Feeding Balanced UNTUNED Open Wire Feeders - A New Experience

[k4kyv]

Ever since I was first licensed, I have always fed my wire antennas with open wire line.  Up to now, the OWL has always been operated as tuned feeders.  I have built many balanced link-coupled tuners using the classic circuit with a split stator resonating capacitor, using parallel or series feed as appropriate.  Designing and constructing an efficient tuner to feed a flat (untuned) balanced OWL is a completely new experience for me, after nearly 52 years on the air. I find this a  little different from the familiar tuned feedline matching network, with somewhat of a new learning curve.

I just completed two prototype matching networks, one to couple the balanced 4-turn link output coils in each of my homebrew transmitters to the untuned 438Ω open wire transmission running to the dog-house at the base of the tower. I experimentally tried several configurations including a link-coupled tuner with various L-C ratios, but what worked best, with the best efficiency and least critical tuning, turned out to be a simple balanced L-network. I could have easily calculated the balanced network as two separate L-networks working in tandem, one on each side of the balanced link-to-balanced feeders, with the coils positioned to minimise mutual coupling between them.  But instead, I used a single split coil, with the transmitter's link directly feeding the gap at the mid-point of the split coil, and a split stator capacitor in parallel with the entire coil, with the feed line in parallel with the split stator cap and coil.  I came up with the best results using trial and error by moving taps around on the coil; I did not attempt to calculate in advance because I was not sure if the normal L-network formulae would work when there is mutual coupling between the two coils.

With the best configuration, I could load the HF-300 rig up to full power, modulate up to 130% positive, and the dual 900/900 pf variable cap I used in the balanced L-network (0.07", 3 KV spacing) never arced over. At about 750 watts carrier output, the rf ammeter read 1.3 amps into the 450Ω dummy load with the transmitter link about 80% meshed with the main coil. The overall DC input to RF efficiency of the final calculated to 75% directly into the dummy load, and about 74% when I moved the dummy load out to the far end of the OWL. The mismatch between the 450Ω dummy load and the 438Ω OWL is so slight that it should cause negligible loss or tuning difficulties.

This appeared to be substantially more efficient on 160m than my present configuration, running the link, resonated with a network of parallel capacitors in a capacitive voltage divider configuration, to feed a 50Ω dummy load or the present unbalanced feed line that runs to the dog-house.  In the latter case, the transmitter loads up fully with the link only about 40% meshed on 160m and the link runs hot to the touch after a few minutes of steady carrier. Apparently, with so few turns, resonating the link with parallel capacitance to get the transmitter to take a full load at that low a frequency, causes the Q and circulating current to be so high as to produce substantial loss in the link coil.  Overall efficiency with the old configuration calculates to about 60% on 160, and slightly higher with a simple variable capacitor in series with the same 4-turn link on 80 and 40.

I won't attempt to use the balanced L-network with the Gates broadcast transmitter, since it has an unbalanced pi-L output, and the input to the tuner would necessarily be grounded at one side without the use of a balun, which I prefer to avoid.  Instead, I threw together another prototype tuner, using off-the-shelf junkbox parts on hand. This one is a link-coupled balanced tuner with split stator capacitor, symmetrically tapping down on the coil to feed the 450Ω resistive load. I experimented with number of turns on the link, number of turns across which the 450Ω line is tapped, and total number of turns on the coil.  I didn't bother to calculate the inductance of the coil, which is about 4" in diameter with around 80 turns, an edgewound coil taken from the variable inductor in a parted-out RCA BTA1-R broadcast transmitter. The coil originally had a massive moveable ferrous (or was it copper?) slug in the interior for varying the inductance, which I long ago removed.  To get a perfect 1:1 match out of the Gates without resorting to a variable capacitor in series with the link, I ended up with 5 turns on the link, and about 13 turns (IIRC) between taps on the main coil to the OWL, using the ENTIRE COIL in the main parallel tuned circuit. It resonates at 1900 kc/s with only about 35 pf (70 pf per section) across the coil, and gives excellent efficiency for a class-C tube final (76% total DC input to rf at the dummy load). But the thing is practically a Tesla coil! A 1/8" spacing(4500v per section) capacitor arced over immediately.  I substituted a 3/8" spacing (11kv per section) 100/100 pf split stator capacitor, and it still occasionally arced over on voice peaks at 130% positive modulation with the carrier running 750 watts. I have a 1/2" spacing cap (13kv/section) on hand that would probably work, but I may eventually try doing the maths to see if I can come up with a better L-C ratio with a more reasonable coil while still avoiding series capacitor at the link, although any change from the above settings of the tap and turns in the link in the trial-and-error version either produced lower efficiency or would not tune to a 1:1 match to the Gates.  However, since I can't complain about the efficiency of the beast, and the tuning is not overly sharp or critical, and it maintains a good match over the entire 160m band,  I might just use the thing as-is in all its glory with the big coil and widely-spaced tuning capacitor. I'd bet Derb would have really been impressed if he had seen that.

