I do not have any experience working with traps in HF frequencies, but it was obvious the main problem will be the capacitor. A trap is a parallel connected LC at the top of an antenna radiating some power, so it was obvious the capacitor would need to withstand very high voltages. How high? I don't know. Maybe 5 kV or maybe 20 kV. A very high value in any case.
After googling a bit, I found three solutions to the problem:
1: Use a "doorknob" type high voltage capacitor.
2: Use a double sided PCB to construct the capacitor
3: Use coaxial cable to construct the trap.
I quickly discarded the doorknob capacitor. I do not have any, they are expensive and bulky. I need something that fits into a restricted area (a pair of cm at most).
Using a double sided PCB to construct a capacitor sounded quite good. I used a 1.6mm thick double sided FR-4 and tested the capacity. I got 2.6 pF / cm², a value in good consonance with this page from G3NGD.
Using the coaxial trap calculator from VE6YP I got the L and C values for a 2cm diameter coil: 30 pF and 0.31 μH, so I cut a small piece of PCB and trimmed it to 30 pF.
Then I added the coil, and using the spectrum analyzer and the poor man's tracking generator I construct some time ago, I quickly tuned the trap to 51 MHz. and everything seem to work ok.
Because I got some time, I tried the option 3: Coaxial cable trap. Using the calculator from VE6YP I found all I need is 4 turns of RG-58 over a 2cm form. It was easy.
Then I tested the resonance frequency. It was a bit off, at 49 MHz, but separating a bit the winding did the job.
Have you noticed it? Yes. The coaxial cable trap has a notch nearly 20dB deeper than the one made with FR-4. Why? I really have no idea. Both traps have the same value, the one calculated with VE6YP coaxial trap calculator: 30 pF and 0.31 μH. Coaxial traps are know to be lousy , but it seems FR-4 based traps are even lousier, at least at 50 MHz. Because I will use no more than 100 watts with that experimental antenna, I think I'll use the coaxial cable made trap for the experiment.
Miguel A. Vallejo, EA4EOZ