Cleaning Up Problems and Impedance Required
jump to Common_Mode_Choke
see balun testing
Jump down this page to dipole models
To understand how a balun operates and why a balun is
needed, we first must understand balance. Most often balance is described only
by current in each conductor of a transmission line. That
can mislead or confuse us. There have been many articles incorrectly describing
measurement of balance. Test equipment improperly measuring balance has been
manufactured and sold.
Perfectly balanced lines and perfectly unbalanced lines all have equal and opposite currents entering and leaving the conductors at each end, as can lines that are out of balance. All properly operating two-conductor transmission lines, coaxial or parallel wire, carry equal and exactly opposite phase currents in the two conductors. Non-radiating coaxial lines and parallel lines (twin lead or ladder line) both have exactly equal and opposite flowing currents entering or leaving each conductor at any given end of the transmission line.
The only thing determining the gender of balance is the electric field to space surround the line, or the conductor voltages to "ground" or space around the line.
We can establish these rules for properly operating transmission lines:
all lines have equal and opposite currents in each conductor
perfectly unbalanced lines have zero electrical field (voltage) outside the line to space or "ground"
perfectly balanced lines have equal and opposite electric fields to space around the line, or to "ground"
See coaxial lines at this link
Currents flowing without equal, very closely space returning currents, are called common mode currents. Common mode currents cause coupling and radiation. In a dipole antenna, or any antenna for that matter, common mode currents in the antenna element(s) are entirely responsible for radiation. Inside the Ham shack or along the antenna feed line, common mode currents are responsible for unwanted noise ingress, RFI, RF burns, and a host of other maladies. Common mode currents effectively bring the radiating part of the antenna system down along the feed line or the antenna's metallic supporting structure. Common mode currents can extend all the way to the desk and station equipment, and even out through power line connections.
We generally do not want our feedlines to act like antennas, and we certainly don't want to listen to the electrical noises in our house wiring. The best idea is generally to keep common mode currents out of the feedlines and control cables. Common mode currents should thus be "dealt with" at the source of those currents, or as close to that as is most effective. This is where the the problem actually is.
Antenna design, layout, and wiring first. Suppression last!
Some sites claim impedances in the thousands of ohms are necessary to reasonably or properly eliminate common mode. Such claims are truly absurd. Even in critical systems, the system has a major design or layout problem if it requires more than a few hundred ohms common mode suppression impedance.
In general, a properly constructed layout and proper
antenna with good connections and cabling will not benefit from suppression
inside the building. Even a severely compromised layout, such as with the
antenna close to RF sensitive devices and/or equipment, the equipment should still
follow good basic layout and wiring principles.
There is a side benefit to a good layout. The very things reducing common mode issues also reduce lighting damage susceptibility. Lightning protection and common mode RF immunity go hand-in-hand together.
Once the station or equipment is properly installed, even relatively small amounts of additional common mode impedance will offer a significant reduction in common mode. A proper system will not need more than a few dozen ohms or hundred ohms additional isolation. A poor layout might not be improved with nearly infinite isolating impedance. One thing is universally sure, if a system needs more than a few hundred ohms CM suppression the system has a major layout or wiring problem. It is best to fix that problem before adding chokes, and worrying about obtaining choke impedances impossible to achieve and maintain in a real-world system is pointless.
Placing things in perspective, on 40 meters just 9 pF of stray capacitance is 2500 ohms. Do we really honestly think, with even a few feet of wire or a system device the size of a small mobile radio cabinet, we have could have less than 25-100 pF of stray coupling to other things? With that in mind, would something like 1000 ohms even make any sense at all? Don't waste time chasing rainbows. Spend your time and money actually doing something useful. Anyone can make something "work" or produce a large number in an uncluttered controlled test bench. The real world and how things work in the real world is what matters. If you need more than several hundred ohms suppression, something is seriously wrong with the equipment, antenna, or feed system design or layout.
All proper installation should have a common point where cables enter for all desk functions, including power, Internet, control, and RF cables.
Shack and power mains grounds must be bonded with fairly low impedance, the
lower the bonding impedance better.
Multiple isolated grounds create dangerous ground path loops, inviting damage or RFI problems.
Cables to the desk should be bundled and/or routed in parallel. Bundling or close parallel routing reduces "open area" of the loop formed by multiple conductors.
A common mode choke alters the common mode impedance of the system. The
isolation added to any system, just like the impedance required, is not predictable.
