LAB NOTES: The 3-Wire CFA
During this last month I have been working with the 3-wire CFA at 20 meters and at 80 meters. This antenna configuration represents the simplest version possible, consisting of a tuner and a 4.57-meter length of #14-2 with ground house wire. The wire is poked up a 15-foot length of PVC pipe to make it a vertical antenna. At the end of last month I learned from GARDS’ member, Heikki, OH2BGC, that this antenna will work on 80 meters. An early version of his tuner is shown in Figure 1. And coincidentally, an article of Heikki’s appears in this month’s issue of antenneX describing the evolution of his tuner and CFA work. It’s titled CFA-Tuner Development. All work was done using this circuit with various adjustments of C1 and C2. I wanted to see how the responses at 80 meters and another one I found at 17 MHz would move as the tuner was adjusted within the frequency range 14 to 14.5 MHz. The hope is that this simple antenna might be a dual band device.
In the middle of the month, a friend and I had a small exercise building J-pole antennas for 2 meters from two articles in CQ. We carefully cut the 1/2-inch copper pipe according to the formulae given and found it didn’t work below 150 MHz. We had to figure out how the J-pole was supposed to work and using the Smith Chart confirmed that it indeed should work as modified. At the end of the month I am downloading some Smith chart software in an effort to include a more elaborate description of the J-pole exercise in this article.
The CFA
I have concluded that the 3-wire CFA is really two different animals at 20 meters and 80 meters. At 3.5 MHz the 4.57 meter wire is electrically short, and primarily capacitive in nature.
At 20 meters the antenna is almost a quarter wave long, and the three highly coupled wires interact strongly (from DLA experiments, the coupling factor is on the order of .82 between wires). Reflections are running around with big phase changes from end to end of the antenna and the impedance presented to the tuner varies with frequency. Insertion of a 50-ohm resistor in series with one wire drastically changes the behavior of the antenna. The best SWR tunings are different if one starts with C2 at maximum capacitance compared with the settings found when C2 is started at the minimum capacitance position. A tuning is a two handed operation where the SWR null is worked down by tuning both capacitors at once.
Table 1 shows the movement and width between the 2:1 SWR frequencies for all the low SWR tunings found at each of six 20 meter settings. Adjusting the tuner 100kHz at 14 MHz moves the 80 meter tuning smoothly about 10 kHz.The 80 meter bandwidth and best SWR are essentially unchanged. The variation of the 17 MHz response is larger and more varied. This response seems to be made up of several SWR nulls very close together – as they merge the shape of the composite null becomes irregular (but below 2:1), going up and down across the band of low SWR.
Table 2 combines three tests. When taking the data in Table 1, I observed that C1 was at maximum value for several data points, so I padded it with another capacitor and tested the end points of Table 1. The response that went with the 14.0 MHz setting seemed to have moved up to the 14.5 MHz setting. With more capacitance on C1 the 17 MHz, response disappeared, leaving only the 80-meter response. With C1 and C2 both at minimum capacitance, the highest frequency 80-meter response was 3.5 MHz. Adding a 50-ohm series resistor in only one wire of the 3-wire antenna killed all responses!!
All of this suggests that the 3-wire “CFA” is more a resonant structure instead of a CFA. I understand the prime action of the CFA structure is to meet the 5 criterions necessary:
E/H ratio = 377 ohms
E and H pointing 90 degrees from each other
Curvature of E and H equal
E and H in time phase which means:
E plate and D plate time phased 90 degrees to each other
When the structure is a large fraction of a wavelength in overall dimensions, then only a small part of the antenna can meet the criteria above since the space and time phase of all the voltages and currents vary widely over the structure. Thus, CFA calculations are essentially those of lumped components, and are appropriate to easy spice circuit calculations. To adequately calculate the 3-wire “CFA” requires modeling the long, shorted or open circuited transmission line as it interacts with the tuner circuits. This requires finding the distributed capacitance and inductance of the 3-wire cable.
For all the data in the various tables in this and previous articles, the 3-wire antenna was hooked up with the bare “ground” as the CFA ground plane and the black and white wires as the E plate and D plate respectively. This put the ground plane between the other two. Heikki wired his antenna with an outside wire as the ground and the bare wire as the D plate. Making this change in wiring of the 3-wire antenna alters all the tuning settings at 3 MHz. The new configuration tunes easily over the entire 80 meter band to a 1:1 vSWR – I haven’t checked the 20 meter tunings yet, but it sure looks better on 80 meters. The fact that there are big changes when switching wires around supports the complex and coupled nature of this device.
In support of PSpice modeling the 3-wire CFA I used the MFJ analyzer to measure the impedance between the 3 wires taken two at a time. Table 3 shows the results. The first group of 3 measurements gives the capacitance between pairs of wires in the 4.57-meter antenna. The MFJ finds the reactance, then interprets the value as either an inductance or a capacitance. I don’t know exactly what goes on in the MFJ, but capacitance is easy to interpret. It is harder to interpret the “inductance” between two open but coupled wires. In any event, it is interesting to take things at face value and find a corresponding “inductance” for each “capacitance” data point, then compute the square root (L/C), which has the dimensions of impedance. The second and third groupings in Table 3 show this: The impedance may be a good analogue of the impedance of the wire pairs as a transmission line for modeling purposes.
The J-Pole
Without going into great detail at this time, the J-pole antenna is a (roughly 1/2-wave long) radiator on top of a “U” that acts as a shorted transmission line (roughly 1/4-wave long). The antenna coax is connected between the “U” uprights part way up that section. It works like this: the open circuit at the top of the radiator is transformed to a different impedance by the time delay associated with its length. Remember, impedance is Voltage/Current at each point and V is the sum of energy coming from the transmitter and passing the point of measurement and energy that went by a while ago, reflected off the open circuit, and is now passing the measuring point going back toward the transmitter.
At the top of the “U” the J pole becomes a 143-ohm impedance transmission line terminated with an open circuit with some phase angle (i.e., a reactance, inductive or capacitive). The impedance across the transmission line with this load is further transformed by the length of the part of the “U” above the coax connection. The result is a load on the coax that is, in general, resistive and reactive. Now look at the lower part of the ”U” below the coax connection. It too is a 143-ohm transmission line, with a short circuit at the end. The short is transformed by the lower section of the “U” to be a pure reactance across a 143-ohm transmission line, but with the opposite sign from the energy from the top of the J-pole. The best SWR happens when the reactances cancel and the two line impedances paralleled equal 50 ohms, the coax line impedance. The antenna is matched by sliding the coax connection points up and down. It turns out that to work, the “U” is a bit longer than 1/4 wave, and the radiator is a bit shorter than a 1/2 wave, both by about 5 inches at 146 MHz.
A neat device that does the above calculations is the Smith Chart. The Smith Chart is a piece of polar plot graph paper that can be used to show any impedance along a wire or transmission line and easily be rotated to see how the impedance is transformed. Free Smith Chart software can be found by entering “smith chart” and the criterion ”the whole phrase” (to avoid getting all the smiths of the world) into an Internet search engine. I found two or three that looked promising, but have not had time to test them yet. It’s a really useful tool that can save a lot of guessing. Try it –you’ll like it!
Originally posted on the AntennaX Online Magazine by Joel C. Hungerford, KB1EGI
Last Updated : 31st May 2024