Link-Coupled Antenna Tuners

L. B. Cebik, W4RNL

We have gone to great lengths to adapt unbalanced antenna tuners (ATUs or transmatches) to service with balanced lines. Among the schemes used are the following most common ones:

1. Float the tuner from ground and install a balun at the input end;

2. Install a balun, usually 4:1, at the antenna side of the tuner, to convert the balanced line to an unbalanced line;

Either system is subject to limitations. Floating the tuner does not guarantee freedom from common-mode currents that defeat balance. A 4:1 balun often reduces the already low impedance at the antenna terminals to a still lower one, and high reactances are normally unfriendly to the cores used in such baluns.

It is possible to build perfectly usable balanced network tuners. However, the parts count goes up, and ganging the components often presents further building and size difficulties. However, in the search for an ATU design that will handle balanced lines naturally, balanced PIs and Tees should not be overlooked by adventuresome home builders. A simple 1:1 balun on the transmitter side converts the resistive balanced input load to the unbalanced line for which the transceiver is designed.

To make a balanced network, chose a good unbalanced network as a model. Look at the series elements. Put one in each line in corresponding positions, but use half the reactance. For a coil, that is half the inductance; for a capacitor, that is twice the capacitance. Now look at the shunt component(s), those from the series line to ground in the unbalanced network. Use the same value across the two lines of the balanced network. This simple conversion leaves the system center floating, but that is rarely a problem. Do not forget to add the balun (1:1, 50 ohms) at the input. And remember that there is no safe place to put your finger when the system is hot.

A more classic alternative is the link-coupled or inductively coupled ATU. The figure shows the basic circuitry. The unbalanced input is inductively coupled to the main inductor. Since the mutual inductance between the coils is critical for maximum efficiency, the coupling is varied either by a movable link or by a series input capacitor.

The output is shown in a typical parallel configuration. To effect the best match, clips are provided to connect the line to the proper turns on the main inductor, as determined by maximum feedline currents. The split- stator capacitor or dual ganged capacitors can be connected to ground so that the system is balance with respect to system ground, as shown at *. However, the tuner will remain balanced if the ground is lifted.

The ** point indicates that the rotators of the capacitors may be separated and lifted from ground. In this case, they can be connected to the output terminals instead of using a coil tap. For further flexibility, the stators can be connected to various taps on the coil for maximum output. This configuration, often called the series connection, permits efficient matching to low impedances presented at the output terminals of the ATU.

To the best of my knowledge, no one in the U.S.A. is currently producing such a tuner commercially. Chapter 25 of recent ARRL Antenna Books has a simple tuner with some limitations in link-main coil coupling efficiency, but not a bad start. The only commercial unit I know of was made in Germany until a few years ago.

Despite their relative scarcity, link-coupled tuners have certain advantages over all other tuners. First, they are inherently designed for use with balanced feedlines. Second, they exhibit very low losses, even at high power levels with reasonable care of construction. Third, they can be configured to almost any balanced line condition that might face the operator.

However, they also have certain disadvantages. First, they do not lend themselves to automated or rapid changes in setting and configurations as one changes bands. Second, they work best when equipped with plug-in coil sets that permit the most optimal coil-link size ratios.

A network tuner provides the most efficient match when its variable components are truly variable rather than switched. This principle applies especially to the inductor. A switched inductor may provide a 1:1 match at one or more settings, but it does not necessarily permit the use of the most efficient (lowest loss) setting of the coil. Likewise, a single coil and link for all HF bands does not provide the best coupling ratios for all possible conditions. Without provision for coil tapping and series connections, the most efficient operating mode may be inaccessible, despite a 1:1 match.

For a "hurry-up" operator, these inconveniences may be worse than the losses inherent in current systems pressed into balanced-line duty. However, for operators seeking the most efficient transfer of power to balanced lines, nothing beats a properly designed and constructed link- coupled ATU.

Will we ever see link-coupled tuners on the market--or even the components necessary for us to roll our own? These include reasonably priced capacitors or capacitor sets, along with plug-in coil systems similar in concept to some of the old Millen sets. (I doubt that, even under the best of conditions, that we shall see reasonably priced swinging link sets.)

