This question is perhaps the one that I am asked most often.
Many "legends" are circulating about; some say that a Balun is useless, and even produces losses; some people think that a Balun serves to eliminate the SWR; and finally, many say that a Balun (or UNUN) can be put together with a piece of wire and an unknown Toroid, found at the flea market; or even rolling up a stretch of the coaxial cable.
First, we clarify the definitions:
Next (image taken from Wikipedia) we see a dipole, with a balanced line connected to a load (that is our station).
The radio waves induce a current in the dipole, that by means of the balanced line arrives at the receiver.
Since the currents in the balanced line are equal and opposite, the line does not irradiate and does not receive, and only the currents induced in the dipole arrive at the receiver.
But in reality, things are not going so well ...
First of all no dipole it is never really balanced. To begin with, there will be objects, around and under the arms, which will unbalance it: houses, support poles, fences ...
Then the coaxial cable comes into play. The center conductor is connected to one of the arms; but the other arm is connected to the braid of the cable, which due to the "skin effect" is as if it were constituted by two conductors separated and independent: the inner face of the braid, and the outer face. In the figure (thanks, David Achilli IZ2UUF) see how the current that goes out from the transmitter divides: the current of the central conductor of the coaxial cable (C) goes all to the left (A), while the current that travels on the inner face of the braid (D) is divided into two parts: one goes on the right arm (B) and another part goes back to the transmitter traveling on the outside of the braid of the coaxial cable (E). This current comes back in the station (and in addition it's at high voltage, from which we get the shock if we touch the radio), will irreparably ruin the radiation lobe of our antenna, and ultimately does not do any good. The opposite also happens in reception, with a dramatic increase of interference and noise.
But here we have a simulation made with EZNEC+ of a G5RV antenna: a dipole (1 and 2) fed with a balanced line
In (5) we connected our coaxial cable (6); and since there is no Balun, on the outher face of the braid of the coaxial cable (6) we will have a good (bad!) common mode current that at the end falls in the middle of the station.
Besides, since the currents in the arms (1) and (2) are never, inevitably, well-balanced, also the currents in the "balanced line" are different; then the line is NOT any more more balanced, for which radiates (and should not) and receives interference and noises as well.
Now, in our simulation, we have added a good Balun in the position (5).
Miracle! The currents in the arms of the dipole are now equal to each other, as well as the currents in the balanced line feed (3).
And on the shield of the coaxial cable (6) the current is NOTHING!
So in summary: If our balanced antenna directly connects to a coaxial cable, it will surely become unbalanced and we will have common-mode currents.
But is it not enough to make a good Choke, perhaps with a score of turns made by wrapping the coaxial cable around a plastic tube (the so-called "Ugly Balun")?
Well, certainly we will block many of the common mode currents ... but the antenna will remain unbalanced, and the radiation pattern will be distort and perhaps ruined ...
Do not forget, then, that many Balun can simultaneously provide an impedance transformation; so if our antenna has an impedance other than 50 ohms (for example, a folded dipole, or an OCF dipole powered off center) at the same time we have a perfect impedance matching between the antenna and the cable.
It is not so simple. If we take an unsuitable toroid, or worse unknown (bought at the flea market) and we roll up, around it, a piece of electrician's
wire, we get something that LOOKS like a good Balun designed with care, and tested having at hand the appropriate instrumentation; and I assure you that an SWR
meter or even an economic VNA are NOT sufficient.
There are at least fifty different Toroid mixtures produced by at least as many producers; and the specifications that the manufacturers diffuse are better suited to a qualified Electronic designer than the average Ham! Our market is not their primary target!
Moreover, each manufacturer uses different names, and NO ONE writes the name on its products; so the toroid bought at the market remains a complete stranger ... AND NEVER trust the color! There are around toroids that Chinese simply repainted ...
This, for example, is a great Balun 4113 of Balundesigns. The toroid is a special compound, made to order, and does not match any of the most
common ones (believe me, I've checked ...).
The winding is made in Polyamideimide wire, insulated with PTFE tube of non-standard thickness.
And why all this?
Because in this way the Balun operates in "Transverse Electromagnetic Mode", for mutual induction between conductors which have exactly matched the impedance of the project; The core intervenes only on the current 's imbalance, and it is much more difficult to became saturated, even with high power and with high SWR.
Well, this one here beside should do the same thing ...
The toroid, smaller, was probably repainted; the wire (electrician's) is wrapped in a "go ahead, which is fine" way; I have doubts even on the circuit ...
And surely the impedance of the conductors is totally random ...
At the end it is cheap; who buys and installs it will not see big improvements ...and will speak ill of Baluns in general ...
Incidentally: if it is sealed, and to open it you have to cut it, maybe it has something to hide!
Then I made the rebuttal.
I built a Balun 4: 1 Guanella (current) with the circuit of Jerry Sewick W2FMI, a toroid T104-2 by Fair-Rite, PTFE insulated wire, tightened with Fiberglass tape.
I made a measurement circuit, as you have to adjust the output impedance (200 ohms) with the 50 Ohm of VNA input, and tried to measure the parameters with a VNA-3E model DG8SAQ made by SDR-Kits ,duly calibrated to Standards Amphenol, and the software VNWA.
Premise ... the matching network (required) adds a constant to the measure, so those are of relative interest (as it were, for a valid comparison), not absolute.
Below are the results ...
Then, going in order:
The SWR is fairly linear, going from 2.72 on 80m, then 2.18 on 40m, and then is quite linear between 1.95 to 30 meters, 1.75 to 20 meters, and down to 1.44 on 10 meters.
The impedance ... it could be better. From 25,55 Ohm, 80 meters it rises gradually, on 20 meters is 41.53 Ohm and arrives at 49.44 Ohm on 10 meters.
The attenuation (here plays the matching network ...) is not linear, is from -3.84 to -7.30 db db.
Conclusion: This Balun, even if carefully constructed, it does not have sufficient linearity to be able to be used on all HF.
Exact same measurement configuration with a balun Balundesigns 4132.
The SWR is a perfect straight line, from 1.07 to 1.19 on the whole band!
The impedance is also a straight line: from 52.32 up to 50.82 Ohm Ohm across the bands!
Attenuation is also perfectly linear, as -9,71db to -9.79 db!
Two words on attenuation: that is NOT the attenuation that the Balun will show on your line, but it is the attenuation of the measuring network for impedance matching!
Conclusion: the linearity is outstanding, the ROS is exceptional even when compared to what we saw above.
A perfect choice of components most suitable, and an impeccable construction.
To clarify the latter point, I connected two Balun 4132 in series and inverted to each other so as to eliminate the matching
network (I obtained directly the output at 50 ohms).
I had to amplify the scale, otherwise we would have had all the straight lines overlapped in the middle and we will not understand anything.
SWR caused by the TWO balun ranges from 1.05 to 1.10!
And the attenuation? ranging from -0.00 to -0.10 db db ... virtually nothing!
The impedance ranges from 50,62 to 52,21 ohms!
Conclusion: It 'so linear, and has an insertion loss so low that I had to expand the scales just to see something meaningful on the graph!