Is "balanced" (as in baluns, etc.) a myth?
Displaying 1-7 of 7
| Sep 5th 2018, 04:32 | |
ag7ovTotal Topics: 0 Total Posts: 0 |
Books and articles say we can have balanced or unbalanced transmitter outputs, feed lines, and antennas. The idea seems to be that somehow balanced things are incompatible with unbalanced, and so some sort of transition circuitry (a balun) is required. I've been trying to make sense of this, but the facts don't seem to add up. I'm wondering whether maybe the whole idea is a bit flawed. It sort of looks to me like in all cases the only real consideration is isolation, and this doesn't really have anything to do with whether something is "balanced" or "unbalanced." Take transmitter outputs. On many modern transcievers, one side of the transmitter output is at the chassis-ground potential. This is said to make the output "unbalanced." But this seems to be totally irrelelvant. To give an analogy, in some audio amplifiers, one side of a speaker output is at the chassis ground potential. In others it isn't. We don't speak of "balanced" or "unbalanced" speaker outputs because it simply makes no difference. The same seems to be true of a transmitter's output. If you connect a 50-ohm resistor from the center pin of the UHF connector to the chassis, you will be able to key up without any RF on the chassis. Full transmit power is converted to heat by the resistor. Is the resistor "balanced" or "unbalanced"? How could it be one or the other? It's just a load. Likewise, if the resistor is at the end of a 50-ohm transmission line, all power not lost in the transmission line is dissipated by the resistor. Does it matter what kind of line? It could be coax, or (if the transmitter output impedance were higher to match) it could be window line. As long as the impedances match, it makes no difference what kind of line. Suppose there are impedance mismatches? Well, then the transmitter sees a mismatched complex impedance, and so it does not operate at full output efficiency. There may be higher voltages and/or currents in the output section. The transmitter may reduce power to protect its circuitry. All this happens without regard for whether the transmission line is coaxial or spaced wires. It would be exactly the same if neither terminal of the transmitter output were at chassis ground potential. It simply does not matter. What does matter is isolation. Suppose the transmitter feeds an antenna through a transmission line. Unless the arrangement is perfectly symmetrical in all ways, there will be common-mode currents on the transmission line. Simply put, the transmission line will end up functioning as both a transmission line and a part of the antenna system. In real-world systems, the simplifying assumptions of transmission-line theory break down, and we end up with unbalanced (and here the term has meaning) currents and/or voltage gradients on a transmission line, meaning they are not equal and opposite as required for normal transmission-line behavior. We can view this as a simple superposition of the normal transmission-line signal composed of equal and opposite voltage and current waves propagating along the line, and a common-mode signal--some other signal that is the same in both lines. With spaced-wire line, a common-mode signal shares the line with the differential transmitted signal by simple superposition. With coaxial line, the transmitted signal is kept inside the line, and a so-called common-mode signal exists as a separate entity on the outer surface of the shield, separated from the signal on the inner surface of the shield due to skin effects. The shield acts as two distinct conductors. Either way, it's this common-mode signal that we are concerned with when we employ a "balun." If one terminal of the transmitter output is at chassis ground potential, the common mode signal on the line can appear on the chassis (which is, after all, an extension of one of the line's conductors). This is the case whether coax or spaced-wire transmission line is used. Now, if the transmitter output did not have a terminal referenced to chassis ground, if both output terminals were well isolated from chassis ground, then it would not be possible for this common-mode signal to energize the chassis. The reason for this has nothing to do with "balancing." It has only to do with isolation. If we suppose that our transmitter has two output terminals, each of which has the same exact impedance with respect to the chassis ground, will this prevent common-mode currents from energizing the chassis? Of course not. There is no "common-mode rejection" going on here. That's a phenomenon in balanced signal lines that is completely irrelevant here. The only thing that matters here is that each side of the line should have a high impedance to the chassis ground. In other words, only isolation matters. Would it help if modern transmitters had isolated antenna terminals? Not all that much, really. Common-mode RF on lines in the shack is still a bad thing, from the standpoint of RF safety and preventing RFI. Common-mode RF on lines in the vincinity of the antenna can alter radiation patterns in undesirable ways and/or increase antenna-system losses. We can't really avoid dealing with the problem of isolating feed