Here is a wide-open field for the entrepreneur who understands both the concepts of link coupled tuners and has the know-how to develop well constructed plug-in coil assemblies and allied components. If parts and designs were available, I'll bet there is a customer base that would make the venture mildly profitable at the 100-200 watt level and at least break- even at the 1500 watt level. (For QRP work, I would use nothing less than a 100-200 watt ATU to minimize losses associated with tiny layouts and components. I would reserve the super-tiny--for example, compact Z- matches--for field work, where size and weight reductions may overshadow potential losses.)

To this moment, I have received no direct word of a return of the Annecke tuner to the marketplace or of a modern replacement for the Johnson Matchbox. Parts, especially coil stock for the inductor and link, may have become prohibitively expensive. As well, there is a fiercely competitive market for single-ended network tuners equipped with various types of toroidal baluns (transmission-line transformers) to handle balanced lines. However, interest in inherently balanced tuner designs, such as the link coupled tuner and some balanced network tuners with 1:1 choke baluns on the 50-Ohm side of the system remains quite high. Interestingly, the latest and most competent of the single-ended network tuners seem to be priced in the range one might expect for a link tuner of the same power-handling capability. Balanced network tuners are even more costly. Indeed, the manufacturing economies inherent in network tuners may likely have removed the possibility of any return to commercially-made link-coupled tuners.

The Johnson Matchbox and the Annecke Parallel-Line Tuners

I have heard the Johnson Matchbox described as a modified Z match, which is not quite right. Having obtained one, I thought I might describe the circuit, which appears to be very similar to the link coupler offered by Annecke in Germany.

The circuit is a straightforward link-coupled circuit. The input with the relay and associated circuitry includes taps for a 50-ohm transmitter connection and a 300-ohm receiver connection, since receivers continued to used balanced input strips long after transmitters had gone to shielding, pi networks, and 50-ohm outputs.

The Johnson Matchbox uses a single link for all bands with no variability. The Annecke design does away with the relay, receiver tap, and other pre-50-ohm transceiver features. Instead, it uses a larger (overcoupled) link with taps for the various bands (80, 40, 30-20, 17-15, and 12-10) mechanically linked to the secondary coil tap switches. In addition, it employs a series variable capacitor to adjust coupling (or input impedance, which amounts to the same thing). The Annecke design is superior in this regard.

The secondary systems of both the Matchbox and the Annecke are almost identical, differing only in output connection options. The secondary coil is tapped at reasonable positions for 80/40/20/15/10 meters, shorting out the unused turns toward the outer ends. Although the Annecke is marked with the preferred settings for the WARC bands, Matchbox users will have to experiment on these bands. Across the outer limits of the coil is a split stator capacitor, center grounded, which is used to set the tank at resonance. The required value of capacitance will vary somewhat as the reactance and resistance at the antenna terminals is varied.

The terminals are not connected directly to the outer limits of the tank. Each side of ground passes to a differential capacitor. The center of each differential goes to the antenna terminal. A differential capacitor is a split stator variable arranged so that as capacitance on one side goes up, it decreases on the other. The antenna terminal on each side of ground is thus set at a certain reactance from ground and certain series reactance from the tank. This arrangement is said to form a voltage divider. It also forms a means of compensating for reactance at the antenna terminal of the tuner, allowing it to match a wide range of R+/-jX combinations that might be present at the antenna terminals and still present the requisite high impedance to the tank circuit ends. Because the entire series combination of capacitance (and capacitive reactance) appears across the tank circuit and the load, the function of the differential capacitors is not so simple as this brief note might lead one to believe.

The design goes back to AM days, so the 275 watt rating is likely not only conservative, but conservative to the power of the carrier plus side bands of a 100% modulated AM signal. Capacitors appear to be spaced for at least 3 kV or better. The KW matchbox uses the same circuit with beefed up components and only some slight connection changes. The Annecke unit is rated at 200 watts and appears (from the photo of the case and the control arrangement) to use slightly lighter components than the Johnson, even though the 200 watt rating is also conservative.

Perhaps the only thing I would do to improve the design is not mechanically feasible: to have a rotary coil with contra-rotating sections (in line) to permit the full span of taps in order to eliminate the switch. The goal is to make available the most efficient coil settings for every possible set of R+/-jX presented, but I have no idea of how to make that work mechanically and preserve the center link and main coil center efficiency.

