I got the informatlon about loop size from the EZNEC User Manual. Depending on the version, the computational part of EZNEC is that of NEC-2 or NEC-4. The User Manual is an interactive Word program that is the same for all versions of EZNEC, including the free version that is readily downloaded.
Under "Building the Model," there is "Some Special Cases," which includes "Small Loops," which is on page 58 in my printout, but can be on another page.
Under "Small Loops," it says, "NEC-2 is unable to model small loop antennas." The minimum loop circumference given for NEC-2 is .05 wavelength; but then there is a lot of waffling about double-precision versions of NEC-2, and NEC-4. Because of the waffling, I had hoped that maybe smaller loops can be simulated, but experiment indicates that they had a good reason to mention .05 wavelength minimum circumference.
Has the FCC really ever put the kibosh on a helical or continuous loaded antenna? Really? Speculation aside..... The idea that there is to much wire does not fly logic wise, if they allow any loading coil. A continuous loaded coil is just a loading coil, s long one. There is no limit on diameter is there? A foot is pretty wide but 4" dia on 9 feet long is not odd. Not that would do this (since it does not exist and too expensive) could you not have a 4" dia copper mono poll pipe for a mono-poll?
I designed a few of these antennas on paper, had hope of a little more efficient (in theory), but a 4" diameter PVC is more of an eye sore. A base loaded coil on 102" CB whip is more stealth. Even 1/2" copper is not going to block out the sun like 4" PVC wrapped in wire. Also it takes lots of wire. Dia of wire is also critical. It's lots of winding.
I played with the numbers of these helical antennas -OR- continuously loaded antennas. No surprise, no free lunch, there are pros and cons. It's not a panacea or super efficient miracle antenna, since the length is still very short. That's the limit; that's the fly in the soup.
You can design a full coil antenna to be close to resonate Freq, but you need some tuning, like a little extendable part at the top and tuning circuit at base. Any antenna can benefit from tuning.
I say go for it... make one and experiment. If it's good put it up. What FCC? Why would you have it inspected. If you meet the 3 meter LENGTH (and your power and ground are legal) then the worst they could do is say stop. It's not like this is way out gross violation. It's less then 3 meter. Wrapping wire or loading coils is not a problem.
OK, thanks. It also appears in the EZNEC help under “small loops”. This is what it says about min loop size.
Double precision versions of both engines permit smaller loops than the standard or single precision versions. Tests with a square loop in free space, one segment per wire, showed reasonably accurate results with loops down to the following minimum circumferences:
NEC-2: 0.05 wavelength
NEC-2D: 0.0005 wavelength (Available in EZNEC+, EZNEC Pro/2 and EZNEC Pro/4 only)
NEC-4: 0.001 wavelength (Available in EZNEC Pro/4 only)
NEC-4D: 10-7 wavelength (Available in EZNEC Pro/4 only)
The results I posted above in this thread for the 115-turn helix came from running EZNEC. Following are the results from both EZNEC and 4nec2 for comparison. EZNEC uses the NEC-2 single-precision engine according to their we site. 4nec2 uses the NEC-2D double-precision engine. Results are very similar for both.
NEC-2 VS. NEC-2D RESULTS FOR COMPARISON :
3 meter vertical helical antenna, generated for 12” diameter helix, 4 1-segment wires per turn.
Frequency = 1.706 MHz, Wavelength = 175.7 m.
Length of each generated wire is actually = 8.488 in = 0.2156 m = 0.001227 wavelength.
Circumference of one 4-wire turn = 4 x 0.001227 = 0.004908 wavelength.
EZNEC (NEC-2 engine):
Source impedance: 10.23 - J 0.1759 ohms
Far Field Gain: -11.14 dBi
Average Gain Result: 1.146 (0.59 dB)
Corrected Far Field Gain: -11.14 - 0.59 = -11.73 dBi
4nec2 (NEC-2D engine):
Source impedance: 10.2288 - J 0.397495 ohms.
Far Field Gain: -11.14 dBi
Average Gain Result: 1.15 (0.59 dB)
Corrected Far Field Gain: -11.14 - 0.59 = -11.73 dBi
When NEC and Medhurst agree, I have high confidence that the results are correct. When they disagree, there is something wrong somewhere, and more investigation is needed.
What would be more desirable than anything else is an actual physical example of a helical antenna. Then both Medhurst and NEC should both agree with the physical example. All we need to know is the number of turns required. This can be found, like in so many other situations, by trial and error.
I mounted a length of B & W coil 2 1/4 " in diameter and 5 5/16" long, .085" in wire diameter, with 32 turns, vertically on an aluminum disc 3' in diameter that I use as a ground plane for testing small coils. I measured the self-resonant frequency of the coil over the ground plane to be 22.7 Mhz.
The use of the Medhurst data gave a calculated resonant frequency of 21.54 MHz, and NEC-2 gave a resonant frequency of 27.84 MHz.
NEC-2 simulation gave an error of + 22.6 % in resonant frequency, compared to the measured value; and the Medhurst method gave an error of only - 5.11 % compared to the measured value. It looks like The Medhurst method gives a more accurate result than NEC-2. The circumference of a loop of the test coil is appreciably smaller than .05 wavelength for all three of the resonant frequencies used (measured, Medhurst, and NEC-2).
A short vertical rod antenna over ground has a triangular RF current distribution, meaning that there is a maximum current on the bottom, zero current at the top, and the current distibution is a straight line between zero and the maximum.
A resonant vertical helical radiator over ground has a convex curved distibution along its axis shaped like a quarter-cycle cosine function, with maximum current at the bottom and zero current on top. The cosine shape of the current distribution along the axis of the resonant helical antenna gives it a slight gain over a rod of the same height, which is (4/pi)^2 = 1.621 = 2.098 dB.
