The AM Forum

Hello to all: (Updated)

Where the ladder line connects to the link antenna coupler I am connecting a variable shunt inductor between the line.  One end of this inductor connects to the left side of the line and the other end connects to the right side of the line.  This inductor will be used to cancel capacitive reactance on the line.  

Please see below two attachments of possible inductor setup.  

Questions:  

(I) Which coil configuration is preferred, figure A or figure B?  

(II) If figure "B" is used do you think there is going to be a coupling/interaction problem between the two coils?

(III)  By using the coil setup in figure "B" do you think the current/voltage balance of the total antenna system (coupler, line and balanced antenna) could be upset?

Chuck

Figure A ought to be better because it may have less magnetic coupling between the coils and the line.

====

If the coils are would opposite direction and are near each other, the inductances may partly cancel. (push-push)
If the coils are would same direction and are near each other, the inductances may partly add (push-pull -might be advantageous).

====

In the Northern hemisphere, electricity flows more easily in a counter clockwise direction due to the Coriolis effect, so the coils ought to be wound counter clockwise.
In the Southern hemisphere, electricity flows more easily in a clockwise direction due to the Coriolis effect, so the coils ought to be wound clockwise.
This assumes the coils are mounted vertically, with turns 'stacked'.

If you mount the coils horizontally, this can be ignored.

Opcom:
Thanks.
Chuck

Eh?

It's not April 1st for a while.....

Please enlighten me!

Any of those will work but floating ends invite unnecessary voltages.

In practice the necessary inductance may seem quite small depending on your situation.

When you tune out the above, how do you know it's tuned out? Is another meter used and would your recommend this for any OWL?   I've heard the term and I guess I should know this being a ham but a person can't know everything.

It isn't necessary to tune all reactance out. Depends on the situation. When the load is outside the capability of the coupler a parallel inductor can be one way to facilitate power transfer.

If one wanted to verify they had tuned all reactance out at H.F. they could substitute an adjustable resistive load at the "coupler" output and check SWR'sss'sssssses. 

Or dip the final with the load and parallel inductor disconnected.  When the load and parallel inductor are re-connected, the PA tuning resonant dip will not appreciably shift if no reactance is introduced. If the dip does shift, try adjusting the inductor so that resonance (plate current dip) returns to the original setting.

I apologize, that part was a joke based on the fallacy of the direction water swirls in the bowl or drain, said joke  may not have been as obvious as I thought. I should not have made it. The rest I meant seriously.

No Worries - Hence April 1st reference - April Fools Day (not sure if the US shares this tradition though)

Makes about equal sense (or should I say nonsense) to speaker cable break-in, £27,000 power cords, $250 ea. wooden knobs, etc.  Hard to tell these days what is intended to be taken seriously; one may be surprised.



I doubt it would make a lot of difference in the case of relatively low-Z feedline impedances. In the prototype tuners I have been working with lately, I could see zero difference in zorch tendency or rf efficiency whether the unused turns were shorted or left floating. With something like a high-Z tank coil, the Tesla coil effect of floating turns might contribute to losses or arc-over.

I am back from visiting my sister for Thanks Giving and got a chuckle from the North and South Hemisphere tip. 
Can anyone get a little specific regarding questions #1, #2 and #3?

Chuck

If everything is laid out reasonably symmetrically, I can't see how one layout would have a greater tendency than would the other to unbalance the line. If you are using two identical inductors, they should be wound in the same direction, so placing them close together end-to-end would increase the amount of inductance for the total number of turns, due to mutual coupling.  Laying them out parallel to each other might be easier, space-wise, and unless they are spaced far apart there still would be some mutual coupling - nothing I would worry too much about. Unless the required inductance and the inductances of the coils have already been closely calculated, I would try one inductor first, and if it is sufficient, there would be no reason to use two in series. Since this is to be a mid-impedance line, stray capacitance would have much less effect on balance than would be the case with something high-Z, like a push-pull final tank circuit or a parallel-tuned antenna coupler feeding a voltage loop of a tuned feeder.

