It seems to me that when we achieve a purely resistive load to an transmitter, we are creating a resonant system of the antenna + feedline + tuner. Of course we aren’t making the antenna itself resonant, but I believe that we are resonating the system as we have caused energy reflected by the antenna to “stay” in the tuner+antenna+feedline system.
Am I thinking about this correctly or is my mental model missing something?
The problem with this line of thought is that there is not a clear, universal definition of “resonant” as applied to an antenna. A resonant antenna can have a high SWR, which isn’t what some people mean when they insist a resonant antenna is best.
But, yes, you have the right idea: an antenna tuner provides an impedance match between the feedline and the transmitter, allowing the transmitter to deliver full rated power to the load. If there is no reflected power on the coax between the antenna tuner and the transmitter, then none of the reflected power in the system is getting back into the transmitter: all the power is either radiated (good) or dissipated as heat in the coax and antenna tuner (not as useful).
Some people may still complain that your antenna “isn’t really resonant”, or you are just “fooling the transmitter”. Ignore them. Use a reputable calculator to show that the coax losses are acceptable for what you want to accomplish, and don’t get tangled in their semantic mess.
Tony,
You have an excellent way of distilling complex explanations into a few sentences. Forgive me for expanding the story.
The ATU transforms the impedance presented to it by the feedline into 50 +j0 ohms. The transmitter knows nothing of the SWR on the feeder so it runs as it should.
The feedline antenna combination is both made resistive in value (resonated) and the resistive component transformed to 50.0 ohms. Often this dual action is achieved by a simple L network with one inductor and one capacitor.
If the feedline is balanced a balun or similar arrangement is used to preserve feedline balance input and unbalanced output.
If we ignore the transient surge when the first few cycles of RF enter the system then we have a steady flow of power from the tx to the ATU and up the feeder to the antenna. Now power flows back to the ATU from the Antenna.
Measurement of the voltage along the feedline will show a periodic fluctuation that can be shown as mathematically the result of two waves flowing on opposition and mutually interfering. However the actual power flow is to the antenna. Losses will of course be incurred along the way.
Tha ATU presents a conjugate match to the antenna end of the feedline so all of the power that reaches the antenna is accepted by the antenna and all the power received, less antenna losses, is radiated.
The radiated power is the transmitter output power less the losses in the ATU and the feedline and the antenna. It will mostly be, or should be, no more than 3 dB less than the transmitter power. Except for QRP operation, higher losses may be unacceptable due to temperature rise in the lossy parts. If you are running 1 kW having hundreds of watts dissipated in the ATU will certainly be noticed whereas 5 watts from your QRP rig will be hard to notice.
Having a simple doublet radiator fed with open wire line and an ATU is an effective way to operate on 80 m thru to 30 m. At shorter wavelengths the radiation pattern breaks up and the likelihood of the DX station being in a null increases. I have used such a system from 3 MHz to 50 Mhz with acceptable. results.
It has much to commend it. The radiator length is less critical and what you can fit in is more important than is it a particular length.
73
Ron
VK3AFW
This is a very important point! I’ve seen the plastic coil supports in tuners melt, or the coils even unsolder themselves, at well below the rated power of the tuner when it has high losses. In fact, an easy way to monitor tuner efficiency at 100W is to use a “digital temperature sensor” (aka “finger”) applied to the coil after a period of operation. (But not when transmitting, of course!)
At QRP levels we don’t have such an indication that the tuner (or feedline or antenna) has high losses, except that our signal is weaker than it might be, and it is easy to wave that off as poor propagation.
Tuner loss isn’t always easy to measure, as it depends on the frequency and the load impedance. With some tuners, such as the common T network types with 3 controls, it is often possible to match an antenna using different combinations of settings, but with different efficiencies.
That’s not to say not to use a tuner - sometimes that is what is needed to get on the air with a compromised antenna, and they can be very effective. But it is worthwhile being aware of, and trying to minimize, potential losses. Here is one evaluation that I did.
Great insights and great article! I appreciate the separation of transient response and steady state response 
I haven’t thought at all about how the different match settings/systems could impact system losses and I have had some fun playing with a doublet into an old johnson matchbox I breathed life back into. I have a little thermal camera and it might be fun to pull the cover off and make some slightly less “digital” measurements!
Following through with this thoughts, are automatic tuners at risk of sub-optimal local minima? E.g. if you happen to start near a sub-optimal network configuration will it descend to that minimum or are the controllers smart enough to select the configuration that will minimize loss?
