A Short Course On Baluns

A Communicator Reprise: April 2012

Before getting into this, it is necessary to describe a somewhat puzzling phenomenon (called "skin effect") associated with radio frequency currents.

SKIN EFFECT

High frequency AC currents in a conductor flow only in that portion of the conductor which is very close to the surface.  Very close being 0.1 mm (the "skin depth") or less.  While a wire has only one surface, the shield on a coax cable, being a cylinder, has two surfaces - an inner surface and an outer surface.  Consequently, you can have a high frequency AC current flowing on the inner surface and an entirely different one flowing on the outer surface, with neither of them affecting each other. It's like there's an insulator keeping them separated.
Weird but true.

BALUNS

Unless precautions are taken to minimize it, all coax fed antennas will have significant current flowing on the outside of the coax shield while transmitting.
If the antenna is a horizontal dipole with the coax hanging straight down then the coax acts as a vertical antenna.  This is because this  current flowing on the outside of the shield will radiate energy exactly as a vertical antenna would.  
For casual ham HF operation, this can be a good thing as radiation will take place over a wider range of elevation angles which in turn makes it possible to make contacts over a larger range of distances.  However, for a communications antenna system designed to have a specific radiation pattern, this unwanted radiation from the coax changes the pattern and reduces the effectiveness of the system.
When receiving, the coax still acts as a vertical antenna and couples whatever it picks up into the actual antenna and from there into the receiver.
An example of these effects on a Near Vertical Incidence Skywave (NVIS) antenna system follows:

TRANSMITTING

For NVIS the desired radiation elevation angles (the range in angles with respect to the horizon over which the antenna radiates most of the transmitted energy) is about 60 to 90 degrees, 90 degrees being straight up.  A horizontal dipole 1/4 wavelength or less above ground will provide this.

If this antenna is fed with coax and no precautions are taken to minimize current flowing on the outside of the shield, the shield will act as a vertical antenna with a range of radiation elevation angles between about 5 to 40 degrees.  The amount of power radiated at these low angles can be 50% of the power radiated at the desired elevation angles.  This means that less power is available to be transmitted at the desired high angles.  In addition, this power radiated at low angles can cause interference to other stations thousands of miles away.

RECEIVING

This is where the coax shield acting as an antenna really sucks.
The shield, acting as a vertical antenna, will pick up those previously mentioned stations several thousand miles away and feed them into the dipole which will then happily send those signals to the receiver where they will interfere with the signals you want to hear.  
In addition, most noise on the lower HF bands is caused by lightning in the tropics.  If your antenna responds only to high angle signals (NVIS) it won't respond to these lightning induced static crashes because they arrive at low angles.  Consequently, you won't hear them.  However, if the coax shield is acting as an antenna you will hear them and they may make it impossible for you to copy the signals you want to hear.
Because the coax comes right into your station, it runs close to all kinds of noise generating equipment such as computer monitors, plasma TVs, switch mode power supplies, etc.  Because the coax is acting as an antenna, it will pick up all kinds of garbage that the dipole won't hear because it's much further away from these noise sources.

BALUNS TO THE RESCUE

Well, not just any old balun. "Balun" is a contraction of the phrase "balanced to unbalanced". An ideal horizontal dipole is said to be "balanced" with respect to ground.  What this means is that:
The voltage between ground and the point where one wire of the feedline is connected to the antenna is exactly equal (but of opposite polarity) to the voltage between ground and the point where the other wire of the feedline is connected to the antenna.
The current flowing out of one wire of the feedline into the antenna is exactly equal to the current flowing into the other wire of the feedline out of the antenna.

This is an ideal which is never achieved in practice.  Things like one end of the antenna being lower than the other, closer to a tree or building, etc., will all cause the currents in the two legs to be different.  Consequently, no antenna is truly balanced.  Some come close, though.
Feedlines such as the old TV twinlead, window line and ladder ("open wire") line are all considered to be balanced as they are symmetrical with respect to ground (as long as one side of the feedline isn't closer than the other to ground, metallic objects, etc).  
Our transmitter outputs are "unbalanced" with respect to ground.  This is because the outer shell of the output coax connector is connected directly to ground through the metal case of the radio. Similarly, the coax cable we connect to the transmitter is unbalanced as the shield is connected directly to ground via the coax connector on the radio.
So now we have a balanced antenna to which we want to connect an unbalanced coaxial cable. Physically, this is easy to do, we just connect them.  But, as soon as we do, we destroy the balance of the antenna and we get current flowing on the outside of the coax shield with all the resulting bad effects.

So, what we need is a box between the antenna and the coax to make the transition between the balanced antenna and the unbalanced coax feed line, and, yes, these boxes exist and they are called baluns.

There are two types of baluns:

Voltage Baluns

These force the voltage between ground and the point where one wire of the feedline is connected to the antenna to be exactly equal (but of opposite polarity) to the voltage between ground and the point where the other wire of the feedline is connected to the antenna.  The idea here is that, if the voltages are equal, the currents will be too.

However, as most supposedly balanced antennas are at least somewhat unbalanced, the currents in the two legs will not be the same.  This causes a current equal to the difference between them to flow on the outside of the coax shield.  i.e. the coax shield now acts as an antenna.  Just what we DON'T WANT.

Current Baluns (also known as "Current Chokes")

These force the currents in the two legs (halves) of the dipole to be equal.  As there is now no difference in these current values, no current flows on the outside of the coax shield.  This is just what we want, so these are the ones to use.

SUMMARY

Every coax fed antenna system designed to have specific characteristics needs a current balun located at the transition between balanced and unbalanced.  These are often built into VHF and UHF antennas (skirts, sleeves, etc.) but are generally separate items in HF systems.

Without a current balun the antenna is strongly coupled to the outside of the coax shield. 

This means that energy transmitted by the antenna will be coupled into the coax shield which will reradiate it and thereby affect the radiation pattern of the antenna system.

It also means that signals picked up by the coax shield acting as an antenna will be coupled into the antenna and passed to the receiver.

It is important to note that, like any other conductor, the coax shield will pick up signals and re-radiate them.  The current balun simply ensures that these re-radiated signals don't get coupled into the antenna and so don't reach the receiver.

APPENDIX

SKIN EFFECT DETAILS

If you pass a DC current down a wire, the current flows uniformly down the wire.  i.e. the amount of current flowing down the centre of the wire is the same as the amount of current flowing near the outer edges of the wire.

This is not true for AC currents.  For a current with a frequency of 1 MHz in a copper wire, the current will be concentrated in a region with a depth of less than about 0.1 mm from the perimeter of the wire.  This 0.1 mm is called the "skin depth".  The skin depth gets smaller the higher the frequency.  It doesn't matter if the wire is 1 cm or even 1 metre in diameter, only this 0.1 mm thick "skin" has significant current flowing in it.

See http://en.wikipedia.org/wiki/Skin_depth but skip the eye-glazing math and check out the chart at the end.

As mentioned earlier, the shield of a coax cable isn't a wire with one surface - it is a cylinder with two surfaces, an inner and an outer.  This means that, provided its thickness is more than a few skin depths at the frequency of interest, you can have a current flowing on the inside of the shield and an entirely different current flowing on the outside!  This is because neither current penetrates far enough into the shield material to affect the other current.  It's just as if there were an insulator keeping them separated.

The point to take away from all this is that, counter-intuitive as it may be, you can have two distinct AC currents flowing in a coax shield, one on the inner surface of the shield and one on the outer.

~Jim Smith VE7FO