Two friends have asked me to assemble dual-band 2.4GHz/10.4GHz patch/horn antenna feeds, known aa a POTY (for Patch of the Year).
The first one I made was from 1.5mm brass sheet, and I made it on the lathe and milling machine, but for these, a group buy was arranged on the UKMicrowaves group to get them cut on a CNC laser. The cuts are very slightly oversize, by about 0.2mm, but the edges are very clean and only needed minimal fettling to remove the inevitable pip at the start/finish of each cut. I had to open up the centre holes using a deburring tool to make the fit perfect. I also had to mark up and drill/ream the hole for the socket. I made a custom collet for the previous one, it supports the reflector at exactly 90 degrees. It is relieved near the solder joint.
It was a lot of effort and actually it would have taken about the same amount of time to machine them from scratch, but for folks without a machine shop, they are excellent.
The Radiall N socket my friend supplied is not the usual small-format N with 4.1mm PTFE sleeve over a 1.27mm gold-plated brass pin. It has a solder spill. I had to cut it back a little and solder in a suitable extension pin. Again, a lot of work for no benefit.
In addition the N socket and the TNC supplier by the other friend for hit POTY both have a 1.5mm raised boss behind the socket.
The N connectors I use are like this:
On the 0.9mm laser-cut material, that means the boss would intrude into the space between the element and reflector. I made up a precision washer to ensure the boss was flush with the reflector surface.
I soldered the reflector to the tube at the right spacing from the machined top edge of the tube (6mm+0.9mm+3mm) using 183C tin-lead solder paste, which makes a neat joint,
I used a 3mm U-shaped spacer to fix the driven element in place and clamped it, then soldered it. to the tube. Then I stuffed paper towel into the tube and soaked it with water to prevent the existing joints from melting, then soldered the N connector and spacer to the rear of the reflector.
After a good cleaning with isopropyl alcohol then hot water and detergent, I fitted the 10GHz dielectric lens and connected the 2,3GHz port to an HP directional coupler, signal generator and spectrum analyser
So far, I have not been able to get the correct double-dip response. I can get a very deep single dip to -33dB (!) but even doing asymmetric bends or dielectric slips, I can’t split the dips. At max RL, the axial ratio is pretty much that of a linear feed, and with it much flatter, the axial ratio is still around 6dB. I am making another of the original patter with 1.5mm brass parts and machining them on the lather and mill, with a good small-format N socket with 4.1mm PTFE over 1.27mm brass pin so see if I can find what is going on. Then I need to make another with these TNCs…