I’m building a homemade small magnetic transmitting loop antenna that I want to ultimately operate at up to 1500 watts over the HF ham radio bands. One of the challenges with these antennas is the high voltages present at the variable tuning capacitor, which range from thousands of volts at 100W to tens of thousands (up to 20kV) at 2KW and vary based on frequency, which can range from 1 MHz to 30 MHz.

The prevailing designs use vacuum variable caps, which are expensive and air variable caps, which are quite large. I need a variable capacitor capable of 10 pF to around 650 pF that can handle 25 kV or more. It must be tunable via a stepper motor to quickly adjust the antenna to resonance.

My design consists of 5 copper plates of at least 16 sq in (4” x 4”) spaced 0.06” apart, using 24 gauge (0.025”) copper sheet with oversized, 0.03” thick plastic sheet (PET) as dielectric, producing variable capacitor with up to 669 pF capacitance at 22.5 kV dielectric strength. The design assumes oversized plastic sheets are glued to each copper plate after plates are attached to copper plate base via brazing and solder. The plastic sheets act as insulators and capacitor dielectric material between each copper plate electrode element. PET should have dielectric constant between 2.8 and 3.4 from sources I have found.

The idea is to then use a 3D printer style stepper motor with a slide screw linear actuator or equivalent to adjust the capacitor plates in and out. The assumption is that the plastic (PET) insulated copper plates can rub closely against one another and should pose no real issue. I believe this design will be smaller and less expensive than the alternatives and provide a much higher voltage rating vs vacuum or air variable caps in the same or less space.

Will this design meet the requirements as expected? Am I on a good track here?

  • 1
    \$\begingroup\$ +1 for good home brew . \$\endgroup\$
    – Autistic
    Aug 28, 2018 at 1:46
  • 1
    \$\begingroup\$ getting 65:1 capacitance ratio is probably challenging. \$\endgroup\$
    – Henry Crun
    Aug 28, 2018 at 2:59
  • \$\begingroup\$ Consider Teflon as well. It's self lubricating and has higher resistivity than PET. Also don't neglect protecting past the sides of the plates so you don't end up producing a thin air gap when they start to separate. \$\endgroup\$
    – K H
    Aug 28, 2018 at 4:13
  • \$\begingroup\$ A "small" magnetic transmitting loop won't have much range unless you are able to get the electric field intensity right - using plate capacitors doesn't achieve this as far as I know. Maybe you have a trick? \$\endgroup\$
    – Andy aka
    Aug 28, 2018 at 12:02
  • \$\begingroup\$ The plan is to use 3’ to 4’ diameter loop made of hardline coax as the main loop. I’m using Heliax AVA5-50 hardline, which is working beautifully in early prototype. The loop efficiency varies from 30% on lower bands up to 90% or more on higher bands. These loops operate primarily off the magnetic H field instead of the electric E field, which makes them much quieter in today’s typically high electrical noise world. The capacitor is used to tune this high-Q resonant LC tank circuit into an effective antenna. These antennas have less gain than dipoles but much better signal to noise ratio. \$\endgroup\$
    – rbraddy
    Aug 28, 2018 at 16:01

1 Answer 1


I made some loops with the objective of as close to zero cost as possible, for use with modest powers. The loop was thick bamboo tied with bicycle tube rubber, tension spokes to an axle rod (think penny farthing bike wheel) and wrapped with aluminium foil.

I tried several capacitors. Two sleeving PET drink bottles with foil glued on them as a trombone capacitor was one. A tubular butterfly capacitor was another (again foil on PET bottles).

A more expensive but much better cap was a (rotary) single vane butterfly capacitor using two sheets of FR4 PCB material with the butterfly copper pattern and Kapton film (used for 3D printer beds) as the dielectric. For my loop it was about 5" diameter from memory, and I put a small amount of dishing preload on it. (push rotor disc into the stator disc) (silicon spray to improve slip) (push-pull dyneema dial cords to rotate it)

The point about butterfly capacitors is that you do not have a moving electrical connection to the rotor.

As long as the area is symmetrical between the two halves (true for rotary , false for linear), then the voltage is split between acoss the two halves of the butterfly, through two sheets of dielectric for double voltage.

Theoretically this is over 20kV for 2x 50um.

Who knows what cheap 3D printer Kapton achieves in practice. If you have two pinholes in a single sheet, and each is on different halves of the butterfly then you have dielectric failure. If you have two sheets of Kapton (one on rotor pcb, one on stator pcb) then pinholes are unlikely to overlap, and you can assume you have 1 sheet of dielectric strength (x2 for equal butterfly)

At a rough calculation, a 200mm diameter single butterfly has approx 5000mm^2 area per half, and with 2x50um Kapton film, gives you roughly 600pF

  • \$\begingroup\$ Very interesting. Looks like Kapton has a higher dielectric constant at 3.9 and slightly higher dielectric strength of 7,700 V/mil. I didn’t realize the voltage rating could double like that. \$\endgroup\$
    – rbraddy
    Aug 28, 2018 at 1:52
  • \$\begingroup\$ Actual plate spacing is higher: some air gap even when pressed together,+ thickness of the acrylic adhesive. I see the 200mm wide on aliexpress is 50um \$\endgroup\$
    – Henry Crun
    Aug 28, 2018 at 1:57

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.