Now thinking through your article, I am thinking that it could be fun to figure out how to effectively model the anticipated loss for a given antenna/feedline combination. I think we would want to map the energy loss surface L(f,C1,L1,C2 | z_ant+feedline) Is the loss model for the inductor/capacitor simple? Are losses primarily resistive losses in the inductive element itself, contact resistance?
-Tony
Without getting technical about it, the “tuner” makes the radio think the antenna is
resonant. The antenna/co-ax, etc. may not be radiating very well on whatever non-
resonant frequency you may be using but the radio will like it and put out power.
That power might be wasted, or a goodly amount of it will be wasted.
K6YK
So staring at this a bit more, it seems that my concept “resonating” the system is at best incomplete. For example, if we are starting with a resonant system at say 450 + 0j we can still use this network type to convert to 50 + 0j but there isn’t a meaningful sense of resonating here since we already had a purely dissipative load.
I don’t like the term ‘resonant antenna’ and have said so a few times on other SOTA threads – it’s misleading, but I don’t want this to be a discussion about semantics. I’m no antenna expert so anyone who is, feel free to correct my understanding: -
My (centre-fed) linked inverted-V dipoles need about 10m of feeder to run from the top of my 6m or 8m poles to my operating position a few metres from the base of the pole. Doesn’t this mean a significant loss of my precious 10W of CW despite the excellent ATU in my KX2?
By contrast, my inverted-L / inverted-7 EFHW needs only a 2-3m feeder (summer) or 4-5m feeder (winter) from the impedance-matching box (near the bottom of the pole) to my operating position, presumably with much less feeder radiation at non-resonant frequencies. Despite the claimed advantage [by some] of the CF HW dipole over the EFHW, doesn’t this feeder loss diminish any practical advantage?
I have also talked previously about how the resonant frequency particularly of wire antennas can change by how and where they are deployed. Using an antenna analyzer, I’ve noticed the resonant frequency can change quite alarming with change of pole height, slope angle of the wire elements, and how tight the wires are (wire sag reduces the effective length). I believe the topology and soil type at the summit, large objects in the near field, etc. affect the impedance as well as radiation pattern.
Your ‘resonant antenna’ may not be as resonant as you think.
I think of the antenna system in terms of impedances. I don’t think its helpful to think about resonance.
The antenna feed point has some impedance, the transmission line has an impedance (hopefully the same as the antenna, or is a low-loss line), and the antenna tuner is a magic box that converts the antenna/feed line impedance to 50 ohms. Or at least gets it close enough you won’t damage your finals.
I’m working on my mental model of that “magic box” for building intuition.
With my understanding I would only expect the feedline (assuming coax) to radiate when their is rf on the outside of the shield, e.g. common mode current, which would be killed by a balun regardless of antenna resonance. Are you thinking of balun performance dropping because impedances are changing and the balun CM impedance is no longer high enough, or am I not thinking about this correctly?
I always understood the increased feedline losses when operating “off resonance” (i.e. with high SWR on the feeldine) were due to increased dielectric loss as the actual voltages increase with increasing SWR and the dielectric loss is typically proportional to V^2.
And aren’t you typically operating the EFHW on resonance, even if not on the fundamental?
Generally not. Most relay-switched autotuners use a reversable L network, which will only match an impedance at one setting. That is a lower Q setting than you will get with a 3-element tuner (like a T or Pi network), so losses are lower. (Except that the component values for loads close to 50 ohms may be inconvenient, and some L-network tuners will add an extra capacitor to resolve that.)
Older auto-tuners (like the one in my TS-450) use a band-switched coil and two motor-driven capacitors. Because there are only 2 adjustable components, there is only one solution, but it may not always be the most efficient one because the coil can’t be adjusted.
Generally the most efficient setting will use maximum capacitance and minimum inductance.
It depends on the tuner topology.
W9CF had a great T-network tuner simulator that ran online, where you could turn the knobs to match various impedances and it would tell you the efficiency. But now you have to download the code and run it on your own computer.
The primary losses in a tuner are the loss resistance of coil, as the Q of good air-dielectric variable capacitors is quite high. That might not be the case with polyvaricons, or some fixed capacitors with ceramic insulation (or pieces of coax or printed circuit board used as capacitors).
