Part V: Components, Construction,
and Measurement

L. B. Cebik, W4RNL





The remaining questions we need to examine with respect to inductively coupled ATUs are component values and ratings, construction practices, and measurements to assure best results. Let's look at these questions, one at a time. We shall be focusing upon parallel-tuned secondaries, with either taps on the inductor secondary or a capacitor-divider to accommodate a wide range of load impedances. Figure 1 provides an abbreviated schematic diagram of both types of circuits. We shall use CP, CS, and CD to sort out the capacitors in the primary, secondary, and (if used) divider circuit. LP and LS will designated the inductor primary and secondary windings.

Component Values

There is no magic set of values for the components in an inductively coupled tuner. Instead, there are a few situations we want to avoid and then a set of reasonable compromises to make. The situation to avoid is having too high a value of CS relative to LS. Lower values of inductance carry with them lower values of inductive reactance. This, in turn, lowers the Q of the coil for resonant combinations on the ham bands, resulting in greater power losses in the coil.

The 1960 ARRL Antenna Book recommends a secondary component reactance value of about 500 ohms as a reasonable compromise. This value tends to yield higher Q inductors whose physical size is not too much different than the physical size of variable capacitors having the desired characteristics. Here is a table of the resulting values, not only for the secondary components, but for the primary components as well (assuming a 50-ohms matching with the transmitter). I have extended the list to include the WARC bands.

              Recommended Component Values for a Link-Coupled Tuner
       Band           LS             CS             LP             CP
       160            42 micro-H     170 pF         4.2 micro-H    1700 pF
       80-75          22              90            2.2             900
       40             12              45            1.2             450
       30              8              32            0.8             320
       20              6              23            0.6             225
       17              4.5            18            0.45            180
       15              4              15            0.4             150
       12              3.2            13            0.32            130
       10              2.9            12            0.29            120

Table 1.  Recommended component values for a link-coupled tuner.

These recommended values tell us several very practical things. First on the list is the reason why commercial ATU manufacturers rarely try to cover 160 meters with an all-band tuner. The 160-meter coil will be roughly twice as large--for any given power level--as he 80-10-meter coil. Likewise, the tuning capacitor will be large. In the primary circuit, the series capacitor may require either ganging two variables or adding fixed capacitors in parallel with a 1000 pF unit.

In general, the component size question for 160-meter tuners also inclines builders to using a tapped coil secondary circuit (instead of the capacitor-divider system). Taps every 2 or 3 turns down to a coil size of 50% of the full length should suffice for most loads encountered. For a specific antenna system, once the correct taps are found, they should not need to be changed. Indeed, it may be possible to use a good ceramic wafer rotary switch to move between taps.

Second on our list of lessons from the recommended values is the practicality of an 80-10 meter link coupled tuner. For a very simple design, coil taps may be used, but the capacitor divider system may prove far superior in terms of its continuous range of adjustment. A dual differential capacitor of about 100 pF per section will generally suffice for the capacitor divider.

If we use the primary series capacitor, we may be able to combine the 12 and 10 meter positions to save one band switch position. Likewise, the 17 and the 15 meter position may be combined. A 3 section, 6 position ceramic rotary switch should suffice to cover all the bands.

Notice that the primary coil should also be tapped for band changing purposes. Depending on the coil construction, some builders may wish to cover more than 2 bands with a single tap, since tapping at partial turns may be physically inconvenient. Tapping the primary is normally done on only one end of the coil. Hence, the position of the primary coil may not be perfectly centered for all bands. This slight imbalance does not produce any significant negative effects.

The series circuit may be altered to use lower values of series capacitance by increasing the size of the primary inductor. Values up to double or triple the recommended value may be used on 80 meters to make use of series capacitors in the 350-500 pF range. As the frequency is increased, the coil taps may be brought closer to their optimum recommended values, since the required resonating capacitance would fall within the range of the smaller unit selected.

