Balanced L-Network for 160m Vertical

[k4kyv]

I am contemplating changing from 50Ω coax to a balanced open wire feeder operating as a flat 500Ω line, for the link between my transmitters and the antenna tuners at the base of the tower that feed the series base-fed vertical and the open wire tuned feeders to the dipole.  This shouldn't be a problem with the dipole, which is inherently balanced, since all I would have to do would be to replace the feed line and adjust the coupling on each tuner to match, probably by symmetrically tapping the open wire line to the coil. An alternative would be to use a separate coupling coil to replace the present link. Each antenna uses a separate tuner switched in for each band.

The 160m vertical is a different matter, since I  would be feeding an unbalanced antenna with a balanced flat open wire feeder. I have come up with an idea of changing the L-network from an unbalanced configuration to balanced, and let one side of the balanced tuned circuit feed the antenna.  I would be interested in any comments.

The present set-up uses a simple L-network to match the 50Ω coax to the series fed vertical antenna, which because of the top loading effect of the dipole attached near the top of the tower and the feed line running up through the interior of the tower, measures as follows across the base insulator.  Without the top loading, the base impedance would be expected to measure about 38Ω +j0. All measurements were taken with a General Radio 915AL antenna impedance bridge.

1800 kHz: 117Ω + j233

1900 kHz: 185Ω + j316

2000 kHz: 320Ω + j425

Click on the attachment to see the circuit diagrams of the present L-network and the evolution to the balanced version I am contemplating.  I haven't attempted any actual calculations.  The original L-network was designed by trial-and-error, and the evolution to the balanced version is only intuitive at this point.

The single section variable capacitor is 1000 pf @ 3500 volts.  The split stator capacitor in the proposed balanced version is 500 pf per section @ 3500 volts.

Fig. 1 is the present L-network. The capacitor is set to approximately mid-scale at 1900 kHz, and for a near-perfect match, varies from about 60% meshed to 40% meshed from one end of the band to the other.

Fig. 2 is the  same network, redrawn to resemble a regular tuned circuit.

Fig. 3 is the approximate equivalent to the  L-network, with the 50Ω coax matched to the coil via a tap instead of series feed.

Fig. 4 is the equivalent balanced L-network, using the split stator variable capacitor with both sections in series, and the coil adjusted to resonate at the desired settings of the capacitor.  Each capacitor section is half the capacitance of the single-ended tuning capacitor. The tuned circuit operates at a higher L to C ratio, with 1/4 the total capacitance of the unbalanced circuit, but since the load is across only half the coil, the loaded Q would be the same as in the balanced case. The rotor and frame of the split stator capacitor is intentionally left floating, with the ground point established by the mid-tap on the coil.

Fig. 5 is the balanced L-network fed with a balanced 500Ω flat line, rather than the unbalanced 50Ω coax.

Fig. 6, the balanced 500Ω flat line is series fed via a split in the coil, instead of by tapping the coil. In this case, the ground point is established  by bonding the rotor of the capacitor to ground.

I would try both the Fig. 5 and Fig. 6 configurations and use the one that performs better if there is any difference.  In any case, an rf ammeter would be inserted in series with the antenna feed point at the base of the tower and the readings noted with the original configuration, at specific frequencies and power levels, to compare to any modifications to the feed system.  A decrease in rf current reading at the antenna feed point would be unacceptable.

[WQ9E]

Don,

Might a link coupled tuner to both match and transform balanced to unbalanced be a better choice for the base of the vertical?  The L network is easier to set up for tuning multiple frequency ranges but that doesn't apply to your 160 meter only application.

[K1ZJH]

Fig. 5 provides a static discharge path.

[k4kyv]

Don,

Might a link coupled tuner to both match and transform balanced to unbalanced be a better choice for the base of the vertical?  The L network is easier to set up for tuning multiple frequency ranges but that doesn't apply to your 160 meter only application.

Link coupling implies a relatively much lower impedance at either the input or output of the tuner, compared to the load impedance on the main tuned circuit.  That would work with the present set-up with the 50Ω feed line, but when I built it 29 years ago, the L-network turned out to be easier to tune and less complicated.  Figs 4, 5 and 6  basically are link-coupled balanced tuners, except that the line is coupled via taps on the main coil or by series feed, rather than by a physically separate coupling coil.  Both alternative methods of coupling are suggested in the original 1930s magazine articles on link coupling.  "Link" coupling doesn't necessarily mean a physical 2-4 turn coil wrapped round the main coil.

