|
Roy Lewallen wrote:
Dr. Slick wrote: Agreed. Then any mismatch loss from PA to the 50 ohm coax is never measured at the meter. So we never really measure the reflected power coming right out of the PA, even if we attach the meter directly to it's output. No "reflected power" comes out of a PA. By definition. I wonder who invented that definition? It seems pretty obvious that not all PA's are Z0-matched so they will always re-reflect 100% of the incident reflected power. But that is exactly what that definition implies. -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 80,000 Newsgroups - 16 Different Servers! =----- |
Dear all, Here's a useful tip -
The loss along any sort of HF transmission line, SWR or not, increases with line temperature according to just ONE HALF of the resistance-temperature coefficient of the conductors. Why ? It's something to do with skin effect. RTC of copper is near enough 0.4 percent per degree C. So loss along any line, in dB or nepers, increases by 0.2 percent per degree C. Now you may not think this matters very much. But if you consider a seasonal change in temperature on the ocean bottom of only 2 degree C along a 2000-mile transatlantic cable which has an attenuation at 5 MHz of 1.6 dB per mile, total attenuation = 3,200 dB, then you will appreciate the responsibilty laid on the shoulders of the design engineers of the first oceanic telephone cable systems. An uncertainty of 0.4 percent of 3,200 dB = 13 dB which is enough to wreck system operation. Omission of a submerged repeater allows signals to fall below noise level at the last repeater. Addition of one more repeater overloads the last repeater causing cross-modulation, cross-talk and noise. Bear in mind repeaters are both-direction amplifiers and the lowest speech channels are at 60 KHz where overall attenuation over the same distance is only about 350 dB. Repeater power is fedover the the inner cable conductor from constant current sources at both ends, maximum voltage = 10 KV. +ve from one end, -ve from the other. During magnetic storms and other disturbances the potential difference between ground electrodes in N.America and W.Europe can rise to several thousand volts. Although the last time I measured it on an AVO-8 it was only 1.3 volts. I did, of course, make use of the safety grounding stick before using the crocodile clips on the ends of the meter leads. Depended on the tides and the flow of the Gulf Stream across the Earth's magnetic line of force. It always struck me as being highly incongruous, even absurd, that in normal operation, cables of the highest possible quality materials, manufactured by automatically controlled, specially-designed precision machinery, laid at great expense by an 8000-ton, specially-design ship over thousands of miles, should end up by being terminated with a foot or so of soldered, screwed-up, cotton-covered 22-gauge wire rescued from the terminal-station scrap bin. This is true. I have seen both ends with my own eyes. On one occasion I even did the soldering after completing overall tests! But I was careful to use a fairly straight length of wire with sufficient sag to eliminate any possible tension beyond that due to its own weight. Dear readers, believe me, there's no time to worry about SWR when loss in revenue amounts to $100,000 per hour + repair-ship expenses every time a flatfish trawler scoops up a cable in its net, cuts it free with an axe, and the skipper sneaks away at top speed without telling anybody in case he has to pay for the damage. Coax cable Zo = 43 ohms. Diameter over polyethylene = 1 inch. Inner conductor = longitudinal overlapping crimped copper tape, laid over the cable's principal strength-member of a number of high-tensile steel wires, overall inner diameter about 1/3 inch. Outer conductor = 6, touching, longitudinal aluminium tapes with a small spiral lay. Sheath = 0.1-inch thick extrusion of black polyethylene if I remember correctly. For shallow water and continental shelves there was a number of protective heavy iron wires laid on a bed of tarred hessian as had been used for 100 years on the first of the Atlantic telegraph cables. I sometimes think of (the later) Lord Kelvin who followed his calculations with the recommendation to investors "Go ahead, make and lay the bloody stuff". But it was Heaviside, a generation later, a genius who died of neglect, who eventually described how the "bloody stuff" and radio propagation really worked. Folks, just a little light-hearted digression, a respite from so-called SWR meters. Please continue with your discussion. ;o) ---- Reg, G4FGQ |
Someone sed:
"The SWR is based on the ratio of the forward to the reflected power." But this isn't the *definition* of SWR. Trivia question: What is the definition of SWR? 73 de Jack, K9CUN |
|
There is NO definition of SWR!!