lines from antenna systems and providing high impedances to common-mode RF currents. This is the true problem. It has nothing to do with "balancing." (Besides, grounded coax shields are probably a good idea from the standpoint of lightning safety.) If I feed a ladder line with my transmitter's PL-259 output, do I need a "balun" to transition from the "unbalanced" output to the "balanced" line. No. What I do need is some sort of impedance matching network, because ladder line is the wrong impedance for most transcievers. And, well, it would be a good idea to provide isolation to keep common-mode currents on the ladder line out of the shack. A so-called "balanced antenna tuner" will provide both, but some isolation closer to the antenna may also be warranted, unless the whole arrangement is very symmetrical (which it is, sometimes). If I feed a half-wave dipole with a coax cable, do I need a "balun" to transition between the "balanced" antenna and the "unbalanced" coaxial line. No. Lots of people feed dipoles directly from coax. But because the outer shield of a coax acts as a sort of separate wire, unless it is placed very symmetrically, a choke is still a good idea in order to keep RF off that shield. The funny thing is that the problem is much worse for an off-center-fed or end-fed antenna. You're much more likely to get RF on the coax shield then. But this is a decidedly *unbalanced* antenna type. So what do you need when going from an unbalanced antenna to an unbalanced transmission line? Why, an "unun" of course! What's an "unun"? I've never seen one that wasn't just a "balun" with another name attached. The true issue is isolation. It has nothing whatsoever to do with balancing. I have a 2-meter Yagi with a built-in "balun" consisting of a half-wave transmission-line segment. One side of the antenna is fed from the end of the feed line, and the other side is fed from the same signal delayed by about 180 degrees. Neither side is fed by the coax shield. Here at last perhaps we seem to be seeing a "balun," in the sense that it's a circuit that creates a differential signal by a phase reversal, which is analogous to how "balanced" signals are created in other types of systems. But what does this accomplish? It is already the case that the shield and center conductors of the coax have equal and opposite (i.e., balanced) currents and voltage gradients. The phasing line just creates a separate equal/opposite voltage/current (which in fact will only really be equal and opposite at one specific frequency). The net result of all this is simply that the coax shield does not drive anything on the antenna. The coax shield simply terminates in an open circuit. And thus, it remains *isolated* from the antenna system from the standpoint of RF currents on the outside of the shield. This type of balun is still just an isolator. (And of course the shield can still pick up RF currents because the cable runs through the antenna's RF field. A choke isolator at the antenna would be another way to accomplish exactly the same thing. Would there be any fundamental difference in antenna-system behavior if a simple choke balun had been used instead? No--thus further proving that the only point of the balun is isolation. Now, I write all this in the hope that some RF guru will come on here and set me straight, and thereafter, I will have learned the error in my thinking. If the "balanced" versus "unbalanced" distinction is actually real, and I'm missing something important, I have yet to find a good description in the books and articles I've read (which include the ARRL Antenna Book and the ARRL Guide to Antenna Tuners, among others). After a lot of poring over text in confusion and trying to figure out what this "balanced/unbalanced" distinction is, I could only finally conclude that it's a myth. But I'm still new to radio (though I have an electronics background), and I could be missing some big, important concept. If so, I welcome correction from a knowledgable source. But if not, I would also welcome someone knowledgeable's responding with, "yes, this concept is a myth." If it's a myth, we ought to start debunking it and thinking in more accurate terms, shouldn't we? Thanks and 73, Dan AG7OV |
| Sep 5th 2018, 08:55 | |
W1VTSuper Moderator Total Topics: 0 Total Posts: 0 |
Yes, this distinction is actually quite useful, based on actual amateur practice. For instance, I chose to wind an unun for my 80M vertical on an iron powder core, which can't practically be re-designed for a high common mode rejection. But, unlike ferrite cores, iron power cores handle a lot of power, so I don't have to worry about it burning up while running an amplifier. Speaking of amplifiers, hams used to build amplifiers with separate high voltage power supplies. Which meant that it was best to bond everything together securely for electrical safety. Safety first! An effective way of getting rid of common mode currents with verticals is to use an extensive radial system. I think it is quite useful to describe an off center fed dipole as on unbalanced antenna. With a conventional dipole, you can eliminate shield currents by using a balun and properly routing the transmission line away from the antenna. But, with an OCFD, even a perfect balun won't eliminate shield currents caused by the asymmetry between the antenna and the way the feedline is routed. This can be easily shown with computer modeling, which allows you to insert a perfect balun between antenna and the feedline. I've had the opportunity to get plenty of feedback on what type of articles people like by publishing a hundred of them. Back around 1996 I wrote some rather math intensive articles for QEX, and found that EEs in the industry preferred that I do construction projects instead. Now that I have other hobbies, such as growing two hundred rose bushes, and cooking for myself, I only have time to do the occasional article in my free time. Zak Lau W1VT ARRL Senior Lab Engineer |
