How does this strangely-shaped horn at Honeysuckle Creek Tracking Station work?

After reading the BBC article Australia to create national space agency and clicking around to identify the dish antenna shown in the photo in the article, I came across an image of an interesting reflector configuration called the 42 foot Cassegrain Horn at Australia's historic Honeysuckle Creek Tracking Station where the famous video of the first footsteps on the moon were received among of course plenty of other things.

The horn looks to me like a big dead-end, but of course it isn't. How does it work optically? How does it collect microwaves entering its large aperture and concentrate them into a single small feed horn? Which way is it actually pointed as shown in the photo? What are the advantages of this unusual design?

below: "Honeysuckle Creek Tracking Station" from here.

• @Hobbes that't the 2nd link in the question. The optics are still a mystery to me. It still looks like a dead-end for microwaves. – uhoh Sep 25 '17 at 7:47
• @Hobbes Bingo! OK Figure 6.5 turned upside down seems to match nicely, and so in the photo this horn is is actually pointed more towards the horizon, rather than almost vertically as I had imagined. So the top surface in the photo is a concave hyperbolic secondary. – uhoh Sep 25 '17 at 8:42
• @Hobbes when I get a chance I'll find a physical copy of the book and sit down and read it (google has too many missing pages). I don't really understand how this can be called a true Cassegrain yet. The hyperbola is concave rather than convex, and its axis at least seems to be orthogonal to the parabola's axis, not coaxial. But in the mean time if you wanted to paste a screenshot of Figure 6.5 and call it upside down, that's the "Aha!" part of the answer and certainly acceptable. – uhoh Sep 25 '17 at 14:59

That antenna is a Cassegrain horn (or 'casshorn'). This is an evolution of the earlier horn-reflector antenna (paper that describes the design), where a horn radiates into a parabolic surface:

The Cassegrain horn reduces the antenna size (long distance between the horn and parabolic dish) by introducing a hyperbolic reflector:

The feed is now at position F', but the hyperboloid is shaped to produce a virtual source at F.

The result is a beam that looks like this:

The photo is upside down relative to the drawings in the paper:

In addition to reduced antenna size, the Cassegrain horn has these advantages over a normal Cassegrain antenna:

• no blockage by the subreflector in front of the main reflector
• less feed spillover (leakage), both in transmission (less interference with other systems) and reception
• good rejection of ground noise

The disadvantages are obvious from the photo:
- the antenna needs more structure. Instead of a few trusses holding up the secondary, you get 3 full "walls". It's cheaper to build a Cassegrain with a slightly larger primary to compensate for the blockage of the subreflector.

The book Practical Conic Sections: The Geometric Properties of Ellipses, Parabolas and hyperbolas also describes the geometry of this antenna.

• Beautiful answer, thank you! Does reduced spillover advantage refer to transmitted power, or to less thermal radiation from the ground reaching the receiver? – uhoh Sep 27 '17 at 0:45
• Spillover mainly refers to transmitted power, I believe. – Hobbes Sep 27 '17 at 7:47
• I remember reading somewhere (possible the google book link in your now-deleted comment, possibly elsewhere) that an important advantage of this design was the fact that it does not view the ground, the contribution to the thermal noise was only about 2K. Irregardless of the usage of the term, the design's benefit is low thermal noise from the Earth. edit: Ah, it's mentioned in the last sentence of the abstract of your linked paper. The receive sensitivity is the motivation for this design. – uhoh Sep 27 '17 at 8:08

I'll add a little bit to Hobbes' excellent answer and detective work. I looked up a physical copy of the book linked there; Practical Conic Sections: The Geometrical Properties of Ellipses, Parabolas and Hyperbolas by J. W. Downs, Dover, NY, 1993 and found it a short but incredibly interesting and informative little book, if you like reading about conic sections.

From reading and enjoying the illustrations, I've learned that the hyperboloidal secondary does not have to be at all coaxial with the primary paraboloid. All sections of the paraboloid focus to a single point, so you can orient they hyperboloid any way you choose, so long as one of its foci coincide with the focus of the paraboloid.

Also, you can use a positive or negative hyperboloid, whichever suits your purpose. Hyperbolae come in pairs and are associated with two foci. Concave or convex, it will redirect the paraboloids rays converging to one of its focal points to a focus at the other point.

Finally, as discussed in the linked paper in Hobbes' answer, a motivation for building this unusual-looking reflector is that it prevents the feed horn from "seeing" the ground or atmosphere near the horizon where thermal noise is a serious problem. According to the abstract:

A new antenna type is described which combines the low noise temperature characteristics of the horn -ref lector antenna with the more attractive mechanical features associated with the paraboloidal reflector. Cassegrain optics used in an off-set feed configuration enables a virtual source to be formed without sub-reflector blockage. An extremely compact structure is realized with a concave hyperboloid which mirrors the actual feed located on the paraboloidal surface. Except for the aperture, the antenna is completely shielded. The design approach is outlined and measurements on an experimental model are presented. Ground noise contribution from minor lobes is about 2°K. (emphasis added)

From: A New Low Noise, High Gain Antenna S. R. Jones and K. S. Kelleher Aero Geo Astro Corporation, Alexandria, Virginia, Reprinted from 1963 IEEE International Convention Record.

below: Figures 6.4 and 6.5 on pages 49-50 of Practical Conic Sections: The Geometrical Properties of Ellipses, Parabolas and Hyperbolas by J. W. Downs, Dover. 1993