Everything Old Is New Again

L. B. Cebik, W4RNL

Sometimes an old article in a ham magazine shows you that some of your finest achievements are only extensions of much older work. That lesson came home to me in an article (actually, 2 articles in one) in QST for October, 1937. Peter Dodd, G3LPO called my attention to the piece, because it had a forerunner of the VK2ABQ square. Actually, the articles had a lot more.

The composite article, called "Concentrated Directional Antennas for Transmission and Reception," (pp. 27-30), was put together because "Rhombics and multiple arrays of conventional form give high gains--but even for the 14-Mc band, they take considerably more yardage than most of us have available. Therefore, concentrated directional systems which are more readily fitted into the usual back yard have a distinct appeal. . .." It is interesting that the editor chose to call these small and possibly directional antennas "concentrated." The term conjures many images, although nothing clearly electronic. Such imagery--sometimes useful, sometimes misleading--was more common in the 1930s than it is today.

The second thread that connects the two articles is that the first is by John L. Reinartz, W1QP. The second is by Dr. Burton T. Simpson, W8CPC, based on some suggestions made to him by Reinartz.

The Half-Wavelength Loop

Reinartz writes on "Half-Wavelength Loop Antennas." One of his basic designs is a square, 1/8 wl on a side. The square is open for the feedline in the middle of one edge, and in the middle of the opposite edge is a gap. Reinartz analyzes the antenna to have a low feedpoint impedance, and uses 72-Ohm line to feed it. The basic design appears in Figure 1.

Interestingly enough, in another notebook item in this collection, The IL-ZX Antenna for 40 Meters, I presented a compact loop of intermediate size. The major difference between the Reinartz design and the one in my notes is the use of a double winding. The The double IL-ZX winding acts as an impedance transformer to raise the very low loop impedance to coax levels. The double loop, if brought to something of a point on each side of the gap, permits easier antenna adjustment for resonance.

Nonetheless, it is good to know where (or perhaps only approximately where) the design had its start. The field strength noted by W1QP is recorded as "about 28%, as compared to a straight half-wave dipole." The double loop shows a better performance level, being down from a modeled free space dipole by only about 1.1 dB and just a little more than that when set up near ground as a vertically polarized antenna. Since it is not possible to devise a perfectly comparable placement vertically over ground for both the loop and a linear dipole, gain comparisons are elusive. The antenna shows about a 4 dB front-to-side ratio, but not more than about 0.1 dB front-to-back. Reinartz believed he got a significantly higher front-to-back ratio, which I have not been able to replicate in models.

In 1937, the typical amateur station tuned both the antenna and the rig together, adjusting the antenna and the coupling to the final tank (parallel tuned) circuit simultaneously. The era of fixed or even narrow range of output impedances for transmitters had yet to emerge. Further adjustment of the impedance match involved simply widening the feedpoint gap, which provided a form of delta match for the feedline. Hence, determining the actual operating conditions of the W1QP 1/2 wl loop is nearly impossible without further precision in the account. However, such precision, even if available, would have been spurious to the ordinary ham of the era, since the means of achieving maximum power output without exceeding (catastrophically) the plate current ratings of the final amplifier tubes varied considerably among stations.

Finding the origins of one antenna design is a nice day's work. Finding two is serendipity.

The 1-Wavelength Square

Between the work of Fred Caton, VK2ABQ, and recent developments employing its offspring, the (Les) Moxon (G6XN) rectangle, there is a continuity. (Some minor improvements appear in other notes of this collection, for example, Moxon Rectangles for 40-10 Meters and An Aluminum Moxon Rectangle for 10-Meters. However, as Peter Dodd has pointed out in the 2nd edition of his Antenna Experimenter's Guide (pp. 99-100), there is a more distant predecessor, "A Square 'Signal Squirter' for 14 Mc." W8CPC credits W1QP for the suggestions that led to his proto-VK2ABQ square. The basic suggestion was that most of the radiation from a standard 1/2 wl dipole emerges from the center half of the antenna wire. The lower current end portions might be bent back with little loss.

The expression "signal squirter" seems today to be almost baby talk. However, in the 1930s. new operators were often called "young squirts," not just because many were very young (which is incidentally true), but because they were new to the game of "squirting" RF around the world. Imagistically, the rig and antenna formed the world's most perfect atomizer, with a mist finer than any perfumer ever dared dream. Note how easy it is to get carried away by the everyday metaphors of 1930s amateur radio.

The Simpson electrical design is shown in Figure 2. We can bypass the extensive reinforced and rotatable wooden square on which the antenna turned.

Simpson specifies a spacing of 16' 6" from front to back, and the wood frame he used, bristling with ceramic insulators to support the 1/4" diameter copper tubing suggest a similar dimension side-to-side. What affect the many metal reinforcement plates had on antenna performance is impossible to tell at this distance.