| Sep 5th 2018, 13:22 | |
ag7ovTotal Topics: 0 Total Posts: 0 |
Hi Zak, Thank you for replying. I don't doubt (and didn't even before your response) that the distinction is made in amateur practice, and that the end result is something useful. But something useful can come from a flawed concept. I have an extensive background in audio engineering, and there are a ton of silly ideas about "ground loops" in audio systems, which are totally contrary to what's physically happening with noise voltages and currents. But the end result is that people employ certain techniques that work in practical terms. The fact that someone arrives at a correct solution to a problem doesn't prove that the concepts that led them to that solution were correct. It may be useful to have concepts that lead to solutions, but it will be more useful in the long run to have *correct* concepts that lead to solutions. Or at very least we can say, "The situation is mathematically complicated, but in practical terms, here's what you need." May I ask you a question? If your unun on your 80-meter vertical doesn't provide isolation, what purpose does it serve? Better still, can you describe to me the electrical difference between a balun and an unun? I've actually tried modeling the OCFD and transmission line with shield. EZNEC shows shield currents much worse than with a center-fed dipole (depending on how far off center). But a "perfect balun" (which would electrically be an open circuit at RF) makes the shield basically into just another wire in the antenna field. Yes, this will pick up RF like any wire. But it can be isolated at the other end on its way into the shack, too. In my modeling experiments, a reactive impedance of about 6k ohms on the shield is sufficient to reduce currents on it to very low, provided it's oriented orthogonally to the antenna. In the veritcal with an extensive radial system, without an isolator ("balun"), the coax shield is just one more radial. Assuming (as is probably reasonable) that the currents on that side are well distributed among the radials, shield current should go down roughly in proportion to the number of radials. If you picked the entire antenna system up and suspended it on its side in free space, this should still be the case, and for the same reason that people employ counterpoise systems in the shack to disperse RF currents. (But ground losses are probably also a factor in the radial case--if the radials are on or in the ground.) An isolator ("balun") still seems to be a good idea here. I don't mean any of this to come across as disrespectful. Please don't take it that way. It's just that if there is some point to a "balun" other than isolation, or if there is some electrical difference between a "balun" and an "unun" besides the name, somebody ought to be able to describe it plainly. Best regards, Dan AG7OV |
| Sep 5th 2018, 13:46 | |
W1VTSuper Moderator Total Topics: 0 Total Posts: 0 |
I use ununs for impedance matching. The 80M vertical has a step up transformer to provide a 200 kHz SWR bandwidth (below 2:1). It worked well enough to get 80M DXCC with a 100 watt solid state rig without a transmatch or ATU in a couple months. A 4:1 voltage balun can often be wired to provide a DC short to ground. This can be very useful for getting rid of static charge that would otherwise build up on an antenna. But, many 1:1 current baluns don't provide this useful feature. Zak Lau W1VT ARRL Senior Lab Engineer |
| Sep 5th 2018, 15:51 | |
ag7ovTotal Topics: 0 Total Posts: 0 |
Thanks Zak. I sincerely appreciate the conversation, and I am learning from your responses. My thoughts of the moment: Impedance matching is impedance matching. Isolation is isolation. A circuit can be designed to provide either or both. Neither really has anything to do with "balancing" or "unbalancing." Do we call something an "unun" if it provides only impedance matching but not isolation? But I've seen advertised 1:1 "ununs" that are simply isolators. The solid-state transciever output provides a DC path to ground on one side, and it seems to be common practice to bond coax shields together and tie them to earth ground for lightning safety. A 1:1 "choke balun" only presents a high shield impedance at RF. It should basically be a wire at DC, so one side of the antenna should remain grounded at DC. Given a low-DC-impedance output from the radio, the same should hopefully be true for the side fed by the coax center. Let me step back to my original point, though. Coming from the audio world, I understand the balanced/unbalanced distinction