Isolation can only be measured with a great
deal of work. We can say one thing with some certainty, large values of common
mode impedance are rarely necessary.
Attenuation with the addition of a choke is a function of the common mode source impedance, the common mode line impedance on both sides of the choke, and the impedance of the line's ground at the equipment side. This is not a simple system. Common mode source and termination impedances are virtually never anywhere around the 50-ohms a typical S21 or S12 measurements are made with. Because of this,. measured or advertised attenuations are meaningless numbers and any blanket prediction of system choke impedance needs are pretty much useless. All that can be said is it is best to use a dissipative impedance (Q<1) rather than reactive (Q>>1) in RFI suppressor systems.
In "A" below, a transmitter drives the coaxial feedline in differential (push-pull). The center conductor in this example A's 50-watt transmitter, assuming matched lines, will be at 50 volts. The shield is ideally at 0 volts to earth, being a commonly grounded point with earth. The current are equal and opposite in properly operating balanced and unbalanced lines, but in normal operation the shield being at zero volts to "ground" (or to the chassis) defines the system as being a perfectly unbalanced system.
If the system were perfectly balanced transmission line currents would still be equal and opposite. The transmitter voltages from each feeder conductor to chassis or ground, however, would be equal and opposite in a balanced line system. The line would also have two parallel conductors equally coupled to earth or ground through space.
Coaxial line behavior, from skin effect and mutual coupling from the
shield's inner wall to the center conductor, pulls all of the transmitter's
differential current to the area inside shield wall. The voltage appearing at the antenna
drives the common mode impedances of both halves of a "balanced" antenna,
pushing one side of the antenna against the other antenna half.
A halfwave dipole, depending on construction, height, and surroundings, has essentially equal resistances. We'll use the values shown as a discussion point. If antenna current was one ampere and no common mode flowed on the coax, each half of the antenna would have 25 volts to earth at the feedpoint. The voltage developed across Rant2 causes the common mode problems, because that voltage also excites the shield on the shield outside.
Allowing that voltage to "float", in this example to 25V, is the same as preventing unwanted outside shield current from developing.
Rant2 was near zero ohms, it would be a unbalanced antenna with a very good ground. The better the ground, the lower "Rant2" below, the less voltage is in series with the resistance driving the shield (B).
This system is complicated by antenna characteristics and other things. The
antenna system common mode impedance and grounding determines the source
impedance and voltage driving the shield with common mode.
The system is further complicated in that cable1's length and surge impedance
modifies the voltage and impedance at the common mode choke. In its most simplified
form, the common mode choke system looks like a pi-attenuator. The common mode
source impedance at the antenna, as modified by the cable common mode surge impedance
shunting leakage and ground paths modifies voltage and impedance driving the
CMC (common mode choke).
Similarly, common mode impedance into the station equipment, likely a pretty low impedance, determines the station side load on the pi attenuator formed with CMC.
The attenuation and required CMC impedance quite obviously can be all over the place. While none of us can predict the real attenuation provided by a given CMC or predict how much CMC impedance is "enough" and how much is a waste of effort, we can make a pretty reasonable generalization. It is very safe to say any system requiring CMC impedances beyond low hundred's of ohms needs layout, design, or re-wiring help more than some extraordinary impedance.
My station, where lowest possible noise floor is paramount, has never required more than a 50-100 ohms of CMC to mitigate all traces of problems. I've actually found grounding and good connections to be far more productive and reliable than dependence on a higher CMC impedances. Once I get into the dozens or hundreds of ohms without complete mitigation of noise, I know I have a major problem with shield integrity. I look for the real problem.
To understand how a balun operates and why a balun is needed, we must understand balance. We tend to think of balance only in the amount of current in each conductor of a transmission line, but that thinking can mislead or confuse us. Perfectly balanced lines and perfectly unbalanced lines alike have equal and opposite currents entering and leaving the conductors at each end!
Coaxial cables with shields more than several skin depths thick always carry equal and opposite flowing currents on the inside of their shields and their center conductors. Current direction and current ratio between the center conductor and inside of the shield in a non-radiating coaxial line is no different than currents in each conductor of a perfectly balanced ladder line. In both unbalanced coaxial lines and balanced lines, the two conductors making up the line carry equal and opposite flowing currents.