But with the future 500Ω balanced line, the input impedance will be so close to the output impedance, that a small coupling coil would not work.  Rather than a large coupling coil wound over the main coil, the taps or series feed should be easier to construct, but I would anticipate that the series feed in Fig 6 would require completely different parameters for the tuned circuit, since the impedance in series with the two halves of the coil would be so high relative to the output load impedance.  Right now I am partial to Fig. 5, since I am not sure how well the split stator capacitor would establish ground and maintain balance in the tuned circuit  with the load on only one side.  Fig 6 would require a static drain choke from the antenna terminal side to ground, which I presently have in the original set-up.  

One thing I like about my current set-up is that it is broad banded.  I have the L-network set to resonance at 1900 kHz with a 1:1 SWR on the coax, and it reaches a little over 2:1 at 1800 and 2000.  I have the capacitor on the L-network locked down and haven't touched it for years.  When QSYing I can compensate for the mismatch in the line by adjusting the coupling network between transmitter and feed line, and I think the SWR loss in the feed line is negligible at less than 3:1.  With the new set-up, I plan to physically gang all the tuners together and drive the whole thing with the reversible motor, so each tuner can be remotely set for a perfect match at any frequency on any band (within the range of the tuners) without the trek to the tower, so if the final result is a little less broad banded that will not be a serious problem.

[K5UJ]

I don't like any of it.  Sorry Don, but I know you want replies with the bark off.  I'm no antenna genius.  for that, maybe Walt will comment.  I looked at the tuner schematics.  Intuitively, it looks to me that you are trying to force feed an unbalanced load with a balanced source.   My hunch with those tuners, 5 and 6 that is, is that you'll have unequal currents on the line and it's gg to radiate.  You might be able to get it to work but I would not be surprised if it is fussy and narrow in frequency or after a geezer transmission something smokes.  Balanced line is relatively cheap and simple.  One of my  MOs with medium wave base fed verticals is that if an idea is cheap, simple and effective, people a lot smarter than me would have been doing it in broadcasting decades ago.   There must be a reason why balanced line was not used at broadcast stations but unbalanced OWL was.  My hunch is that balanced line ----> unbalanced vertical = problems.  But, this is opinion; not fact.  I would be delighted if you proved me wrong and I'd be very interested in hearing the results.

[Steve - WB3HUZ]

Snag some free hardline and be done with it. No tuner required, no loss, and far more environmentally stable.

[WA1GFZ]

I agree with HUZ and stick with an L network KISS

[W2DU]

Hi Don,

I agree with Rod, Steve and Frank about using the simple L network and 50-, or even 75-ohm hardline. Perhaps you could find a cable tech who could steal you some 75-ohm stuff.

I agree with Rod on the unbalanced condition resulting from using the circuit shown in Fig 5. In addition, I don't believe the split-stator cap would have been helpful as long as the center of the inductor is grounded. A single-stator cap would work as well. Fig 6 would have the same unbalanced condition as in Fig 5.

I've been trying to come up with a circuit to add to Fig 5 that would allow the LC tank circuit to remain balanced, but still feed the base unbalanced. So far nil, but I'll keep workin on it.

I like the circuits in either Fig 1 or Fig 2.

Walt, W2DU
 

[k4kyv]

A balanced tuner can feed an unbalanced load, as long as the internal resistance and reactance of the tuner are negligible or successfully compensated for.  It's basically feeding the load off a tap on a coil, instead off the entire tuned circuit.  Similar to the concept of a balanced tank circuit on a single-ended final, as in the case of the BC-610. Or the case of an end-fed zepp, where only one side of the balanced feeder is attached to a load. Of course, it remains to be seen what the unbalanced reactance from the antenna does, but I think it is a better idea than the widespread practice to-day of force-feeding a highly reactive balanced load from an unbalanced tuner, such as a T-network, through a balun.

I don't intend to  rip out the present system until I get the new one working to my satisfaction, and all proposed plans are subject to revision.  I would like to eventually get it down to one feedline with a remote switching system to select the tuner and the antenna it feeds, either with a series of relays or a motor-driven switch.  I am already working on the ganged antenna tuners.  That is turning into a major construction project because of all the precision drilling, cutting and other metal work involved: mounting  components exactly in position, getting shafts to line up, set-screws to hold tight, bearings to not bind up, etc. Probably on par in difficulty with building a KW AM transmitter ground-up from scratch.

I suppose one reason I am contemplating going this route is I have all this stuff on hand, plus a big roll of #8 copperweld, and I'm kind of itching to use it.

[k4kyv]

I agree with Rod, Steve and Frank about using the simple L network and 50-, or even 75-ohm hardline. Perhaps you could find a cable tech who could steal you some 75-ohm stuff.

