|
Strange question about SWR on HV lines
ON5MJ wrote: Hi there, A friend of mine asks what happens on high voltage distribution lines about SWR at the distribution frequency (50/60 Hz). I'm stuck. I understand that HV generators/transformers behave like voltage sources and not power sources. This means that conjugate impedances don't apply here but what about the existence of SWR on those kind of lines and the possible consequences on very long lines. Anyone has an idea ? I understand that excessive SWRs on the long-haul transmission lines have brought entire grids down. 73 de ON5MJ - Jacques. w3rv |
On Sat, 11 Dec 2004 20:54:07 -0600, "Nick Kennedy"
wrote: Impedance matching a generator to the load would be a bad idea .... Each station generates in phase with the voltage that happens to be present at its location. Hi Nick, Phase is necessarily found in Impedance. As you allow that multiple generators share a line, they are also across the load as a load if they do not present the right phase. As for the "bad idea" of matching, you are appealing to Edison's old and deliberate misreading of Thevenin. Matching does NOT require a resistance, this is a mis-read of conjugate matching that follows the fact (in antennas). It does not drive the need (in power delivery). Power stations only need perform a Z Match, not a Conjugate Match. Any form of X is sufficient to accomplish the task and they do it far simpler through field excitation control. Back when Edison was battling Westinghouse/Tesla in the DC vs. AC distribution system wars; Edison tried to confuse the issue with his munged up version of Thevenin's Theorem insisting that his competitors would have to burn up half their power to deliver half their power. He thus claimed his DC system to be more "efficient." New York bankers didn't know the difference between Thevenin or Copernicus. In fact, it was Thevenin's proof that crippled Edison in the marketplace. The only way to cut losses was to lift potentials into the stratosphere. Edison also had a campaign trying to prove AC was lethal, but DC was survivable (largely true). But with a low loss system running in the KV and no way to convert it to residential use, the writing (about lethality) was on the wall. AC, on the other hand, could deal with that easily. Edison's business/technical logic would have to wait for nearly 100 years to be resurrected for the ENRON bubble to coincide with New York banker IQ phasing. 73's Richard Clark, KB7QHC |
Richard Clark wrote:
"---the distance was far enough to observe the effects of SWR - in exactly the same manner we observe them at HF or VHF etc." Yes, that was anticipated. The wavelength is a little less than 186,000 miles per second divided by 60 cycles per second, or 3100 miles per cycle approximately. The reduction in wavelength is due to the velocity factor of the transmission line. Construction determines the velocity factor. The phase delay ib transmission over the actual distance between Hoover Dam near Las Vegas and Los Angeles is only a few degrees. For example, a 60 Hz transmission line slectrical length of 310 miles would be 1/10 wavelength or about 36 degrees. Surely noticible but not crippling. Now, many high-voltage transmission lines are transporting d-c. The rule of thumb is that you need a kilovolt per mile of trangmission line length to get efficiency. So, hundreds of miles require hundreds of KV. At these voltages, the difference between rms and peak voltage becomes important. RMS = DC. Now, conveersion from a-c to d-c and back again is fairly easy and efficient. So, we have HV, DC power transmission. Tesla had the first laugh. Now, maybe Edison has the last laugh after a hundred years of development. Best regards, Richard Harrison, KB5WZI |
On Sun, 12 Dec 2004 04:21:55 GMT, Richard Clark
wrote: On Sat, 11 Dec 2004 20:54:07 -0600, "Nick Kennedy" wrote: Impedance matching a generator to the load would be a bad idea ... Each station generates in phase with the voltage that happens to be present at its location. Hi Nick, Phase is necessarily found in Impedance. As you allow that multiple generators share a line, they are also across the load as a load if they do not present the right phase. As for the "bad idea" of matching, you are appealing to Edison's old and deliberate misreading of Thevenin. Matching does NOT require a resistance, this is a mis-read of conjugate matching that follows the fact (in antennas). It does not drive the need (in power delivery). Power stations only need perform a Z Match, not a Conjugate Match. Any form of X is sufficient to accomplish the task and they do it far simpler through field excitation control. Back when Edison was battling Westinghouse/Tesla in the DC vs. AC distribution system wars; Edison tried to confuse the issue with his munged up version of Thevenin's Theorem insisting that his competitors would have to burn up half their power to deliver half their power. He thus claimed his DC system to be more "efficient." New York bankers didn't know the difference between Thevenin or Copernicus. In fact, it was Thevenin's proof that crippled Edison in the marketplace. The only way to cut losses was to lift potentials into the stratosphere. Edison also had a campaign trying to prove AC was lethal, but DC was survivable (largely true). But with a low loss system running in the KV and no way to convert it to residential use, the writing (about lethality) was on the wall. AC, on the other hand, could deal with that easily. Edison's business/technical logic would have to wait for nearly 100 years to be resurrected for the ENRON bubble to coincide with New York banker IQ phasing. 