Antenna Tuners

[K5UJ]

I decided to run it with the cover off and see exactly where it was unhappy.

I always run the L network with no cover.  (I love it when I get told that lets all the RF out.)  I can just look and see how much L and C I am using.  Adding fixed C is easy, no rotary switches or anything fancy, just plain old copper straps and plugs.  Very reliable.

1/4 inch cap spacing gets you 9 KV rating.


http://www.cardwellcondenser.com/PAGES/jcsp14.html

Bob, it wasn't clear in your original question that you were asking about RF final output networks.  With pi networks the h.v. is on the tune cap, the one right at the anode(s) just after the blocking caps.  I don't know about class C PAs but with leenyar amps (what I use) the tune cap v. rating should be at least twice the B+.  The v. on the load cap is substantially lower because of the inductor storing and releasing energy so on its output you have less v. and more current.  The load cap can be lower v. rating and wider range.   Usually what arcs in a pi output network is the tune cap.  But the spacing all depends on what the B+ is.

[W4RFM]

The reason all this came to mind was, being a broadcast eng for too many years, I remembered that the coils & caps ( the caps were mostly door stop fixed jobs), were always the same size as the "dog house" had. That is the matching network at the base of the tower. So, the question came to me, if those same size pieces could match 570 KHz to 300' of steel, why do I need the increased spacing and robust-ness of componets in my little matching network for 300 watts of AM.

After reading and re reading the posts here, I see your points, and they are well taken. Thanks everyone.

[The Slab Bacon]

The reason all this came to mind was, being a broadcast eng for too many years, I remembered that the coils & caps ( the caps were mostly door stop fixed jobs), were always the same size as the "dog house" had. That is the matching network at the base of the tower. So, the question came to me, if those same size pieces could match 570 KHz to 300' of steel, why do I need the increased spacing and robust-ness of componets in my little matching network for 300 watts of AM.

Actually, Bob,
                   If you look at the final tank/antenna matching section of even a 250 or 500w BC transmitter, they are pretty robust in comparison to "Ham Grade" stuff. When building a tuna, OVERKILL is your best friend. No ifs, ands, or buts! ! If you build it massive and capable of 2 or 3 times your power level, you will never regret it!

[K1JJ]

Bob,

Think of a link coupled tuner as a backwards pi-network, as far a impedance transformation and voltages are concerned. Forget about the balanced output for a minute.

Simply stated, the 50 ohm end sees low voltage and high current --  while the ~ 2K end has high voltage and low current.  Just like in your transmitter. The caps need to be rated for their jobs.  The tuner 50 ohm end WILL see about  the same current and voltage as your pi-net loading cap. And yes, no input tuning cap for the 50 ohm tuner link is generally needed. It can help increase overall matching range, but most matches can be configured at the output tank to give a 1:1 match with a fixed link.

The key here is that each capacitor handles the same power in whatever service it does - it's just a different E * I  ratio.



Moving on, Patrick brought up an interesting question about circulating current in the tank.

Normally, IIRC, in a parallel resonant tank, the circulating current is = Q * E/R.  IE, if the Q is 1 then let's say the current is 1A. If the Q is 10, then the circulating current is 10A.  Or at least Q has a big effect on circulating current.  ( I haven't looked up the formula lately)

My question is, is RESONANCE necessarily a requirement for a link coupled tuner to show a 1:1 match?  I think not but am not sure for all situations.

What if we have a feedline with R= 2,000 ohms and with Xc=0 and XL =0?    Will the tuner combo of L/C be necessarily tuned to resonance  when matched?

And how about a feedline that shows R= 2000 ohms with Xc= -500 and Xl=+500?  By adjusting C, L  and taps to cancel out the feedline reactance and a match transformation is made, does this necessarily mean L/C resonance?  If so, this may mean higher circulating currents since Q would increase with resonance.

IE,  how does L/C resonance figure into a link coupled tuner's adjustments after the feedline Xc and XL are tuned out and all is matched?

T

[Steve - K4HX]

My question is, is RESONANCE necessarily a requirement for a link coupled tuner to show a 1:1 match?  I think not but am not sure for all situations.

Yes, or no energy would be coupled through the link.

