Computer Assisted Low Profile Antenna Modeling II Computer Design of a Low Profile Horizontal Loop Antenna for a Limited Space Backyard by Dr. Carol F. Milazzo, KP4MD (posted 25 September 2010) E-mail: kp4md@arrl.net SUMMARY This article describes the application of computerized modeling to design and analyze the performance of a low cost, low profile horizontal loop antenna that exhibits gain over a dipole within a limited space backyard using household materials and low cost speaker wire zip cord as the transmission line. INTRODUCTION I first experimented with antenna modeling in 19971 when I moved to a residential housing development where restrictive covenants did not permit outdoor antennas.  The space available for antennas at that time was inside a peaked roof attic.  In 2006, I moved to a mobile home park with covenants that restricted antennas visible from the street.  My first antenna at this location was a remotely tuned base loaded 6 foot vertical whip antenna (High Sierra Sidekick) mounted on the roof at the rear of the home.  The aluminum siding served as its counterpoise.  Its observed performance was fair on 21 and 28 MHz, mediocre on 7 and 14 MHz with little reception except to the north, and overall quite poor on 3.5 MHz.  Seeking better performance on the lower frequencies, I later added a 100 foot end-fed antenna of 20 gauge stranded wire with PVC insulation supported by two 15 foot PVC poles at the south corners of the property.  That antenna performed fairly for contacts within 200-300 miles on 1.8 and 3.5 MHz but quite poorly on the higher frequency bands.  Additionally, it was highly susceptible to local noise and radiated strong radio frequency fields within the home's living space. Recently, I decided to analyze the performance of these antennas with computer modeling in an attempt to design a more effective antenna system. NEC AND MININEC Most modern antenna analysis programs have their origins in a very large and complicated FORTRAN program called the Numerical Electromagnetics Code or "NEC." NEC analyzes wire antennas by dividing them into a number of segments, calculating the current in each segment and summing the results. This provides information on the radiation pattern and impedance of the antenna for any selected frequency. NEC was written in the 1970's and was composed of tens of thousands of lines of computer code requiring the use of a mainframe computer inaccessible to most radio amateurs. In 1980, the team of John Rockway and Jim Logan successfully wrote a very simplified version called MININEC that had about 500 lines of BASIC and could run on a personal computer. Since that time, MININEC has evolved through several versions and enhancements to take advantage of the increased power of modern personal computers. MININEC, NEC-2 and NEC-4 provide the basis for a large portion of the amateur radio literature concerning antenna analysis.

CONTENTS Introduction NEC and MININEC Modeling the existing antennas Designing the loop antenna model Analyzing the loop antenna model Building the loop antenna Adjusting & measuring the loop antenna On the air testing Analyzing as a top loaded vertical Antenna Feed System Improvements Noise levels Loop Antenna Raised References 4nec2 input files

Since my work with the antenna modeling program NEC4WIN in 1998, several new user interfaces for MININEC and NEC have become available and lists and comparisons of their features are available elsewhere.  For this study I chose to use 4nec2 by Arie Voors due to its functionality (3-D graphics of antenna model, radiation pattern plots, graphs of impedance, VSWR, etc. vs. frequency) and its availability as a freeware download at http://www.qsl.net/4nec2/.²

MODELING THE EXISTING ANTENNAS

Fig. 1.  Aerial photo of property (north is right)	Fig. 2.  Survey map of property	Fig. 3.  2-D model	Fig. 4.  3-D model

Figures 5 through 16 are the calculated far field radiation patterns comparing the vertical antenna and the random wire antenna on 7 and 14 MHz (North is at the top of all azimuth and 3-D patterns).