Feedline losses are primarily due to the AC resistance of the conductors, at least below the UHF range (as long as it has polyethylene or PTFE dielectric: PVC insulation has higher dielectric losses). The AC resistance increases with frequency due to skin effect.
I agree: there is nothing “magical” about a resonant antenna. What’s important is that the transmitter sees a load impedance that it can deliver sufficient power into. The transmitter doesn’t “think”: it doesn’t know what causes the particular load impedance that it delivers power to.
The other issue is the potential losses in operating a feedline at high SWR. For that we have some good calculators available, such as AC6LA’s TLDetails. (Calculating power lost due to SWR is not as simple as it seems - sometimes operating a line at high SWR can result in lower losseso than when it is matched.)
it is even more general than that: a non-resonant antenna with a mis-matched feedline will present a resistive (resonant) impedance to the transmitter at some frequencies. How far the frequencies are from being harmonically related will depend on the velocity factor of the feedline, among other things. Such a combination can provide a good match to 50 ohms (low SWR, no tuner required) if the lengths are properly chosen, even if they are non-resonant.
On the other hand, a resonant antenna feed with a non-matched feedline will present a reactive load to the transmitter for most line lengths. (It will only be “resonant” for multiples of 1/4 wavelength, corrected for velocity factor.) Just having a “resonant” antenna (zero reactance at the feedpoint) doesn’t guarantee a low SWR at the transmitter.
True. Resonance or not has nothing to do with whether a feedline radiates.
The ATU, if properly adjusted, provides a low SWR to the transmitter. If the SWR is low, then the reactance can’t be very high. (The reciprocal is not true.) The purpose of the ATU is not to balance the feedline, but some times a balanced ATU may help.
That’s a tricky question: it depends on the antenna. I’ve seen SOTA operators connect a dipole directly to the transmitter and elevate the ends, to avoid feedline loss. But using enough feedline to get the feedpoint higher in the air will actually improve your signal, in spite of the added loss. On the other hand, a delta loop antenna with the point down permits maximum radiation at a greater height with zero feedline.
End-feeding a wire directly from the transmitter, via an ATU or some other approach, can be effective if it gets the part of the wire with the highest current up in the air - for example, an EFHW installed as an inverted vee - or for a vertical with low ground losses. Having the ends close to the ground tends to increase losses.
Then we get to the point of having to decide how much loss is acceptable for our particular operation. On single-hop 20m or 40m paths, signals will generally be strong enough to work plenty of stations (at least for those operating from or near populated areas). If you are trying for trans-oceanic S2S contacts, then loss becomes more important, but so does low angle radiation, which is often enhanced by height. There is no “one right answer”: those are trade-offs for each individual to make given the relative weight, performance, and how they have to install the antenna. I lose 1/4 of my power in 8m of RG-174 on 10m with my dipoles (less on the lower bands), but decided that was acceptable for my own operating style and preferences. Others make other trade-offs based on different circumstances.
Andy,
Feedline radiation is in that other can of worms over there.
The primary cause is lack of symmetry of the feedline and radiator element. It isn’t due to SWR.
It is a feature of some EFHW antennas!
Nothing can be done at the transmitter end to stop this. The best that can be done is to prevent the common mode current getting to the rig or your computer. Common mode chokes on coax are the usual fix in conjunction with RF grounding.
I use a seperate primary winding for my EFHW coupling transformer which reduces the CM problem.
73
Ron
VK3AFW
Yes, I agree. On reflection (no pun intended) my ‘BUT’ in point 3 is probably wrong. A decent balun plus a choke in the feed line near the balun should stop CM current in the feeder and therefore prevent the coax outer from radiating RF even at non-resonant frequencies.
The issue is – without a remote ATU at the antenna end of the feeder – reflected power (due to the mismatch with the antenna impedance having reactive component at non-resonant frequencies – causes standing waves as well as less RF power into the antenna.
A tx-end ATU can prevent the resultant standing waves affecting (e.g. damaging) the tx - by presenting a 50-ohm resistive load to the tx - but doesn’t stop the standing waves. And the length of the feeder doesn’t seem to affect that. So re my point 4, maybe the only argument for using a shorter feeder is to minimize attenuation (e.g. 4m of RG174 at 28MHz is better than 10m of RG174).
Andy,
A small correction to your comment.
Chokes won’t stop the Common Mode current from being present and radiating. They stop the stray rf from wandering about the shack doing things that are annoying.
The feeder is in tbe doublet field and, unless perpendicular to a balanced radiator,will have some induced CM current. That’s independent of whether the doublet is resonant or not.