Don:

Very good.  I'll see what I can do about the shunt coil(s). 

Thank you,
Chuck

We may have already discussed this, but have you ever thought of going the balanced L-network route for the highly reactive load?  You might be able to tune out the reactance AND match impedances all  in one throw.  I tried both the link coupled tuner and balanced L-network to work my homebrew link coupled transmitters to the 450-ohm resistive dummy load, and found the L-network to be almost 25% more efficient and broader tuning.

Using trial-and-error with components on hand, I successfully built an efficient link-coupled tuner that works with the unbalanced output of the Gates (looking for an alternative to the input balun).  It turned out to work efficiently only with a ridiculously high L-C ratio (only about 35 pf total capacitance across an 80+ turn coil for 160m) and the feeders tapped down on the coil, but the capacitance needs to be rated for well over 20 kv. A 100/100 pf split stator capacitor with 0.375" plate spacing and the two sections in series across the coil, arcs over on modulation peaks at full power.

Don:

When trying the balanced-L approach I tested many baluns.  I duplicated the baluns manufactures were using and what authers recommended.  Some of these baluns became so hot that you could not touch them.  These test were conducted at the 100 watt level, 160 - 10 meters.  At times current balance was poor.  

I read an article regarding the balun approach and the author concluded a balun placed in a highly reactive environment was not the best approach for feeding balanced line.  

I believe it may have been QST, they ran test on "Balanced" antenna couplers and experienced over heating problems with baluns.

I have found if one leaves all ceramic out of the antenna coupler (door knob capacitors and baluns) you then have eliminated a source of potential heating or failure.  I think TenTec ran in to this problem with one of their couplers that used a door knob capacitor.

The same goes for roller inductors, antenna switches and antenna relays.  Having just a tangent point (one degree contact point) between roller and wire or RF contact to RF contact promotes the environment for arcing in some antenna systems.  

Solid brass clamping taps, on link antenna couplers, provide over 300 degrees of RF contact from tap to coil(s).  

I know what you mean when it comes to tapping down to much to obtain the low impedance point on the tank coil.  I know you know this, but this is when I switch to series tune for a better LC ratio for low impedance matching.

Chuck





  

I believe you misunderstood what I meant regarding a "balanced L-network". I am NOT talking about force-feeding a balun between an unbalanced L-network and the OWL, the bogus configuration used by the majority of the appliance manufacturers in what they advertise as "balanced" OWL tuners. As you said, a balun placed in a highly reactive environment is not the best approach for feeding balanced line.

What I am  referring to is an authentic balanced L-network directly feeding the OWL. It is formed by splitting the inductance into two sections.  It may consist of two coils, one in each side of the line, or a single coil split at the mid-point with the low-Z input fed to the gap where the coil is split.  For even better balance, the variable capacitance would be formed using a split stator capacitor with the two stator sections in series, each connected across the coil(s) and in parallel with the OWL, leaving the rotor/frame floating.

The balun, if used, would go between the transmitter and the input to the L-network, where it would see a purely resistive, low impedance load when the L-network is properly adjusted.  In my case, since the link-coupled output in both my homebrew transmitters is already balanced, I don't need the balun.  The balun would be necessary for a transmitter with unbalanced output, such as a conventional pi-network or pi-L network.  One type of balun for this purpose would be what is  described as an "ugly balun", formed by neatly coiling up a length of coax in a single layer with uniform close spacing between turns much as you would wind a coil out of copper tubing.

You may recall this post I recently submitted on the subject in another thread:
A genuine balanced tuner is either link coupled or has the balun between the transmitter and tuner, and uses a symmetrical tuned circuit, either with two separate identical tuning capacitors or a single split-stator one.

Many of the commercially manufactured "balanced" tuners on the market to-day are totally bogus.  They use an unbalanced L- or T- network to feed the open wire line through a balun. That might work OK when the OWL is working as a flat, untuned feed line, but is a bad idea for working into a tuned feed line, if the load has any substantial reactive component, or is a high impedance (something that would normally require parallel tuning with a balanced link-coupled tuner).  One of those tuners MIGHT work OK if the OWL resonant feeder is being fed right at a current loop where the impedance is low with little or no reactive component. But it probably wouldn't be satisfactory on any other band with that antenna.