Fig. 5.  Vertical 7 MHz azimuth pattern	Fig. 6.  Vertical 7 MHz elevation pattern	Fig. 7.  Vertical 7 MHz 3-D pattern
Fig. 8.  Random wire 7 MHz azimuth pattern	Fig. 9.  Random wire 7 MHz elevation pattern	Fig. 10.  Random wire 7 MHz 3-D pattern
Fig. 11.  Vertical 14 MHz azimuth pattern	Fig. 12.  Vertical 14 MHz elevation pattern	Fig. 13.  Vertical 14 MHz 3-D pattern
Fig. 14.  Random wire 14 MHz azimuth pattern	Fig. 15.  Random wire 14 MHz elevation pattern	Fig. 16.  Random wire 14 MHz 3-D pattern

Table 1 below lists the calculated major lobes of radiation, the direction of the major lobe, and overall radiation efficiency.  The radiation efficiency is a measure of the overall proportion of power that is radiated into space after losses in the structure and the ground are subtracted.

The calculated directionality of the vertical antenna corresponds with the observed performance with nulls in the south and east directions.  From 65% to 95% of the transmitter power was wasted in structure and ground losses.

DESIGNING A HORIZONTAL LOOP ANTENNA MODEL

I chose to raise the feed point to the top of the 20 foot mast that supported a VHF/UHF J-pole antenna. This additional height was needed for the loop to clear the top of the tree in the backyard. A rope and pulley would be used to raise and lower the feed point for maintenance and antenna adjustments. A new 15 foot PVC pole would be tied to a carport upright to support the fourth corner of the loop.  The resulting loop geometry would be a horizontal trapezoid with two sides sloping up to the feed point (see Figures 17 and 18). In order to have a wire segment length on the order of .05 wavelength on the highest frequency (28 MHz), 80 segments were required. To have all segments of nearly equal length, the two sides nearest the feed point were given 18 segments each, and the far sides were given 22 segments each.  By trial and error, it was found that locating a corner of the loop over the center carport upright yielded the desired resonant frequencies in the 7 and 14 MHz bands (Figures 19 and 20). Download the NEC input file.

Fig. 17.  Horizontal loop 2-D model	Fig. 18.  Horizontal loop 3-D model	Fig.19.  Reactance/resistance over 7 MHz Resonance near 7.01 MHz.	Fig.20.  Reactance/resistance over 14 MHz Resonance at 14.25 MHz.

ANALYZING THE LOOP ANTENNA MODEL

Fig. 21.  Frequency sweep of SWR over 6-29 MHz	Fig. 22.  Frequency sweep of reactance & impedance over 6-29 MHz

Figures 23 through 34 show the calculated far field radiation patterns for the horizontal loop antenna on 7, 14, 21 and 28 MHz.

Fig. 23.  Loop 7 MHz azimuth pattern	Fig. 24.  Loop 7 MHz elevation pattern	Fig. 25.  Loop 7 MHz 3-D pattern
Fig. 26.  Loop 14 MHz azimuth pattern	Fig. 27.  Loop 14 MHz elevation pattern	Fig. 28.  Loop 14 MHz 3-D pattern
Fig. 29.  Loop 21 MHz azimuth pattern	Fig. 30.  Loop 21 MHz elevation pattern	Fig. 31.  Loop 21 MHz 3-D pattern
Fig. 32.  Loop 28 MHz azimuth pattern	Fig. 33.  Loop 28 MHz elevation pattern	Fig. 34.  Loop 28 MHz 3-D pattern

Table 2 below compares the calculated major lobes of radiation, the direction of the major lobes, and overall radiation efficiency of the vertical, random wire and horizontal loop antennas.

On 7 MHz, the loop antenna offered omnidirectionality with mostly high angle radiation, 5.5 dB more maximum gain and over double the radiation efficiency of the vertical antenna, and 10 dB more maximum gain and 10 times the radiation efficiency of the random wire antenna..  On 14 MHz, compared to the other antennas the loop antenna offered near omnidirectionality, low angle of radiation, up to 3.3 dB more maximum gain and up to twice the radiation efficiency.

The random wire antenna was removed, and Table 3 compares the 4nec2 calculations for the horizontal loop antenna (SWR at the feed line characteristic impedance of 114 ohms) with the vertical antenna on all amateur radio high frequency bands.

