I plan on environmentally sealing the acoustic horn array plate (the SketchUp model) with this (500nm!!!! What???) mylar film https://www.ebay.com/itm/Polyester-Mylar-film-5-microns-thin/113169978510
The horn plate will be CNC'ed out of an aluminum plate with v-bits and a quick SolidWorks session.
Given the bitstream is 4MHz, the single 12mm spacing phase shift corresponds to an angle of 0.4degrees. Across a row of 16, that's a maximum of 0.025 degrees or 90 arcseconds. If I wanna hit 1 arcsecond steering, I'm gonna need to expand the array to, you guessed it, 60 feet wide. Yikes. Assuming I stay with the current size, across a 30 degree field of view, I could pull off a 1024x1024 image, but I might need to increase the FPGA size to a 100K chip.
What are the benefits of such a dense array? Well, imagine the beam-forming to be similar to fitting points to a polynomial curve. The more points you have, the more bounded the curve becomes. Similarly, you can get finer grain measurements from such a dense array. The same can be said for patch antennas.
I'm a bit terrified of the design- what was supposed to be a clean fanout to the mezzanine connectors turned into a swastika-looking motif. I'd be delighted with any ideas about how to eliminate that pattern!