I agree with Rod on the unbalanced condition resulting from using the circuit shown in Fig 5. In addition, I don't believe the split-stator cap would have been helpful as long as the center of the inductor is grounded. A single-stator cap would work as well. Fig 6 would have the same unbalanced condition as in Fig 5.

I've been trying to come up with a circuit to add to Fig 5 that would allow the LC tank circuit to remain balanced, but still feed the base unbalanced. So far nil, but I'll keep workin on it.

Thanks for the input, Walt.  The main reason for the split stator in Fig. 5 is the fact that I have one, an exact split-stator version of the single-section one in Figs 1 and 2, which I am presently using.  I  could just as easily use a single-section cap @ 250 pf/7000 volts. The closest thing I have to that is physically different in size, shaft diameter, etc, and would be more difficult to  gang together with the other caps.  All the variable caps I have set aside for this project are EF Johnson bread-slicers, compatible in shaft and frame size with the BC-610 plate tank cap (which is what I  now use in the 40m tuner). I have tested the worm drive and the 12 vdc motor, and the motor simultaneously drives all of the 5 capacitors I plan to use, without grunting.

Let me know if you come up with any more ideas regarding Fig 5.

Don

[W2DU]

Don, seein you're itchin to use the #8 copperweld, here's another idea. If there's enough length of the #8 make a four-wire unbalanced line, with three wires at ground potential in a triangle formation looking end-on, with the fourth wire hot and located in the center of the triangle.

The CD of the Laport book is in the mail, which contains info concerning many versions of wire feedlines, data on spacing vs Zo, etc.

I see you're using a GR-415AL impedance bridge, the low freq model. But are you sure it's 415? The model that preceded my 1606A was the 416AL. I'm curious concerning where you measured the 1.8, 1.9 and 2.0 MHz impedances. Right at the base of the tower? If so, then at what freq did you expect to get 38 + j0?

However, I goofed on my comment concerning 75-ohm cable--totally forgot to consider the lack of power handling capacity. Me bad!

Walt

[W2DU]

Arrgh!! Me bad again. In the post above I said the GR-416AL was the predecessor of the 1606A--wrong--I mean't to say the GR-916AL, not 416.

So you're probably correct Don, when you said 415, which is a GR model I'm not familiar with.

Walt

[k4kyv]

38Ω+j0 is what I figure I should have at the resonant frequency of the tower, whose total height from the common point of the radials is 127'.  36Ω for the quarter-wave resonant vertical, plus a couple of ohms for ground loss resistance.  I have 120 1/4λ radials, and most of the books say that should leave about 2 ohms equivalent series ground lead resistance.  But what I actually measured as given in  the table is radically different as you can see.  I attribute the difference to the 80m dipole that hangs from the tower at the 119' level.  Even though the flat-top is not electrically connected to the tower at any point, the open wire feed line goes up through the interior of the tower, centrally spaced geometrically inside the triangle.  The close proximity of the feedline to the tower along its full length appears to provide substantial coupling, so that the antenna on 160 is more akin to a vertical tee than to a true 1/4λ vertical. When I first built the system, I had not anticipated the dipole having that much effect, since the total  length of the feedline plus one  leg of the dipole is about 192', approximately 1/8λ away from a self-resonant length .

I think my meter is a 916AL.  It has been several years  since I have used it, and I'd have to dig out my book or else pull the meter off the shelf, so maybe I didn't recall the model # correctly. Age affects our memory, doesn't it?   It's a big black cube, a little more than a cubic foot, and requires an outboard signal source and receiver, and is good from below the broadcast band up to about 3 mHz, IIRC.  I took the measurements right at the base of the tower.  I use a single wire copper lead from the base of the tower to the tuner box, located about 6 ft. away.  I took a set of measurements at the feed point right at the tuner box, and that extra 6' of wire noticeably affected the readings, although not radically.  All the readings were with the open wire feed line to the dipole disconnected from the balanced tuner and floating free from the tower. I verified the base impedance readings by running a known amount of power from the transmitter and measuring the RF current at  the tower base.  The readings were consistent with the measured tower impedance at each frequency.

I took a set of measurements with both sides of the feed line grounded to the tower near the base, and that brought the impedance much lower.  In fact, at 1810 kHz I recall the SWR was exactly 1:1 with no tuning elements between the coax and the tower whatever, and the j factor went from negative to positive somewhere in the band.  But I had already built the L-network to match the tower with the feeders floating, and it worked very smoothly, so I never pursued the grounded feedline configuration any further, although I have a knife switch at the base that shorts the feed line directly to the tower to help with lightning protection.