*High standing wave ratios were expected to cause increased losses on frequencies other than 7, 14, 21 and 28 MHz.
**Replacing the random wire with the loop antenna significantly increased the maximum gain and radiation efficiency of the vertical antenna on 7 MHz where near field coupling induced significant currents in and radiation from the loop antenna.

BUILDING THE LOOP ANTENNA

The wire loop antenna is made of 140 feet of CTI-20 gauge insulated stranded wire. Cost $10.	The balanced feed line was this 40' roll of 24 gauge speaker wire. (Replaced with 18 gauge speaker wire: see Update - 04 Dec 2010)	The feed point insulator is made from half of a ball point pen barrel and a paper clip.	Five holes are drilled in the pen barrel and the clip is fashioned into an eye hook for the support rope.	The feed point is half assembled. Wire nuts splice the feed line to the ends of the loop antenna.
This tie string tension pulley is mounted atop the mast that supports the feed point with a machine screw eye bolt.	Here is the assembled feed point raised 20 feet to the top of the supporting mast.	Here is the southeast corner of the loop antenna. The screwdriver antenna base is visible on the roof of the house.	At the corner supports, the loop wire passes through a zip tie secured through a hole in the PVC pole.	The feed line is suspended away from the mast with twin lead standoff insulators.
A standoff insulator maintains the feed line (twisted along its length) distance from the porch awning.	A nylon monofilament suspends the feed line where it bends to pass through the vinyl window frame.	4:1 Ruthroff (voltage) balun matched the coax cable to the balanced load. Later replaced with 1:1 current balun: see Update below	The feed line & balun are installed.  The copper braid exits the window to the ground rod.	After the blinds are closed, the feed line and balun are out of view.

ADJUSTING AND MEASURING THE LOOP ANTENNA

Ideally, antenna impedance measurements should be taken directly at the antenna feed point.  To attempt this in the air 10 feet above the roof of the house would be unsafe, so the measurements were taken with an MFJ-269 antenna analyzer at the transmitter end of the feed line with the caveat that there would be limitations in accuracy.  Immediately, there was an obvious incompatibility of the analyzer's unbalanced RF connector and the balanced feed line.  So I first connected the analyzer through the 4:1 Ruthroff (voltage) balun.  I was able to determine frequencies of minimum standing wave ratio, but the measured resistances and reactances did not approach the predicted values and the SWR reading never exceeded 5:1 throughout the full range from 6 through 30 MHz.  These inaccuracies appear because the 4:1 voltage balun introduces inductance across the load, which also lowers the resonant frequency of the loop antenna to which it is connected.  The analyzer requires a 1:1 Guanella (current) balun to measure a balanced load.  I built one such balun by passing 12 turns of RG-174/U miniature 50 ohm cable through a stacked pair of 20 mm OD junk box toroid cores and another one by wrapping 30 turns of RG-174/U around a 4-3/4" x 3/8" ferrite rod salvaged from a transistor radio (Figure 35).  Either of these gave comparable results.  I then measured the SWR, resistance and reactance over the 6 to 29 MHz range and entered them into a Microsoft Excel spreadsheet that produced the graphs in Figures 36 through 38.

Fig. 35.  Homemade 1:1 current baluns	Fig. 36. Measured SWR over 6-29 MHz	Fig. 37.  Measured resistance over 6-29 MHz	Fig. 38.  Measured reactance over 6-29 MHz

Comparing these to the theoretical curves in Figures 21 and 22, the fundamental resonant frequency was shifted up about 200 kHz (since the antenna had been tuned to resonance at 7 MHz with the inductive load of the 4:1 balun), otherwise the visible similarities in the overall curve patterns would support the validity of the model.  Note that the MFJ-269 antenna analyzer does not distinguish between inductive and capacitive reactance, so it represented both as positive values.