Transformer-type baluns like the ones used in those tuners are  designed to handle a purely resistive load at a narrow impedance range.  Force-feeding a reactive load of random impedance through a balun may result in excessive losses and cause the ferrous core to heat up (sometimes to the point of self-destruction), and even worse, may drive the core to saturation, causing it to become non-linear, resulting in spurious radiation products, exactly like what happens when antenna-to-feedline connections become corroded or certain dissimilar metals come into contact.

An excellent article on the subject appeared in the February 1990 issue of QST.  Here is the pre copy-edited version, updated in May, 2003, in a web page put up by the author. Click on the embedded links for the referred circuit diagrams.

This might specifically apply to your case: (quoted from the article) The reactance formulas give exact values only for non-reactive, purely-resistive loads. If the reactances of the L-network are adjustable, a wide range of load reactances can be cancelled by adjusting the L-network to create an equal and opposite reactance. This is accomplished by tuning the L-network for zero reflected power while using a minimum power level.

http://www.somis.org/bbat.html

"  April Fools Day (not sure if the US shares this tradition though) "

Well, we have one over on the UK;  we've expanded the tradition, we're fools most of the time.


klc

Nothing bogus about it. It's balanced under most conditions. You are confusing symmetrical with balanced.

Maybe balanced, but highly inefficient, and sometimes source of spurs.  Otherwise, why would the balun heat up?  A transformer type balun is designed to work into non-reactive load within a certain impedance range.  The balun should be on the input side of the tuner, not on the output side.

Contemporary antenna tuner circuits claim to be able to operate into an unbalanced load or a balanced load such as ladderline. In actual use, most of the contemporary "matches everything, balanced or unbalanced" antenna tuner circuits produce a semi-balanced output when used with a balanced load. Although the antenna will radiate in this situation, a semi-balanced output is like having a semi-balanced checking account. It is less than wonderful.

A look at the diagram for the contemporary "matches everything" antenna tuner circuits reveals that they are usually unbalanced, high-pass filter characteristic, T-Network circuits with an add-on balancing device hooked to the output of the unbalanced tuner circuit. This is a compromise design which, not surprisingly, also has compromise performance when used with a balanced load.

The imbalance in these "balanced" tuners can be easily confirmed with a RF voltmeter or RF amperemeter(s). When the actual current or voltage is measured at each output terminal, the observed imbalance gets progressively worse above about 7MHz. At 28MHz, it is not uncommon to have 50 (percent) more current or voltage in one of the legs than in the other leg.

http://www.somis.org/bbat.html


bogus

adjective
1. not genuine; counterfeit; spurious; sham.

Have you ever used such a set up? How much temperature rise did you see?

Having used such a set up for about five years, I never saw any heating or had any spurs. That's not to say such could not occur, but I've seen heating in link tuners too. Any properly designed/rated system will be OK. Any one that is not, will be more likely not to be OK.

KB3AHE has much more experience with such a set up. Once he rated things properly, the tuner has worked flawlessly for years. His signal speaks for itself.

I never used one here, but my friend across town (Vic, K4SSD) had problems with his.  Don't recall what company made it, but it was a T-network and the balun would get hot to the touch if he tried to run AM. He showed me the burnt yellow tape used to insulate under the winding, that had turned brown.  Worked OK with slopbucket, but crapped out when he tried to run AM through his pair/3-500Z leen-yar.  He finally took the balun apart, acquired a second toroidal core, stacked it on top of the original one (or maybe he used two new ones), rewound the turns with new wire and tape, and after that it ran only warm, not hot. He still had TVI until everyone round him had finally changed over to satellite or cable.

