Tuesday, July 22, 2014

Feeding a Multi-band Dipole Antenna

During recent on the air discussions we revisited the design options for a multi-band dipole antenna with balanced feeder. Design guidelines for a slightly-shortened dipole antenna are published in the document Choosing the Correct Balun by W8JI. The recommended lengths of balanced feedline are based on odd multiples of 1/8 wavelength on the lowest frequency of operation. This is covered on pages 7 & 8 of W8JI's document.

Some of our group wondered if the same feedline recommendations would apply to a standard half-wave dipole as they do to the slightly-shortened dipole. To explore this topic I used the 4nec2 antenna modeling software to analyze a half-wave dipole of total length 134 ft (12 ga copper) at 30 ft above rocky ground. This length is close to resonance at 3.5 MHz. The shortened dipole in the article, to cover 80 m and above, has length 110 ft.

I collected data for the feed point impedance at one frequency on each of six bands. I also expanded the model to use a 450-ohm feedline of length 30.48 ft (close to the 1/8 wave length of 30.9) and collected the impedance data at the end of the feedline for the same 6 frequencies. The results are shown in the graph and table below.

The graph clearly shows the wide variation of impedance (orange bars) at the feedpoint, while the impedance at the xmt end of the feedline (blue bars) shows that the impedance variation is considerably narrowed. This supports the comments on page 7 of the W8JI document.  At 7 MHz the dipole becomes a full wavelength, with very high impedance, but the feedline -- 1/8 wavelength at 3.5 MHz -- becomes 1/4 wavelength at 7 MHz, thus transforming to a much lower impedance.

The table below gives the data used in the graph. I did not include the resistance and reactance data, but they are readily available.


f (MHz)
Z (xmt)
Z (fp)
3.5
338
56
5.3
539
1156
7
105
5970
14
839
2433
21
453
1637
28
225
1180

This analysis covers only the 1/8-wavelength feeder case, but any other lengths (e.g 3/8, 5/8) can be analyzed easily.

John

No comments: