Rob's web

Home - Techniek - Electronica - Radiotechniek - Radio amateur bladen - Radio Communication - Persistent SWR myths

Despite the excellent explanations given a decade or so ago by Walter Maxwell, W2DU, in a series of articles in QSTand later in his book Reflections - Transmission Lines and Antennas published by ARRL, many of the myths that have long surrounded SWR and the use of SWR meters continue to persist. For example, the idea that in the presence of a mismatch between the antenna element and its feeder, reflected power represents lost power dissipated in the transmitter. There is no doubt that an SWR meter is a useful accessory, and virtually essential in the case of a transmitter with a solid-state power amplifier where protection circuits will begin to reduce the output power when the SWR reaches or exceeds about 1.7:1. However it should always be remembered that it is often misleading to attempt to evaluate antenna performance on the basis of 'the lower the SWR the better'.

Fig 1
Fig 1: Graph showing by how much (or at HF usually how little) additional feeder loss occurs with a moderate VSWR on the line compared with the same line having an accurately matched 1:1 SWR.

In the past, based on the writings of W2DU. TT has emphasised such points as:

  1. Reflected power does not represent lost power except for the (usually modest) increase in line attenuation over the matched line attenuation (Fig 1). In a lossless feeder line no power would be lost because of reflection no matter how high the SWR. At HF with low-loss cable, reflected power loss is usually insignificant; at VHF it may become significant: at UHF it may be extremely important.
  2. Reflected power does not flow back into a transmitter and cause excessive dissipation and other damage. The damage often blamed on a high SWR is usually caused by improper output-coupling adjustments and not by the SWR.
  3. Attempts to reduce SWR below 2:1 on any coax line on HF generally represent wasted effort from the viewpoint of increasing the radiation from the antenna - although may be needed to prevent output reduction of solidstate amplifiers arising from the action of the protection circuits.
  4. A low SWR is not evidence that an antenna system is a good one or that it is working efficiently. On the contrary a lower than normal SWR over a significant bandwidth is reason to suspect that a dipole antenna or a vertical antenna is being affected by resistance losses that may arise from poor connections, poor earthing systems, lossy cable or other causes.
  5. The radiator of an antenna system need not be of self-resonant length to achieve maximum resonant current flow, nor need the feed line be of any particular length. A substantial mismatch at the junction between feeder line and radiator does not prevent the radiator from absorbing all the real power that is available at the junction. Where suitable matching (ATU etc) cancels out the reactance presented by a non-resonant radiator and a random length of feeder, mismatched at the antenna junction, then the system is matched and all the real power may be radiated effectively.
  6. The SWR on the feed line is not affected by any adjustment of an ATU (or the length of the cable see below). A low SWR achieved by this means is usually an indication of a mismatch between the transmitter and the input to the ATU.
  7. With an effective ATU and a good open-wire feeder a 132ft centre-fed dipole does not (contrary to general belief) radiate significantly more power on 3.5MHz than an 80ft dipole fed with the same transmitter power. A dipole self-resonant on 3750kHz does not radiate significantly more power on 3750kHz than on 3500kHz with any normal length of feeder, although the SWR may rise to about 5:1 and the coax cable will then, in effect, be working as a tuned feeder.
  8. High SWR in a coaxial feeder resulting from a severe mismatch at the antenna junction does not in itself produce common mode currents on the line or cause the line to radiate.
  9. The SWR in a feeder cannot be adjusted or controlled in any practical manner by varying the line length.
  10. Of the various types of dipoles (thin wire, folded, fan, sleeve, trap or coaxial) none will radiate more field than another, providing that each has insignificant ohmic losses and is fed the same amount of power.

It is however worth noting that where a mismatch exists between a feeder and the input impedance of a receiver, then reflected incoming signals will be re-radiated from the antenna element and lost. Optimum SNR of received signals thus does depend on good matching between feeder and receiver.

It seems worth publishing again an ingenious diagram stemming from a French amateur illustrating the effects of a mismatch between feeder and antenna element: Fig 2. This shows incidentally that forward power in a feeder. as measured in the feeder may actually be significantly more than the real output power of the transmitter!

