|
|
|
Related articles at Balun Test contains model of "perfect" dipole currents. Sleeve Balun shows how a sleeve adds impedance, useful for VHF and higher baluns Receiving Common Mode Noise shows how lack of a balun can contribute to system noise (it applies to transmitting antennas as well) Longwires, Verticals, and Baluns shows how unbalanced antennas can have similar problems Balun and Core selection for transformers and baluns Transmitting baluns on testing transmitting baluns Common Mode Noise IsolationCommon-mode currents can be detrimental to antenna system noise or directional performance. A quick look at systems using common-mode currents demonstrates how effective they are at causing radiation. If common mode currents can radiate effectively; they can also transfer unwanted signals and noise into our antennas when receiving. Many antennas actually function because of common mode currents. Two popular examples are the CFA and EH antennas. Both become significantly poorer radiators if common-mode currents on feedlines are eliminated. Another recent example appearing in Antennex's compact antenna articles was a thick stub "vertical" with no counterpoise. You find an example on the baluns and verticals page of how poor some large grounds can be. A receiving example of an antenna that works because of common-mode excitation is the "Snake" antenna. This system accidentally (or intentionally) induces common-mode on a cable shield in order to receive signals. The entire shield picks up signal, the Snake is simply a reverse-fed random wire lying on the ground. There are many examples where designers intentionally use common-mode currents. Examples are found in textbooks, such as the "Antenna Engineering Handbook" by Jasik on and around page 22-6. The antennas below, copied from Jasik's textbook, outline the
derivation of a skirt collinear antenna from a simple feedline with the open end
terminated by a "stinger". The center conductor termination in these drawings could easily be a ground rod (in the case of a Snake) or an antenna like a Beverage or loop. The termination does not have to be an "open circuit" 1/4 wl stinger that intentionally radiates! Looking at (a), we find by hanging any low impedance on the end of a coaxial cable the shield is excited by common-mode current. The electrical equivalent is just as if the transmitter or receiver (generator symbol in the drawings) is located at the end of the shield. This causes the outside of the shield to act like a longwire antenna. Unless the coaxial shield connects to a zero resistance ground, current with flow on the shield. Looking at (c), we find even multiple sleeves appearing as parallel tuned high-impedance circuits do not fully decouple a shield! It takes grounding and series impedance to do a good job. Common Mode Currents and Receiving AntennasAnalyzing our antennas, we often forget grounds are not perfect. We make assumptions that four radials, or worse yet two radials, form a perfect groundplane. Even a groundplane antenna many wavelengths from earth with four radials has considerable common-mode currents on the feedline. Consider the following model of a "perfect" Ten Meter groundplane using four perfectly horizontal 1/4 wl radials spaced every 90-degrees with a 1/4 wl feedline hanging vertically and attached to the radials. The main element current was set at 100. 
EZNEC ver. 3.0 A glance at radial current shows the bulk of ampere-feet (ampere-feet, or current over spatial distance, determines E-M radiation levels) is on the feedline shield, not the antenna! Radiation from the feedline would be severe, yet most amateur antenna designers claim with only four radials, or worse yet two radials, no balun is needed! The claim that four radials makes a "perfect ground" is false. Why do we depend on a simple ground rod with 50 or more ohms RF resistance to clamp a coaxial cable shield to ground? Receiving SystemsAdmittedly the above antenna is a worse-case example of feedline length and grounding, but even the best cases could cause problems. A best-case system might be "nearly perfect" when transmitting, but it could be a disaster receiving when even minor amounts of conducted noise are present on the station ground. Noise paths exist through station wiring. Only the shunting impedance of ground connections and series impedance of the feedline shield prevent excessive unwanted noise ingress at the antenna feedpoint. Very small levels of conducted unwanted noise often go unnoticed in large high-level transmitting antennas. Noise ingress would not be an issue if local noise levels on power lines are very low, especially if the antenna has substantial common-mode feedline noise rejection. If a feedline is very long and lies directly on or is buried in the earth, ground losses can attenuate conducted noise or unwanted common-mode signals. Unfortunately, we almost never know if the feedline shield is contributing