for signaling. The point of balanced signaling is rejecting line noise and common-mode power-supply noise using differential input topologies, etc. In that world, there is a clear distinction between balanced and unbalanced signaling at outputs, inputs, and along signal cables. In the world of ham radio antenna systems, there seems to be a desire to draw the same kind of distinction. But it seems a false distinction, made by analogy, but not holding up to scrutiny. I still have yet to hear a description of the balanced/unbalanced transition that doesn't just amount to some combination of isolation and impedance matching. Now, both of those are important, but neither has anything to do with balancing. Understanding what's really going on would seem to help in terms of understanding what is needed in a particular antenna setup. By the way, it's really interesting that you achieved a wide SWR bandwidth on 80 meters with an impedance step-up. I imagine the principle is that large changes in the higher impedance are translated to small changes in the lower impedance? That's an interesting idea that I hadn't heard before. I'll have to look into that on 80-meters. Thanks for the discussion! Dan AG7OV |
| Sep 10th 2018, 15:52 | |
W2WDXTotal Topics: 0 Total Posts: 0 |
I am a retired concert sound engineer, so I know a thing or two about audio. That being said, an amplifier to speaker connection is unbalanced. The input of that amplifier can be either balanced or unbalanced, however in the industry we tend to used balanced for low-level signals, principally to maintain higher s/n ratios. The levels on loudspeaker outputs are so high-level this is of little concern since the signal is significantly way above noise. Additionally, a loudspeaker motor device is an unbalanced device. Balanced, unbalanced. This is basic electronic theory. Whether audio frequencies or RF, when carried in electronic circuits all are electromagnetic waves with the same fundamental characteristics. They are all AC in nature. A balanced signal is one where there is positive and negative, referenced to ground. The positive and negative are usually at equal and opposite amplitude referenced to ground. An unbalanced signal is one where there is positive and ground. ALL signal amplitude is referenced either above or below ground, but never both. So as an example, a balanced 2V AC signal could be 1V positive, 1V negative. On a 2V unbalanced signal it would be 2V positive only. On both the signal is 2V AC, however how the signal is referenced to ground is different. This is a generalization, but is the basic concept. Balanced and unbalanced differ mainly in performance. In general a balanced signal is less prone to noise, and also exhibits total rejection of common-mode currents. This is due in theory to the equal and opposite poles which create cancellation of anything not part of the signal. Unbalanced signals tend to be more prone to noise and CMC. These differences by the way are the same regardless of frequency. The idea that somehow audio differs from RF is an incorrect assumption. It is when there is no proper consideration between balanced and unbalanced systems where things seem odd in ham radio. For instance, many hams use unbalanced tuners and open-wire feeders on balanced doublets, which creates problems in terms of actual balance (equal and opposite) on the feed-line regardless of whether or not an impedance match at the transmitter can be acheived; many times RF in shack is the result. Furthermore, the thinking that somehow impedance, SWR and resonance are somehow related to balanced/unbalanced systems is also incorrect. Those parameters are a matter of tuning, regardless of the system employed. However, how balanced/unbalanced systems REACT to SWR and impedance do differ, primarily due to the effects of CMC, reflection and impedance and how each system deals with each. As a general rule an antenna with a wide SWR curve is most often not always a good performer. It also does not say the resonance of the antenna is wide bandwidth. That's a misnomer. What it does say is the transmitter will pass maximum current over that range of frequencies, but what the total antenna system (tuner, feed-line, radiator) can do with that power is a different matter. Much of it may be lost, literally from losses outside of actual resonance. Think of it this way, a dummy load has a very wide bandwidth where an antenna with a high ERP (not created by directional gain) is usually very narrow. Again this is a huge over-simplification, but essentially it is factual ... a guideline. For instance, a resonant dipole which is brought closer to the ground (with all of the associated ground losses) has a broader curve as far as SWR. This is directly correlated to those losses, not because the resonance point of the antenna is broader. Also SWR is not an indicator of resonance or ERP. It's merely a measurement of forward & reflected power specifically at the point on the feed-line the measurement is taken. As far as impedance ... there is no such thing as a "DC impedance", impedance requires some frequency, meaning AC. Impedance is non-sequitur at DC. There is no "DC impedance" in an antenna, which is a tuned AC circuit. A choke balun does not create a ground on