The most significant advantage of the helical antenna over the rod antenna is that it can have high Q if the diameter of the helix is large. A latger diameter helix allows larger diameter wire, or copper tubing, with fewer turns, to be used. This will result in less loading coil loss resistance compared to that for a rod antenna.
Ermi,
Here are some additional results of simulations of your small helix antenna using 4nec2 with NEC-2D engine.
All with minimum-segment-length .014m (.55in) vertical wire from ground to bottom end of coil, 10 ohm ground resistance, perfect ground type.
4 segs per turn, segs 129, seg. len. .0404m, 26.19MHz, 10.12-j0.41, Gain –12.64 dBi, AGT 1.31 (1.18 dB), adj. gain -13.82 dBi
6 segs per turn, segs 193, seg. len. .0286m, 22.82MHz, 10.12-j0.51, Gain –13.74 dBi, AGT 1.19 (0.76dB) adj. gain -14.50dBi
8 segs per turn, segs 257, seg. len. .0219m, 21.75MHz, 10.12+j0.68, Gain –14.12 dBi, AGT 1.12 (0.49 dB) adj. gain –14.61 dBi
10 segs per turn, segs 321, seg. len. .0177m, 21.25MHz, 10.12+j0.18, Gain –14.31 dBi, AGT 1.08 (0.33dB) adj. gain –14.64 dBi
12 segs per turn, segs 385, seg. len. .0148m, 20.98MHz, 10.12+j0.73, Gain –14.40 dBi, AGT 1.05 (0.22dB) adj. gain –14.62 dBi
Comments:
- Gain converges at 8 segments per turn.
- Resonant frequency is still converging at 12 segments per turn, but difference is getting very small.
- Actual physical implementation measurements may differ due to ground configuration, measurement equipment capacitance and tolerance, and the method used to connect the bottom of the coil to the ground plane (coil end soldered to ground vs. short wire between coil and ground.
- The simulations with larger number of segments per turn seem to agree very well with your Medhurst formula calculated resonant frequency of 21.54 MHz.
The same method used for evaluating continuously-loaded helical antennas can be used for analyzing an inductively top-loaded antenna. I am not satisfied with the results I get with NEC-2 (the resonant frequency is higher than expected); but since PhilB likes his result, he might consider evaluating a coil 12 inches in diameter, 12 inches high, with 60 turns of #18 wire along its length, on top of a rod 106 inches high, using his NEC-2 program. This gives a total height of nearly 3 meters for the inductively top-loaded antenna.
A polyethylene bucket may be an actual low-loss coil form for a real top-loading coil. I expect the actual resonant frequency of the antenna I just described to be near 1.6 MHz, but NEC gives me a resonant frequency of 2.1956 MHz.
Inductive top-loading has been done with flat, horizontal, spiral coils. This has usually been done by winding the spiral basket-weave fashion on insulating radial spokes. It's too bad that NEC-2 does not have an automatic feature for creating spiral coils. The spital can be modelled manually, but that is very laborious.
The helical, also called the spiral antenna, is much like a metal slinky toy, although the toy doesn't extend to 10-feet.
Having a slinky type metalic material, however, would allow the experimenter to adjust spacing between windings.
A horizontal metal ribbon was mentioned... what happens if the flat metal ribbon is angled?
A standard Slinky is 2 3/4" in diameter with 90 turns. Supposedly, it can be stretched to 15 feet without deforming. The number of turns for a single Slinky would not be enough for resonance in the AM BCB, however, and you would have the ridiculous situation of needing an additional loading coil to resonate the helical antenna. Two Slinkys in series might be OK. With multiple Slinkys, you might have to clip or off turns, or use a tap, for resonance.
The fact that the steel wire of the Slinky is flat doesn't matter much compared to using round wire. The angle of the flat wire also doesn't mater much.
There are different size Slinkys, and even plastic ones.
I have also read of hams experimenting with flexible dryer duct, a plastic wrapped helix available in several lengths and diameters.
"Because of the waffling, I had hoped that maybe smaller loops can be simulated, but experiment indicates that they had a good reason to mention .05 wavelength minimum circumference."
That would more-or-less agree with Phil's 12" dia. model.
The question I have regards how the 12" dia. is set in Phils' model, which is based upon a square form ... namely: Is the 12" reference the diameter of a circle measured tangent to the flat sides or across the corners of the square form?
IOW, were I to build such an antenna for experiment, should I make the square form with 12" sides, or 12" corner to corner?
NOTE: I checked the local recycle yard ... they want $100 for an 11' piece of 12" PVC drain pipe ... more than I want to spend, so I was thinking about a wooden or 1/2" PVC pipe frame.
TIA ...
Ken,
The 4nec2 software has a "geometry builder" section that allows you to build a helix. The inputs are "diameter" and the number of "wires per turn". As it works out, the wire junctions lie on a 12" circle. Increasing the "wires per turn" results in more wires, but their intersections still lie on the 12" circle. Since the "geometry builder" feature isn't part of the standard NEC software, the "diameter" may be defined differently in various programs, but my results were based on the 4 points of the square lying on a 12" diameter circle.
Hope this is clear.
EZNEC gives the junctions of the segments on the circle. For four segments/turn, the length of each segnent is the radius times SQRT(2). As the number of segments increases, the circle is more closely approximated, but the area of the polygon formed by the segments is always smaller than the area of the circle.
I would prefer it if the area of the polygon formed by the segments were made the same as for the circle, but EZNEC does not make that correction.