[W3GMS]

Don,
Very interesting.  It sounds like the one that worked the best was very similar to the Rick Measures tuner.  The difference being, you did not need a 1 : 1 current balun on the input since your link provided a balanced driving source.  Simply stated, a coil in each leg off the transmitter link with a capacitor bridging end to end of each coil or said another way, the load side of the tuning network.  Electrically you have 2 coils even though yours is wound on one form but broken in the middle to serve as the feed point input for the matching network.  With optimized Q, the losses as you saw are extremely low. 

Joe, W3GMS       

[k4kyv]

The Rick Measures tuner looks very similar to what I came up with.  Another very similar example is found in the mid-1950's era ARRL Handbooks.  See Transmission Lines > Coupling the Transmitter to the Line > Tuned Coupling > Alternative for use with small links, p.333 in the 1957 edition.

In both cases, two separate resonating inductors are used, isolated from each other.  In mine, all the inductance is contained in one coil, split in the middle, but with each half mutually coupled to the other half.  As Dave suggested, the calculations should come out the same if the total inductance is the same.  Mine should have less loss, since it takes fewer turns to achieve the same total inductance with one coil, than it would with two isolated coils in series, because the total inductance increases approximately as the square of the number of turns. The total split coil can built with the optimum diameter vs length of a single coil, so this should result in higher unloaded Q and less loss than what you would get with two coils.

Ideally, a Rick Measures tuner could be built with a special rotary inductor divided into two closely adjacent sections, with a mechanical coupling arrangement that would rotate the shafts of the sections  in opposite directions, allowing the two coils to be mutually coupled while the inductance of each increases symmetrically about the mid-point as the control shaft is rotated. Something that shouldn't be too  difficult to do if one is mechanically inclined and equipped with some basic machine tools and access to a variety of gears and bearings. Reversing the pitch of one of the windings, using a common rotating shaft wouldn't work, since the inductances would then buck each other rather than be additive.

[K5UJ]

I printed this thread out to study later.

The TMC TAC balanced tuner somehow uses a link coupled inductor that has pinch wheel taps and rotates.  As far as I know, it is the only manufactured balanced feedline link coupled tuner to employ this instead of fixed taps and a rotary switch like the KW Matchbox.  I have never seen one up close in person so I don't know exactly how the mechanical arrangement works.

[WA1GFZ]

Don, The so called Measures tuner, that I built in 1983 that I called the balanced L network uses two matched variable inductors ganged to a common control shaft. I've driven it with coax baluns and BB transformers depending on the antenna I connected it to.
This tuner configuration produces exactly the same feeder current as a KW Match Box but my "fugly" will handle a lot more power.
The TMC tuner has an interesting inductor. I would love to own one because it is such a cool design.

[k4kyv]

Don, The so called Measures tuner, that I built in 1983 that I called the balanced L network uses two matched variable inductors ganged to a common control shaft.

Are they inductively coupled to each other, or far enough apart/oriented so that there is minimal magnetic coupling between the two? Most of the examples I have seen in the articles show the two rotary inductors laid out  side-by-side, parallel to each other, and coupled with pulleys and something that looks like the timing belt on a car.

You could get a little less loss if they are closely coupled so that the inductances are additive. Two isolated 30 μHy  inductors (for example) wired in series will give 60 μHys total inductance.  Placing the same two in close proximity so that they act like two sections of the same coil would theoretically produce 120 μHys total inductance. If you are looking for only 60 μHys, you would theoretically need only 40% more turns in a single coil than what you have in one of the original coils, thus reducing resistive losses and any additional losses due to skin and proximity effects.