I have the figures for those other measurements in my note book.  I'll have to dig them out.

[W2DU]

Sounds like yer doing everything right, Don, except for the 6' wire from the base to the tuna. My suggestion would be to get the 916AL as close to the base as possible so the lead from the base to the input of the bridge is as short as possible. You may want to dig out the manual for the bridge to refamiliarize with the effect of the lead connecting the bridge to the base. If you look at the curves for compensating for the lead you'll see that the length is more critical than you might think with respect to the actual measured impedance of the tower. I can't help you there, because my 1606 manual is in FL.

You didn't comment on the idea of the four-wire unbalanced line. Not enough length of #8 wire? With #8 copperweld the loss would be almost zilch (might even get some gain, juss kiddin).

Walt

[K5UJ]

Hi Don,

If you can't wait for Walt's CD (which is a nice thing to have) and you have a high speed connection, you can get the Laport book online:

http://snulbug.mtview.ca.us/books/RadioAntennaEngineering/

I was hoping it was divided up into chapters because the whole thing is 25 Mb but it doesn't look like it is.    
If your connection is as slow as mine you might want to wait for the CD.

There are a lot of great photos of big broadcast towers in various stages of construction.  I think they have one or two of the WJZ (later WABC) radials getting plowed in and there's one of the old top loaded WMAQ 670 tower.  It's fun to just flip through it and look at the pictures.  Each chapter is dedicated to one sort of service or frequency range.

Rob

[k4kyv]

Walt,

I connected the G-R impedance bridge directly to the base of the tower and ground for the initial readings, with maybe 18" leads.  The  readings with the 6' piece of #4 solid wire were taken right at the output of the L-network.  I use that wire to run from the doghouse to the base of the tower.  There is also a piece of 3" ground strap running parallel to it on the ground from the doghouse to the radial common point for ground return.  So the figures with the 6' wire are what the L-network actually sees, since I wanted some space between the ATU box and the tower.

I have enough of that copperweld to make a 4-wire balanced line, or even an open wire coax line if I wanted, so those are always possibilities.  In fact, somewhere round there (I haven't been able to find it for years) I have information on multi-wire balanced lines that use, IIRC, up to 16 wires, to attain super low surge impedances using plain wire instead of unduly large solid conductors or copper tubing. I am hoping the Laport book contains that info.  That may have even been the original source of what I had before it evaporated.

Anyway, I found my notes.  The readings were taken in August 1995.  I didn't think it was that long ago, but I haven't changed anything with the antenna since then, so they should still be the same to-day.

Dipole feed line floating:

@ tower base                                              @ doghouse entrance (6' wire lead)

1800 kc/s     117Ω + j233                                 140Ω + j269

1900            185Ω + j316                                 225Ω + j363

2000            320Ω + j425                                 420Ω + j475


Dipole feed line shorted to tower near the base:

1800             28.8Ω - j10.5                               36.5Ω - j16.1

1900             45.5Ω + j30.5                               52.5Ω + j62

2000             70Ω + j99                                    76.5Ω + j130.5


With the feed line shorted, no matching network, SWR was 1:1 at 1812 kc/s.

With normal operation, i.e. dipole feed line floating and L-network (Fig 1, 2) in place, at 300 watts carrier output from the transmitter (not counting feed line losses between shack and tower), rf current as measured right at the output terminal of the L-network:

1800   1.225 amps

1900   1.05 amps

2000   0.75 amps

[K5UJ]

What is the gauge of the six foot wire?

I don't have that wonderful antenna to experiment with but if it were me, I'd be interested in three measurements:

1.  a similar current measurement as was already done, but with the dipole feed shorted to the base of the tower

2.  A field strength measurement a bit above 1800 with a 50 watt carrier, the measurement taken a mile or two from the tower and the dipole floating.

3.  As above in two, but with the dipole shorted, same power, reading location and frequency and measurement setup of course.

[W2PFY]

I need a little help here. I see that figures one & two have a ground symbol at the bottom of the variable cap. Is this an earth ground or a ground that is common with the shield of the coax? Figure three shows the cap tied to the cold end of the inductor. I want to put up a simple tunable vertical that will work on 160 80 40 etc.

I am not trying to high jack Don's post, it's just that I'm not a brainchild when it comes to some of these matters.

[k4kyv]

Rob, the 6' wire is #4 solid copper.

I have never tried field strength measurements with the feeders floating vs grounded to the tower, but that might be an interesting experiment.  It wouldn't be totally conclusive, however, since one configuration might generate more ground wave, which would indicate higher on the FS meter, but actually generate less sky wave signal.  What I need is a relay switching set-up where I can instantly change from one to the other.  It would require two separate tuners, since the impedances of the two are so different. That way I could gather signal reports from every direction and every distance, kind of like Timtron's shaded dipole vs 160m inverted vee tests.

