Helical vs. Linear Monopole Antennas

[rfry]

1. The first and foremost problem is that Fry's coil is not self-resonant. The cosinusoidal current distribution I talked about in this thread occurs only at resonance.

2. Closely related to # 1 above, the segments are not long enough for the analysis. When analyzing a helical antenna using NEC, overly-short segments cause an error in the analysis that causes the imaginary portion of the input impedance to be negative imaginary.

My 3-m helical model is not self-resonant in the AM broadcast band, but that does not negate the conclusions in my comparison.

Changing the models used in the comparison as suggested by Mr Roos changes operating conditions, and the conclusions of the original comparison won't necessarily apply to the changed versions. Neither will his conclusions about a self-resonant, normal-mode helix necessarily apply to my model.

Both of my NEC models were analyzed using NEC's Average Gain Test, and were reported to be "highly accurate."

 

[edarmsttrong]

My GOD! You people need to quit the pissing contest and get some lives!!!

 

[Ermi Roos]

I tried out a 0.3 m diameter (about 1 foot) helix with 90 turns along a 3 m height, which is self-resonant at the upper end of the AM BCB, according to my manual calculations. There were no segmentation error messages using NEC-2, but adding turns to simulate increased inductance did not reduce the capacitive input reactance to achieve resonance. I saw that 0.3 m is too small a loop diameter to successfully model a vertical helix monopole 3 m high at around 1.6 MHz with NEC-2.

Next, I tried a coil with with 3 m diameter with 7 turns along a 3 m height, which is also self-resonant at the upper-end of the AM BCB according to my manual calculations. The NEC-2 simulation of this coil indicated resonance at 1.676 MHz, but the radiation resistance was calculated to be .3019 ohms at resonance. A small-diameter rod 3 m high would have a radiation resistance in the vicinity of .11 to .13 ohms, but I don't know what it would be for a cylindrical drum 3 m in diameter. It would be necessary to model a cylinder 3 m in diameter. Experience has shown that fat vertical monopoles have more radiation resistance than thin monopoles, but the increase is not great. If the predicted 2 dB gain for the helix applies to this case, the radiation resistance of the cylindrical drum would be .3019/1.621 = .186 ohms. This value of radiation resistance seems plausible, but I don't know how close it is to the true value.

The circumference of each turn of the helix is 9.425 m. 0.05 wavelengths is 8.95 meters. So, in this case, the minimum .05 wavelength guideline for the circumference of a loop analyzed by NEC-2 has been met. From my observations, it is not useful to analyze a small loop antenna using NEC-2. My program has the double-precision feature.

The indicated radiation resistance for the large-diameter helix antenna described here decreases appreciably when the operating frequency of the simulation is only slightly off resonance. So, the results are applicable only at, or very close to, resonance. For a vertical rod, the radiation resistance hardly changes at all if the simulated antenna system is not at resonance; but this is not true for the simulation of the helical monopole.

 

[rfry]

A comparison of a helical and a monopole radiator at the first self -resonance of both forms is revealing (link below).

The NEC model of the helical used in this comparison is the same as was used in my previous links in this thread.

Under these conditions, a linear monopole has higher directivity,. and higher system radiation efficiency for a fixed loss in the r-f ground connection than a helically-loaded monopole.

Notice that the current distribution relative to the overall physical apertures of these two radiators is the same. Distribution is not sinusoidal along the linear while "cosinusoidal" along the helix. It is sinusoidal in both cases.

http://i62.photobucket.com/albums/h85/rfry-100/Helix__Linear_1st_Resonance.gif

 

[PhilB]

I couldn't resist this (sorry everyone).

I got suckered into running an EZNEC model of a 3 meter helical antenna just to see how my results compare relative to this thread.

Resonant Freq: 1.705 MHz.
Helix configuration: 4-sided square loop per turn, helix diameter=12"
Helix Turns: 145
Segment Check: Pass
Wire Size: #14
Height: 3 meters
Total Segments: 461
Ground Resistance: 10 ohms
Helix Coil Loss Resistance: 0
Ground: Perfect
Impedance at antenna base: 10.23 - J 0.1759 ohms
Radiation Resistance: .23 ohms
Far Field Gain: -11.14 dBi
Accuracy Check: Average Gain = 1.146 = 0.59 dB
Corrected Far Field Gain: -11.14 - 0.59 = -11.73 dBi

The model for a 3 meter base-loaded monopole gives these results:

Resonant Freq: 1.6 MHz.
Segment Check: Pass
Wire Size: 0.625 in. (1/2" copper pipe)
Height: 3 meters
Total Segments: 16
Ground Resistance: 10 ohms
Coil Loss Resistance: 0 (not possible, but chosen for comparison to helix above)
Ground: Perfect
Impedance at antenna base: 10.1 - J 0.02625 ohms
Radiation Resistance: .1 ohms
Far Field Gain: -15.09 dBi
Accuracy Check: Average Gain = 0.999 = 0 dB
Corrected Far Field Gain: -15.09 - 0 = -15.09 dBi

Important results:

Radiation Resistance:
Helix = .23 ohms, Monopole - .1 ohms

Gain:
Helix = -11.73 dBi, Monopole - -15.09 dBi

3.36 dBi gain advantage goes to helix, and in reality, the advantage may be even higher compared to a base-loaded monopole. The results for the monopole assumed zero coil loss to focus the comparison on a zero coil loss helix vs. monopole. An actual monopole antenna may have a coil loss of 5 to 15 ohms, so the gain advantage for the helix would be even higher. I don't know what the coil loss would be for the helix, but I think it would be very low.