73's Richard Clark, KB7QHC I had never thought of the power industry having "matching" problems until I picked up a magazine ( the New Yorker of all things ) in a doctor's office years ago and there was a story about fluctuating SWR's etc and what the power company did to compensate. Unfortunately I got called in to see the Doc before I could finish the article. Thanks for sharing some light on the subject. Gary K8IQ |
The VERY first thing to remember about power transmission is that generators
must NOT be conjugate matched to loads. We CANNOT have half the power dissipated in the generator! === Reg. |
The following details will give you some idea of what you are waffling
about. A 2-WIRE, WIDE-SPACED, POWER TRANSMISSION LINE. At a frequency of 60 Hz Length = 100 miles. Wire diameter = 1 inch. Wire spacing = 10 feet. Nominal RF Zo = 650 ohms. Actual Zo at 60 Hz = 650 ohms. Angle of Zo = -2.6 degrees. Velocity factor = 0.99 1/4-wavelength = 767 miles. Resonant Q at 60 Hz = 11 Inductance = 3.54 milli-henrys per mile. Capacitance = 0.00836 micro-farads per mile. FOR A LOAD OF 500 OHMS - Input impedance = 520 + j*51 Line loss = 0.1 dB. Power Loss in line = An economical 2.3 percent. Reflection coefficient = 0.133 Angle of reflection coefficient = 170 degrees. VSWR = 10.5 Economics rules the roost at power frequencies. The normal transmission voltage on such a line is measured in terms of 100,000 volts. Note that, with a resonant Q of 11, should an open-circuit fault occur at a distance of 1/4-wavelength the voltage at the fault can rise to a million volts or more. Electrical power engineers have far worse problems than mere radio engineers have on the popular 40m band. They too are concerned with reflection coefficients and SWR. ;o) But the technicalities were all exactly sorted out in the Victorian age by the young, self-educated, recluse and hard-of-hearing genius like eethoven - Oliver Heaviside who was derided by the old-wives and silly guru university professors of his age. The above technical details, and more, can be computed and studied, from power frequencies up to UHF, by downloading, practical, easy-to-use programs RJELINE2 or 3 from the following website. Download in a few seconds, not zipped-up, and run immediately under common-or-garden DOS/Windows. ---- .................................................. .......... Regards from Reg, G4FGQ For Free Radio Design Software go to http://www.btinternet.com/~g4fgq.regp .................................................. .......... |
Jaques, ON5MJ wrote:
"I am not aware of this but it makes sense that efficiency of AC/DC + DC/AC conversion must be higher than the use of pure AC transmission." Yes. It makes no sense to lose more in conversion than on the transmission line. The problem with extreme high voltage power transmission is insulator flashover and corona. Alternating current and voltage have effective values which are their peak values divided by the square root of two. This means that peak volts times peak amps divided by two is the same average power as rms volts times rms amps. The same average power transmission requires peak values 1.414 times the rms value, effective value, or d-c value. In d-c tramnsmission, peak and effective values are the same 100% of the tiime, so the required d-c voltage is only 0.707 times the a-c voltage peak for the same power transmission. Jaques also wrote: "By definition loads vary all of the time but voltage must not vary accordingly." Use of d-c eliminates reactance as a cause of voltage variation. It also eliminates "skin effect" as an impefance so that the entire cross-section of the line is used. The case for extremely high voltage d-c transmission is pretty good. Best regards, Richard Harrison, KB5WZI |
Note that, with a resonant Q of 11, should an open-circuit fault occur
at a distance of 1/4-wavelength the voltage at the fault can rise to a million volts or more. How do you compute this ? -------------------------------------------------------------------- It is computed in exactly the same way at 60 Hz as it is at 60 MHz. A 1/4-wavelength line behaves as any other tuned circuit. The voltage at the open end rises to Q times the voltage applied at the input end. With a short-circuit line the current at the short circuit rises to Q times the current at the input end. The Q of a tuned circuit is the reactance of the inductance divided by the wire resistance. Q = Omega*L / R. The Q of a transmission line is the reactance of the line's overall inductance divided by the line's overall resistance. And again, Q = Omega*L / R. Since both L and R are directly proportional to line length, for a given line, Q is a constant and is independent of length. Omega = 2*Pi*Freq. There are other more complcated ways of calculating the resonant rise in open-circuit voltage from the line's transmission and propagation properties. But they all give the same answer of course. Such calculations provide a means of checking for program software bugs. The hardest part of the exercise is calculating the inductance and resistance from the line's physical dimensions and operating frequency. Which are needed anyway to calculate all the many other output quantities from the program. Q is just a spin-off. ---- Reg, G4FGQ |