There is a definition for VSWR! There is a definition of ISWR! But, unfortunately, there is NO definition for SWR!! Why don't you offer one? Deacon Dave, W1MCE JDer8745 wrote: Someone sed: "The SWR is based on the ratio of the forward to the reflected power." But this isn't the *definition* of SWR. Trivia question: What is the definition of SWR? 73 de Jack, K9CUN |
You must not have seen my posting yesterday on this thread, where I gave
the definition. Do I need to post it again? Roy Lewallen, W7EL JDer8745 wrote: Someone sed: "The SWR is based on the ratio of the forward to the reflected power." But this isn't the *definition* of SWR. Trivia question: What is the definition of SWR? 73 de Jack, K9CUN |
On Thu, 14 Aug 2003 08:30:20 +0100, "Ian White, G3SEK"
wrote: The subject is why SWR meters might read differently with different lengths of coax. Your statements about mismatch uncertainty are true, but not relevant. Hi Ian, Look at the subject line above. Everything reported by me responds directly to it. I notice you have nothing relevant to add in that regard. To this point NO ONE has responded to the data, nor to the authoritative citations. It would be more useful for you folks to point out error rather than simply arm-chair your way through this with denial. To this point NO ONE has offered any synopsis of my error in method. Such close examination appears to be confined to re-evaluations of the CFA which lack both, and thus make a more amusing target that does not test anyone's skill beyond debate. In the words of Ho Chi Minh: "Paper Tigers." Also to this point, the only "critics" have been those with a voiced interest in not assigning a value to Source Z, but again demonstrating a more than ample capacity to arm-chair their way through byzantine explanations of what it is "not." Bench work seems to be anathema, however I do enjoy the zen-cartwheels being performed, thanks. If I were to offer this with the infinite regress of Cecil's logic, hide my data, and embellish my method, you would all be compounding this thread to 600 entries. I don't play that game, sorry to disappoint you fellows, but I don't write to entertain (even if I can do it better than you without really trying - sometimes, like now, opportunity begs). The repetition of data, like denial, is not debate, so I am satisfied to post real bench work once, and pull paper tails until that goes stale. I am particularly amused by protestations that examination of data is not worth anyone's time, but offering editorialization is. :-) Clearly most of you should take more pleasure in your amusing recreations. You guys worry this out of all proportion, you act like this is especially important and it certainly seems to set your teeth on edge. As many would point out, this is only a hobby; it's not like your job is on the line, or that you have to meet a customer's expectation. :-) 73's Richard Clark, KB7QHC |
Richard Clark wrote:
I am particularly amused by protestations that examination of data is not worth anyone's time, but offering editorialization is. :-) For those of us who missed the data, what date and title does that posting possess? -- 73, Cecil, W5DXP |
|
|
Richard Clark wrote:
How curious of you to have missed this pæan to your style: Not curious at all, Richard. My Mother is in the hospital with terminal leukemia. -- 73, Cecil, W5DXP |
Richard,
Thank you very much for posting that reference to John Juroshek's article. It happens to cover a topic very near and dear to my heart, and I called John and we had a nice chat about it. In addition, I told him where I saw the reference, and asked him if he would say that the SWR on a line, in steady state excitation, with a source at one end and a load at the other, depended on the source's source impedance. He said it of course does not, and cannot understand how you would have interpreted it that way. Should I check with any of the other authors? Cheers, Tom Richard Clark wrote in message . .. On Tue, 12 Aug 2003 02:48:49 -0700, Roy Lewallen wrote: Almost correct. The transmitter output impedance has no effect whatsoever on the line's SWR. Roy Lewallen, W7EL Hi Roy, Entirely incorrect. Transmitter output impedance that does not conform to transmission line Z, when presented with a mismatched load through that line, adds mismatch uncertainty in the form of an indeterminate SWR and indeterminate Power to the load. This has already been demonstrated twice. This has long been documented with NBS/NIST references going back 4 decades. There is nothing mysterious about it at all, and it conforms to the rather simple principles of wave interference so poorly presented by Cecil in months past. The authoritative site: http://www.boulder.nist.gov/div813/index.html Direct reference: "Juroshek, J. R.; A Direct Calibration Method for Measuring Equivalent Source Mismatch; Microwave J., pp. 106-118; October 1997 Obscure references: http://www.boulder.nist.gov/div813/r...00S_n2nNet.pdf "With vector measurements of the generator and meter reflection coefficients Ãg and Ãm, respectively, the power of the incident signal am can be related to the power of the source." http://www.boulder.nist.gov/div813/r...FRad_ARFTG.pdf which describes radiometer calibration (perhaps too exotic for this group) "tests are based on two assumptions. First, the network responds linearly to our signal ( no power compression), and second, the radiometer is sufficiently isolated from the source impedance." ... "One of the assumptions made in deriving eq. (2) was that the output from the radiometer is not dependent on the source impedance. In the construction of the radiometer, two isolators are inserted at the input of the radiometer to isolate the radiometer from the source." ... "The mismatch uncertainty depends strongly on the poorly known correlation between uncertainties in the measurements of different reflection coefficients, and so we use the maximum of the uncertainties obtained