one side of a dipole. That is an incorrect assumption. A dipole is balanced with equal and opposite voltage poles, not positive and ground. Coaxial is not balanced. The shield is referenced to ground, and the center conductor has the full AC waveform above ground. The balun takes the balanced antenna and unbalanced coax, and coverts the signal to be appropriate ground reference for each. The impedance choking on the shield is simply an additional effect due to the nature of inductors and choking impedance, and unrelated to the bal-un conversion. Further more an UnUn is for feeding an unbalanced/unbalanced systems, like coax to a vertical or end-fed. Its purpose is solely to choke feedline CMC and transform impedance. Like 125ohm to 50ohm or some such. Do not confuse impedance matching, SWR and choking impedance with converting balanced to unbalanced. They are different parameters handled by a common device. For instance, on my balanced open-wire feeders I use two 340uh inductors (one on each pole) wired to earth ground. DC is shunted to ground on each leg, but with RF above about 1.2MHz the inductor is invisible, due to the high impedance presented. Also two capacitors are in series (one on each leg) which pass RF, but block DC. This effectively creates not only a bleed to ground for DC static but also blocks DC impulses. Of course, the devices are rated for the power levels involved. I use a balanced tuner and no baluns, since bal/unbal is not needed. Impedance matching is effected by the tuner alone, and the fully balanced system exhibits no feedline radiation since there was careful consideration of maintaining the balance (equal and opposite) of each pole of the system. Noise is very low on receive; due to a combination of very low-loss overall, CMC rejection, and very low-impedance (not just low-resistance) earth grounding. I care very little for feed-line SWR since it doesn't matter as long as I pass maximum current when transmitting when using the tuner, regardless of the SWR on the feed-line itself. A lot of the confusion with antennas is usually the act of propagation; in the case of antennas converting the energy from one state to another. In this case, electrons to photons. Loudspeakers are similar devices, they convert electrons into electromotive force which propagates as compression waves in air. Whether loudspeakers or antennas, impedance matching is critical in order to have current at maximum; regardless of whether the system devices are configured as balanced or unbalanced devices. John, W2WDX |
| Sep 10th 2018, 18:36 | |
ag7ovTotal Topics: 0 Total Posts: 0 |
Hi John, Thanks for responding. The thing is, the balanced/unbalanced distinction requires a ground reference, which I think you'll see if you look at your own description of it. The distinction is meaningless without a ground reference. In fact, on a signal cable, there is no difference between a balanced and unbalanced signal without a ground reference. Otherwise, we can't say, "This conductor has thus and such a voltage, while that other conductor has thus and such a voltage." A voltage can only be measured across two points. Or, "This conductor comes from thus and such a source impedance, while that other conductor comes from thus and such a source impedance." Impedance to where? The impedance of either conductor with respect to the other must be the same. So when we speak of a single conductor having a voltage, that's shorthand for saying there is a voltage between that conductor and some understood reference point (which is often called "ground," but just as often has little to do with the physical ground). Likewise, when we talk of balanced or unbalanced source or sink impedances, in each case we mean with reference to an external "ground" point. The signal between a power amplifier output and a speaker is not spoken of as "balanced" or "unbalanced" (as I'm sure you know, since you have a background in audio). The distinction has no meaning, because the speaker does not have any kind of a "ground" reference. The amplifier delivers power to the speaker via a voltage across and current through two conductors. By Kirchoff's law, the current must be "equal and opposite" in both conductors. The voltage is measured from one conductor to the other. There is no third point of reference. Hence there can be no balanced/unbalanced distinction. Now, at the amplifer, one of the output terminals could be at the same potential as the power-supply ground, or not. It doesn't matter from the perspective of the signal or from the perspective of the speaker. Either way, the speaker sees one voltage and one current. In practical reality, sometimes one terminal at a power amp is at "ground" potential and somtimes it isn't. It makes no difference, and it's a distinction we simply do not draw. There are no "balanced" versus "unbalanced" speaker inputs. (I mean to the speaker itself, not to an integrated amplifier.) There are no "balanced" versus "unbalanced" speaker cables. If someone were to ask whether a particular power amplifier put out a "balanced" output, generally that person would be laughed at in the industry (again, as I'm sure you know). It's a nonexistent distinction. (Well, it can become important if you want to bridge two amplifiers