Thursday, February 28, 2013

Antenna: Mini-loop for 20 meters

I've heard so much about small transmitting loops that I needed to eventually try one even if I have no need for a compact antenna.  I first tried it a number of years back.  I thought I'd try 40 meters since that would be appropriately challenging.  Actually, it was overly challenging and I never got it to work for transmitting, as heating of the capacitors from RF current would change the resonant frequency and drive the SWR through the roof, even at 5W.  My problem was using compromise capacitors instead of butterfly, trombone, or other high quality components.

So I'm trying again, this time on 20 meters.  My loop is made from a 10 foot piece of 1/2 inch soft copper tubing, shaped more or less into a circle of about 38 inches diameter.  It's 0.14 wavelengths long on 20 meters. This kind of antenna is fairly simple technically -- the big loop forms an inductor and a capacitor selected to resonate the inductance at the operating frequency is connected across the open gap where the ends (almost) meet.  A slight bit of technical complexity occurs in coupling the radio to the loop, but it's simple in practice.

There's lots of software to help with the design.  I used RJELOOP1.EXE by Reg Edwards, G4FGQ.  You input the loop dimensions and it gives you the capacitance required, radiation efficiency, loop current and capacitor peak voltage for a given power level, etc.  Mine shows that it will be about 52% efficient, not too bad for a small antenna.  It will have 1800 peak volts across the capacitor while operating at 5 watts(!), And the loop current will be 5.6 A.  Also, I need about 44 pF capacitance for 20 meters.

The Capacitor:

It's a known fact that you don't want your variable capacitor to have contact points such as bushings, bearings or sliding contacts that the large loop current must pass through.  But I read G4ILO's loop antenna page* as part of my research and he claimed that a good quality variable of conventional design and in good condition would work fine.  So I picked a nice one from my junk box and decided to measure the resistance across the bearings before proceeding.  I hooked up a power supply and limiting resistor to shaft and frame and passed 0.5 A DC though it. I used my DMM to measure the voltage drop across the bearings, while rotating the blades to look for bad spots.  Voltage drop was in the low millivolt range and my contact resistance was about 2.5  worst case and 1  typical.  I'm good to go!

The Other Capacitor:

My capacitor is about a 10 pF to 60 pF unit.  I mounted it in the Plexiglass plate that secures the ends of the tubing and stuck on a knob.  In initial testing, I found that I couldn't zero in on a frequency with this system.  Now the resonant frequency is 14.200 MHz, I make a small adjustment and now it's 13.900 MHz.  Playing with my RJELOOP1 program some more I see that 0.165 pF change in capacitance moves the resonant frequency by 25 kHz.  Too sensitive!  One option would be to have a 10:1 reduction drive on the capacitor.  Another would be to have a large fixed capacitor and a small value variable in parallel.  The third is just to parallel a smaller capacitor with the one I have now.  I chose the third method.

I didn't have any suitable capacitors small enough, so I decided to use a 5.5 pF to 23 pF variable and put a 2 pF fixed silver mica in series with it.  A spreadsheet showed me that this arrangement would give me 0.32 pF change for the first 30 degrees of travel from fully un-meshed, but only 0.025 pF change from 150 degrees to 180 degrees (fully meshed).  So an initial setting of 50% or so gives me a good tuning range and slow tuning.


Main & fine tuning capacitors on Plexiglass plate

OK, it's ugly but remember this is an experiment!  The big one with the knob going off to the right is the fine tuning capacitor and the smaller one with its shaft going through the Plexiglass is the main unit.  If you knew where to look, you might see the 2 pF silver mica in series with the fine tuning capacitor.

The Coupling Loop:

Literature on the subject says you can use a wire loop or a ferrite toroid threaded onto the main loop as a one-turn secondary with an appropriate number of turns on the radio side.  I'd already tried the wire loop, so I threaded a FT-114-43 toroid onto the loop. RJELOOP1 says the primary should have an impedance of two to three times that of the feeder (50 ohm coax).  So I used two turns on the radio side.

This didn't work well at all.  My null was over a 2:1 SWR and looked weird -- it came back up slowly when I tuned to the low side, instead of the sharp high-low-high null I expected.

So I went back to the wire loop.  It's just that - a loop of wire connected across the coax feeder. The loop is positioned just inside the main loop, diametrically opposite the capacitor. Nothing connects to the  main loop here, so I used a piece of insulating board secured to the main loop to support the small loop and to mount a coax receptacle on.

The sources I read said a circumference of 1/5th that of the main loop would be good, so I used 24 inches of solid #12 AWG THWN.  I had a hard time getting a good match.  Rotating the loop out of the plane of the main loop is supposed to be one tuning method, but it made things worse.  Not sure what else to try, so I tried squashing the small loop down into a football shape and got some improvement -- down to 1.5:1 at resonance.

My RJELOOP1 program had given a value of 19 inches for the small loop, so I took off five inches.  That made things even worse.  I made a new loop, this time going UP to 26 inches wire length and was able at last to get 1.1:1 at resonance.  I've read that the size of this loop isn't too critical, but I guess I have to say that's not my experience.

Now everything is ready ...


Matching loop, 26 inch circumference


Hanging from a fiberglass pole
off the upper deck of my house

Trying it out:

I made my first declaration of success before I even managed a QSO.  I was able to adjust it to the desired frequency and it stayed there over the course of a couple hours as I transmitted and was idle and it hung out in the sun and cold February wind.  Eventually I did manage a QSO with another QRPer out in New Mexico.  He was weak here and I was weak there but we managed a 20 minute QSO discussing small loops and EFHWs.

I measured my 2:1 SWR bandwidth at 22 kHz, which looks adequate to me.

What next?

If I want to actually use this antenna, I should get back to one capacitor and put on a reduction drive.  Also it would be good if the capacitor were in a weatherproof box.  As-is, I need to keep this antenna out of the weather.  A way to do remote tuning would be good -- maybe use a screwdriver type drive.  Manually adjustable would be OK, but for a big project, servo controlled nulling would be really nice.

I also need to see if it will play on 18 MHz and 21 MHz.  My capacitor should have enough range, but I'm not taking it for granted.

73-

Nick, WA5BDU 
Russellville, AR
February 28, 2013

* http://www.g4ilo.com/wonder-loop.html