by assuming either complete correlation or no correlation whatsoever." "Forthcoming Paper: Influence of Impedance Mismatch Effects on Measurements of Unloaded Q Factors of Transmission Mode Dielectric Resonators" IEEE Transaction on Applied Superconductivity "Analysis of Interconnection Network and Mismatch in the Nose-to-Nose Calibration Automatic RF Techniques Group , June 15-16, 2000 , Boston, MA - June 01, 2000 "We analyze the input networks of the samplers used in the nose-to-nose calibration method. Our model demonstrates that the required input network conditions are satisfied in this method and shows the interconnection errors are limited to measurement uncertainties of input reflection coefficients and adapter S-parameters utilized during the calibration procedure. Further, the input network model fully includes the effects of mismatch reflections, and we use the model to reconcile nose-to-nose waveform correction methods with traditional signal power measurement techniques." As I mentioned, obscure references. However, given the impetus of their discussion is long known (and that I have already provided the original references they rely on), NIST presumes the investigators already have that basis of knowledge. 73's Richard Clark, KB7QHC |
|
|
On Thu, 14 Aug 2003 14:56:27 -0700, W5DXP
wrote: Richard Clark wrote: You among the many have done more in less time. I can accept negative, informed response. Can he explain my posting? That is the more telling as it is MY statement, not your summary of it. A couple of questions about your data. What kind of coax and connectors are used for the BVT? hardline of RG-58 dimension; only two connectors, the gozinta, and the comesoutta (never changed throughout any variation nor calibration). What kind of wattmeter was used for the readings? Bird. The load seems to be 50/3 = 16.67 ohms. I would like to see the same data for a load of 50 ohms. That is already discussed in the report. Seems you are getting the same resulting effect that I do (on purpose) when I vary the length of my ladder-line while watching the 50 ohm SWR. Your results may repeat every 1/2WL. What if you had gone out to 20 feet instead of stopping at 15 feet? I have no reason to think it would offer any change in a periodic variation already in evidence. One last thought. Did you consider common-mode current possibilities? Yes. 73's Richard Clark, KB7QHC |
W5DXP wrote in message ...
By definition. I wonder who invented that definition? It seems pretty obvious that not all PA's are Z0-matched so they will always re-reflect 100% of the incident reflected power. But that is exactly what that definition implies. Your vocabulary is very confusing here, Cecil. "Incident" usually refers to the forward power, so when you say "incident reflected" it's extremely confusing. Slick |
Roy Lewallen wrote in message ...
Dr. Slick wrote: Agreed. Then any mismatch loss from PA to the 50 ohm coax is never measured at the meter. So we never really measure the reflected power coming right out of the PA, even if we attach the meter directly to it's output. No "reflected power" comes out of a PA. Roy Lewallen, W7EL Maybe that why you can't measure it. Slick |
Roy Lewallen wrote in message ...
The observation that changing line length changes the measured SWR is regularly reported in this newsgroup, and the explanation is as regularly provided. There are at least three ways this can happen. 1. The SWR meter is designed for an SWR that's different from the line impedance. It's easy to show that this will result in different readings for different line lengths. I've measured RG-58 at over 60 ohms characteristic impedance, so this can happen even with a perfect 50 ohm SWR meter and "50 ohm" line. In this case, changing line length isn't really changing the line's SWR, just the meter reading. But a 60 ohm transmission line transformation from a non-50 ohm load will certainly change the SWR, as it won't be on the constant VSWR circle anymore. 2. There's significant loss in the cable. In that case, the longer the distance between the meter and the load, the better the SWR. That's the obvious one. 3. There's current on the outside of the coax. This means that the outside of the transmission line is actually part of the antenna. When you change its length, it changes the effective length of the antenna, which really does change the SWR. Current on the outside of the cable can also get into a poorly shielded SWR meter and modify its reading. And this is all in agreement with established theory. So you see, theory does say you can change the SWR reading, and in some cases, the actual SWR, by changing the coax length. But only under very specific circumstances. When observations don't match theory, chances are overwhelming high that either the observation is erroneous or misinterpreted, or theory is being misapplied. Roy Lewallen, W7EL I believe the source of confusion is he PA----+----50 ohm line----+SWR meter+----50 ohm line----+50 ohm dummy load 1 2 3 The "+" are connector points. You folks are saying that as you change the PA (source) impedance, that the SWR you read will remain the same, even if the incident power changes due to the change in reflected power at point 1. This i can agree with, as no matter how much incident power makes it past the impedance discontinuity at point 1, the system after this point will theoretically always reflect the same ratio of reflected power to incident. But, if you place an SWR meter of the same impedance as the output of the PA at point 1, you will definitely see a change in SWR at point 1 as you change the PA impedance, as you are changing the reference impedance (center of Smith re-normalized). This is what i thought you meant when you said "change the source impedance", but you meant to say "change the source, but keep the reference impedance the same". Understood, assuming this is what you guys mean. Slick |
Roy Lewallen wrote in message ...
