MARTIN - G8JNJ

When using an autotuner you don’t necessarily have to use a resonant length of wire. In fact if you wish to use one for multiband operation it may be best to pick a length of wire which isn’t resonant on any of the required frequencies. Let me explain.

 

Auto tuners tend to fall into three categories. The first type is often referred to as auto coupler and is usually built into the transceiver. They have a limited impedance matching range, and can only cope with VSWR’s of 3:1 or less. The main purpose is really just to increase the bandwidth of resonant antennas to provide full band operation.

 

The second types are those which can match a wider range of impedances, and are able to cope with VSWR of up to 10:1. These are more versatile and can be used to match a variety of coax fed antennas or random wires when used with a 4:1 transformer. A good example of this is the LDG Z-11 (6 ohms to 1000 ohms) or the MFJ-991 (6 ohms to 3200 ohms)

 

The final types are genuine random wire tuners which should be able to cope with almost any impedance, although in practise they have a finite matching range. Examples include the SGC range such as the SG-237 or MAC-200.

 

Some test results can be found at :-

 

http://ham.srsab.se/pdf/test_aut_tuners.pdf

 

All of these tuners use a switched LC network to provide the best match to between the 50 ohm nominal output impedance of the transceiver and the complex impedance presented by the antenna and feeder. In order to reduce the size and complexity of the design most tuners use a compromise in the maximum values of capacitance and inductance which can be switched into circuit. Unfortunately it is not always apparent what the actual matching range of a tuner is. Some manufacturers give a range of frequencies and impedances the autotuner will handle. Others simply give limited frequency ranges, maximum and minimum wire length, or say avoid using ½ wavelengths of wire. Care is required when interpreting these descriptions.

 

Further info can be found at:-

 

http://www.k0bg.com/couplers.html

 

As there is a harmonic relationship between the amateur bands, if an antenna is resonant at 3.5MHz it will be presented with a high impedance on frequencies which are even harmonics of the fundamental such as  7, 14, 28MHz  etc. In these cases the impedance can become as high as 5000 ohms which is well beyond the matching range of most autotuners. Alternatively if a very short antenna is used, such as a mobile whip, the impedance is likely to be much less than 50 ohms. Typically it can be 15ohms or less, which is also outside the matching range of many auto tuners.

 

Many manufacturers suggest using a 4:1 transformer in order to bring the antenna impedance within the matching range of a tuner. Although this works reasonably well, it is compromise, as impedance transformers only perform correctly when the termination impedances are close to the design values. Outside of this operating range, a mixture of antenna and transformer impedances interact in an unpredictable manner which can result in less than satisfactory performance. In some cases the only reason this technique works, is because the additional losses introduced by the balun mask the reactive component of the impedance to a level that the tuner can handle. My own experiments with a variety of baluns suggest that system losses of 2-6dB may not be uncommon with high, low or very reactive terminations. However this is still likely to be more efficient than connecting the antenna directly to a remote tuner via a length of coax if the VSWR exceeds 10:1 on a desired frequency.

 

I have recently been trying to quantify the losses through baluns and transformers when used with remote tuners in this way, as I suspected that the losses were much higher than expected. I also suspected that simple auto-transformer style transformers may be more suited to this application than more complexs transmission line versions.

 

In order to test this theory I used the configuration shown in the diagram below

 

 

 

 

 

The basic idea is that a second ATU or L-network is set-up to represent a complex load impedance. During my experiments I found that the greatest loss seemed to occur when trying to match to a low impedance load, as I was unable to measure much loss when I set-up impedances of 200 to�� ohms.

 

I was also unable to perform tests with load impedances much lower than 15 ohms, when testing 4:1 baluns. This was because the load impedance, which was transformed down by the balun, was outside the range ATU 1 could match to.

 

However the concept seems to be valid and I was able to make a set of measurements with several different types of design.

 

Test freq 3.6MHz	Tuner in by-pass	Tuner + Txfmr	Tuner + Txfmr + X	4:1 Tm line 6t bifilar	4:1 auto-Txfmr 6t bifilar	4:1 Coax Tm line	4:1 T130-2 14t bifilar auto-Txfmr
Through loss	20			19.5	19.5	19	18.5	Watts
Through loss	Ref			-0.11	-0.11	-0.22	-0.34	dB
50+J0	20			19.5	19.5	19	19.5	Watts
50+J0	Ref			-0.11	-0.11	-0.22	-0.11	dB
15+J25	20	18		16.5	17.5	14	17	Watts
15+J25		Ref		-0.84	-0.58	-1.55	-0.25	dB
15-J100	20	18	12	11	4	3.5	10.5	Watts
15-J100			Ref	-0.38	-4.77	-5.35	-0.58	dB
15+J100	20	18	16	14	15	9.5	15	Watts
15+J100			Ref	-0.58	-0.28	-2.26	-0.28	dB

 

 

The most useful comparison is between the 4:1 Tm line 6t bifilar and 4:1 auto-Txfmr 6t bifilar, both of which used the same set of windings on the same ferrite core, but were configured either as a Transmission Line, or Flux coupled auto-transformer. In both cases the measured losses are very similar, except when feeding a low resistance load with a relatively large capacitive component. This is typical of the sort of load presented by a short vertical antenna (10m) on 3.6MHz. Which is why I used it as one of the test impedances. However all is not as it may seem. I suspect the high value of loss for the auto-transformer configuration, is simply due to ATU 1 not being able to cope with the load it was being presented with.

 

Also note that the 4:1 coax transmission line transformer has a high loss figure when feeding the 15 -J100 load, this may be due to ATU 1 not matching, or it may be a genuine problem associated with the type of construction and the coax introducing a highly capacitive parallel component.

 

This brings another aspect of this concept into question. If you can match to a low impedance load via a 4:1 transformer, is it because you have a good ATU, or is it because the transformer is lossy, or is converting the load impedance into something the ATU can handle ?

 

More work needs to be performed on this subject.

 

 

Because of the harmonic relationship of the Amateur bands, you are likely to encounter problems if you wish to obtain multiband operation with a random length of wire. One solution is to select a length which has a natural resonance outside the Amateur bands.

 

For example, good lengths of wire to use for vertical antennas, permitting operation from 80m to 6m are around 7.2m (tune for resonance on 10.4MHz) or 9.2m (tune for resonance on 8.2MHz). Note that the shorter length gives better results on the higher frequency bands. This is because the radiated energy tends to tilt up towards the horizon, when the length of the antenna becomes greater than 5/8 wavelength long. This results in poor performance for long distance contacts.

 

Using these wire lengths the impedance rises to a maximum value of around 1000 ohms on 14 and 18MHz and to a minimum of  approximately 78ohms at 7MHz, all of which are fairly manageable. Even my own LDG Z-11 Pro, which is not a true random wire autotuner, can work with this length of wire without an additional 4:1 transformer.

 

If you are using a longer length of wire, perhaps as a Sloper or inverted ‘L’ good lengths to choose are either 19.4, 22.8 or 34.3m long, as these avoid high impedances on most of the Amateur bands from 160 to 6m.

 

Even if you are using a true random wire tuner, it may still be advantageous to use a length of wire which does not exhibit a very high impedance at the required frequencies, as this can result in very high voltages being present at the feed point. This can place additional stress on the matching components inside the tuner and components such as insulators on the antenna.

 

Martin Ehrenfried – G8JNJ – 21/04/2008 – V1.4