Richard Clark wrote:
Hi Nick, Phase is necessarily found in Impedance. As you allow that multiple generators share a line, they are also across the load as a load if they do not present the right phase. As for the "bad idea" of matching, you are appealing to Edison's old and deliberate misreading of Thevenin. Matching does NOT require a resistance, this is a mis-read of conjugate matching that follows the fact (in antennas). It does not drive the need (in power delivery). Power stations only need perform a Z Match, not a Conjugate Match. Any form of X is sufficient to accomplish the task and they do it far simpler through field excitation control. Hello Richard, Even with your redefined version of matching, individual generating stations don't explicitly match to the load reactance or (certainly not) load resistance. A little generation 101. (This goes a little beyond the scope of discussion, but I throw it out because I think it's interesting.): Generators operate in two modes: Isochronous (single generator supplying local bus) and parallel (multiple generators in parallel; connected to a grid). The operator has two main controls: Steam (or whatever) flow to the turbine or other driver, and current to the field. In the isochronous mode, varying steam flow causes the speed of the generator to change. It also varies the amout of power delivered. In the parallel mode, the generator can't measurably push the speed of the grid, so increasing steam flow only increases the electrical power output. In the isochronous mode, varying field current changes the terminal voltage of the generator. In parallel mode, varying field current can't significantly change grid voltage. But it does change the reactive power output (MVAR or kVAR) of the generator, as you said. The normal mode of operation for a large generator is in parallel with the grid, so the operator is using steam (diesel, water, hamsters, etc.) to regulate real power output and field current to regulate reactive power output. Now some anecdotal stuff about how generators are operated. The system dispatcher requests individual generators to adjust their power and VARs to match load. This isn't impedance matching, it's simply supplying the demand. In the case of VARs, the goal is both to supply the demand and to equalize voltage across the system, not to cause any kind of mathematical match between the generator's internal X and the system's X. Oh yeah, I said earlier that individual generators don't appreciably affect grid voltage. That's true, but locally they do have an effect, like tent poles in a big canvas. So the local stations are both supplying their share of the total reactive load and propping up voltage in their area. (The operator increases VAR output by taking his excitation switch to the "raise voltage" position.) Anyway, I digressed from my anecdotal stuff. At my plant, the generator puts out 1050 MW 24/7, but MVAR may vary between 0 (or slightly negative) and 200 MVAR. So we're not matching to any specific impedance, but supplying load and maintaining voltage. A story transmission guys like to tell is how they may use open ended transmission lines as a kind of capacitor bank. Say there's a line 100 miles long from my plant to somewhere that's not needed to carry load. The system controller might connect it at my plant's end but leave the breakers open at the far end. A line has both capacitive and inductive reactance of course, but when unloaded, the capacitive dominates. So the trick of the trade is to use it to supply reactive MVARs. The point of the story in this context is that the controller isn't concerned about SWR on this extremely mismatched line. Another possibly relevant story. We connect our emergency diesel generator to the grid for testing and load it to about 3000 kW and typically from 0 to 100 kVAR. But to fully test the excitation system, the kVAR is at some point raised to 1400. The point being that the generator can be operated anywhere within its rating, with no need to match to any mysterious impedances out there in the world. Makes sense when you think about it. Who would want a generator that was constrained to operate at some fixed ratio of real to reactive power? 73--Nick, WA5BDU 73's Richard Clark, KB7QHC |
Thanks very much for the interesting and informative tutorial from
