily for Feed antennas used on reflectors historically.
Nowadays UWB antennas need phase versus frequency stability or at a minimum phase center location predictable in some simple algorithm (based on new thru-the-wall-imaging antenna requirements I've worked with).

GPS antennas for differential GPS need phase center stability versus angle over their operating band and can be much more difficult to design than just throwing a patch on a ground plane.

kch