noise, because it is nearly impossible to measure the common-mode noise contribution of the feedline! Measuring Common-Mode NoiseWe sometime hear suggestion that we test a system for noise ingress by disconnecting and replacing the antenna with a dummy load. This idea actually has no theoretical foundation at all. The dummy load would change system common mode greatly. The only real test would come from a dummy load with the same connections and impedances (both differential and common mode) as the actual antenna. In other words the test load has to be the actual antenna to keep feedline common mode ingress the same! Obviously this is a useless test! The best approach is to use preventative measures in initial system design and installation. Quite often the cost of being safe is less than a few percent of the initial system expense. Analyzing SystemsThis circuit is simplification of typical common-mode paths in Beverage, EWE, and other similar antenna systems:
R_Source and V1 represent the source creating voltage across R_Station_Gnd, the station's ground impedance. Feedline_R is the equivalent series-impedance of the feedline shield. Current through the feedline shield path develops a voltage across R_Ant_gnd, which represents the earth connection ground impedance at the antenna. V2 is a voltage source representing desired signals, while R_ant is an impedance representing the sum of the coaxial differential input impedance presented to the antenna (from the desired signal path into the coax) and the actual antenna impedance. Using the circuit below, we can find the attenuation. Assume: R_source is 90-ohms R_station_ gnd is 10 ohms R3 (the coax shield) is 500 ohms R5 is the combined series resistance of antenna impedance and impedance presented by the feedline matching system, is 1000 ohms In a typical system where a single six-foot or deeper rod (the earth's skin depth prevents deeper ground rods from decreasing resistance substantially) is driven into typical soil, R_ANT_GND will typically be between 40 and 120 ohms, assume 100 ohms. We have the following results:
Using the model above, only ~1 volt of common-mode voltage across the station ground results in .152 volts driving the feedline exactly as a signal from the antenna would. Path attenuation from station ground to the feedline's differential input at the antenna is 20log 151.5/985 or 16.26dB. Changing the ground resistance to 10-ohms results in: 19.1/982.7 or ~34dB attenuation of common-mode noise. Increasing R3 by adding beads has a similar effect. If R3 is effectively made ten-times larger, attenuation is in the 30dB range. Obviously it takes a combination of reducing ground resistance and/or adding series impedance on the cable shield to significantly isolate any low-noise receiving antenna from conducted ground noise over the feedline's shield. We sometimes observe much less noise on transmitting verticals after installing a large effective ground system. Decreasing ground impedance at the antenna reduces common-mode excitation of the antenna feedpoint and reduces noise ingress, although adding a feedline choke would sometimes help. SolutionsA typical isolation scheme would be to use an isolated primary and secondary in the matching transformer, and ground the feedline shield some distance away from the antenna's signal ground. This will introduce several thousand ohms of reactance in the common-mode signal path, as well as provide another path to earth for common-mode noise. Another method, in cases where the feedline can not be isolated through a floating primary in a matching transformer, is the use of multiple independent ground rods with a series of choke baluns between each. This forms a multi-section pi attenuator, making even modest choke impedances effective. As an additional benefit, lightning paths are disrupted by this method. SummaryNoise contribution can vary with time. A receiving antenna's ground connection resistance varies with soil moisture, and sources of noise come and go. As noise levels and grounding changes noise contribution as a ratio to antenna noise will change. The fact we can not readily measure noise contribution by substituting dummy loads further complicates the issue. Real systems are vastly more complex than the simple analysis above. Since we can't easily measure noise contribution, we shouldn't take chances. It makes no sense to gamble that unwanted signals (from wrong directions) or noise are so low that they will never contribute to noise in a special antenna installed to reduce noise and interference. While isolating feedline common mode effects from the antenna and antenna's ground may not reduce noise, isolation can generally be achieved at virtually zero time and material cost. With the low cost of prevention in mind, it is shortsighted at best and foolish at worse to not isolate a feedline shield from any low-noise antenna's signal ground path. Follow these rules:
|
|
