Impedance matching in antennas

[geoduck]

Why is so much emphasis placed on matching the impedance of a feed line with an antenna?

Shouldn't the emphasis be placed on matching the internal generator impedance AND feed line with the antenna instead?

That is, you take the impedance of the source, transform it using the transmission line equation to get the source impedance as it appears at the terminals of the antenna, and only then do you try to match impedances, except this time it's trying to match the impedance of the source (as seen at the terminals of the antenna) with the impedance of the antenna?

Also, it seems that impedance matching amounts only to cancelling the reactive part of the antenna. You can't add a resistor to the real part of the antenna to help you match the real part of an impedance, because then the extra power you get is going to be used up in the resistor you added and not the antenna, so what's the point? Whereas if you add a reactive part to your antenna, no power is used in the reactive part, so it makes sense to add reactive parts to increase the power, but not resistors. Is this right?

 

[jsgruszynski]

Why is so much emphasis placed on matching the impedance of a feed line with an antenna?

Shouldn't the emphasis be placed on matching the internal generator impedance AND feed line with the antenna instead?

That is, you take the impedance of the source, transform it using the transmission line equation to get the source impedance as it appears at the terminals of the antenna, and only then do you try to match impedances, except this time it's trying to match the impedance of the source (as seen at the terminals of the antenna) with the impedance of the antenna?

Also, it seems that impedance matching amounts only to cancelling the reactive part of the antenna. You can't add a resistor to the real part of the antenna to help you match the real part of an impedance, because then the extra power you get is going to be used up in the resistor you added and not the antenna, so what's the point? Whereas if you add a reactive part to your antenna, no power is used in the reactive part, so it makes sense to add reactive parts to increase the power, but not resistors. Is this right?

Both matches for source-to-line and line-to-antenna matter. The source-to-line is generally easily matched and less of a problem. The bigger mismatch is generally the line-to-antenna. Hence the focus.

You can sometimes improve a match using something called a "pad match" which is basically using a resistive attenuator to improve a mismatch. It works by attenuating the reflected signal created by the mismatch. This is effectively what "adding resistance" does in terms of match.

This works ok in a lab environment when you control the power levels and they are fairly high but not too high: mismatch improvement is directly related to the attenuation inserted which means better match but lower signal levels at the load at the price.

For antenna applications this isn't usually the optimal result. You are typically dealing with signals received that are near the noise floor so adding attenuation is bad news for weak signals. Or you are dealing with high power in a transmitter and the attenuation basically throws away hard-gained antenna drive power.

The typical values you use for pad matches are 10-30 dB so that's 10x to 1000x signal reduction or more clearly and soberingly from 90% power dumped to heat to 99.9% of power dumped to heat.

So in the case of a high power transmitter, putting a even a 10 dB attenuator in line means you are dumping 90% of the power away as heat dissipated in the attenuator rather than being transmitted as usable broadcast signal.

Clearly if you could tune out a mismatch using reactive components that do not dissipate energy, that would be far better, which is what using tuning transformers, loading coils, etc. is actually doing.

 

[skeptic2]

That is, you take the impedance of the source, transform it using the transmission line equation to get the source impedance as it appears at the terminals of the antenna, and only then do you try to match impedances, except this time it's trying to match the impedance of the source (as seen at the terminals of the antenna) with the impedance of the antenna?

The problem with this approach, using the transmission line to transform the generator impedance to the antenna impedance, is that if the transmission line length is ever changed the antenna is no longer matched. A transmitter designer won't know how much transmission line will be needed to connect to the antenna. By using the same impedance for the generator output, the transmission line and the antenna, parts can be interchangeable and the transmission line length doesn't matter.

 

[geoduck]

Both matches for source-to-line and line-to-antenna matter. The source-to-line is generally easily matched and less of a problem. The bigger mismatch is generally the line-to-antenna. Hence the focus.

I think I'm getting confused over two different types of impedance matching. One is conjugate matching, where if you have a source with a load, then you get the most power transferred to the load if the load has resistance equal to the source resistance (and reactance equal to the negative of the source reactance).

The other type of impedance matching is terminating a transmission line with impedance equal to the characteristic line impedance.

The latter involves making the reflection at the line-antenna boundary equal to zero, and the former doesn't take into account reflections but just uses Ohm's law.

So I understand matching the antenna to the feeder in order to avoid reflections.

But is the matching of the source-to-line because of reflection, or is it because you are trying to conjugate match? I thought there is only a reflection when a signal tries to escape the line. But when a signal enters the line because of a generator, there should be no reflection.

The problem with this approach, using the transmission line to transform the generator impedance to the antenna impedance, is that if the transmission line length is ever changed the antenna is no longer matched. A transmitter designer won't know how much transmission line will be needed to connect to the antenna. By using the same impedance for the generator output, the transmission line and the antenna, parts can be interchangeable and the transmission line length doesn't matter.