If you ever do see an equation which shows transmission line SWR as a function of source impedance, please let me know where you see it. I'll add it to my list of untrustworthy sources. A simple experiment could be set up to disprove it. Roy Lewallen, W7EL I believe we are both right, and there is simply a misunderstanding here. I believe the source of confusion is he PA----+----50 ohm line----+SWR meter+----50 ohm line----+50 ohm dummy load 1 2 3 The "+" are connector points. You folks are saying that as you change the PA (source) impedance, that the SWR you read will remain the same, even if the incident power changes due to the change in reflected power at point 1. This i can agree with, as no matter how much incident power makes it past the impedance discontinuity at point 1, the system after this point will theoretically always reflect the same ratio of reflected power to incident. But, if you place an SWR meter of the SAME IMPEDANCE AS THE OUTPUT of the PA at point 1, you will definitely see a change in SWR at point 1 as you change the PA impedance, as you are changing the reference impedance (center of Smith re-normalized). This is what i thought you meant when you said "change the source impedance", but you meant to say "change the source, but keep the reference impedance the same". Understood, assuming this is what you guys mean. Slick |
W5DXP wrote in message ...
I was thinking about how you changed the impedance of your SWR meter, and I doubt that simply changing the terminating resistors to ground on both directional couplers is all you have to do. I believe this because the SWR/power meter needs to be a Zo ohm Thru, with a microstrip line of a certain dielectric thickness, and particular width and spacing from the couplers. And you cannot change these very easily. Do you have a schematic for us to see? Slick |
You must not have seen my posting yesterday on this thread, where I gave the definition. Do I need to post it again? Roy Lewallen, W7EL ============ No I saw it right *after* I posted. DUH Jack K9CUN |
|
Dr. Slick wrote:
Roy Lewallen wrote in message ... The observation that changing line length changes the measured SWR is regularly reported in this newsgroup, and the explanation is as regularly provided. There are at least three ways this can happen. 1. The SWR meter is designed for an SWR that's different from the line impedance. It's easy to show that this will result in different readings for different line lengths. I've measured RG-58 at over 60 ohms characteristic impedance, so this can happen even with a perfect 50 ohm SWR meter and "50 ohm" line. In this case, changing line length isn't really changing the line's SWR, just the meter reading. But a 60 ohm transmission line transformation from a non-50 ohm load will certainly change the SWR, as it won't be on the constant VSWR circle anymore. Once again, you're confusing the SWR meter reading with the SWR on the line. When the line and SWR impedances are different, the two are *not* the same. When I speak of the SWR on the line, I mean the SWR on the line, not the meter reading. Now look at the sentence you wrote. It's not very clear to me, but it would certainly make more sense if you replaced "the SWR" with "the SWR meter reading". Do *not* confuse the two. 2. There's significant loss in the cable. In that case, the longer the distance between the meter and the load, the better the SWR. That's the obvious one. 3. There's current on the outside of the coax. This means that the outside of the transmission line is actually part of the antenna. When you change its length, it changes the effective length of the antenna, which really does change the SWR. Current on the outside of the cable can also get into a poorly shielded SWR meter and modify its reading. And this is all in agreement with established theory. So you see, theory does say you can change the SWR reading, and in some cases, the actual SWR, by changing the coax length. But only under very specific circumstances. When observations don't match theory, chances are overwhelming high that either the observation is erroneous or misinterpreted, or theory is being misapplied. Roy Lewallen, W7EL I believe the source of confusion is he PA----+----50 ohm line----+SWR meter+----50 ohm line----+50 ohm dummy load 1 2 3 The "+" are connector points. You folks are saying that as you change the PA (source) impedance, that the SWR you read will remain the same, even if the incident power changes due to the change in reflected power at point 1. I'm saying that the line SWR doesn't change when you change the source impedance. I didn't say anything about incident or reflected power anywhere. And I won't. Cecil probably said something about the incident and reflected power, but explanations in those terms are strictly up to him. If you were to change the impedance of the left hand line (the one between the PA and meter), then the SWR on the left hand line would change, but the SWR on the right hand line wouldn't, and the SWR meter reading would remain the same. For that matter, you can do anything you want between the SWR meter and the PA -- add an impedance of any kind in series or parallel, change the left hand transmission line length and/or Z0, change the power, whatever you want, and it won't change either the meter's indicated SWR or the actual SWR on the right hand line. The rule is that whatever you change, it won't affect the SWR on any line that's "downstream" (toward the load) from the change you made. This i can agree with, as no matter how much incident power makes it past the impedance discontinuity at point 1, the system after this point will theoretically always reflect the same ratio of reflected power to incident. Be really, really careful when you start talking about forward and reflected power. It can very easily lead you to wrong conclusions about what's going on. Just check the postings on this group for the past few months for evidence. All the phenomena you can observe and measure can be fully explained by looking at forward and reverse voltage and current waves, and it's a whole lot less