Fig 2
Fig 2: An ingenious diagram originated some years ago by a French amateur that Illustrates the effects of a mismatch between antenna and transmission line. Although with even moderate SWR a significant amount of power may be reflected, most of this subsequently returns back up the feeder. Note that forward power, as measured in the feeder can be appreciably more than the true output power of the transmitter.

My excuse for attempting, once again, to destroy some of the common myths about SWR is that misleading information continues to be published in some periodicals and books. A letter 'SWR and the feed line' from Phil Winter. KM4OD in OST, September 1992. pp88-89, expresses concern that an article in QST, April 1992 on a five-band, two-element quad - although generally extremely useful, seemed to imply that adding three feet of coax to a quarter-wave feed line 'cured' a 4:1 SWR on 7MHz. KM4OD shows that on 7MHz three feet of coax would have a negligible effect on impedance and no effect on SWR.

He writes: "I strongly suspect that the real culprit was not the quarter-wave feed line or the reactive load, but current flowing on the outside of the coax shield. This could lead to a 4:1 SWR indication on the shack SWR. Maxwell states 'Since there is no practical way to determine the impedance of arm 3 (outside of the coax shield), the true antenna impedance and SWR cannot be calculated from the measured data'. Adding three feet of coax could have altered the impedance on the outside of the shield so that a lower SWR reading was indicated .... If we are to get the most out of our stations. we must make sure we understand what's going on in that black stuff that runs out the wall and up the tower. Firstly, SWR does not change as the feedline length is altered, discounting attenuation effects. If we do measure an SWR change as the feed-line length is changed, then current must be flowing on the outside of the coax shield, rendering any SWR measurement inaccurate. Secondly, line loss aside, no single length of coax is any better or worse than another. Make your feed line long enough to reach your antenna, turn on your rig and make contacts."

Perhaps a more typical example of the misleading nature of antenna measurements brought about by RF current on the outer shield was reported in the Technical Correspondence column of QST a few years ago by Scott M Hower, K7KQ, as follows:

"Antenna construction and experimentation can sometimes be very confusing, with measurement yielding results that change and do not seem to make sense. Most amateurs use an SWR indicator or perhaps a noise bridge, both of which generally provide useful data for HF antenna measurements. There is one situation, however, that will result in meaningless readings from either device. that being when the outer shield of the coaxial transmission line becomes part of the antenna system. My antenna is a 14MHz vertical ground plane, consisting of a quarter-wave vertical radiator mounted at the peak of the roof. Four radials, each a quarter-wave-length long, run out from the base of the antenna along the roof. A short length of coax connects the antenna to the radio equipment on the top floor, immediately beneath the ground plane system."

"Two major problems I encountered were constantly changing SWR-meter (or noisebridge) readings and lots of RF in the shack. creating a hot chassis, microphone and so on. Several attempts to detune the transmission line by using different line lengths had no effect on the problem of RF in the shack. Changing the transmission-line length did. however, have an effect on the SWR readings. The clincher occurred when I observed that the SWR readings changed when the headphones were plugged into the rig!"

Finally, I realized that the outer shield of the transmission line was contributing to the composite load, made up of the shield and the antenna itself. Further, all of my radio equipment - SWR meter, linear amplifier, coaxial cable jumpers, even the headphones and microphones - would add to the total length when connected.

"One might think that a good earth connection would solve the problems. but a good earth is hard to obtain from the third floor of a house. Turning to The ARRL Antenna Book, the solution became obvious. A portion of the coaxial transmission line was wound into an RF choke, 5 turns, approximately 6in in diameter - right at the base of the antenna. The addition of the choke cured both problems completely."

It may be useful to add that the problem of a 'hot' chassis in an upstairs shack can be successfully overcome by the connection of a quarter-wave length of wire to the 'earth' point of the ATU - a dodge that has been mentioned several times in TT. Similarly the choke current-balun could be of the ferrite-ring type. using the constructional details shown in the September TT.