An interesting finding was that the measurements were significantly affected when the loading coil of the vertical antenna was tuned to the same frequency at which the loop was being measured.  This effect was most prominent on 7 MHz and decreased at the higher frequencies.  This confirmed others' observations that loop performance is affected by other resonant objects within the near field of the antenna.  For this reason, the loading coil on the Sidekick vertical antenna was set to minimum inductance during all measurements of the loop antenna.
 

ON THE AIR TESTING On frequencies below 7 MHz, the loop antenna often provided clearer reception of signals with lower noise levels, although signal reports from other stations showed that the vertical antenna was the more efficient radiator.  Later, received noise levels on both antennas were quantitatively compared.  On 7 MHz received sky wave signals were predominantly up to 12 dB stronger on the loop, with few signals favored by the vertical antenna.  Above 7 MHz received and transmitted sky wave signals were predominantly stronger on the loop antenna. In the week after the loop antenna was erected on 11 September, 2010, my online log recorded solid contacts on 40 meters with stations in the eastern USA, Australia, Brazil, Guatemala, Japan and South Korea, areas that I could rarely contact with my previous antennas.  Within 12 hours on 10 November 2010, the horizontal loop antenna yielded confirmed contacts on 6 continents using 5 watts with WSPR mode on 7 and 14 MHz.  Starting in January 2011 WSPR data was used for further analysis of antenna performance.8 Comparisons between the antennas in contacts with local stations via ground wave were variable, favoring either the vertical antenna or the loop antenna depending on polarization and antenna directivity.

Fig. 39.  6 continents in 12 hours using 5 watts WSPR mode on 7 & 14 MHz.

ADDENDUM - 07 October 2010 - The Horizontal Loop as a Top Loaded Vertical Antenna

ANALYZING THE LOOP AS A TOP LOADED VERTICAL ANTENNA

Fig. 40.  Frequency sweep of SWR over 1.6-4.2 MHz	Fig. 41.  Frequency sweep of reactance & impedance over 1.6-4.2 MHz

Figures 42 through 47 show the calculated far field radiation patterns for the antenna used in this manner on 1.8 and 3.5 MHz.

Fig. 42.  Loop as vertical 1.8 MHz azimuth pattern	Fig. 43.  Loop as vertical 1.8 MHz elevation pattern	Fig. 44.  Loop as vertical 1.8 MHz 3-D pattern
Fig. 45.  Loop as vertical 3.5 MHz azimuth pattern	Fig. 46.  Loop as vertical 3.5 MHz elevation pattern	Fig. 47.  Loop as vertical 3.5 MHz 3-D pattern

Table 4 below shows the calculated low gain and overall radiation efficiency of the horizontal loop antenna used in this manner.

ON THE AIR TESTING

Update - 04 December 2010 - Antenna Feed System Improvements Described in detail separately in Zip Cord Transmission Lines and Baluns.9  In December 2010, the 24 AWG transmission line was replaced with lower loss 18 AWG speaker wire (Figure 48), common mode chokes were added (Figure 49) and the 4:1 voltage balun was replaced with a current balun (Figure 50) in order to reduce the antenna feed system noise and attenuation. Figure 54 shows the present station antenna configuration.  At T3, a 1:1 current balun sufficiently matched the transmitter to the transmission line across most of the 7 and 14 MHz bands to allow operation with a direct connection that bypassed the MFJ-949B antenna tuning unit.  On the 21 and 28 MHz bands, the antenna tuning unit achieved best impedance matching with the 4:1 current balun at T3.  The antenna tuning unit and current baluns allowed operation with the loop antenna on frequencies other than 7, 14, 21 and 28 MHz with variable decreases in signal strengths.	Figure 48.  Pfanstiehl 18AWG AS-18/50Z Speaker Wire	Figure 49.  1:1 choke at feed point - 10 turns on FT114-43 core	Fig. 50. 4:1 current balun made of 18 gauge  speaker wire on two FT140-43 toroid cores

NOISE LEVELS

As expected, the noise levels decreased with increasing frequency.  The noise levels were comparable on the loop antenna and the vertical antenna on 7, 14, 21 and 28 MHz.  Both noise and received signal levels were lower on the loop antenna on other frequencies due to mismatch losses at the antenna end of the feed line.