hazardous. One of the several problems with thinking in terms of forward and reflected power is that it's universally meant to refer to average power. So you've lost all time and phase information, making it impossible to clearly see how the traveling waves interact. If you must deal with "forward power" and "reverse power", do your thinking and calculations with voltage and current waves, then calculate the power when you're all done. As I said before, the ratio of forward to reflected voltage or current is independent of the source impedance. That ratio, when measured at the load, is simply the reflection coefficient at that point. But, if you place an SWR meter of the same impedance as the output of the PA at point 1, you will definitely see a change in SWR at point 1 as you change the PA impedance, as you are changing the reference impedance (center of Smith re-normalized). No, you won't see a change in the SWR at point 1 as you change the PA impedance. All the fiddling you do with your Smith chart just won't make it happen. Sorry. The SWR, voltage, current, impedance, power, reflection coefficient, waves, or anything else don't change in response to your Smith chart exercises. This is what i thought you meant when you said "change the source impedance", but you meant to say "change the source, but keep the reference impedance the same". No. When I said change the source impedance, I meant change the source impedance. Surely we don't need a discussion about what "impedance" means? When you get out your grease pencil and change the reference value of your Smith chart, it doesn't magically change the waves on the line on your workbench. Understood, assuming this is what you guys mean. It's time for me to leave this discussion. I've tried to make my statement as clearly and simply as I know how, but somehow people have decided that I really meant something else, or that there's this condition or that condition that cause exceptions to it, or that it all depends on what you scribble on your Smith chart with a grease pencil. It bears a striking resemblance to a political science (what an oxymoron!) course I took, in which we could make up any definition for anything, or any interpretation of anything anyone said or wrote (and were encouraged to do so), and all were equally valid. I've spent too much time interacting with engineers and not nearly enough with politicians and philosophers to know how to deal well with this fuzziness. Anyone who really cares can look up the equations in a couple of minutes. I'm sure they're on the web, if you have an aversion to paper media. Look up the equations, study them, understand them. If you don't believe them, make up your own equations. Then set up a couple of simple experiments to test them, and see which are right. That's how science and engineering are done. Roy Lewallen, W7EL |
Dr. Slick wrote:
Your vocabulary is very confusing here, Cecil. "Incident" usually refers to the forward power, so when you say "incident reflected" it's extremely confusing. "Forward power incident upon the load" or "Reflected power incident upon the source" makes sense to me. Given the definition of "incident", the reflected power has to be incident upon something. Simply replace "incident upon" with "arriving at". -- 73, Cecil, W5DXP |
Dr. Slick wrote:
W5DXP wrote: It seems pretty obvious that not all PA's are Z0-matched so they will always re-reflect 100% of the incident reflected power. But that is exactly what that definition implies. Your vocabulary is very confusing here, Cecil. "Incident" usually refers to the forward power, so when you say "incident reflected" it's extremely confusing. The point was not to confuse. So allow me to re-word it. Some say there is zero power reflected from a PA. For that to be true, all PA's must exhibit an impedance of Z0, i.e. all PA's must be Z0-matched. Doesn't that seem a little far-fetched? Some say that 100% of the reflected power is re-reflected by the PA. For that to be true, all PA's must exhibit an open, short, or pure reactance to the reflected waves. Doesn't that seem a little far-fetched? The problem lies in the definition of "generated power" which is forward power minus reflected power. A mental exercise will illustrate. XMTR---one second long feedline-----mismatched load For the first two seconds, a directional wattmeter at the XMTR reads 100W forward, zero watts reflected. During steady-state, the directional wattmeter reads 100W forward, 25 watts reflected. If the XMTR is a 100W signal generator equipped with a circulator and load resistor, we have no problem deciding that the signal generator is generating a continuous 100W and dissipating whatever reflected energy arrives. If the XMTR is a ham transmitter, we say it generates 100W for two seconds, and after that, it generates 75 watts, by definition. This seems to me to be just a useful shortcut that avoids opening Pandora's Box (in which the source is located). :-) If a transmitter only generates (forward power minus reflected power), it follows that the transmitter always re-reflects 100% of the reflected power arriving at its terminals. Does that sound reasonable? Or, if reflected energy is never re-reflected from a PA, then the PA must be dissipating the reflected power, which it previously generated, just like a signal generator with circulator load does. Does that sound reasonable? Or, if there is a circulator load, the reflected waves contain energy, but if there's not a circulator load, the reflected waves don't contain any energy. Does that sound reasonable? What sounds reasonable to me is that the reflected waves arriving back at the source obey the rules of the wave reflection model as described by Ramo and Whinnery. But since the impedance presented to the reflected waves by the transmitter is usually unknown, we are again up that proverbial creek without a paddle. So, by all means, use the shortcuts, but recognize that they are definitional shortcuts which may or may not represent reality. -- 73, Cecil, W5DXP -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
W5DXP wrote in message ...