someone in the industry. I have one question: Nick wrote: . . . Another possibly relevant story. We connect our emergency diesel generator to the grid for testing and load it to about 3000 kW and typically from 0 to 100 kVAR. But to fully test the excitation system, the kVAR is at some point raised to 1400. . . If your customers' loads were, for the sake of argument, purely resistive as seen at your power plant output, then the voltage and current would be in phase at that point. But in order to make your generator produce "reactive power", the voltage and current have to be forced out of phase at the generator. How is this resolved? Is that reactive power "delivered" to (actually swapped back and forth between) other generators in the system -- that is, do the other generators in the system shift their own phase angles so that the V and I can be at some angle other than zero at your generator output (and, necessarily, also at the outputs at other generators in the system) yet in phase at your customers' loads? Or do you have some local bank of reactance that you can switch in to feed the "reactive power" back and forth to when you run this test? Roy Lewallen, W7EL |
Richard Clark wrote:
"The trig is identical as are the results." Yes, but the equipment often takes different forms. The best place to get rid of circulating current in the transmission line is at the load, before it causes additional line loss. For signal lines a capicitance or an inductance often is formed by a line stub. For power lines a capacitance is often produced by an over-excited synchronous motor or motors. Some constant speed loads are suitable for sychronous machines. Such a machine drawing a leading current has been called a rotary capacitor. Its current draw and capacitance are controlled by its excitation. Most induction motors and industrial loads have lagging currents. Power factor correction requires the production of an offsetting leadng current. Best regards, Richard Harrison, KB5WZI |
|
Hello Roy,
Good question and one I had considered addressing in my already over long post. In general "the grid" is viewed as an idealized source or sink of both real and reactive power. So we can theoretically supply it as much power as we wish, and supply or take in as much reactive power as we wish. No reactive load banks needed. So when I said generation (of both watts and VARs) is matched to demand, that's not necessarily *exactly* the case when it comes to VARs, as you guessed. Generators can both supply and absorb them to meet the need, and the net VAR output doesn't necessarily have to equal whatever the customers are offering as the load at any given time. BTW, in the power biz, we have the convention of "supplying", "outgoing", or positive VARs to describe reactive power out from the generator to a lagging (inductive) load and incoming, or negative VARs to leading (capacitive) loads. Incidentally, real power must flow *out* only. We have reverse power (anti-motoring) relays to trip the unit off line if this rule is broken. The tendency of generators to exchange VARs when in parallel leads to a stability problem in excitation control. A slight mismatch in excitation systems can lead to a huge exchange of VARs and resulting overcurrent. So excitations system incorporate what is known as a "droop" feature which essentially provides a negative feedback based on reactive current. Increased VARs out tends to reduce excitation, stabilizing the system. Droop is typically switched "off" in isochronous (one generator isolated) mode. There's an analogous "droop" feature on the governor for speed control when in parallel. Not sure if your question included this, but it's interesting to consider just how a generator produces out of phase current when connected to what we're essentially considering to be equivalent to an ideal voltage source, since by definition the generator's terminal voltage must equal that of the source (grid). As I see it, the key is that the generated voltage, Eg, is n ot the same as the generator's terminal voltage, Et. There's a drop across the armature reactance, so Et equals Eq minus that drop. Interesting that out of phase currents produce drops in phase with Eg ... Well, I thought so anyway. Current is Et minus Eg divided by Za (armature impedance). Changing excitation changes the magnitude of Eg (Et is fixed by the grid and so is an anchor point). By fooling with the phasors, I think you can see how changing excitation changes the phase angle and therefore controls VARs. How *power* is controlled is beyond the scope of this discussion (and maybe of my understanding). But it actually is related to the angle of the rotor's physical position relative to the rotating field of the armature. That angle is dependent upon the torque supplied by the driver. 73--Nick, WA5BDU in Arkansas "Roy Lewallen" wrote in message ... Thanks very much for the interesting and informative tutorial from someone in the industry. I have one question: Nick wrote: . . . Another possibly relevant story. We connect our emergency diesel generator to the grid for testing and load it to about 3000 kW and typically from 0 to 100 kVAR. But to fully test the excitation system, the kVAR is at some point raised to 1400. . . If your customers' loads were, for the sake of argument, purely resistive as seen at your power plant output, then the voltage and current would be in phase at that point. But in order to make your generator produce "reactive power", the voltage and current have to be forced out of phase at the generator. How is this resolved? Is that reactive power "delivered" to (actually swapped back and forth between) other generators in the system -- that is, do the other generators in the system shift their own phase angles so that the V and I can be at some angle other than zero at your generator output (and, necessarily, also at the outputs at other generators in the system) yet in phase at your customers' loads? Or do you have some local bank of reactance that you can switch in to feed the "reactive power" back and forth to when you run this test? Roy Lewallen, W7EL |