I thought those who have never seen a Matchbox might be interested in the basic design. (Love those elegant Johnson panel colors used after about 1960, similar to the colors on my own QSL, but not intended that way.) I have come across two photos, one of the 275 watt unit without meter for SWR through the external directional coupler, and the other with the available couplers on top and the meter in place. The difference in the panel lettering is a clue as to which is older and which is younger.

The rearside detail is also interesting for those who might wish to home brew a link tuner or for those wishing to know what to expect from the unit. Jan Axing, SM5GNN, provided photos of the rear of his unit, which does not have the coupler. My thanks for his thoughtfulness. The 275-watt Match Box is about 10" wide, 10" deep, and 7.5" high.

Phil Grill, KA2IWR, sent me photos of the Nye-Viking version of the Matchbox. Nye obtained rights to the Matchbox when Johnson ceased production. The Nye version omits the relays from the start, since the transceiver had already become the main type of amateur rig. However, Nye either used surplus Johnson cases or the dies for punching cases, since the relay terminal strip cut-out is evident. Due either to increased difficulties in obtaining parts or the prevalent hype about single-ended network tuners, Nye ceased making the Matchbox clone and turned to a single-ended design. So Nye Viking Matchboxes are not common.

The circuitry of the KW version of the Match Box is virtually identical to that of the smaller 275 watt version, although considerably beefier. Instead of a ceramic terminal for single-ended antennas, there is a coax connector. For identificatioin purposes, here are photos, front and back, of the KW Match Box, but they appear to be of different units. The front shows the later paint scheme, while the back shows the earlier logo without the "J" behind the Viking. You can gauge the relative size difference between the 275 watt and KW versions from the directional couplers on top of the respective units. My thanks to Peter, SM5HUA, for the photos.

A copy of the Johnson Manual--used for both sizes of the Match Box--can be obtained from K4XL's BoatAnchor Manual Archive. Incidentally, the 1963 prices for the Johnson Matchbox were $64.95 for the 275 W unit less the indicator, $94.94 for the same unit with the SWR indicator, and $154.50 for the 1 KW Matchbox.

At some point in the production of Matchboxes, Johnson added an extended range tuner covering about 2 to 30 MHz, with obvious applications to the 160-meter amateur band. Harry MacLean, VE3GRO, sent me the following panel close-up of the extended range version that he has. It is a 275-watt model, and I do not know if there was an extended-range KW version of this design.

Note the "auxiliary" control at the top left corner of the main panel. As Harry describes the interior, "The coils--both main and link--are the same as in the 5-band ham matchbox, but the variable capacitors all have double the number of plates. The auxilary capacitors are switched in on only the three lower ranges. Ranges are 2-3.5, 3.5-5, 5-7 7-9, 9-14, 14-21.6, and 21.6-30 Mhz." Harry has used the unit on 160 meters be adding a couple of turns to the ends of the main coil.

My photocopies of the Annecke case are black and white, so I do not know the color scheme (although it appears to be black and white). The outer case thickness gives the impression of battleship construction. Catalog price (now out of date) was 495 DM. I have since learned (Jan. 4, 1999) that the Annecke company is no longer in business due to the severe illness of its owner. So far as I know, the designs have not been bought or otherwise picked up by any other company. It is indeed regretable that the only surviving major commercially-made link-coupled antenna tuner is no longer available. Now European hams can feel the loss of these units as keenly as we U.S. hams feel the loss of the old faithful Match Box. However, any entrepreneur will tell me that now the market is clear for the first manufacturer who presents us with a new match box. (Late note: Annecke may have been purchased by a Dutch ham. Will post more as it is learned.)

Courtesy of Jan Anker, ON4CAF/PA0LBN, I can show something of the inside of the Annecke unit. The first photo is of the inside from the bottom rear, showing the split-stator tuning capacitor to the left. The input tuning capacitor is to the right.

The second photo is of the interior from the top rear, showing the differential capacitor to the right (with the input capacitor nbow to the left). Since the two large capacitors are vertically aligned, a single photo cannot show both.