As one increases frequency toward the high end of the HF spectrum, the range of the capacitors is not optimum for smooth tuning. Even though circuit Q may be satisfactorily low, control positions may be very sharp. Ideally, one should consider separate tuners for 80-20 meters and for 20-10 meters. The latter unit may use smaller capacitors and more widely spaced inductor turns. However, for all-band doublets and similar wide-ranging antennas, the 80-10 meter ATU is usually the design of choice.

In principle, if space is available, paralleling a high value and a low value variable capacitor can provide more optimal control of capacitance. Both capacitors should have the same voltage rating. At 14 MHz and above, the larger capacitor is switched out of the circuit. Below 20 meters, the smaller capacitor is set at mid-range, with main tuning done with the larger capacitor. The smaller capacitor may then be used as a "fine tuning" control. Since this system requires extra space, an alternative is to use a reduction drive with the main tuning capacitor.

The capacitance values given in the chart are the total capacitance across the parallel- tuned secondary circuit. Split-stator or ganged capacitors are normally used to preserve balance across the circuit. Each section should have a capacitance of twice the listed value so that the series combination equals the recommended value.

Varying the circuit values does little harm to coupling efficiency so long as sufficient flexibility is maintained in the primary series resonant circuit and in the secondary load impedance transformation circuit. Limitations show up chiefly in the range of load impedances that the coupler can effectively handle, and careful line-length adjustment can usually provide load values within the capabilities of the coupler.

Component Specifications

The best way to look at the components, is one at a time.

Inductors: For moderate power levels up to 200 watts or so, air-wound inductors with a diameter of 2 to 2.5" and 8 turns per inch (tpi) are very practical. They provide reasonably sized inductors with good Q across the HF range. Inductor Q will normally decrease at the highest frequencies in the range. For 160 meters, a coil of larger diameter may be required to hold its length within reason.

Placement of the link at the center of the secondary usually follows one of two physical designs. In simpler designs, the link may be several turns of the main coil stock. The turns adjacent to the limit of the link are tied together to provide continuity in the secondary. Although this system will work, it limits the range of coefficients of coupling. Tighter coupling can be obtained by placing the primary inductor over the secondary. For fixed links using typical air- wound inductor stock, a nonconductive adhesive can bind together the support bars of the coils.

The current which the parallel-tuned secondary coil must handle can be estimated from the following simple equation:

where IC is the estimated maximum circulating current in amps, QL is the loaded or working Q of the coupler, P is the power level in watts, and RL is the load resistance in ohms at resonance. For a maximum Q of about 10, a power level of 100 watts, and load resistance of 5000 ohms, the maximum current will about 1.4 A. If we use the ideal case in our running example of a load resistance of 1500 ohms and a Q of 2.8, the current is only 0.7 A.

Increasing the power to 1500 watts from 100 watts increases the circulating current by the square root of the power ratio. For the two sample cases, the maximum current will be 5.4 A and 2.7 A, respectively. For high power applications, #12 wire is generally satisfactory for these levels, while #14 wire may be used at mid-level powers.

For very low power levels, such as those encountered in QRP work (5 watts or less), it is theoretically possible to use wire as fine as #20 or #22. However, at very low power levels, every effort should be made to minimize power loss. The high power wire sizes do not guarantee minimum loss, but only that power losses in the inductor wire are not problematical. Minimum loss wire sizes would be larger, and at QRP levels #18 wire or larger is always in order for inductors with minimal losses.

The current in the primary winding is a function of the coil reactance and the power level. For example, if we use a coil with a reactance of about 50 ohms, the current at 100 watts power will be about 1.4 A. Viewed another way, the current will be about the same as the maximum circulating current in the secondary tank circuit. Hence, wire size recommendations applicable to the secondary are also applicable to the primary.

Capacitors: The chief problem facing inductors is heat from the conversion of RF currents. This problem has a time domain, and brief periods of excessive current are often without harm. With capacitors, the chief problem is arc-over, which may result from virtually instantaneous peak voltages across the capacitor plates.