The common mode signal and noise rejection were tested by shorting both sides of the balanced feed line together and observing for quieting of the receiver. Among the baluns, common mode signal rejection was greatest in the current baluns and least in the 4:1 voltage balun.  At 3.5 MHz, the 4:1 voltage balun showed no measureable rejection of common mode noise at all.  The toroid current baluns with their increased common mode signal rejection also improved reception in the low frequency and medium frequency ranges that were previously covered by strong intermodulation products from nearby medium frequency AM broadcast stations.

UPDATE - 07 June 2011 - Loop Antenna Raised

With the horizontal loop antenna at this increased height above ground, the modeling program predicted increases in radiation efficiency on all frequencies and overall increases in gain at both low and high elevation angles.  Figures 51 through 53 compare the performance of the horizontal loop antenna against the vertical screwdriver antenna.  The loop antenna yielded greater gain at the desired high radiation angles (near vertical incident skywave or NVIS) for short distance paths on frequencies below 14 MHz and at greater gain at low radiation angles as desired for long distance paths on and above 14 MHz.  The loop antenna had significantly superior radiation efficiency on all frequencies above 3.5 MHz.

CONCLUSION Computer antenna modeling was useful in significantly improving my station antenna performance.  Adding some wire and supports converted an inefficient random wire antenna into a much more efficient full wave horizontal loop antenna. The $15 spent for the loop antenna wire, feed line and balun yielded significantly superior results to the screwdriver antenna which cost $450.  Compact antennas like the Sidekick screwdriver vertical compromise performance but have their applications for their portability and when space is especially limited, as on a vehicle. NEXT PROJECT: Comparative Antenna Analysis with WSPR ACKNOWLEDGEMENTS Many thanks to Dutch engineer Arie Voors for sharing his excellent 4nec2 antenna modeling program with the radio amateur community, Magnus Beischer, Don Lucas, Matt Pyne for their freeware TinyCAD program used to draw the schematic diagram, Dan Maguire for the Transmission Line Details program, and to Larry Sutter, WD6FXR, who very graciously permitted me the use of his MFJ-269 antenna analyzer for this project. REFERENCES "Computer Assisted Low Profile Antenna Modeling I", Milazzo, C, KP4MD, 13 June 1998. 4nec2 antenna modeling software, Voors, A. "A Beginner's Guide to Modeling with NEC", Cebik, LB, W4RNL, QST, November 2000, pp. 35-38. "The Loop Skywire", Fischer, D, W0MHS, QST, November 1985, pp. 20-22. "Zip Cord Antennas - Do They Work?", Hall, J, K1TD, QST, March 1979, pp. 31-32. "Zip Cord Antennas and Feed Lines For Portable Applications", Parmley, W, KR8L, QST, March 2009, pp. 34-36. "Portable Antenna Notes", Wiesen, R, WD8PNL "Comparative Antenna Analysis with WSPR", Milazzo, C, KP4MD, 13 January 2011. "Zip Cord Transmission Lines and Baluns", Milazzo, C, KP4MD.

Figure 54.  The present station configuration.  The radio is connected through two-foot RG-58/U coaxial cable jumpers to the MFJ-949B antenna tuning unit and the 4:1 Guanella current balun (T3).  A total of 4 feet (1.2 m) of the AS-18/50Z speaker wire are wound on the the toroid cores (T1, T2 and T3) for an approximate 36 feet (11 m) of total transmission line length of speaker wire to the full wave 40 meter horizontal loop antennna.  The antenna tuning unit is omitted when operating WSPR in the frequency hopping mode.

APPENDIX: 4nec2 INPUT FILES

Return to KP4MD Home Page