He had rf in the shack until he followed my suggestion to disconnect the external ground from the tuner case and let it float.  That reduced the rf, but he still had the TVI and the balun still ran warm (but no longer burnt up) on AM. I suggested that he relocate the balun from the output side to the input side, and mount the whole case of the tuner on a well-insulated support well away from anything else in the shack.  Don't know if he ever tried that.

If you are working into OWL tuned feeders on the lower bands, and feeding it at or near a current loop, it probably will work flawlessly. It will probably work OK to a flat untuned balanced line, 300-600Ω.  But I dare say if you try running high power, using such a JS configuration to feed a voltage loop, or worse still, to what Chuck is trying to load, fed at an odd 1/8 wavelength (midway between a voltage and current  loop), it won't work quite so flawlessly.

I'm not saying that oversizing will fix anything technical (loss, efficiency, etc) but it may avoid a 'tuner meltdown' so you have time to get things right according to your needs.

Go bigger? - parts from a 5KW BC rig may make good tuner parts.

Most high pass T network hamateur tooners use a 4:1 balun. Not a good way to transfer QRO am to a high current load.

Don:
I did make the mistake in thinking you were trying the forced feed an unbalanced L-Network and expecting it to pass as a balanced L.  I should have known better. 

Chuck

Don:

If you have time please draw the schematics of post number 16 of this thread please.

Chuck

See Fig. C

http://www.somis.org/bbat.f1.jpg

L1A and L1B can be a single coil, split at the mid-point, instead of two separate coils.

C1 can be a split stator, with each section twice the capacitance of C1 as shown in the diagram.

The choke-balun is not needed if the transmitter uses a balanced link-coupled output.

Mine is without the balun, a single split coil, and split-stator capacitor.

Or you can use one coil and a single section cap and skip the mechanical complication.   ;)

When you say split you mean cut the coil in half electrically?
Chuck

Instead of two separate coils, use only one coil, with the same total inductance as the two separate coils in series.

Electrically and physically split it into two identical sections without widening the turns spacing between the sections at the split. Actually, slightly widening, or even doubling the spacing at the split point won't hurt if you have more than a just a few turns, but it is best to maintain uniform spacing between all the turns if possible.

Theoretically, you will need only 70.7% of the total number of turns (1.414 times the number on one of the separate coils), although practically, the number count will likely vary, due to such factors as turns spacing, coil diameter, length vs diameter, etc.  See the coil formulae in the handbook, or use this handy calculator at http://hamwaves.com/antennas/inductance.html . Over the years I have developed enough of a feel for the dimensions of a coil for a particular application that I can usually hit the number of turns by trial-and-error in one or two attempts, more quickly even than using the calculator, when building from existing coil stock on hand with some means of temporarily attaching taps to the turns to the coil and moving those taps around.

Once you have determined the number of turns and the dimensions, and constructed or tapped the coil, count turns from each end and determine the exact mid-point of the coil. Think of each turn in terms of 360°, and roughly count degrees if the start and finish points of the coil are not exactly in line with the axis of the coil. (If you plan to use an existing high quality commercially built edge-wound coil, I would suggest throwing together a mock-up as a prototype, using heavy-gauge scrap copper wire or small diameter tubing to make sure your estimate is close to correct before you cut and damage a good coil only to find it won't work.)

When the exact mid-point is determined, cut out a section of the conductor to form a gap in the winding at that exact point. Make the width of the gap wide enough to withstand the peak rf voltage on the line at the transmitter output, with ample safety margin. Connect the wire leads from the transmitter (or balun if you use one) to the cut ends of the coil stock at the gap, and mechanically secure the cut ends.  With a homebrew coil, it is merely a matter of winding it in two identical, symmetrical sections. Make sure that the spiral of both coils continues in the same direction.   The split in the coil serves as the low-Z input, and the outer ends serve as the hi-Z output where the variable capacitor goes and where the output line is attached.

