Variations on a Theme

Glen E. Gardner, Jr.



I have edited this page for correctness. Some of the things I said in the original page were just wrong and really bothered me a lot. This page is old and is mostly left up for historical reference. If you want to see the latest version of this antenna, with a lot more information, and some of the math behind it, click HERE.

The task at hand

I am an apartment dweller in an urban area. Until recently I had ruled out HF operations at my apartment due to an inability to have an outdoor antenna. A few years back, I had built a magnetic loop antenna for receiving that covered 600 KHz through 4MHz. It worked extremely well, so I published it on the internet. Later I built some small, two-foot single turn magnetic loop antennas that covered 40m through 15m, and later 80m through 15m. Recently the bug to have an HF station bit me again and I began experimenting with a variety of small indoor antennas for HF, with results that varied from marginal to poor. After a few months of ruling out things that did not work well , I revived a 24 inch single turn magnetic loop antenna that had been languishing in the closet for several years. Amazingly, I began making a few contacts, but the results, while better, were still not very good. However I was hearing a lot of dx on HF with it, and the antenna seemed to show a lot of promise. Eventually I decided to build a more efficient design intended from the beginning for transmitting on the HF bands.

So what kind of antenna is this?

This antenna is a single turn magnetic loop antenna. It is optimized for transmitting on the HF bands. It makes use of parallel conductors in the main loop to improve efficiency. Other than the parallel conductors it is a rather generic magnetic loop antenna. If you were looking for a page about multiple turn receiving loop antennas for HF, I have a page for that too. In that event, go HERE.

Killing the Sacred Cows

I wanted to build a new antenna, I wanted it to work well, and I wanted to explore some new ideas. I really wanted to get outside the box on the thinking of magnetic loop antenna design and construction and to learn a thing or two. This prompted me to set the following requirements;

  1. The antenna must be efficient, with good efficiency on the bands in use.

  2. It must use all aluminum construction.

  3. It must use off the shelf hardware, available from hardware stores online and in brick-and-mortar businesses, or hardware that can be hand fabricated from readily available materials.

  4. The antenna must use bolt-together construction. No custom machining, no welding, just hardware, nuts, and bolts.

  5. The antenna must be capable of reliably handling 100 watts, minimum.

  6. Where possible, use novel ideas, and challenge the established rules.... don’t follow what others are doing, but learn what others are doing and build the antenna based on research, testing, and understanding, not popular mythology.

  7. The antenna must be small enough to fit into my small apartment, but large enough to be a good performer.

With the above requirements in mind, I set out to determine what materials were available. After a bit of searching, I found a source for 30 inch long 7/16 inch diameter aluminum rods that were tapped internally for 3/8 X 24 threads. Since 3/8X24 was a more or less standard size for antenna hardware , and the length of the aluminum rods were close to what I had in mind, this seemed an appropriate starting point for my new loop antenna.

I wanted the antenna to be efficient, so this meant that I had to consider problems involving skin effect and dc resistance. For small single turn magnetic loop antennas, the radiation resistance of the loop is often extremely low. The end result is that a high percentage of the transmitter power can end up being wasted as heat in the ohmic losses of the antenna if they become significant compared to the value of the radiation resistance. The dc resistance, and losses due to skin effect must therefore be kept as low as possible. Usually the resistive and skin effects are reduced by using large diameter conductors and one-piece, welded, soldered, or brazed construction. Bolt-together construction is taboo according to many small transmitting loop aficionados, but the number of people who have built small transmitting loop antennas using bolt-together construction and getting good signal reports is strong reason to dispute that claim.

One-inch diameter and larger conductors made of copper in small HF transmitting loops appear to be a common thing to use. The problem is that copper is heavy, not very strong, and often expensive. My solution was to use multiple aluminum conductors in parallel to reduce the resistive components and to reduce weight. I settled on two parallel conductors connected to the tuning capacitor, with a third conductor to be used as a pickup loop. The connecting bits of aluminum, and associated hardware were made more massive than mechanically necessary in order to reduce the dc resistance and skin effect losses. The three loops were mounted at the corners on five-inch long insulators made from two-inch by 1/4” thick electrical grade fiberglass angle stock. The 30”, 7/16” diameter rods are connected at the corners of the loop by one-inch wide pieces of 1/4” thick, 2” aluminum angle stock drilled for a 3/8” bolt, providing a loop that is 32” wide. The connecting aluminum angles were mounted on the fiberglass insulator (the three loops are insulated from each other) and the whole thing was bolted together with 3/8 X 24 nonmagnetic stainless hardware.