Like the VK2ABQ square, the W8CPC antenna separates two folded-back dipoles with a small end gap. The tubing ends were 18" apart, but each end had a brass rod insert to adjust the gap. Tuning up the antenna involved simultaneous adjustments to both the transmitter and antenna, using an RF ammeter across a temporary gap in the reflector. Again, the feedpoint gap was fairly large by current standards, allowing a variable length spread in the feeders as an aid to matching--where matching means maximum power transfer to the antenna. In this case, the transfer was measured in terms of reflector element current levels. When tune-up was complete, another brass rod closed the gap left by removing the RF meter.

Having modeled a number of VK2ABQ designs, I am aware of the critical nature of the gap in the end of the two elements. There are actually two gap settings that provide forward gain and a front-to-back ratio. The narrower gap provides more gain but what we would today consider mediocre front-to-back ratio, while a much wider gap provides a current on the reflector that is nearly perfect in magnitude and phase to yield a maximum rear null (>30 dB), but at the cost of over a dB of gain. (Similar results accrue to the Moxon rectangle, but with considerably greater ease of finding the right dimensions to use with no post-assembly adjustment required.)

The gap required for a maximum front-to-back ratio exceeds by a wide margin the range of adjustment available in the W8CPC design. Hence, I have to conclude that he adjusted his antenna for a quite narrow gap, something close to 1". The trick is the position of the gap. Using a 16' 6" square, one can terminate the driver 4" forward of the side-to-side center line. If the reflector is then brought within an inch of the driver, the loop can yield about 5.1 dB forward gain and about 5.0 dB front-to-back ratio. If the same gap size is transferred to the 8.5 and 7.5 inch positions, the gain drops to about 4.7 dB, but the front-to-back ratio exceeds 8.0 dB. This is about the maximum movement forward of the center-line within the +/-9" range of adjustability in the antenna. For both cases, the source impedance is less than 150 Ohms, considered low in those days, and both settings are within 18 Ohm reactance of resonance at the target frequency.

I cannot be certain that my models have captured the Simpson design exactly, especially in view of his claim of about a 7 S-unit difference in a DX report of his signal forward and reverse. However, they are indicative of the kind of performance that might be typically achieved by the antenna type. The numbers from these models are consistent with models of the 1" coat-button spaced wire antennas in the VK2ABQ collection. Clearly, the Simpson square is a close match to the VK2ABQ squares, apparently without direct lineal connection. And it follows that all contemporary variations on the square and the folded-back rectangle share a common root going back at least to 1937.

Appreciation of the Past

Am I finished with these antenna designs from the past? Not completely. If nothing else, there is that elusive front-to-back ratio claimed by both authors. So far I have not been able to reproduce it an models. Rather than conclude hastily that it is not there, I prefer to think that perhaps there are more models to try. If I find anything, I shall certainly add the results to this note.

Some antenna designers and analysts might be disturbed--even discouraged--by the discovery that their seemingly new offspring are intimately related to designs of the more distant past. I do not share that view. Instead, I am happy to pay homage to past work and to grow in my appreciation of how well basic design principles were known 6 decades ago. The 1930s comprise a fertile decade for design thinking in antennas, and vast strides were made in pressing wire antennas to their limits.

It would take the 1980s and 1990s to yield new techniques of software to allow some of the older design principles to be implemented with assured performance. Moreover, new ground is being broken daily in antenna design. However, we should not let contemporary development rates lead to the impression that "they knew hardly anything back then." They knew a very great deal--and developed ingenious ways of making it work in the absence of the complex antenna analyzers and other test equipment we have at our disposal today.

It is no slur against current design and analysis efforts to suggest that they mostly carry out the details of principles long known. Those details are crucial to making antennas perform to the peaks that the principles promised but that older measurement and calculation limitations could not assure. (One might get a sense of the importance of recent developments by asking when the 1920s work of Yagi and Uda started yielding really good antenna designs instead of really good antenna claims.) Perfecting the implementation of basic principles is a worthy enterprise in any era.

At the same time, new principles and techniques are also emerging today. Someday, they will be the more distant past to which we look with admiration. But at just this moment, October, 1937, will do as a past worth at least my own appreciation.

I want to thank W6TOY, W6ZH, W3/V31FA, and KE6RIE, all of whom came to my rescue in locating a copy of these studies. Several of these gentlemen expressed the hope that I could do something good with the articles. However, it is truer to say that the articles did something good for me.

Updated 10-28-98. L. B. Cebik, W4RNL. Data may be used for personal purposes, but may not be reproduced for publication in print or any other medium without permission of the author.

Go to Amateur Radio Page