Dr. Slick wrote: Your vocabulary is very confusing here, Cecil. "Incident" usually refers to the forward power, so when you say "incident reflected" it's extremely confusing. "Forward power incident upon the load" or "Reflected power incident upon the source" makes sense to me. Given the definition of "incident", the reflected power has to be incident upon something. Simply replace "incident upon" with "arriving at". "Forward Incident" a bit redundant in my opinion. "Incident" usually refers to the power moving towards the load, away from the generator. And your previous quote was: "It seems pretty obvious that not all PA's are Z0-matched so they will always re-reflect 100% of the incident reflected power." So here, you don't say where it is "incident" upon, of where it is arriving at, athough I assume you mean simply the reflected power. I'm just trying to make your vocabulary less confusing. Slick |
Dr. Slick wrote:
"Forward Incident" a bit redundant in my opinion. "Incident" usually refers to the power moving towards the load, away from the generator. I don't think that is true. The HP ap note, AN 95-1, refers to "the voltage wave incident on port 1" and "the voltage wave incident on port 2". Those two waves are moving in opposite directions, one toward the load and one toward the source. -- 73, Cecil, W5DXP -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
On Fri, 15 Aug 2003 11:39:51 -0700, W5DXP
wrote: You can't measure phase or magnitude? I simply cannot differentiate what part of the forward wave has been reflected back from the source Vs what part is actually generated by the source. That's the crux of the problem. Hi Cecil, But since the impedance presented to the reflected waves by the transmitter is usually unknown You have a forward part from the transmitter, a reverse part from the reflection. What more do you need unless you are discarding phase? If so, recover it (if this is all about a SWR meter, then take out the diodes and use a comparator). 73's Richard Clark, KB7QHC |
Richard Clark wrote:
You have a forward part from the transmitter, a reverse part from the reflection. We are talking about re-reflection FROM THE SOURCE! The forward part and re-reflected part from the source are coherent and traveling in the same direction so they cannot be separated for measurement purposes. That's why Bruene tried to determine the source impedance by bouncing another separate signal off the source. -- 73, Cecil, W5DXP -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Dr. Slick wrote:
Roy Lewallen wrote in message ... This is typical of you, Roy, for you to back out of a discussion when you don't want to admit that maybe someone else has a logical point. I've agreed with you on many things (antennas as transducers and such), but like many intelligent but close-minded people, you cannot accept someone elses points. I guess you think other people can't teach you anything, eh? NO ONE knows it all, even about a specific topic as impedance matching. Slick Let me explain why I leave these discussions. I certainly don't know everything, and am constantly learning. But not from threads like this one. I can change the SWR on a transmission line by renormalizing my Smith chart? Is that a "logical point" I'm running away from? The reason I post in the first place isn't to try and convince the party I'm directing my posting to. Nor is it an ego trip. What I hope to accomplish is to provide a counterpoint to what I see as incorrect information. This group is read by a very large number of "lurkers" who seldom or never post. I know this for a fact, because many of them introduce themselves to me at Dayton and other places I appear publicly. When someone posts misinformation on this group, I try to present what I consider to be correct. There are people, some of whom post here, who won't be convinced regardless of the evidence. It's as much a total waste of time to argue with those people as it is to try and convince someone his religion is wrong. When I encounter a person like that, I'll post my point of view, present what evidence I can, then withdraw. I have much better things to do than continue flogging a dead horse. I feel that the lurkers, who are really the people I'm addressing, should be able to make up their minds on the basis of what's been presented. There's more than ample evidence to back up what I've said that's easily available to anyone with a real interest in learning. Anyone who really cares and is willing to invest even a modicum of effort can search out the information and reach a conclusion. You've chosen not to go to that effort(*), but rather interpret what I say in a way that suits your preconception. Sorry, I just won't waste more of my time trying to talk you out of it. If lurkers are convinced by your arguments and find them more compelling than the ones I've made, then so be it. I've done what I can. To continue posting over and over again the same thing isn't my choice of a way to use my time. There are people who feel that the person who posts the last message "wins", and so anyone who withdraws has "lost". You can see the result of this philosophy in the threads that have run to literally hundreds of postings without ever resolving the issue. It's simply not a game I play. Roy Lewallen, W7EL (*) For example, have you ever looked up the equations for calculation of SWR? Noticed that there's no term for the source impedance? And no term for your Smith chart normalization factor? |