The fasteners and materials used in the antenna mount were made of non-conductive materials. One exception being the four ¼ bolts securing the bottom pipe flange in the pedestal, which were made of stainless. The antenna was mounted to an 8” X 12”, 1/4” thick garolite plate using four 1/2” fiberglass U-bolts. The garolite plate was bolted to a 1 1/2” threaded PVC pipe flange with eight 1/4” nylon bolts , and the pipe flange threaded onto a pedestal made using 1 1/2” PVC pipe.

The tuning capacitor is a surplus vacuum variable (5-500pF, 5KV) connected to the two outer loops by metal plates in one corner of the antenna. The middle loop was the pickup loop/driven element and was originally connected at the opposite corner to an appropriate matching transformer that provided an impedance match to the transmitter. This approach to impedance matching later proved to be problematic, and was later changed to a small pickup loop because the extremely low impedance (measured as approximately 280 milliohm at 7MHz) presented by using the full size pickup loop was difficult to match on all bands. It also turned out that the frequency response of the rf matching transformer did not roll off appropriately, and the feed impedance of the antenna increased with frequency, possibly because the radiation resistance of the main loop went up as frequency increased. Furthermore, the large number of turns on the primary of the ferrite core rf step-down transformer had resulted in several unwanted self resonances in the HF and VHF frequency range. Using a small pickup loop, 1/5 the diameter of the main loop, resulted in a reasonably good impedance match from about 5 MHz through 29 MHz and eliminated the unwanted resonances. A slightly asymmetrical placement of the pickup loop helped allow for an unbalanced to balanced feed transition and simplified the construction of the pickup loop. The rf transfomer idea will be revisited later.

Serendipity is a wonderful thing

I was concerned about proximity effect when using parallel conductors for the main loop. If the spacing was too close, losses might become excessive. Too far, and the conductors might begin behaving as independent radiators. I had initially settled on two inches as the spacing between the conductors of the loop antenna. It was an arbitrary choice that had left me a little apprehensive about the conductor spacing in general. By accident, I stumbled onto a US Navy research paper published in 1971, on this very subject. It turned out that they had studied the problem that I was concerned about, and that they had arrived at a simple means of determining an appropriate spacing between parallel conductors. It turns out that when it comes out to choosing the spacing between parallel conductors, more space is better, and that the number of parallel conductors has an effect on the optimal spacing between the conductors. They went a little further and pointed out that proximity effect is a real consideration and will interact with skin effect to significantly increase the effective resistance of the conductors if the spacing is too close.

Based on their work, I found that while two inches was an acceptable spacing, a spacing of one-inch between centers was determined to be optimal for three 7/16” diameter conductors. Since I was to use one-inch wide aluminum brackets to connect the corners of the loop, a one-inch spacing between the centers of the metal rods was impossible because it would short circuit the parallel conductors at each corner of the antenna. After some thought, I decided that a spacing of 1.5 inches between centers would be more than adequate for the expected potential differences between conductors, providing ½ inch insulation between the adjacent aluminum angle brackets that connect the corners of the loop conductors. After switching to a much smaller pickup loop, the spacing between the parallel conductors of the main loop was increased to 3 inches in order to use the existing hardware.

Weebles wobble but....

After construction, the antenna proved to be quite strong but so flexible that it wobbled excessively during tuning or when rotating the antenna. Tensioners were added by means of using 80-pound test braided dacron fishing line (catfish line) strung diagonally across the sides of the loop to provide additional stiffness. The catfish line helped the problem a little, but it is still too flimsy for my liking. This is another issue that needs to be worked out a little better.

Tuning the antenna

The antenna tunes from about 5 MHz through 29 MHz. Tuning is sharp, and becomes progressively sharper as frequency increases due to the nonlinear tuning presented by the linear motion of the vacuum variable capacitor’s screwjack. The antenna is difficult to tune at frequencies above about 21 MHz. Holding Q constant and increasing frequency results in a bandwidth that gets progressively wider as frequency increases. This behavior is present in this antenna. On 40 meters a 2:1 SWR bandwidth of 16 KHz was noted. On 20 meters, a 2:1 SWR bandwidth of 67 KHz was found, and on 15 meters the 2:1 SWR bandwidth measured 100 KHz.


The initial test setup provided a low 1.1:1 SWR from 7 through 25 MHz. After the final assembly was completed and the bulkhead mounted coaxial connector was installed the VSWR climbed to about 1.6:1 on 15m. This was apparently caused by the additional length of the aluminum strapping that connected the SO-239 to the pickup loop. Reverting to an earlier configuration improved things a little. The swr is still a little higher on 15 meters, but drops off to 1:1 at 29 MHz.


This antenna is deceptively quiet. The very sharp, high-Q tuning seems to make tuning extra difficult.

