Rob's web

Home - Techniek - Electronica - Radiotechniek - Radio amateur bladen - Radio Communication - The 8JK revisited and the new BRD-zapper


The major breakthrough in the design of amateur rotary beams came in 1937 from Professor John D Kraus, WBJK (my apologies for giving him last month the callsign W8XK, that of the old Pittsburgh HF broadcast station, later KDKA). Both his work on the first close-spaced 'flat-top' bidirectional driven arrays (often implemented as fixed beams), and that of Walter Van Roberts, W3CHO, shortly afterwards on the first close-spaced, two-element Yagi-Uda unidirectional parasitic array, were based on recognizing the significance of the work of Dr George Brown (RCA) on vertical monopole arrays for medium-wave broadcasting. Until then it had always been assumed that a Yagi parasitic reflector should be ¼λ behind the driven element, requiring a boom length, at least on 14MHz, that virtually ruled out flat-topped arrays, although a few amateurs had built Yagi rotary beams based on vertical elements spaced ¼λ apart

By the time that UK amateur operation was closed down on 1 September 1939, knowledge of the family of 8JK bidirectional arrays and close-spaced Yagi arrays had crossed the Atlantic (see for example brief references in the first edition of the RSGB's The Amateur Radio Handbook (1938) but were still rare. It was not until the early postwar period that the three-element Yagi array began to establish itself worldwide. Due to their unidirectional properties, Yagi arrays became, and remain, far more popular than the 8JK driven bidirectional arrays, although these have some important advantages, including multi band operation with a single set of two elements: Fig 1.

Fig 1
Fig 1. Basic single section WBJK flat-lop antennas. In 1981 Dr Kraus pointed out that the centre-led arrangement with a typical spacing, S, of about λ/8 on the lowest frequency used, L can range from less than λ/2 to more than 1.5λ, permitting its use over a continuous frequency range of 3:1. With L equal to 7.3m (centre gap G forms part of the measurement) and S equal to 2.6m, the bidirectional gains are: 14MHz 5.7d81; 21MHz 6.7d81; 28MHz 7.701; 50MHz 8.2dBi. Values for 18 and 24MHz bands can be interpolated. The end-led version should provide roughly similar results if correctly tuned.

In QST October 1981 (see TT, November 1982) Dr Kraus restated the very real attractions of the centre-fed design when it comes to multiband use, showing that a compact single-array with 7.3m elements can cover 14, 18, 21, 24. 28 and 50MHz bands. Among the plus points he listed in 1981 were:

  1. continuous frequency span of more than 3 to 1 (with suitable matching/tuning arrangements);
  2. no traps or loading coils;
  3. no critical dimensions, as the entire antenna and feed system can be resonated;
  4. it can be used horizontally or vertically for optimum radiation angle;
  5. it is ideal for finding round-the-world (long path) openings:
  6. it has theoretically zero radiation off the ends of the elements:
  7. it can be fed with low-loss, inexpensive twin line (or verylow-loss cut-away 300Ω line or open-wire line); and
  8. a compact array can cover many bands including the non-harmonically related bands.

Minus points are that the bidirectional pattern does not give protection against continental European signals when the array is pointed towards North America: the forward gain is lower than with a good monoband Yagi array; and the low radiation resistance does not make it a good system to install indoors in the presence of metal structures etc (but in practice some amateurs have found the 8JK most satisfactory as a fixed indoor array). The end-fed form is particularly easy to install in, for example, a roof space or as an outdoor antenna supported from a mast or tree using two light wooden spreaders.

For those who require unidirectional patterns, it is possible to convert the 8JK into this form by using the fact, shown in Dr Brown's classic paper, that when two driven elements are led 135° out-of-phase with equal amplitudes of RF power, a cardioid radiation pattern results. giving forward gain and a deep rejection null in the reverse direction Exploitation of this principle has led to such antennas as the ZL-Special, the HB9CV and G8P0 antennas as described in TT, October 1981.

The ZL-Special was so named and described in print by Fred Judd, G2BCX, (Shortwave Magazine July 1950, pp337-9) where he, commented on its origins "Data on the aerial came to the writer from New Zealand, hence the name ZL-Special. Little is known of its origin save that it was designed in the USA just prior to the war, for commercial purposes. (This is probably a reference to the work of Dr Brown on MP broadcast antennas - G3VA). Since the war it has been modified and developed for amateur use by W5LHI, WOGZR and ZL3MH. Further tests and measurements made by the writer may be of interest "

The 'G8P0 Special', as described by J E Ironmonger. G8PO, in the RSGB Rulletin (November 1947), provides a 'reversible' unidirectional fixed beam, originally developed as a result of a 'mistake' in cutting feeder lengths when erecting an 8JK. It uses two feeders of different length, ie arranged so that either feeder can be 1/8λ longer than the other by having two sockets into which the transmitter feed-tine is connected: see Fig 2.

Fig 2
Fig 2. G8P0 reversible unidirectional 14MHz array as described in 1947. Direction depends on whether transmitter feed Is connected at S1 or 52. Note the delay line section is twisted once to provide the 135° out-of-phase drive.

The same basic approach is also used in the antenna developed by Rudolf Baumgartner, HB9CV, which in recent years has become a popular 144MHz portable antenna, with unidirectional characteristics.

Now Rod Newkirk, W9RRD, has come up with an antenna that combines the relatively little-used 8JK end-fed arrangement with the 135° out-of-phase approach. This has been implemented as an indoor 21MHz fixed wire-beam pushing signals into Europe from the Chicago area with less interference from the western States than with the basic 88JK end-fed array (Fig 3) from which the 'BRD Zapper (QST June 1990, pp28- 29: 'The 'BRD Zapper A Quick, Cheap and Easy "ZL Special" Antenna') was developed with the same basic dimensions.

Fig 3
Fig 3. W9BRD's WBJK gamma-fed wire beam for 21MHz made to fil the dimensions of the bedroom it occupies. Dimensions are not critical for the elements but they should be close to λ/2 (total each), with the stub about λ/4 or multiples. Phasing point is fairly critical and should be selected as described.

Instead of feeding the elements 180° out of phase, they are driven 135" out of phase by feeding the stub λ/16 from the stub's shorted end so that the feed path fo one element is 1/8λ (45°) longer or shorter than the other, which, after the stub's 180° phase reversal, produces the required (180-45 = 135° ) phase shift. W9BRD points out: "You can find the proper feed point on the stub by 'sniffing' signals of known origin along one side of the stub with the insulated centre conductor of some coax hooked to a receiver. (Start looking for this point by measuring λ/16 up from the bottom.) Directivity is reversed by selecting the opposite side of the twin-lead stub. Pattern distortion through incidental radiation and pickup must be minimized. The coupler must be built in the most compact form possible, mounted right at the stub, and isolated to keep the feed line from distorting the pattern. Such isolation is done at W9BRD via a home-brew coaxial choke made of 30 turns of the antenna's RG58 feed line wound on a ferrite rod just before the matching network: see Fig 4(b). The single wire run from matching network to antenna had better be no more than an inch or two. Here, I'm borrowing on the single-wire-feed theme by Windom. Since any circuitry above the choke will be hot with RF a bulky matching unit will not do. System Q is lower with wider spacings but gain is maximum with the elements spaced 1/8λ."

Fig 4
Fig 4. W9BRD's WBJK converted Into the unidirectional 'BRD Zapper. L consists of 10 turns of No 16 wire, 1.5in dia, space-wound. Broadcast-type variable capacitors can be used at powers up to about 100W. Adjust inductor tap for lowest SWR. Coaxial choke made of 30 turns of RG-58 cable wound on a ferrite rod.