On Mar 23, 2007, at 1:22 AM, Tom Clark, K3IO wrote:
Back in Feb, Bob Davis wrote: Jim,
I would like to add to the agenda: top-level antenna configuration. I've begun to seriously consider what to build for the Eagle mockup at Dayton. Several people exchanged antenna configuration sketches late last year (September 2006), and don't know if there was resolution on A sketch. Something must be baselined soon (maybe by mid-March)), if we're to have the mockup ready for mid-May.
bob
Robert Davis
KF4KSS
Bob -- I finally "got a round tuit" on this topic :-[
This applies to both the C-band and S2 phased arrays. Instead of drawing the possibilities, I figured it was easiest to build (spare no expense!) simple models and three possibilities are attached. When we computed the needed array gain, we came up wanting about 35-40 individual patch antennas (15-16 dB of array gain added to ~5-6 dB of element gain for a single patch). The individual patches will need to be about 0.7 wavelengths in diameter; my guess is that the patch itself will be about ½ wavelength in size and it will sit inside a "cup" cavity that is about ¼ wavelength deep. This cavity (a.k.a. choke ring) will help minimize the mutual coupling between the antennas. For the two frequencies, we have
- S2 = 3400 MHz = 8.8 cm wavelength. Therefore the "cup" will be ~6.2 cm diameter and ~2.2 cm tall.
- C = 5800 MHz = 5.2 cm wavelength. Therefore the "cup" will be ~3.6 cm diameter and ~1.3 cm tall.
We will want good symmetry in the array to make the phasing as painless as possible. This led me to the 3 models I've attached:
- "Square" is a 6x6 array (36 elements). I think I have "invented" a nice way to do the phasing for a square array based on the Butler combiner matrix (the antenna analog of the Cooley-Tukey FFT) which looks like it will work well (meaning easy to phase to electronically de-spin the antenna) on an 8x8 array and I think it can be trimmed down to 6x6. Since the rows/columns are spaces 0.7 wavelengths, the diagonal element spacing will be just one wavelength, so we won't have much power wasted in sidelobes.
- "Hex" is 37 element array that is the tightest possible packing. It approximates a 4.2 wavelength dish in its properties. So far, I have not yet "invented" a Butler-like matrix for a hexagonal array, so I don't know how well it can be implemented.
- "Ehex" extends the Hex geometry by adding 6 elements (total=43) for a bit more gain, with a pattern more like a 4.6 wavelength dish.
Stan and Lew have suggested that the honeycomb structures might be done on an NC mill, which could give us a good thermal radiating surface. For the transmit phased array we have thought in terms of a ~1 watt (RF) power amplifier at each patch. Depending on what kind of DC => RF efficiency we can achieve, this will mean that we will have 20-30 watts of heat to dump from the array.
My idea, instead of milling a big block, was to use extruded aluminum cups (something like http://www.zerocases.com/catalog/deep-drawn-enclosures/round-precision-miniatures.html).
Maybe we can chat about this Saturday at the DC AMSAT meeting --- 73, Tom
<hex.jpg>
<Ehex.jpg>
<square.jpg>
_______________________________________________
Via the Eagle mailing list courtesy of AMSAT-NA