The standard trick of peaking the tuning capacitor for maximum noise simply does not work very well with this antenna. You pretty much have to peak on a signal or adjust the antenna using low power for minimum SWR. Initially, I thought there was something wrong with the antenna, and perhaps the Q was too low. Tests of the 2:1 SWR bandwidth and the definite sharp peak of the tuning capacitor when tuning up clearly indicated that the Q of the antenna was more or less what it should be. At the time that I began testing this antenna I had just put a Kenwood TS-480 into service, and it does appear to have a really low noise figure that plays well with a magnetic loop antenna, so maybe that is what I am reacting to. I can’t prove it, but I think that maybe my obsessive adherence to the use of non-metallic fasteners, and the use of non-magnetic stainless hardware everywhere that metallic fasteners were needed has possibly provided a lower than average noise figure in this antenna.

On the air

My first contact was on 40 meters CW. I called an individual in Chicago. When I informed him that I was using a 32 inch mag loop, he admonished me to “get a real antenna”. Interestingly enough, my second contact on 40 meters CW with this antenna was in Switzerland.

The next day, I tried 20 meters and was pleased to discover that I could easily make contacts in Europe, with good signal reports. Later, I tried 15 meters SSB with good results into the Caribbean.

The antenna is not very efficient on 40 meters. Signal reports are often weak, or I get a 599 signal report initially, but later told that QSB was a problem. However I am able to use the antenna on 40 , where I once had no antenna at all for 40 meters. On 20 meters and 15 meters I am able to make contacts easily, and signal reports are usually good. The antenna clearly works better on the higher frequency bands. Simulations using online magnetic loop simulators place the efficiency of the loop at about 20% on 40 meters, 70% on 20 meters and 95% on 15 meters. My experiences with using this antenna seem to support the information supplied by the online simulators, with the exception of self resonant frequency and usable tuning range.

More Tests...

I reworked the antenna to have four parallel conductors in the main loop instead of two.

I was expecting to see the Q of the antenna increase by as much as a factor of two. Instead, Q only went up by about 25 percent on the 40m band.

With four parallel conductors in the main loop, preliminary measurements of the 2:1 SWR bandwidths were; 40m: 13 KHz, 20m: 60 KHz, 15m: 100 KHz, and 10m: 160 KHz. Taking the 2:1 SWR bandwidth as a measure of Q, there seems to be a significant increase in Q on the lower HF bands with a four conductor configuration being measurably higher for LF and low HF. The interesting part is that the apparent Q of the antenna is better on the lower HF bands than on the higher HF bands. Adding more conductors did not measurably change the Q of the antenna at frequencies above about 14MHz, but did improve the Q significantly at the lower end of the tuning range, well into the 40m band.

Experiments on 6m....

This antenna is only 8.5 inches wide! It is not exactly a VHF DX antenna, but I have had great fun making contacts on 6m during sporadic E openings. I live in a third floor apartment and the antenna sits on the window ledge on the opposite side of the room from my operating position.

Using spare parts from the HF loop and a 3200V, 15-150pF transmitting air variable tuning capacitor from my junkbox, I built a small, 8.5 inch diameter magnetic loop that tunes from below 21MHz through 54 MHz. All the parts for the little antenna are the same as the HF transmitting loop except the length of the aluminum rods are shorter. The rods in this case are only 6 inches long (spare rods for the pickup loop for the HF antenna) instead of 30 inches used as the main loop for the HF antenna, and the pickup loop is made from a piece of flat 15 gauge copper wire 8.5 inches long instead of the 24 inch circumference pickup loop used in the HF loop. It comes out to 1/5 the size of the main loop after accounting for the mounting plate, the straps connecting the tuning capacitor, plus the main loop circumference. The SWR is 1:1 at resonance. The tuning is so sharp, I really need to add a vernier drive to the tuning capacitor. The 2:1 SWR bandwidth on 6m is about 200 KHz.

I was particularly interested in the six-meter band. After a few days of casually checking in on 6m with the little antenna placed on the window sill, the band finally opened and I immediately worked a station in Georgia. I was given a 599 signal report on CW. A number of W6 stations were plainly heard, and many, many W4 stations. An EA2 prefixed station was faintly heard. I am a little amazed that this tiny antenna works so well.

Using the 2:1 SWR bandwidth as a measure of Q, I discovered that just like the HF loop, the dual conductor 6m loop exhibits a higher Q at the lower end of the tuning range. In the case of this antenna Q appears to be about 400 on 15m and about 250 on 6m. I need to make more measurements to be sure I have it right. I’ll post more info as things develop, so expect these numbers to change a bit.