When currents flow without close-by opposing currents, we call the unopposed portion of current common mode current. Common mode currents promote or encourage external coupling and radiation. In a dipole antenna, or any antenna for that matter, common mode currents in the antenna element are responsible for radiation. In the hamshack or along a feed line, common mode current is responsible for unwanted noise ingress, RFI, RF burns, and a host of other maladies. Common mode currents, in effect, bring the radiating system into the feed line or station equipment.
Common-mode currents, or currents flowing in the same direction, cannot exist inside a coaxial cable at any frequency where the shield is several skin depths thick. Shield skin depth serves to isolate the inside of the shield from the outer wall of the shield. Common mode (same direction) currents can only flow on the outside of the coaxial cable shield. Differential mode currents, or normal transmission line currents, flow on the inner surface of the shield wall. Currents entering and leaving the shield and center conductor at each end of a coaxial line must be equal and opposite or the cable will radiate. If a coaxial line is not radiating, currents in the shield and center conductor are exactly balanced and opposite flowing. Both types of transmission lines, balanced and unbalanced, will have equal and opposite currents entering and leaving each conductor when they have minimal radiation.
What then defines an unbalanced line, source, or load? The answer lies in the voltage or electrical potential between line conductors and the environment around the line. In the ideal balanced line, the electric potential of each conductor is equal and opposite in relationship to the environment surrounding the line including the chassis or cabinets of our equipment. In the ideal coaxial line, the outside of the shield has no electrical potential difference to the environment around the line, including the chassis or cabinets of our equipment. The shield of our coaxial cables, as we commonly accept and understand, is at ground potential. We say the shield is “grounded”.
With real-world antennas, the coaxial shield connection point almost never has zero electrical potential to the environment around the shield or points further along the cable’s length. Being a less-than-ideal zero-voltage termination, shields almost always have common mode current, even if a small percentage of differential (normal transmission line mode) current. For example, the four radials of a groundplane antenna, no matter how configured or tuned, are never truly at the same electrical potential as the environment around the antenna or shield potential further down the feed line. Experimenting with a groundplane antenna, we find the feedpoint is mostly but not perfectly unbalanced. The shield is not connected to an electrically zero point. Significant current can and often does excite the outside of the shield on a groundplane antenna, with outside shield current 20% or more of antenna base current under some feed line grounding and lengths! We consider the groundplane antenna “unbalanced” and it is definitely not balanced, but it is not perfectly unbalanced.
A coax fed dipole, or a vertical with a single radial, is much worse for voltage balance. Because both antenna halves have finite and nearly equal common mode impedances, both sides of the feedpoint want to have nearly equal voltages between themselves and the environment around the feedpoint. If we could magically make a perfect single point ground appear at the feedpoint, both legs of these antennas would have very similar voltages to that reference point. Of course we can’t make that perfect reference point appear, but the feed line brings a “ground” or reference connection to the feedpoint. Third path impedance (the unwanted common mode path impedance) varies with feed line length and routing, the environment the feed line routes through, and how that feed line is grounded. While it can affect SWR and the current flowing in that path be affected by SWR, high SWR does not cause and low SWR does not prevent unwanted common mode current or RF in the shack.
Most antennas are neither perfectly balanced nor perfectly unbalanced. Most antennas are in a nether-world someplace between perfectly balanced and perfectly unbalanced. This is why a current balun, a device that floats each balanced terminal to the voltage necessary to drive balanced currents into the load, is such a desirable type of transmission line to antenna interface.
Feedlines and Balance
Traditional two-conductor feedlines or transmission lines should have very little, if any, radiation. Typical coaxial lines shouldn't have noticeable unwanted signals or noises leaking into the cable, and there shouldn't be noticeable radiation out of cables. This is true at all radio frequencies from the AM broadcast band upward through UHF, and applies to any cable with a reasonably thick single or double shield. Even cheap 80% coverage shields are adequate throughout HF in all but the most critical applications.
small amount of
radiation from the line, or
leakage into the
line, exists. The leakage
amount should be
very low in properly installed
lines. Problems can occur with an unwanted EMI
source, or sensitive systems,
close to the
feed line. Problems with unwanted signal pickup or radiation can also occur with
very long open wire or unshielded balanced lines.
Ladder lines and unshielded balanced RF lines should be isolated from other objects. Ideally, the only air should be allowed within several conductor spacings of balanced or unshielded lines. Significant levels of electric and magnetic induction fields surround the line for distances of several transmission line conductor spacings. Line radiation (electromagnetic radiation is a different mechanism) also extends out in a line through the conductors. Unwanted radiation primarily occurs in directions aligned with the plane of the line conductors, nulling at right angles to that plane.