The voltage across the primary series capacitor will be a function of the capacitor reactance and the power level. If a 50-ohms capacitive reactance is used, then at 100 watts, the voltage peak will be 1.4 times the r.m.s. voltage across the capacitor or about 100 volts. At 1500 watts, this peak voltage will increase by the square root of the power increase to about 385 volts. These levels are within the abilities of various sizes of receiving capacitors, which are often employed to achieve the high values of capacitance needed at 80 meters.

If the 80 meter capacitance is reduced to 1/3 of the recommended 1000 pF, the capacitive reactance will increase by a factor of 3. In this case, the peak voltage at the 100 watt power level will be about 170 volts. At 1500 watts, the peak voltage reaches about 665 volts. More widely spaced capacitor plates will be required.

In the parallel-tuned secondary circuit, the peak voltage is simply 1.4 times the r.m.s. voltage across the tuned circuit. The line voltage is a function of power level and load resistance, where

For the example using a load of 5000 ohms at 100 watts, the peak voltage is about 1000 volts. Where the load is 1500 ohms at the same power level, the peak voltage drops to 540 volts. At the 1500 watt power level, these peak voltages increase to 3870 and 2100 volts, respectively. However, off-resonance voltage peaks may be considerably higher.

The actual arc-over voltage for a capacitor depends on many factors, including the sharpness of the edge of the capacitor plates and the air quality and humidity around the capacitor. In general, capacitors are chosen with a good reserve. 1500-volt units are common at 100 watts, 3 kV units at 250 watts, and 7 kV of higher units for the legal amateur power limit. These values provide about a 2:1 safety margin in balanced circuits using split stator or ganged capacitors, where each unit of the whole capacitor sees only half the total peak voltage across the line.

An often overlooked aspect of capacitor construction is the size of the capacitor frame and its materials. Large, closely space metallic frames can restrict the minimum value of capacitance obtained by a variable capacitor. E. F. Johnson units used in their Match Box series of link tuners employed the minimum metal frame to support the capacitor and permitted very low values of minimum capacitance. These or similar units are desirable in tuners designed to cover the entire 80 to 10 meter range of amateur bands, especially in the parallel tuned secondary of the coupler.

Capacitor construction is equally important in maintaining circuit balance with respect to ground. Split-stator capacitors provide an inherently balanced structure, with roughly equal influences on both sides of the circuit from stray capacitance to a metal case or other metallic objects in the circuit. Single-section capacitors, while usable, tend to unbalance the circuit by coupling more capacitance through the larger structure of the frame than through the set of plates not connected to the frame.

For capacitor-divider circuits, the capacitors form a series chain of 4 units across the line. Each unit sees about a fourth of the total line voltage. However, with considerable reactance on the line, the voltage across each unit may be somewhat higher. Therefore, the voltage rating of the dual differential capacitor is usually set to be the same as for the split-stator tuning capacitor.

When an inductive coupler is undergoing initial tuning, the control combinations may result in very high voltages across capacitors in the circuit. Therefore, initial tune-up should always be done at the lowest power possible to prevent component arc-over.

Rotary Switches: Rotary switches used to change bands or coil taps should have large, well-spaced contacts. The material should be ceramic, rated for RF service. For medium power levels, standard 1.25" wafers are normally satisfactory. For high power, use larger switching wafers with more widely spaced and large contacts.

Shorting switches--that is, switches that connect together all preceding switch positions-- are preferable to simple switches that leave preceding switch positions open. Shorting out the unused turns in the secondary coil normally results in fewer problems with power losses from circulating currents in those turns. However, only experiment can usually determine whether the interturn capacitance in combination with the inductance of the unused turns may result in a resonance at some harmonic of the operating frequency. For this reason, some designers add an additional coil position on multi-band tuners, roughly tuned to 5 to 5.5 MHz. Although this tap might prove useful on some occasion with particularly troublesome 80 or 40 meter loads, its chief function is to reduce the size of shorted inductor sections when operating above 40 meters.

Terminals: The input terminal for most ATUs will a standard coaxial cable fitting. Output terminals should be ceramic feed-through types. Ring or U terminals are normally used for both inside and outside connections to the threaded shaft that runs through the dual ceramic pillars. Steatite, developed three quarters of a century ago, is still the usual ceramic of choice for RF service in the HF bands.