The last photo is of the coil assembly, which is remarkably similar to the scheme used in the Johnson Match Box. For maintenance of circuit balance, symmetry is essential.

The 200-watt Annecke unit is about 5" high, 10" deep, and 12" wide. Personally, I would have left a bit more room in the cabinet to reduce stray capacitance, but I certainly would not refuse to use the Annecke unit.

Again, courtesy of ON4CAF/PA0LBN, here is a photo of the schematic on the back of the Annecke unit. In case the cvalues are hard to read, the split stator os 155 pF per section, the differential is 100 pF per section, and the input capacitor is 270 pF.

One bad habit of all antenna tuner makers is to list a simple impedance range for the antenna matching capabilities, for example, 50-3000 ohms for the Annecke. Actually, the ability of any tuner to effect a match is determined by the combination of R and jX at the antenna terminals. Whatever the reactance-compensating scheme, it will have some limitations, usually being more effective with large values of either jX or -jX, but not both. Moreover, the range of impedances for which any given tuner design can achieve maximum efficiency is ordinarily quite a bit narrower than for achieving a mere 1:1 SWR on the transmitter side. However, bringing the impedance presented to the antenna terminals within the range of the tuner's maximum efficiency potential is usually only a transmission-line length change away.

In early 2004, Anthony (Tony) Brent, WD7G, sent me some useful notes on the availability of components for anyone wishing to replicate either of the classic link tuner designs. With his permission, I shall simply quote what he wrote.

I thought I would pass along some information to you that I have found that will be of interest to anyone who wants to build such a tuner.

I recently ordered new manufacture Johnson variable capacitors which are duplicates of those used in the Johnson Matchbox. Cardwell Condensers in New Jersey is manufacturing these units, and also Hammarlund variable capacitors under the original Johnson and Hammarlund part numbers, so these nice variables are indeed available again brand new in the box.

http://www.cardwellcondenser.com

I spoke with Paul Meyer, who was very helpful. He informed me that Cardwell can also manufacture custom variable capacitors to customer's specifications.

I ordered two Johnson 154 series capacitors, a single section 450 pF 2000 volt for $75, and a dual section 330pF 2000 volt for $150. See the attached photo. Not exactly ham fest prices, but not that outrageous either. A builder ought to be able to come up with a finished Matchbox for somewhere in the neighborhood of $250 to $300, I would think, and that isn't bad given the price of most of the mid- to high-power T-network tuners on the market today.

Cardwell also has various fixed and variable inductors, and the old stand-by Velvet Vernier drives.

Jackson Brothers in England also have variable capacitors available at their web site: http://www.mainlinegroup.co.uk/jacksonbrothers/index.htm

They also have the nice Jackson Brothers vernier drives and other parts that may be useful.

The remaining hurdle is the coils. B and W are still in the coil stock business, but they are quite expensive, and some times out of stock. I have experimented with a method of coil winding that the British and Australian amateurs have been using that works very well, and is quite easy to do.

Their trick is to use an acrylic or polycarbonate sheet laid out and drilled for the proper coil diameter and spacing. One set of holes is offset by half the coil spacing from the other so the coil will advance at the proper pitch when wound through the holes. One extra hole is added at each end of one line of holes for the start and finish of the winding. The other row has the same number of holes as the number of turns in the coil.

The wire is wound on a former a little smaller than the finished diameter, and then is threaded through the holes in the acrylic, which hold the coil to its proper dimensions and keeps everything in place while only touching the wire at two points along the coil. It also provides a secure means of mounting the coil by using brackets fastened to the acrylic. It works very well.

I imagine other materials could also be used. I happened to have some acrylic on hand so I gave it a try. I was very pleased with how well it worked.

See the attached photo, which is my attempt at a single coil Z match tuner following the design work of Lloyd Butler, VK5BR (http://www4.tpgi.com.au/users/ldbutler/) Interesting circuit that I am still playing with, but I think I still prefer the Matchbox circuit.

The same type of coil former ought to work well for winding a Matchbox style coil and link, and perhaps with some extra holes to hold the wires for the coil taps, so they don't depend only on the solder joint.



Updated 5-23-1999, 12-30-2000, 08-16-2002, 09-01-2003, 03-01-2004, 09-12-2004, 08-04-2006. � 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