[K1JJ]

My question is, is RESONANCE necessarily a requirement for a link coupled tuner to show a 1:1 match?  I think not but am not sure for all situations.

Yes, or no energy would be coupled through the link.

I assume you mean the secondary L/C components where the feedline is connected must be resonant?


However, the input link can be just a coil with no capacitor and never tuned, thus not necessarily resonant.


T

[Steve - K4HX]

Yes.

[K1JJ]

Yes.

HO-K.

I think what is cornfusing is that we can use a coaxial balun (a 1/2 wavelength 1:1 coaxial match using a stub short or open) and tap the open wire to find a perfect 1:1 match to 50 ohms. This is not resonant but works FB.  Or is it?

When using a link coupled tuner in the past I found there were many settings of L/C that would achieve 1:1 just by moving the feedline taps around. These items caused me to wonder if resonance is really a requirement.  

Hmmmm... then again by definition, resonance is when Xc=0 and XL = 0. This is what we are doing with matching.  No one said the Q had to be any particular number too...   (Edit correction -  resonance is better stated as XL=Xc)

T

[Steve - K4HX]

No matter how you match, ultimately the entire system must by resonant or it won't take power.

[K1JJ]

No matter how you match, ultimately the entire system must by resonant or it won't take power.

That says it all.

So the more reactance a system has, the less and less power it will take, with pure resistance taking maximum power.  Any kind of matching system basically cancels out the reactance and matches the source to the remaining resistance for max power transfer. Resonance is a byproduct of this process.

It's tough getting old..

T

[KA2DZT]

Yes.

HO-K.

I think what is cornfusing is that we can use a coaxial balun (a 1/2 wavelength 1:1 coaxial match using a stub short or open) and tap the open wire to find a perfect 1:1 match to 50 ohms. This is not resonant but works FB.  Or is it?

When using a link coupled tuner in the past I found there were many settings of L/C that would achieve 1:1 just by moving the feedline taps around. These items caused me to wonder if resonance is really a requirement.  

Hmmmm... then again by definition, resonance is when Xc=0 and XL = 0. This is what we are doing with matching.  No one said the Q had to be any particular number too...

T

Resonance is when XC is equal to XL  each is not zero.

Moving the taps around alters the inductance of the coil, I would think.  Changing the inductance would require a different setting of the cap, I would think.

To me the way the tuner works is that it tunes out the reactance on the feeders bringing the whole system (tuner plus antenna) into resonance.  The link does not have to be resonant.  The link excites the coil.  The load on the coil (antenna) absorbs power from the link which is drawing the power from the xmtr.  Just like any xfmr when the load on the coil is correct to present a 50 ohm load on the link you would get maximum power transfer.  The turns ratio between the coil and the link determines this.  There is some flexibility in this because a pi network can match a range of loads not just 50 ohms.

The problem is finding the correct locations on the coil (taps) that matches the feeder resistance (R).  The question is, what is the resistance (R) on the feeders at the tuner?  One would think it's the nominal impedance of the balanced line (450 ohm or 600 ohm or whatever).  But I don't think it is.  It would be the nominal impedance if the load (R) at the end of the line was the same as the nominal impedance.  In most cases the load (R) of the antenna feed point is not the same as the nominal impedance of the OWL.

With a 1/2 wave dipole, you can attach the feeders at some distance from center that was the same as the feeder resistance (R).  Center being the usual 70 ohms and some distance along the antenna from center you can locate the higher nominal impedance of the feeders.  You would now find that the resistance (R) at the tuner would be the nominal impedance of the line.  This works great if you're using the resonant antenna on just one freq.

When you try to use a fixed length antenna on different bands many different conditions can exist at the tuner.  Try to add something about this in another post.  Right now my brain is smoking.

Fred

[ke7trp]

There lies the trouble.  You can tap all over that coil while turning that cap to find a great match.  The only way to know its tuned correctly is with current meters and a Big Plus is a IR temp gun.  Once its cold and the current is high, you know its tapped correctly. 

C

[KA2DZT]

There lies the trouble.  You can tap all over that coil while turning that cap to find a great match.  The only way to know its tuned correctly is with current meters and a Big Plus is a IR temp gun.  Once its cold and the current is high, you know its tapped correctly. 