My 160m vertical is not like a broadcast antenna, in that the goal is NOT to produce the strongest ground wave possible.  I rarely work anyone within near-by ground wave range on 160, so I am looking more for sky wave but with low enough angle of radiation to cover most of N America.  For semi-local work, the 80m dipole tuned to 160 as a 1/4λ dipole seems to work best.  In Nashville, for example, the dipole gives me about 30 dB more signal than the vertical, and makes the difference between the signal being buried under urban radio smog and having a full quieting carrier.  Out beyond 100-200 miles, the vertical begins to perform noticeably better than the dipole, and at far distances like New England and the west coast, the vertical is usually far superior.

Terry, in fig 1 & 2, the coax braid, variable capacitor and ground lead to the radial system are all tied together.  It's just the way I drew the schematic. 

Sometimes when a tuner is link coupled using a separate link coil, you get less noise pick-up and rf-in-the-shack when the link is allowed to float ungrounded at the tuner, since grounding the far end of the coax allows it to act like a Marconi antenna, and this can generate common-mode currents at the outer surface of the coax braid and bring garbage into the shack. I would try it both ways, and see if one works any better than the other.  This is something hard to predict, since unknown factors may creep into antenna installations. Mine is grounded, since I use the same feed line for the unbalanced vertical and the balanced dipole, and I didn't want to bother with a separate switch or ganged wafer for the ground side of the circuit.

With a balanced link-coupled tuner feeding open wire line to a dipole or zepp, I never ground the midtap of the coil or the rotor plates of the capacitor, since this can cause the dipole feed line to have common mode antenna currents in reference to ground with the whole thing trying to act as a vertical tee.  The result would show up as unbalanced current in the feeders, and may or may not affect the actual transmitted signal.

BTW, to answer a couple of others' question regarding why I don't just acquire some heliax, bury it, run coax feed and be done with it, I want to try out an open wire feed system, since I already have plenty of wire, insulators, etc.  This is for pretty much for the same reason one would go through all the work and trouble to build from scratch a homebrew plate modulated transmitter or a class-E rig, instead of simply buying a Flex Radio or other SDR, or one of the riceboxes known to perform well on AM, and use it with a leen-yar.

[K5UJ]

Don, okay on the wire.  No. 4 has pretty low inductance.  I was wondering given your 6 foot run, if something with a lot of surface area would bring down the Z at 2 MHz a bit and allow for more current to transfer.    I noted that you get more current lower in the band with a lower Z.   But with no. 4 I am not sure something like 1 inch o.d. copper tubing would get you much.  I know broadcasters go with that from the doghouse to the tower but it probably has more to do with power handling than anything else.  Some put a 1 turn loop in the pipe which I've heard is supposed to attenuate lightning strike RF but I have doubts about the efficacy of that if true.

Understand your interest in skywave but a field strength measurement isn't intended to find which has the stronger groundwave (although that is unavoidably what happens) but to find which setting causes the antenna to transfer the most energy to the atmosphere by removing the ionosphere variable.    Let's say setting A gives the strongest field strength (it is always possible there will be no difference between A and B).  You could interpret that to mean strong groudwave lousy skywave, but it could also mean strong groundwave and even stronger skywave.  Of course a pessimist/optimist view could be applied to setting B as well, but if A has the stronger ground based field we know that in that way, A is getting out better.  Short of renting a helicopter, you'd have to take that information and consider it along with skywave testing data, but the latter is tough due to QSB.  Maybe those internet linked receivers can be employed for that testing.

[W2PFY]

Some put a 1 turn loop in the pipe which I've heard is supposed to attenuate lightning strike RF but I have doubts about the efficacy of that if true.

I think that one turn loop was just for rain and moisture to drip off the bottom of the loop rather than run down the pipe right into the network or accumulate on the  hardware.

[k4kyv]

The loop is for lightning protection.  The lightning pulse is extremely steep, so it behaves like rf.  The one turn loop acts like an rf choke.  The hope is that the lightning pulse will find the spark gap across the base insulator an easier path to ground than following the lead-in into the ATU.

Positioning the arc gap is critical, regarding rain and ice.  The gap should be horizontal, not vertical so that the drip doesn't dwell across the gap and short it out.

[K5UJ]

Ah, okay got it thanks.  I did not consider the idea of obstructing one path just enough to make the other one (the arc gap to ground) look better.   Yeah Terry, may as well make it with the loop hanging down so water drips off it.