Your description is very good in how to make the coil and apply the connections for the capacitor.  Have you experimented with a tapered line from your rig to the balanced L coil vs. the traditional balanced line?   I have a home brew balanced L on the shelf that I could experiment with.  Or, I believe the home brew link/tank coils would work here if needed.
Chuck 

From my notes on the prototype:

Transmitter output circuit is link coupled, with 4 turns on the links of both transmitters, although the HF-300 rig has a larger diameter link than the one on the 8005 rig. Data was collected mostly using the 8005 rig.

The leads between the output link and the split coil must be as short as possible.  With longer leads (approx. 4 ft) the physical placement of the leads had a drastic effect on tuning; apparently the series inductance of the leads was a substantial percentage of the total inductance in the high-C/ low-L tuned circuit; lead inductance is strongly affected by spacing between the leads. With short leads, the link inductance is in series with the gap in the coil, so the link + each side of the split coil (plus any inductance contributed by the leads) make up the total inductance in the tuned circuit of the tuner.

160M: 8⅔ turns on split coil (4⅓ turns each side of gap).  Variable capacitor is 900/900 pf split stator. Each stator section connected to one side of the coil with frame/rotor left floating, so maximum total variable capacitance is 450 pf. 840 pf fixed capacitance (large transmitting micas) shunted across variable cap/ far ends of the split coil/ 450Ω dummy load. A short section of 438Ω OWL runs between the tuner output and the DL.

80M: As above, 8 turns total (4 turns each side of gap).  No fixed shunt capacitance.

40M: Works with either 1 or 2 turns each side of gap (2 or 4 turns total), no fixed shunt capacitance.

This is essentially the same circuit as the "alternative for small links" shown in the 1957 ARRL Handbook, as mentioned earlier in the thread, except that a single split coil is used instead of two separate isolated coils for the series inductance.

I believe this circuit could be thought of either (1) as a balanced L network matching the link to the load while cancelling the inductive reactance of the link, or (2) a high C parallel tuned circuit, feeding the load as an ATU running parallel tuning, with the split coil in series with the link to form the total inductance of the parallel tuned circuit.

From my notes on the prototype:

Transmitter output circuit is link coupled, with 4 turns on the links of both transmitters, although the HF-300 rig has a larger diameter link than the one on the 8005 rig. Data was collected mostly using the 8005 rig.

The leads between the output link on the transmitter, and the split coil in the tuner, must be as short as possible.  With longer leads (approx. 4 ft) the physical placement of the leads had a drastic effect on tuning; apparently the series inductance of the leads was a substantial percentage of the total inductance in the high-C/ low-L tuned circuit; lead inductance is strongly affected by spacing between the leads. With short leads, the link inductance is in series with the gap in the coil, so the link + each side of the split coil (plus any inductance contributed by the leads) make up the total inductance in the tuned circuit of the tuner.

160M: 8⅔ turns on the split coil (4⅓ turns each side of the gap).  Variable capacitor is 900/900 pf split stator. Each stator section is connected to one side of the coil with frame/rotor left floating, so maximum total variable capacitance is 450 pf.  840 pf fixed capacitance (large transmitting micas) is shunted across the variable cap/ far ends of the split coil/ 450Ω dummy load. A short section of 438Ω OWL runs between the tuner output and the DL.

80M: As above, 8 turns total (4 turns each side of gap).  No fixed shunt capacitance.

40M: Works with either 1 or 2 turns each side of gap (2 or 4 turns total), no fixed shunt capacitance.

This is essentially the same circuit as the "alternative for small links" shown in the 1957 ARRL Handbook, as mentioned earlier in the thread, except that a single split coil is used instead of the two separate isolated coils shown in the Handbook (one wired to each side of the link) for the series inductance.

I believe this circuit could be thought of either (1) as a balanced L network matching the link to the load while cancelling the inductive reactance of the link, or (2) a high C parallel tuned circuit, feeding the load as an ATU running parallel tuning, with the split coil in series with the link to form the total inductance of the parallel tuned circuit.

I found the "Alternative For Use With Small Links" information on the bottom of page 333, 1957 ARRL Handbook.  I am going to read this again.  I recognize this material but it has been a while.

Chuck