Richard Clark wrote: The conjugate argument is unnecessary and in error as a response to my posting. Good. Glad we've come to an agreement on that one. In the isochronous mode, varying field current changes the terminal voltage of the generator. In parallel mode, varying field current can't significantly change grid voltage. But it does change the reactive power output (MVAR or kVAR) of the generator, as you said. This is matching explicitly. Not quantifying the load does not make it something other than R =B1iX Ohms When quantified, it would undoubtedly lead to very small Rs and Xs, but all the while, the angles they resolve to are always significant. In every sense of the term Matching, there is not a jot of difference between these applications (AC/RF) except frequency and magnitudes of voltage and current (and not always that). This isn't impedance matching, it's simply supplying the demand. Absolutely no difference between applications. Yes there is. The operator of the generator has the freedom to adjust power output from 0 to 100% of rated and VAR output between the maximum incoming and outgoing rated values. No matching required. I sense that you are beginning to argue my side of the case for me. Please give me appropriate credit. not to cause any kind of mathematical match between the generator's internal X and the system's X. Not demonstrated, in fact your entire recitation argues to the contrary. My Power Transmission handbooks say quite explicitly that manual or automatic operation attends to the phase shift by necessity. Even if you don't calculate any quantified value it remains as a mismatch until intervention. Trying to draw this back into the Conjugate is, again, a misread of the distinctions between Conjugate and Z Matching. The two are frequently mixed in discussion (through error), but they are not the same. Agreed; please stop bringing conjugate matching into this. You've already accepted the fact that it doesn't apply. So we're not matching to any specific impedance, but supplying load and maintaining voltage. This statement is simply unquantified Matching. A story transmission guys like to tell is how they may use open ended transmission lines as a kind of capacitor bank. Say there's a line 100 miles long from my plant to somewhere that's not needed to carry load. The system controller might connect it at my plant's end but leave the breakers open at the far end. A line has both capacitive and inductive reactance of course, but when unloaded, the capacitive dominates. This is classic matching technique at ANY frequency and has been part of the canon for more than 100 years. No it is not. You are making oblique reference to the use of stubs in RF matching. In that application, the length of the stub in degrees is critical. In the one I describe, the length of the line is random; it is being used for its capacitance only. It could be replaced by an equivalent capacitor to produce the same effect. The same is not true of a matching stub. So the trick of the trade is to use it to supply reactive MVARs. The point of the story in this context is that the controller isn't concerned about SWR on this extremely mismatched line. Actually, the concern is quite fundamental and has also been part of the canon for more than 100 years. Who would want a generator that was constrained to operate at some fixed ratio of real to reactive power? Hi Nick, Who would want a generator that was constrained to supply only toasters? Such strawmen arguements can be lined up from here to the moon. Kind of like canons and "known for more than 100 years"? Empty supercilious statements that say nothing? One of my Power distribution handbooks (ca. 1907) is not shy to the matter of Generators seeing the products of mismatches: "Thus a wave passing from one part of a circuit to another having a greater ratio of inductance to capacity will develop an increased voltage and decreased current. This explains the breaking down of windings, due to surges entering them." I don't have to say SWR for it to be evident in the nature of the description above. I don't have to say Z matching for it to be evident in the nature of the corrective action. I don't have to say X for it to be evident in the myriad of phase drawings and calculations that are offered page after page. The old practices could measure Gamma or Rho as we describe it in this forum. Calling it VAR does not make it a mysterious process confined to 50/60 Hz, it is simply a term that describes the same thing and follows the same dynamics and is reduced by the same operations. We shift the phase using a variable capacitor or a variable inductor. In the plant the same thing is done through adjusting field excitation (or any number of tricks that are available to the RF craft too). The trig is identical as are the results. Finally, the use of the term VARs is not a power engineer's sly attempt at obfuscation. It is a common and well defined term in daily use. 73--Nick, WA5BDU =20 73's Richard Clark, KB7QHC |