For maximum mutual coupling, the rotary inductors would need to be lined up end-to-end using a special shaft coupling, so that the ends of the coils are very close to each other, and so that at minimum inductance the rollers are at the ends of the coils closest to each other and move apart as the shaft is rotated to increase the inductance.  Ideally, the inductor in each would be somewhat "fat", so that the two coils together would form the optimum L/W ratio of a single coil. The turns pitch in each would have to spiral in the same direction in order for the inductances of the turns to all be additive (otherwise the inductances would cancel each other out and you would have a large fractional-ohm non-inductive resistor). IOW, except for the direction of the spiral, the two adjacent inductors would be mirror images of each other and the  rollers would move in opposite directions as the shaft rotates.  For this to work, the shafts of the two inductors would have to rotate in opposite directions, using a system of gears and/or pulleys, preferably somewhat removed from the near fields of the coils if metallic parts are used.  That is probably something like what the TMC tuner uses. Question is, for the average builder, would the added efficiency be enough to make worthwhile the additional complications in a homebrew tuner, unless one already has machine tools and an accumulation of mechanical widgets on hand.

In mine, since each tuner couples a specific transmitter to a specific load, a universal tuner is not needed, so fixed taps selectable with a switch will do, using a single fixed coil with optimum L X W, split at the mid point so that the gap serves as the input port to the tuner from the balanced output of the transmitter (or from the current balun if one were used).

[WA1GFZ]

I mounted my inductors end to end in the fugly and fed the center
Each inductor is a 22uh 5kw broadcast part. I couple them with a HD 3/8 inch insulated shaft coupler. The coils are about 6 inches apart so coupling is light

[W2XR]

Here's how I built my balanced-L tuner. I completed it a number of years ago.

Note that the perspective of scale in this photo is way off for some reason, and the tuner appears much smaller in the photo than it really is. In reality, the baseplate measures about 30" deep, with a width of 23".

I used a pair of E.F. Johnson edge-wound roller inductors that came from the old Mackay Radio/WSL site here on Long Island (I forget the model no., but they are easily rated at 5 KW CW), and a 10KV vacuum variable for the output capacitor. The roller inductors are mechanically coupled and synchronized together with a 1/2-inch wide cogged timing belt. My tuner is configured for impedance step-up only. The turns counters were pulled from a decrepit Gates BC-1J 1 KW BC rig I had parted out.

I deliberately did not use a metallic baseplate or enclose the unit within a metallic housing, so as to maintain the Q of the inductors.

The coaxial balun shown is long gone; I replaced it with a commercially-made 1:1 balun rated for 5KW CW, and I found that this resulted in less RF in the shack, although this was never really a problem to begin with. The balun is built by Array Solutions and is of very high quality, and runs at room temperature when I run the rig at full-throttle plate-modulated AM (a pair of 4-400As in the final).

Using a 126-foot dipole fed with the W7FG 600 ohm open wire transmission line, I can maintain a 1:1 VSWR anywhere from 3.5 to 30 Mhz.

I would recommend this type of tuner to anyone.

73,

Bruce

[K5UJ]

That's a nice building job.  my stuff never looks good enough to show (I show it anyway) so I'm glad I'm in radio and not TV hi hi.

The belt driven inductors work if you don't wear out the belt.  Palstar's first try at such a tuner resulted in them going to coils end to end on a common threaded shaft so I gather customers broke or wore out the belts.  Bliss did the same thing with his Matchmaster line--his twin coils were on a common shaft and tapped with finger stock  rings one each surrounding the coil.   the twin coils moved together in the same direction passing through the fixed fingerstock rings. 

I had problems at my location with that design balanced tuner--out of band strong RF would pass through to flumox my swr analyzer.  a link coupled tuner stopped that.   

I made at one time a coax coil common mode choke for 160 m. using 100ft of 213 on a plastic garbage can 18 inches diameter.   The coil had plenty of XL for 160 but at the expense of lengthening the feedline 100 feet.

Don, you must be correct  about the TMC inductor design.  I never gave it much thought.   There must be some sort of opposing mechanism to cause the pinch wheels to move apart for the inductor must be one spiral. 

I am reading p. 333+ in my '57 handbook but it is slow going at night when I am tired.  Vy interesting about flat lines.

[W2XR]

That's a nice building job.  my stuff never looks good enough to show (I show it anyway) so I'm glad I'm in radio and not TV hi hi.

The belt driven inductors work if you don't wear out the belt. 

Hi Rob,

Thanks for the nice compliment.

Gee, I don't see how that timing belt could ever wear out. It would probably last a gazillion years considering the application it is used in, along with the essentially nonexistent stress imposed on it. My guess is will rot away from age before wearing out.

And the only tuner I would replace my homebrew one with would be the TMC TAC-1.

73,

Bruce

[k4kyv]

And the only tuner I would replace my homebrew one with would be the TMC TAC-1.

Here is the manual

http://bama.edebris.com/manuals/tmc/tac1

Free software to read djvu files:
http://sourceforge.net/projects/windjview/files/latest/download?source=files

Looks like a standard link-coupled balanced tuner to me.