RIP 6M Loop! Last fall I dropped the 6m loop while moving it and destroyed the tuning capacitor's insulators. It was a fun antenna , but I have no plans to rebuild it for now.

Parts List for the HF Antenna

NOTE: The 30 inch internally threaded rods are no longer available from mcmastercarr.com

They still sell the same rod in 24 inch lengths, which should be suitable for those wishing to build a smaller loop.

Alternatively M^2 Antenna company does sell a 1/2 inch diameter , 36 inch long mast (meant for use with their HoLoop products) tapped with 3/8 X 24 threads. They are a bit pricey at nearly $50 each, and you might find it a lot cheaper to have them made at a local machine shop.

Main loop components

8 (eight) 7/16 inch dia, 30 inches long, internally threaded rods (3/8 X24 threads).

6 (six) one-inch wide angle bracket made from 2" long X 1/4" thick aluminum angle stock. Drilled with a 3/8 dia hole, 3/4 inch from the tag end on both legs. DO NOT oversize these holes.

4 (four) five-inch long fiberglass insulators. Made from two inch electrical grade fiberglass angle stock, 1/4 inch thick. Drilled to accept above angle brackets. Two holes in each leg, spaced 3 inches, 1 inch from edge of fiberglass angle. The insulator opposite the tuning cap has three holes, spaced 1.5 inches (middle hole is to mount the pickup loop). See the pics on the web site. NOTE: I found that assembly was made a little easier by oversizing the holes in the insulators a bit to allow for loose tolerances.

1 (one) 4X4 inch 1/8 inch thick aluminum plate (rear tuning capacitor mount).

1 (one) 4X7 inch, 1/8 inch thick aluminum plate (front tuning capacitor mount).

1 (one) 4X4 inch, 1/8 inch thick aluminum plate (bottom capacitor mount).

1 (one) four inch long, 1 3/4 inch aluminum angle, 1/8 inch thick (capacitor mount).

16 (sixteen) 3/8 X 24 stainless steel bolts. 1.75 inches long.

12 (twelve) 3/8 lockwashers.

32 (thirty-two) 3/8 id , 13/16 od, 1/16 thick stainless steel flat washers.

6 (six) 10-32 screws, and hex nuts, 3/4 inch long (for attaching the front tuning cap mounting plate and angle bracket).

Pickup Loop components

2 (two) 7/16 dia, six-inch long, internally threaded rods (3/8 X 24 threads)

1 (one) one-inch wide, 1/8 inch thick, 7.5 inches long aluminum strap, drilled with 3/8 inch holes, 3/4 inch from the ends.

1 (one) one-inch wide, 1/8 inch thick, 5.25 inches long aluminum strap, drilled with 3/8 inch holes, 3/4 inch from the ends.

1 (one) angle bracket, one-inch wide, made from 2 inch aluminum angle stock, 1/4 inch thick, drilled with 3/8 inch holes 3/4 inch from tag end.

6 (six) 3/8 X 24 stainless steel bolts. 1.75 inches long.

6 (six) 3/8 lockwashers.

8 (eight) 3/8 id , 13/16 od, 1/16 thick, stainless steel flat washers.

2 (two) aluminum mounting brackets for connecting pickup loop to SO-239 connector mounting plate, make to fit.

1 (one) so-239 connector, mounting plate, insulator, hardware, 4.5 inches square, 1/4 inch thick black garolite, batelite, phoenolic, fiberglass, wood, etc.... make to fit.

Pedestal and Mount

2 (two) plastic 1.5 inch threaded pipe flange.

2 (two) 1.5 inch PVC to threaded adapters

1 (one) 24 inch 1.5 inch PVC pipe

4 (four) 1.5 inch PVC pipe

1 (one) 1.5 inch PVC X fitting

4 (four) 1.5 inch PVC pipe caps

4 (four) 1/4 inch X 20 stainless steel machine screw, 4 inches long.

4 (four) 1/4 inch X 5/8 OD stainless steel flat washer

4 (four) 1/4 inch X 20 stainless steel wingnut

8 (eight) 1/4 inch X 20, 1.75 inch long, nylon hex head bolt

8 (eight) 1/4 X 20 nylon hex nut

4 (four) 1.5 inch fiberglass U bolts

4 (four) 1.5inch X 2.5 inch, 1/4 inch thick, garolite spacers, drilled for the U bolts.

1 (one) 7 inch X 12 inch, 1/4 inch thick garolite plate. Drilled with holes to bolt to flange , and to secure antenna to the plate with U-bolts and spacers.

Where to get stuff

You can get everything except the tuning capacitor and SO-239 connector at mcmastercarr.com

The tuning capacitor that I used came from surplussales.com

You can get the aluminum from metalsonline.com