C

Not really,  there will be one set of tap locations that will match the resistance (R) of the feeders (at the tuner) to reflect a 50 ohm load by to the link.  Retuning the cap will tune out the reactance and bring the system back into resonance.

Fred

[K1JJ]

Resonance is when XC is equal to XL  each is not zero.

Roger on that. I woke up this morning realizing that error....  XL = Xc, thanks -  not Xc and XL =0



Now to answer Patrick's question of how to calculate tuner tank circulating current. Now we know there must be resonace.  


Someone tell me if this is correct:

First we must figure the Q. This will require knowing the RF resistance of the coil and capacitor.     At matched resonance, the tank  I_circulating would be Q* E/R     OR     I_c= Q*Sqroot(P/R)      OR     I_c=  Q* P/V

How's that for a start?

T

[KA2DZT]

Tom,

Trying to figure the current in the tank coil of the tuner may not be as easy as it looks.  The circulating current in the coil without a load may be one thing, but when you add on the load of the feeders, attached somewhere along the coil, the problem gets more complex.

Fred

[WQ9E]

I remember QST reviewed a number of tuners several years ago and the place that most of them got into trouble was matching a low impedance load on 160.  As I recall some of them provided less than 50% efficiency which means a lot of RF getting turned into heat in the tuner.  I think MFJ has improved their tuners but some of them were pretty famous for coils that got hot enough to burn the spacing insulators back in the 90s.

I use a simple L network with a Gates roller inductor.  The C comes from a vacuum variable and I use a BC-375 tap switch to move the capacitor from one side of the coil to the other for matching either high or low impedance loads.  A second tap switch allows adding additional padding caps.  I use turns counter dials for the inductor and capacitor along with a chart so changing frequency is fast.  This is used with a 80 meter full wave horizontal loop (fed as a top loaded antenna on 160) for the vintage gear in the house.  I have never observed any heating (or shift in match/SWR) while using various rigs including the Desk KW on AM.  The more modern gear in the barn radio room feeds a Hy Gain Hy tower for 160-40 and a four element quad for the higher bands and no tuner is needed with these.

[W4RFM]

If you look at the final tank/antenna matching section of even a 250 or 500w BC transmitter, they are pretty robust in comparison to "Ham Grade" stuff.

 
 I said or meant to say that the size of the components in the final of the broadcast transmitter were the same physical size as the matching network in the tuning unit.
Thus the question why do I have a receiver spaced loading cap connected to a tuner with a 9 KV input cap (exageration).

[flintstone mop]

I melted a lot of tuner parts with just 300 watts from a T-368..........A lot of wire for a dipole and OWL just 30 feet from the ground...musta been 20 amps of antenna current ouch!

[ke7trp]

That T3 is no joke.  If something will fail it will fail behind that T3. It comes up to power real fast.

C

[W7TFO]

BC ATU matching hardware handles a lot of mod power + carrier, but it never has to  QSY by design.

Ham stuff is all over the dial, and some overlaid Smith charts would look like a kaleidoscopic bunch of ink.

One needs to be able to encounter many different impedances for amateur service.

To wit, any piece of metal can radiate on any given frequency IF (big if) you can live with the resultant impedance. 

IOW, the 'teeter-totter' gets crazy lopsided in the real world, and lots of hams' antennas run on the ragged edge of physics.

73DG

[The Slab Bacon]

To wit, any piece of metal can radiate on any given frequency IF (big if) you can live with the resultant impedance. 
IOW, the 'teeter-totter' gets crazy lopsided in the real world, and lots of hams' antennas run on the ragged edge of physics.
73DG

that definately sez it all! !  Any length radiator can work on any frequency IF you can deal with the resultant impedance and reactance. I have preached this for years and some just refuse to believe me. A real short antenna may not be efficient or perform very well, BUT it CAN be made to work!! You just have to make it look to the transmitter like something it wants to load into. That is where the tuna comes in.

As you get into wierd situations, your tuna can be dealing with extreems of current or extreems of voltage. If you are using a "non standard" skyhook, build a tuna that is ready to deal with either and you will live happily ever after. As I said before, build it to handle at least 3 times the power level you plan to run and you will never regret it! ! !

[K1JJ]

Any length radiator can work on any frequency IF you can deal with the resultant impedance and reactance.