A device usually described as a 1-to-1 choke balun is amongst the most simple of all radio components. Actually, 1-to-1 has nothing to do with impedance-matching or transformation, or anything else. The choke simply allows a balanced circuit, of no particular impedance, to be connected to an unbalanced circuit, of another no particular impedance, without any significant interaction between them. It is just a very short length of balanced-twin transmission line, like speaker cable, of no particular impedance, wound on a ferrite ring to behave as a bifilar-wound RF choke. Loss is negligible. There's only copper loss. The ferrite plays no part in transmission along the short line, only in the longitudinal choking action. If there are any ferrite losses they only occur due to the very small longitudinal current - which is what the choke is doing its best to get rid of anyway. When used at the end of an antenna feedline a choke balun is just a short continuation of that line, albeit of a different impedance. Which is of no consequence. At the junction of the balanced-to-unbalanced lines, such as coax to open-wire, there's going to be a large mismatch anyhow. But that's what the tuner is for. The balun does indeed have an impedance transforming property as does any other short length of line. But in the case of multiband antenna systems it merely transforms one set of random-value impedances to another random set. Which, in effect, leaves things as they were. The average impedance of speaker type cable is about 140 ohms which, for perfectionists, fits very nicely between 50-ohm coax and 450-ohm ladder line. But it hardly matters. The length of line wound on a balun should not exceed 1/8 wavelength at the highest frequency of interest, at the lines own velocity. Choke action at the lowest frequency of interest depends on number of turns and permeabilty of the ferrite. For multi-band operation a choke balun should be used. It is far better than fixed-ratio 4:1 and 9:1 baluns which involve wishful thinking and are best used over relatively narrow bands. Choke baluns also make balanced tuners redundent. Who wants to crank two roller-coasters when one will do. ---- Reg, G4FGQ |
Sorry, my previous message was placed in the wrong thread.
I can add that 4:1 and 9:1 fixed-ratio baluns should be used only between known, well-defined, relatively narrow impedance ranges. They are best NOT used on the transmission line side of tuners. Otherwise the tuner is forced into dealing with reactances in the balun itself. This restricts the range of antenna impedances which can be handled by the tuner. Fixed-ratio baluns, with their excellent frequency response, are ideal on the transmitter side of tuners, i.e., between precisely known impedances. However they are seldom needed in that unusual position if the required impedance transformation takes place in the tuner itself. As it always is with a standard type of tuner and a 50-ohm transmitter. ---- Reg, G4FGQ |
Thanks once again for the excellent explanation. What little I absorbed
in the required year of power systems coursework has pretty much faded completely out, so I appreciate your taking the time to educate me and the other readers. Roy Lewallen, W7EL Nick Kennedy wrote: Hello Roy, Good question and one I had considered addressing in my already over long post. In general "the grid" is viewed as an idealized source or sink of both real and reactive power. So we can theoretically supply it as much power as we wish, and supply or take in as much reactive power as we wish. No reactive load banks needed. . . . |
The last chapter.
There's yet another point in favour of twin-balanced line rather than coax for a choke balun. As previously stated, the length of line wound on it is best not allowed to exceed about 1/8-wavelength at the highest frequency of interest at the line's own velocity factor. Solid polyethylene coax has an appreciably lower velocity factor than twin line such as figure-of-eight speaker cable. Or two separate wires wound alongside each other. Consequently, for the same length of balun line in wavelengths, more twin-line turns can wound on the ferrite core than coax turns. This increases LF inductance and extends the lower frequency downwards. Or alternatively, with a shorter physical length of line on the balun, the higher frequency is extended upwards. The usable bandwidth of the twin-line version therefore increases roughly in the ratio of the two velocity factors. A choke balun may indeed be amongst the most simple of radio components to construct - its true complexity being hidden. But as always with radio, almost anything will work! ---- Reg, G4FGQ |
| All times are GMT +1. The time now is 12:25 AM. |
Powered by vBulletin® Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
RadioBanter.com