This not related to how well matched the antenna is to the feedline.
One more point.
You can’t reduce the feedline loss below its matched loss. That’s always there. You will get extra losses due to the SWR creating higher currents and voltages as compared to the matched case. Most coax manufacturers give handy graphs showing losses vs frequency and SWR.
73
Ron
VK3AFW.
Good point. Some years ago I made two 10m RG174 feeders for portable use, one with an in-line choke and one without. I never noticed a difference between them. Since then I’ve not bothered with in-line chokes. I suppose any stray RF is just wandering about the hilltop and not annoying my KX2.
[I didn’t notice this on first reading]
Yes, My most-used one is an EFHW for 40/30/20 (EFT-MTR from LNR-Precision). You remove a SMA link to reduce the length for 30m. I recently trimmed/added length to minimize the SWR at my favourite 30m and 40m CW frequencies when configured as an inverted-7 on my usual pole.
Actually, you can, in limited conditions. As described in the “power lost due to SWR” article from AC6LA that I linked to previously, power dissipation along a feedline operated at high SWR is NOT a linear function. Most feedline losses at HF are due to the current through the AC resistance of the conductors. By choosing a section of feedline where the impedance is high (that is, low current), the actual loss can be lower than the matched line loss, because the current is less for the same power.
That only applies over the high impedance portion of the line - in other parts the current will be higher, and the losses per unit length will be greater. When the line is a multiple of 1/4 wavelength, they all average out, and for long lines (in terms of wavelengths), assuming constant loss per length isn’t a bad estimate.
You can verify this by measuring the return loss of a short piece (less than 1/8 wavelength) of feedline with the far end open and shorted. The shorted case will have higher losses, because the currents are higher for the same power level. Taking the average of the two gives a better estimate of the actual loss.
Other than measuring feedline losses, this usually isn’t very significant in day-to-day operation. But it is a good check on our understanding of how feedlines work with high SWR.
They might not be perfect, but they can do a very good job of it., especially if placed at the antenna feedpoint.
The greatest source of imbalance with a coax-fed dipole is because the outside of the coax shield is connected to one side of the feedpoint, and becomes part of the antenna, just like the antenna wire is connected to the same point. Of course, that changes the antenna, and its impedance, so it can affect the SWR. The current divides between the two conductors based on their impedance, just as it does with two resistors in parallel. When the outside of the feedline has a much higher impedance than the antenna wire, most of the current flows on the antenna and common mode current on the outside of the feedline is reduced. A choke / current balun at the feedpoint does this.
Without the choke, the impedance of the outside of the coax shield depends on the length and what is connected to it at the far end. If it is 1/4 wavelength long and connected to a battery-powered radio with no ground connection, the coax looks just like another quarter wave wire, and can radiate significantly (with high RF voltage at the rig). With a half wavelength feedline under the same circumstances, it presents a high impedance at the antenna, and little current flows on the outside of the coax, even without a balun. Grounding the radio in those cases would change the current distribution, and likely the SWR.
In one case, plugging my headphones into my radio changed the SWR, because it added length to that leg of the antenna.
Adding a choke at the radio can help reduce the stray RF floating around the station, and help avoid that case the SWR changing as you connect other items to the transmitter. But it also changes the impedance looking down the outside of the coax from the feedpoint, which may either raise or lower the common mode current along, and radiation from, the coax.
Yes, there is some coupling between the dipole/doublet and the feedline if they aren’t exactly perpendicular, but it generally is small compared to the effect of the outside of the coax serving as a third wire connected to the feedpoint.
In my experience, when hams set up a coax-fed dipole without a balun (as I frequently do), most of the time they don’t notice any symptoms that they attribute to the lack of a balun. That last clause is carefully worded: there may be symptoms that they don’t notice, or that they blame on something else, like a higher noise level, or unstable SWR. Especially for SOTA work, where the weight and fragility of a ferrite core can be problematic (especially at the top of a thin mast), it is often a reasonable approach. But it helps to be able to identify the various quirks that it can cause, and know how to deal with them if needed.
Same here. For QRP portable I see no difference that the choke makes, for QMX and QCX, efhw or random wire. Direct feed, no coax, but cp.
It’s situation dependent. With an efhw, a kx2, and 5m of feedline I had issues with rfi on the mic and less stable swr as I moved around. Choking @1m from the feed point resolved both.