That makes sense about the length of the line. But does the generator output the voltage at the beginning of the transmission line, or does the signal have to travel a distance inside the generator before reaching the transmission line? Because if the generator outputs the signal directly at the beginning of the transmission line, then it seems you don't have to match the generator output to the transmission line, because the signal is placed at the beginning of the line and not before it, so there is no interface.

 

[Averagesupernova]

Impedance matching to the transmission line is generally not an issue if the length from the transmitter to the antenna is a small fraction of a wavelength of the signal being transmitted. These cases are few and far between of course. However, concerning transmission line inside the transmitter such as between the connector and the output stage, this could be one of those cases that the line is so short compared to the wavelength.
-
It can be mismatched in several ways. Firstly, if the antenna is not resonant it will appear as an inductive reactance or capacitive reactance depending on length. This reflects power back to the source in the same way that a reactance has current out of phase with the voltage. Alot of people have a hard time grasping an antenna actually reflecting power but it is no different than trying to drive a reactive load. When this happens the output stage of the transmitter can seriously overheat since maximum current appears at the crossover point on the AC cycle (current and voltage 90 degrees out of phase). So then we adjust the length of the antenna and get it resonant, but it presents a feedpoint impedance of 100 ohms instead of 50 ohms. This is still a mismatch and a different way of feeding the antenna will fix this. Here's and example:
-
A number of years ago I found out of surplus a whip antenna that had probably been mounted on a jeep or tank. I measured it out and found it was about 5/8 wave for the ten meter ham band. Fed at the bottom a 3/4 wave antenna is resonant. Resonant meaning that it appears neither inductive or capacitive at the feed point. So, a loading coil needed to be added if I were to mount this on the back of my truck and put it on 10 meters. What was done was a coil was wound with one end grouned and the other attached to the bottom of the whip. The number of turns was experimented with until I found a condition that was resonant. I HAD NOT ATTACHED TO A FEED LINE AT THIS POINT! Something called a dip meter was used to find that it was resonant where I wanted it. Basically this is a device that loosely couples to loading coil (transformer type coupling) and is tuned around over the frequency of interest until a meter on the device actually dips telling me at what frequency the antenna is best absorbing power. The antenna was shunt-fed by the feedline. This means that the feed line was attached between ground and part way up the coil. The coil formed a sort of autotransformer. Where this tap is placed on the coil determines what the feed point impedance is. Using another simple piece of test equipment called a noise bridge (think very simple and inexpensive network analyzer) it was determined what position this tap should be to present a 50 ohm feed point impedance.
-
I hope this gives you an idea of the steps involved in matching a simple antenna. Basically 2 steps, finding resonant, and if possible getting the correct feed point impedance. Alot of times we live with a mismatch of feed point impedance as long as the antenna is resonant. For instance, a straight half wave dipole presents a feedpoint impedance of around 70 ohms. Close enough.

 

[vk6kro]

Also, it seems that impedance matching amounts only to cancelling the reactive part of the antenna. You can't add a resistor to the real part of the antenna to help you match the real part of an impedance, because then the extra power you get is going to be used up in the resistor you added and not the antenna, so what's the point? Whereas if you add a reactive part to your antenna, no power is used in the reactive part, so it makes sense to add reactive parts to increase the power, but not resistors. Is this right?

Antennas can be adjusted to some extent to control the feed point impedance, however there is often only so much you can do.

After that, it is a matter of tuning out any reactance to at least get the antenna resisitive.

Then you can use Matching networks, T-Match, Z-Match, ferrite cored transformers, short transmission lines etc to change the impedance to match the line. These are fairly lossless methods of impedance transformation.

Doing this ensures minimum standing waves on the line.
Standing waves produce oddball impedances at the transmitter end of the line as well as large voltage and current peaks in the feedline. These oddball impedances will vary with frequency, so the transmitter will have to be readjusted if you change frequency.

Assuming you can get the standing wave ratio down to something reasonable, there is still a likelihood that the feedline will not look like the charachteristic impedance of the line. ie a 50 ohm coax feedline may look like 30 ohms to the transmitter, and the transmitter has to be adjusted to suit this new impedance.

So, it is better to start at the antenna end, where you have least control, and match the antenna to the feedline and then try to match the feedline to the transmitter.

 

[sophiecentaur]

This is Engineering. You need to work with 'black boxes' where possible so that anyone component can be changed without needing to look at the other. If you make sure that your antenna is matched to the feeder then you know that the transmitter will 'see' 50Ω (or whatever), however long the transmission line is.

Afaik, it is often difficult to combine transmitter efficiency and ideal matching and different transmitters may use different compromises. The impedance out of an High Power AM transmitter is going to vary as the RF level changes during the modulation cycle, for instance. But the designers just aim for the best performance 'into' the specified feeder impedance and leave it up to the antenna and feed designers to get their bit right.