Even with perfect balance in a two-conductor unshielded line, some radiation occurs. The small spatial separation prevents perfect cancellation of far field radiation. The conductors carrying out-of-phase currents are not occupying the same physical space, causing a very small spatial phase delay. This means in two directions radiation fields are not precisely 180 degrees out-of-phase. The amount of phase error, and thus the level of radiation, is a function of conductor spacing in wavelengths and the direction from the line. This effect is minimized by twisting the feed line at small fractions of a wavelength.
Radiation from a perfectly terminated six-inch spaced 50-foot long two-wire transmission line on 80 meters.
Radiation of the same line at 30 MHz is 24 dB stronger. This is because the conductor spacing in wavelengths is wider.
In order to be balanced, a balanced transmission line must have both equal and opposite voltages at any particular point along the line as well as equal and exactly opposite currents at any particular point along the line. If the voltage is not equal and opposite, current cannot remain equal and opposite along the balanced transmission line. This will result in a very large increase in feed line radiation because the imperfection causes common mode currents.
In order to be unbalanced, an unbalanced transmission line must have equal and exactly opposite currents entering and leaving at every point along the line. The voltage gradient laterally along the outside of the transmission line has to be zero. If either the lateral voltage gradient is not zero, or currents entering the line are not equal and opposite on the shield and center, current will not remain zero on the outside of the shield. This will result in common-mode shield currents and feed line radiation.
To avoid feed line radiation every balanced to unbalanced transition has to be properly treated for level and phase of voltage and current.
To be properly balanced, the following must occur:
Voltages from 1 to A, and from 2 to A, must be equal and opposite
Currents into 1 and 2, at the source, must be equal and opposite
Voltages from 1 to C, and from 2 to C, at the load must be equal and opposite
Currents out of 1 and 2, at the load, must be equal and opposite
Voltages all along the line, at any point, to B must be equal and opposite
Common Mode Excitation
Common-mode current is current that is not opposed or counteracted by an equal and opposite phase current flowing at every point along the line in closely-spaced conductor or conductors, and the outside of the shield has current flowing in a coaxial line.
Any transmission line becomes at least partly, a radiating conductor if we make a poor balanced to unbalanced transition. This can be useful when we wish to use a feed line as an antenna or as a conventional conductor, but it can be detrimental to a system if we do not want radiation or reception by our feed lines. When we excite a cable as shown below, we have common mode current:
The common mode source is end-to-end on one or more conductors of twinlead
The common mode source is end-to-end on one or more conductors in a twisted pair of wires
The common mode source is end-to-end on the center, the shield, or both conductors in coaxial or shielded cables.
When we excite a transmission line as shown below we create common-mode current:
The shielded coaxial cable (top) and the parallel conductor cable (bottom) in this diagram radiates just like a single wire would do. Objects surrounding the line, like dielectrics or other conductors, couple to or interact with these lines when they are fed or excited this way. For example, adding a ferrite sleeve over the lines will add loss and/or make the lines behave differently. The impedance of the system will change, and if we are watching system SWR the SWR will change.
This is true even when currents are equal in the two conductors, and can even be true when currents are equal and opposite at one point in the system, so long as the line is excited this way.
The key to having a line behave like a transmission line is feeding it differentially (across the two conductors) at one end, having a load that maintains the differential excitation, and not applying a voltage or a potential difference across the length of one or both conductors.
This is a transmission line as we generally know it, and as dozens of reputable engineering textbooks define it:
This is differential mode, or TEM mode. This is the normally desired excitation mode when a two-wire line behaves like a normal non-radiating line that transfers energy from one point to another.
The above configuration shows a direct wire connection from source to load. It transfers voltage, current, or impedance directly along the conductors.
1/8 wave high dipole
A 1/8th wave long (35 feet in this case) coaxial feed line to the ground point on a dipole often does not need a balun! Here are feeder common mode currents for this case:
Maximum feeder common mode is only .079 amperes (out of a 1 ampere source current) with very good antenna current balance. This is without any balun!
The same dipole 1/4 wave high:
|shield at dipole||0.018|
Common mode currents are also low with a 1/4 wave feeder. In many or most cases of dipole height between 1/8th and just over 1/4 wavelength, a balun is NOT necessary provided the feed line drops straight down through open air to ground, and the feed line is grounded when it reaches the earth.
Dipole 1/2 wave feeder to ground
|shield at antenna||0.521|