Richard Clark wrote:
W5DXP wrote: The forward part and re-reflected part from the source are coherent and traveling in the same direction so they cannot be separated for measurement purposes. That separation is unnecessary as it represents a sufficiently fixed value for every moment beyond the first millisecond for usual applications. So what part of the generated signal is actually generated and what part is merely re-reflected energy? -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
On Fri, 15 Aug 2003 15:58:08 -0500, W5DXP
wrote: Richard Clark wrote: W5DXP wrote: The forward part and re-reflected part from the source are coherent and traveling in the same direction so they cannot be separated for measurement purposes. That separation is unnecessary as it represents a sufficiently fixed value for every moment beyond the first millisecond for usual applications. So what part of the generated signal is actually generated and what part is merely re-reflected energy? Hi Cecil, Separation is unnecessary after the first millisecond. Impedance from that time on is sufficiently fixed to measure. You can answer your own question from the resultant, but it is of no particular interest in the determination. 73's Richard Clark, KB7QHC |
Richard Clark wrote:
Separation is unnecessary after the first millisecond. Impedance from that time on is sufficiently fixed to measure. You can answer your own question from the resultant, but it is of no particular interest in the determination. I don't have a question, Richard, I am trying to answer yours. What difference can it make in the SWR if "separation is unnecessary"? -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Tarmo Tammaru wrote:
Try this out. You have an amplifier of unknown source impedance connected through a directional meter to a 1/4 wavelength line that is shorted at the far end. Without knowing about SWR, you know (because you are a ham radio operator) that the amp is driving an infinite impedance, and delivering 0 power. Now if you adjust the amp so the meter reads 100W in both directions, the amp is still delivering 0 power, and 100% of the reflected wave is rereflected. Where did the 100W come from? the amp delivered it during the first 1/2 cycle after it was turned on. It didn't "know" the line was shorted until the first reflection came back. How do you know that the amplifier is not "delivering" 100W of forward power and dissipating 100W of reflected power (as it would with a circulator+load)? How do you prove that the impedance looking back into the amp is zero, infinity, or purely reactive? Doesn't it have everything to do with the arbitrary definition of "delivered" which doesn't necessarily match reality? -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
"W5DXP" wrote in message ... How do you know that the amplifier is not "delivering" 100W of forward power and dissipating 100W of reflected power (as it would with a circulator+load)? Aha, I will give you special dispensation. You are allowed to bias the amp for true class B, and measure the DC drain current. No way is the drain current going to be different from the case of a plain open circuit. Power (RF) can not be more than Power (DC) How do you prove that the impedance looking back into the amp is zero, infinity, or purely reactive? Use any RF impedance meter you want. Doesn't it have everything to do with the arbitrary definition of "delivered" which doesn't necessarily match reality? I think the only alternative in un nice; namely that there is no reflection in steady state Tam/WB2TT -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Richard Clark wrote:
In the scope of Z determination, I am wholly unconcerned with what precedes the first millisecond. Perhaps you should phrase your question along other lines. You theorized uncertainty because of reflections from all directions. You seem now to be "unconcerned" about that uncertainty. -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Tarmo Tammaru wrote:
I think the only alternative in un nice; namely that there is no reflection in steady state There are an infinite number of possibilities between the rails of 100% re-reflection and zero re-reflection. I suspect a PA obeys the rules of the wave reflection model set forth in Ramo & Whinnery. -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
On Fri, 15 Aug 2003 21:19:19 -0500, W5DXP
wrote: Richard Clark wrote: In the scope of Z determination, I am wholly unconcerned with what precedes the first millisecond. Perhaps you should phrase your question along other lines. You theorized uncertainty because of reflections from all directions. You seem now to be "unconcerned" about that uncertainty. Hi Cecil, In the first millisecond, yes. Do you have a question about it following that point in time? 73's Richard Clark, KB7QHC |
Dr. Slick wrote:
Well, this is clearer than what you wrote earlier, as they have included the "on port 1" part, whereas you stated just "incident reflected", which told us nothing. I stated: If ..., PA's will always re-reflect 100% of the incident reflected power. That's obviously reflected power incident upon the PA. -- 73, Cecil http://www.qsl.net/w5dxp -----= Posted via Newsfeeds.Com, Uncensored Usenet News =----- http://www.newsfeeds.com - The #1 Newsgroup Service in the World! -----== Over 100,000 Newsgroups - 19 Different Servers! =----- |
Ok, I'll try once more.