Yes!

Here is an extreme case example to remember:

If we had use of so-called "super conductors," metal would then have a resistance approaching zero ohms.  We could then construct a 3el 75M rotary Yagi that could be held in the palm of our hand. This antenna would have the same transmit gain and pattern as a full size Yagi with 140' rotary elements.  The elements would probably resemble resonant loop-stick ferrite elements.  This tiny horizontally polarized Yagi antenna would still need to be supported at 1/2 wavelength above the ground to achieve a good DX pattern, however.

IE, there is no requirement demanding full-size 1/2 wave elements, etc., if losses are controlled.  A point source will work FB.

So, as already said, a short antenna can work FB as long as the losses are held in check from A- Z.

T

 

[W2DU]

that definately sez it all! !  Any length radiator can work on any frequency IF you can deal with the resultant impedance and reactance. I have preached this for years and some just refuse to believe me. A real short antenna may not be efficient or perform very well, BUT it CAN be made to work!! You just have to make it look to the transmitter like something it wants to load into. That is where the tuna comes in.
 

Right on, Slab and Tom, exactly what I've also been preaching !

Walt

[k4kyv]

My question is, is RESONANCE necessarily a requirement for a link coupled tuner to show a 1:1 match?  I think not but am not sure for all situations.

The link does not have to be resonated with an external capacitor in series or parallel. It may be resonated by the capacitive reactance presented to it by SWR on the feed line.

By true "link", we are talking about a coupling coil that is a few turns of wire wound round the cold spot on the main coil. The coupling coil may also serve as the untuned primary of a transformer.  First, think of a balanced tuner, with the main  coil tuned to resonance, working into a load.  If we want to feed it with, say, a 50Ω line, we can find two points, symmetrical about the mid-point of the coil, that will measure 50Ω resistive.  We could use a couple of coil clips at this point, and feed the coil directly. Or we could wind the same number of turns close-coupled round the coil, to achieve the same effect. But the coupling must be almost unity to fully load the transmitter. The old-fashioned swinging link HVDL-type coils won't quite cut it.  Even with the link all the way in, the final will probably not load up to full plate current.  By inserting a series tuning coil to resonate the link, full load can be achieved with the link about 2/3 way in. For those who say they can fully load the transmitter with nothing but the coax connected directly to the link, I'd say what is really happening is that the coax is running with substantial SWR, and the length just happens to be such that the line presents the amount of reactance needed to approach or achieve resonance with the link at the operating frequency. For the "link" to serve as the untuned transformer primary, expect to need more turns than the normal 3 or 4.  As I recall, the ARRL handbook gives information on how to figure out the number of turns needed for each band without a lot of mathematical calculations.

Back in the early 60s after hams had become obsessed with SWR but many still did not possess a reflected-power meter because they were expensive or beyond their technical expertise to build, I used to chuckle whenever I heard Hammy Hambone on the air explaining that he had "trimmed" the length of his coax to get the SWR down to 1:1 and allow the transmitter to fully load.  Of course, he wasn't changing the SWR at all; by trimming the length of the coax he was presenting the link with just the right capacitive reactance to resonate the link coil. This works similarly with a pi-network and has resulted in the same mistaken belief about trimming the coax and "getting the SWR down".  In fact, if just the right (or I should say wrong) reactance is presented to the output of the pi-net, the loading capacitor works in reverse; heavier plate loading is achieved by increasing the capacitance of the loading cap instead of decreasing it.

I had a problem loading my homebrew link-coupled rig on 160 to a 50Ω transmission line. With the series resonating capacitor, I could peak the plate current, but with the 4-turn link shoved all the way in, it would still run only about 1/3 the full plate current. I tried the variable  capacitor both in series and in parallel with the link coil, with the same result.  Then, by a long episode of trial-and-error I came up with what worked.  Two heavy-duty fixed mica transmitting capacitors, several thousand pf each, the kind you see in BC transmitters, connected in series, and the whole thing connected in parallel with the link, with an additional 1800 pf variable paralleled across one of the fixed micas (to serve as fine tuning for resonance). The 50Ω line connects across one of the fixed capacitors. When the circuit resonated with the variable, full loading is achieved with the link a little over half-way in.