The 12" diameter of the helix may not pass muster with the FCC. A smaller diameter helix with more turns is certainly workable. Unfortunately it can't be modeled accurately with NEC-2 due to minimum segment size restrictions.

 

[PhilB]

Two lines were in error for the helical antenna model results in my last post.

Resonant Freq: 1.705 MHz. should be 1.706 MHz.
Helix Turns: 145 should be 115 turns

I checked all the data again and these appear to be the only errors in the post.

 

[rfry]

... John Kraus: From what I have seen, Kraus was mostly interested in analyzing the helical antenna operating in the axial mode. For helical antennas operating in the normal mode, he mostly deferred to the previous work by Wheeler. To simplify analysis, Kraus assumed additional end loading, or top loading, external to the helix.

And later, he wrote:

I doubt the usefulness of engineering references in this particular argument, because we both have access to the work of John Kraus, and we reached completely different conclusions. I think that you misread Kraus, and you undoubtedly think the same about me.

John Kraus gives much discussion of the helical antenna (which he invented) in the first edition of his textbook ANTENNAS, including the fact that the current distribution along a normal-mode helical antenna with no top loading is sinusoidal, just as it is along a monopole having no top loading -- which he calls a "long straight antenna" in that portion of the text.

Kraus discusses a normal-mode helix with capacitive top loading in an earlier part of this chapter, in the development of the performance of helices.

It is difficult to mis-read what Kraus wrote about this, but perhaps Mr Roos has not differentiated between these two parts of Kraus' text.

Here is a link to a scan of the section from Kraus about normal-mode helices with no capacitive top loading (yellow high-lighting is mine):

http://i62.photobucket.com/albums/h85/rfry-100/Helix_Current_Distr.gif

 

[rfry]

As for NEC simulation, I already mentioned overly-short segments, but there is also a concern that very small loop antennas (less than .05 wavelength in circumfurence) can't be simulated accurately by NEC-2 at all. NEC-4 supposedly does better.

A loop antenna is not the same as a helical antenna.

My NEC model of the 30-turn helix uses 10 straight sections in each complete turn of the helix. Each section is 0.03248 meters in length. I used NEC-Win Synth to construct, and NEC-Win Plus+ to evaluate the model, and the instruction books for these programs state "Minimum segment length: 10-4 wavelengths," which is 0.0001 wavelengths.

A wavelength at the 1.7 MHz used in the study is 176.353 meters, and 0.03248 / 176.353 = 0.000184 -- which is a segment length longer than the minimum length. It should also be noted that NEC-Win Synth tests models for segment and geometry errors, and none was reported for this model.

NEC-Win Plus+ uses NEC-2D, the double-precision version of NEC-2. The Average Gain Test of the helix model in this study reported that it was "highly accurate."

Below for verification is the output list of the first 13 lines of the segmentation data of the NEC model The first three rows show the three equal-length segments in the vertical conductor used at the base of the antenna. The source is placed on the center segment of that conductor in accordance with the NEC requirement that the segment lengths on each side of the segment with the source have the same length as the segment with the source (it is unknown if the models constructed by others observed this rule).

There is no legitimate reason to doubt the integrity/accuracy of this NEC model.

NEC-Win Plus+ OUTPUT DATA

C:\Program Files\NEC-Plus\RJF Antenna Files\3-m Helix.NOU

SEGMENTATION DATA
-----------------

COORDINATES IN METERS

I+ AND I- INDICATE THE SEGMENTS BEFORE AND AFTER I

SEG. SEG. CENTER COORD SEG. ORIENTATION ANGLES WIRE CONNECTION DATA
TAG

NO. X Y Z LENGTH ALPHA BETA RADIUS I- I I+ NO.

1 0.00000 0.00000 0.01667 0.03333 90.00000 0.00000 0.00250 1 1 2 1

2 0.00000 0.00000 0.05000 0.03333 90.00000 0.00000 0.00250 1 2 3 1

3 0.00000 0.00000 0.08333 0.03333 90.00000 0.00000 0.00250 2 3 -333 1

4 0.04523 0.01469 0.10500 0.03248 17.93209 107.99989 0.00250 -304 4 5 2

5 0.02795 0.03847 0.11500 0.03248 17.93179 143.99934 0.00250 4 5 6 3

6 0.00000 0.04755 0.12500 0.03248 17.93177 180.00000 0.00250 5 6 7 4

7 -0.02795 0.03847 0.13500 0.03248 17.93179-143.99934 0.00250 6 7 8 5

8 -0.04523 0.01469 0.14500 0.03248 17.93209-107.99989 0.00250 7 8 9 6

9 -0.04523 -0.01469 0.15500 0.03248 17.93209 -72.00011 0.00250 8 9 10 7

10 -0.02795 -0.03847 0.16500 0.03248 17.93179 -36.00066 0.00250 9 10 11 8

11 0.00000 -0.04755 0.17500 0.03248 17.93177 0.00000 0.00250 10 11 12 9

12 0.02795 -0.03847 0.18500 0.03248 17.93179 36.00066 0.00250 11 12 13 10

13 0.04523 -0.01469 0.19500 0.03248 17.93209 72.00011 0.00250 12 13 14 11

//