Construction and Operation

The inductively coupled tuner is essentially a combination of passive circuits and produces no power of its own. Therefore, the shielding practices normally used for power producing equipment are to a large measure optional with antenna tuners. Perfectly operational tuners may be laid out on breadboards or placed within attractive wooden or clear acrylic cabinets. Perhaps the only operational caution with unshielded layouts is to insulate control shafts in order to prevent shock haxards and hand-capacitance effects. Safety to shack visitors (or to the operator) is a strong reason for enclosing the tuner.

If a metal case is used, it should be large enough to permit all components to be well spaced from metallic surfaces. This precaution reduces the introduction of stray capacitance, which can reduce the flexibility of the variable controls, especially at higher frequencies. With metal cabinets, long, sturdy ceramic stand-off insulating posts must be used for components that require isolation from ground.

Component layout should follow good RF practice, with attention to maintaining secondary circuit balance. Hence, leads to and from comparable points on either side of the center of the secondary should be as short as possible and of roughly equal length and proximity to adjacent components. Leads in the secondary should use wire of the same size as the coil winding. It is possible to space switch wafers to achieve this physical balance, and the rest is largely a matter of component placement.

The primary side should use short heavy leads. If the coil primary assembly is at some distance from the coax fitting, a length of coaxial cable may be used to connect the two.

A common ground of the smallest spread provides the least potential for excessive current circulating in this path. Even when metal cases are used and form the ground buss, care should be taken to use contact points in closest proximity to each other, commensurate with the use of short leads.

Input and output measurement circuits, if internal to the tuner, should be isolated from the fields surrounding the main components of the tuner. The DC and meter portions of the measuring circuits are best isolated by placing them in grounded metal boxes as far as possible from the main coil and capacitors.

The input side of the coupler is normally an unbalanced circuit. It requires a common ground. Ideally, this ground should be common with the other station equipment ground as well as the station earth ground.

The balanced secondary of the tuner presents the user with some options. Ordinarily, the center of the coil is left ungrounded (or "floating"), largely due to difficulties presented by the centering of the link over the center of the secondary inductor. The center of the split-stator tuning capacitor and the junction of capacitor-divider differential units are often grounded as a matter of construction convenience when using metal cases or chassis. This connection is, however, optional. (However, it is good practice to connect the junction of the differential capacitors to the common rotor of the main tuning capacitor.)

Grounding the center of the of the secondary circuit provides a common reference for the two sides of the feedline and the secondary circuit. However, it also provides a point of direct coupling for out-of-band signals. Such coupling is significantly reduced if all secondary components are left floating. Home constructors may wish to experiment to determine the better system for their individual situations. A floating secondary may be especially useful with 160- meter tuners, where strong AM broadcast band stations require all the filtering possible.

Tuning up the coupler is a matter of finding the correct control settings for maximum output and a 1:1 SWR at the input. Initial tuning is largely trial and error in the absence of definitive knowledge of the load resistance and reactance. For a tuner using a tapped secondary inductor, a trial balanced pair of taps is the starting point (at low power, of course). The secondary capacitor is resonated, as indicated by a dip in the SWR metering circuit that is in the primary line either inside or outside the tuner. The series capacitor is then adjust to the lowest SWR, followed by a series of alternate tweakings" of the primary and secondary capacitors. If the initial taps do not result in a 1:1 SWR, the next adjacent set should be tried--and so on until the match is obtained.

The goal is to use the set of taps closest to the outer limits of the secondary inductor that permit a perfect match. These taps represent the lowest working Q for the circuit and thus provide the largest bandwidth for satisfactory operation without further control adjustment. Ordinarily, they also provide maximum power output. However, as noted along the way, it is possible--although unusual--to find a perfect match while most of the current is circulating within the components rather than going to the line.