together, but I don't think that has bearing on this discussion.) RF and audio are very different in a lot of ways. For example, at the lengths we normally use, audio cables do not practically function as transmission lines. We don't care about their characteristic impedance. We don't worry about impedance matching line-level signals because reflections on the cable are not an issue. There is no skin effect to speak of, so the cable shield doesn't act as two separate conductors. Audio signals as a general rule don't radiate electromagnetic waves when running through the wires of normal audio systems (or not enough to concern ourselves with). Yes, both are electrical signals, but there are many considerations that only become important at higher frequencies, where conductors are a significant fraction of a wavelength. Now, I submit that in the case of an antenna system, what we have is like a speaker. What I mean is that there is no external reference point against which to meaningfully measure the individual voltages on the two conductors of the transmission line. We can measure them against the power-supply's DC "ground" potential in the radio, but that is a reference that has no meaning at all from the standpoint of the antenna. From the standpoint of the antenna (as with a speaker), there is simply a voltage across the feed point lines, and there are equal and opposite currents flowing in the feed point lines, and (unlike with an audio signal) there are voltage gradients along the lines, meaning it's possible to measure a significant voltage from one point of a conductor to another point on the same conductor. In any transmission line, the currents and the voltage gradients are equal and opposite so that the E and H fields cancel each other, which is what keeps the line from radiating. At the antenna end of the line, we just have two equal and opposite waves on the two conductors. It is unimportant whether one of the lines is at the power supply "ground" potential in the radio. The result would be exactly the same whether was or it wasn't. (In actual fact, it typically won't be, because of those voltage gradients I mentioned, even when the line is fed from a radio that references its output to ground.) If we feed the coax through a "balun" or an "unun," the result is the same as if we'd fed it from the radio--equal and opposite currents and voltage gradients as referenced to each other. There is no meaningful external point of reference. Even if we wanted to use the radio's "ground" as a point of reference, it wouldn't be correct except at certain precise points along the transmission line. That's not to say the "balun" or "unun" isn't doing something important. It is--isolation. Isolation is an important function, but it has nothing to do with balancing or unbalancing, because the distinction simply doesn't exist in a transmission line. It may be tempting to say that earth ground is the external voltage reference, but the problem is that there is no one earth ground potential. Different parts of the earth are at different potentials, and we know that RF current flows through the earth in the vicinity of many antennas (and elsewhere too). The earth is not a magical "zero" point for voltages. Also, what would we do in space? The same transceiver you use in your shack can be placed in a satellite orbiting the earth, and then where is "ground"? In the world of audio, we don't care about earth ground (except for lightning safety). "Ground" means signal ground, an established common "zero" point of DC power supplies in the system. There's no such thing in an antenna because it doesn't have a power supply. What we're sending to the antenna isn't so much a "signal" (conveying information via a changing voltage). What we're sending to the antenna is power, most of which we hope will be radiated. "Ground" is meaningless here, just like with power amplifiers feeding speakers. By the way, I also made a living running concert sound, owned one recording studio, managed another, worked as a staff engineer in yet another, and built many PA systems, large and small. I say all that not to be competitive. I only want to try to establish my credibility to speak about audio systems. One last clarification: Impedance is a complex-valued function of reactive and resistive elements in a system and frequency. Since impedance varies with frequency, it's fairly common to speak of DC impedance, meaning impedance at the zero point on the frequency curve. In the case of a usable balun, DC impedance should be as close to zero as possible. If not, it's a resistance that dissipates power as heat, which is of course undesirable. Once again, I don't mean to be in any way disrespectful or merely argumentative. I write this because I think there is value in getting our concepts straight. I recognize that this is a difficult thing to do, as these are not simple concepts. As I said before, I am new to ham radio and open to being set straight in my thinking. In the meantime, I don't know any other way of trying to get to the bottom of something like this than to try to clarify my views by hashing them out. I'm hoping this will come across as a cordial discussion, because I mean it that way, sincerely. 73 Dan AG7OV |
Displaying 1-7 of 7