Look again at what I wrote earlier: No, you won't see a change in the SWR at point 1 as you change the PA impedance. All the fiddling you do with your Smith chart just won't make it happen. Sorry. The SWR, voltage, current, impedance, power, reflection coefficient, waves, or anything else don't change in response to your Smith chart exercises. To which you replied: Absolutely incorrect. Time for you to review your Smith Chart again. Here's an experiment for you to try. On your workbench, measure the characteristic impedance of a cable, or connect it to a source or load and measure the SWR on it. Now go over to your desk, take out your grease pencil, and change the reference impedance of your Smith chart. Go back to the bench, check the SWR and the cable Z0. Has it changed? If the answer is yes, I'll admit having a serious shortcoming in my knowledge of the power of the Smith chart. If *you'll* review the Smith chart again, you'll find that *if* you set the reference impedance to the Z0 of the transmission line you're analyzing, then the SWR, impedances, and so forth that your read from the Smith chart are correct. If you set the reference impedance to some other value, an SWR read from the chart certainly isn't the SWR on the transmission line. Z0 in the equation you quote refers to the characteristic impedance of a transmission line. If you change the impedance of the transmission line, given the same load impedance Zl, you change the SWR, as the equation indicates. (Changing the reflection coefficient changes the SWR on the line.) But changing the reference value on your Smith chart doesn't change the characteristic impedance of the cable. Now let's see what you wrote this time: Dr. Slick wrote: . . . When you change Zo, you change the normalized center of the Smith, and therefore the Ref. Coeff. and SWR, looking into the same load. Yes. When you change the line's Z0, you should change the normalization of the Smith chart. But changing the normalization of the Smith chart doesn't magically change the cable dimensions to give it a correspondingly new Z0. (Or does it? Something I'm missing here?) The Smith Chart is an extremely powerful graphical RF tool, which has become part of the basis for communicating in the RF world, as well as a standard for displaying impedance on most RF measuring devices. You need to read up on this if you want to understand me. Thank you, although I'm not an expert at using the Smith chart, I know my way around the circle. But perhaps more study would reveal the mechanism by which changing the chart reference causes spontaneous redimensioning of the cable. Even when the other person is correct too? I think it IS an ego trip if you can't admit someone has a point, and are too scared to discuss it further for fear of looking weak or exposing a lack of knowledge about something. All right, I was scared to admit that you have a point, that all you have to do to change a line's SWR is to go over to the desk and change your Smith chart. But I've overcome my fear now, and am exposing my ignorance, or weakness as you say, for all to see. I'd like to learn more about the mechanism, though. When you renormalize your Smith chart, is the line's Z0 changed by a spontaneous change in inner conductor diameter, outer conductor diameter, dielectric permittivity, or some combination? Is it some sort of telekinesis, or perhaps something to do with Chi? I never could quite get the hang of Feng Shui, so maybe that's it. I can't find any reference to the phenomenon you're claiming in my engineering texts -- perhaps an occult or New Age bookstore would be more fruitful? From Pozar's Microwave Engineering: Reflection Coefficient = (Zl-Zo)/(Zl+Zo) Where Zl is a purely real load impedance, and Zo is the purely real characteristic impedance reference. When you change Zo, you change the normalized center of the Smith, and therefore the Ref. Coeff. and SWR, looking into the same load. Almost right, but a misplaced "therefore". The reflection coefficient and SWR don't change because you change the Smith chart normalization. They change because you've changed the cable's Z0. If you know how to use a Smith chart correctly, you can then go and renormalize your Smith chart to the new Z0 (as you've said), and from it you can read what the new reflection coefficient and SWR are. Or you can forget the Smith chart altogether and measure them, or calculate them from an equation. But if you normalize your Smith chart to some value other than the cable's Z0, you're no longer reading the transmission line impedances and SWR from it. It sure looks to me like you're confusing reality with what you read from your arbitrarily normalized Smith chart. They're not the same. Just like the reading on an SWR meter isn't the same as the SWR on a transmission line of a different impedance. And actually for exactly the same reason. Hm, maybe that provides even one more way to say it. Set up your source, cable, and SWR meter like you have it in your earlier posting. Replace the 50 ohm SWR meter with a 75 ohm SWR meter. Renormalize your Smith chart for 75 ohms. Presto! The SWR meter reads the same as the Smith chart! But y'know what? YOU DIDN'T CHANGE THE SWR ON THE TRANSMISSION LINE. Put the 50 ohm SWR meter back in just on the load side of the 75 ohm one. It reads just the same as it did before, and it's reading the actual SWR on the 50 ohm line between it and the load. Again, changing the Smith chart's normalization does not change a cable's Z0 or SWR. Feng Shui and voodoo notwithstanding. A Smith chart is simply (not to disparage in any way its ingeniousness or utility) a polar plot of reflection coefficient on a special scale. (The trick is, of course, generating the scale, an exercise in conformal mapping I recall doing in fields class.) Check it out -- measure the length of the vector from the center to any impedance point (with the chart radius equalling one), and the angle from the main axis (zero ohms being the positive direction), and you'll see you have the reflection coefficient. I can't figure out any more ways to say this. I am always interested in learning. I don't use a Smith chart a great deal, but I know Tom Bruhns, a regular poster here, does. I'll gladly defer to him on issues of Smith chart use, and hope he'll feel free to correct me on any errors I've made in the above discussion -- as he has a number of times in the past, and which I've greatly appreciated. Walt Maxwell, an occasional poster, is truly a Smith chart expert, and I'd also welcome any corrections or amplifications he'd care to make. Or any other knowledgeable Smith chart user. Roy Lewallen, W7EL |
| All times are GMT +1. The time now is 12:31 PM. |
Powered by vBulletin® Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
RadioBanter.com