For the capacitor-divider coupler, the coil tap is usually fixed by a band switch. Adjustment involves alternate changes to the series primary capacitor, the secondary resonating capacitor, and the differential capacitor. However, the aim is the same: the lowest Q (and broadest tuning) that still permits a perfect input SWR reading. Since an opaque case and simple reference marks tend to obscure what is happening with the capacitor-divider, the bandwidth of an adjustment set may be the only indication that the most satisfactory match has been obtained.

All final setting should be logged and attached to the tuner case. Not only do they ease the adjustment when changing frequencies, they also provide a reference for antenna system diagnosis. If the required settings drift with time, vary radically with certain weather changes, or change suddenly to a new permanent set, antenna and feedline maintenance is indicated.

Measurement

Two measurements are important to any coupler: the match of the input circuit with the line to the transmitter (and receiver) and the relative power output. Ordinarily, we make the first of these measurements and simply presume that the second needs no measurement. The ordinary is simply not good radio practice.

The input monitor usually consists of an SWR metering circuit mounted within the tuner cabinet or placed in the line between the tuner and the transmitter. Either system works well. Since SWR circuits are legion, no further comment on them is required, except perhaps for the reminder that the DC metering portion of the circuit should be well shielded from the fields of the tuner components.

Output measuring has long been unnecessarily difficult, since standard recommendation call for RF ammeters, which are difficult to find, especially in the ranges useful for the wide variety of amateur power levels. For higher power, light sources coupled to high-voltage positions along the line have provided an alternative power output indication.

Since the coupler may encounter a wide range of load impedances whose actual values are not easily determined, true power output is not the measurement of choice. Rather, efficient operation of the coupler is a matter of achieving the highest possible power output for any given input. Hence, a relative power output indicator is sufficient for most non-laboratory applications.

For any given load impedance, both the current and the voltage will rise as power is increased. Both rise as the square root of the power increase. Therefore, a simple voltage sampler may be connected to the output terminals of the coupler and left in the line permanently. The sample voltage can be rectified, filtered, and then measured with a voltmeter. For relative readings, where we wish to track the rise and fall of the voltage as we adjust tuning controls, an analog meter is preferable.

Figure 2 shows the basic schematic diagram of a relative output meter. High value voltage dividers go from each terminal to ground. The resistors should be in the Megohms, but the exact values should be determined experimentally relative to the range of impedances and power levels used by a particular station. Diodes provide full-wave rectification of the sampled RF. At levels higher than QRP, use high-voltage diodes as a protection from reverse voltage breakdown. The potentially high impedance of the output circuit dictates a higher level of filtering than usually applied to low impedance circuits.

In this basic circuit, a simple potentiometer may act as a sensitivity control and thus be the only element within the "Amp. & Control" box in Figure 2. An FET op amp in a voltage follower circuit provides the voltage and current the voltmeter. It is wise to add zener diode or other protection to prevent meter damage from excessive voltage levels.

The circuit is open to innumerable improvements in the "Amp. & Control" section, especially for automating the meter range for the wide variety of sample voltage levels that the tuner may provide. A second voltage divider might be used as a reference to one of the bargraph LED driver chips to obtain up to ten decades of control. The chip output would not only light an LED telling the operator which decade was in use, but as well trigger a control chip that inserted the correct voltage reducing resistor for each step rise in power level. Of course, such improvements require a modicum of power, and all such circuitry requires excellent shielding.

The point of this exercise is to move us into contemporary means available for monitoring coupler output without disturbing the balance of the tuner at any power level. By tracking relative power output from the tuner, we can assure that we not only have achieve a correct input-side match, but as well have achieved maximum power output. It is the output that does the work of communicating, not the match.

This small tutorial has drawn on a large number of sources in an effort to bring together the principles and practices of inductively coupled tuners. If such tuners are better understood, they may once more take their proper place beside network tuners, each doing the job for which it is best fitted and not being forcefully adapted to a task which it is not well-fit to perform. Despite nearly a half century's effort to make coaxial cable the only ham cable, amateurs are discovering that parallel transmission line has an important place in the array of antenna- transmission line combinations. And wherever parallel transmission line is used, the inductively coupled tuner has a natural home.



Updated 11-28-97. © 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.



Return to Index