Talk:Distributed-element filter

Rewrote the lead
The old lead had some problems.
 * A distributed element filter is an electronic filter designed for frequencies above the VHF band or thereabouts. At these frequencies the wavelengths of the operative signals are shorter than the size of the device being constructed and it is no longer possible to use the more familiar lumped element model. Filters and other passive circuitry are instead designed using the distributed element model in which the components of the circuit are treated as being transmission lines, which indeed, they effectively are.

Note that at the top of VHF band the wavelength in air is ~1 meter. Lumped elements crap out about here but clearly "wavelengths of the operative signals" are not "shorter than the size of the device being constructed". I've also added some helpful wikilinks. I think the new language flows a bit better as well but YMMV. JPatterson (talk) 20:34, 15 February 2010 (UTC)

Things to improve

 * Front figure: "for suppressing non-TEM modes of transmission" needs a cite. Adding a note on image width would help understanding the size - not everyone knows this connector type. Image file needs a better description, preferably explaining why rubber seal (moisture ?).
 * I have addressed the "through holes for TEM mode" issue. Image size is a bit problematic; it is not my image so I cannot retake it with a rule included.  It would, of course, be possible to estimate the size from some of the circuit structures, but that is almost the definition of OR. I am against explaining the rubber seal (it is to give the enclosure an IP rating because the product is intended for outdoor use) - this is going way off-topic; there are any number of features that could be explained here.  The only one mentioned, the through holes, are only described because they can easily be mistaken for a filter structure.  Sp in ni  ng  Spark  20:48, 8 April 2010 (UTC)
 * Regarding the via fences: there will be an internal metal lid cast with a pattern of walls on the underside, which corresponds to the pattern of via-fences. This lid is clamped down tightly onto the circuit by a large number of bolts - notice all the screw holes associated with the via fences (it can't have been screwed down very well in this example as the walls have not made a visible indentation in the metallization). The via fences connect the base of the walls to the ground-plane, so forming a number of near perfectly isolated enclosures, communicating only where the microstrip passes through a notch in the cast walls. This internal lid is in addition to a simple external lid which makes a seal with the rubber gasket.--catslash (talk) 10:41, 29 April 2010 (UTC)
 * Regarding the photo: I've just put a more close-up view of a (just) a filter at File:Microstrip-bandpass-filter.jpg. I was going to use this to illustrate edge-coupling on the microstrip page, but never got around to writing about it. I haven't put the picture on this page, as you may feel that the existing one is better - take a look and see what you think. --catslash (talk) 16:35, 29 April 2010 (UTC)
 * Nice picture. That would be great for the parallel-coupled lines section.  I don't think we should expand on via fences in this article, it is only mentioned for clarification purposes.  I have red-linked it with the intention that an article should go there someday.  Sp in ni  ng  Spark  07:35, 30 April 2010 (UTC)
 * OK. I just felt that the phrase for suppressing non-TEM modes of transmission though strictly speaking true, is rather misleading. --catslash (talk) 11:55, 30 April 2010 (UTC)


 * There is some imbalance between figure captions - one cites references and others don't. I probably understand why, but this could be picked up at FAC.
 * There is usually no need to give citations to image captions unless they introduce facts not in the main text. All the figures here are described in the text where a citation is given.  I have made an exception for figure 2 because it is a composite from a number of different sources and I wished to make it clear which filter structure came from which source.  Sp in ni  ng  Spark  20:30, 9 April 2010 (UTC)


 * Add history subsection?
 * I am actively looking into this. The history of the various technologies (co-ax, stripline, microstrip, waveguide etc) is well documented, mostly of them starting in radar in WWII and the filters in these technologies followed shortly after their first use.  However, sources dealing with distributed element filtering as a general topic are a bit harder to find and I need to see if enough material can be gathered to write a worthwile section.  Sp in ni  ng  Spark  20:30, 9 April 2010 (UTC)


 * I have now started a history section. There is still a bit more that could go in, but I think the point is now addressed.  Sp in ni  ng  Spark  18:13, 11 April 2010 (UTC)


 * Need to clarify or wikilink "quarter wavelength stubs", "LC (anti-)resonators" (even LC needs some wikilink at first appearance), "impedance or admittance transformers" Materialscientist (talk) 04:21, 6 April 2010 (UTC)
 * done  Sp in ni ng  Spark  12:21, 11 April 2010 (UTC)

Prose comments by Cryptic C62

 * "Its purpose is to allow a range of signal frequencies to pass, but to block others." What is a "signal frequency"? I would be satisfied with a simple wikilink, but there doesn't seem to be an article with such a title.
 * Wikilinked to frequency spectrum.  Sp in ni ng  Spark  21:58, 17 June 2010 (UTC)


 * "There is no precise frequency beyond which distributed element filters must be used but they are especially associated with the microwave band (wavelength less than one metre)." The use of "beyond" seems a bit imprecise. Either "above" or "below" should be used instead. Also, why does the main sentence discuss frequency while the parenthetical uses wavelength? I'm aware of how the two are related, but this will probably be a source of confusion to those who are not already familiar with the terminology.
 * Fixed "beyond". I think the mixed use of frequency and wavelength is unavoidable in this article.  There are good reasons why signals are invariably described in terms of frequencies, but circuit structures are described in terms of wavelengths.  Length, in fact, is an additional element in distributed circuits, although the article has not delved into this aspect (but maybe a future one will).  The way to deal with this is head-on in the lede, establishing that frequency rises as wavelength falls.  Math markup could be used to give the exact relationship but I have tried to avoid that as much as possible in this article and I don't think that it is necessary for understanding.  On a pedantic note, one metre is the formal boundary of microwaves.  Many books will give this as 300 MHz, but the two are, of course, only approximately equal - and not equal at all in a high permittivity medium.  Sp in ni  ng  Spark  21:58, 17 June 2010 (UTC)


 * "All of the well-known filter classes used in lumped element designs (Butterworth, Chebyshev, etc.) can be implemented using a distributed element approach." I am of the opinion that the phrase "well-known" should always be avoided in an encyclopedia. In this case, the sentence implies that the lesser-known classes can't be implemented using the distributed approach, which is silly. Perhaps just "Most filter classes used in lumped element designs (such as Butterworth and Chebyshev) can be implemented using a distributed element approach." would work better.
 * I think we can safely use "all filter classes" here.  Sp in ni ng  Spark  18:03, 18 June 2010 (UTC)
 * I strongly believe that the first and third paragraphs of General comments should be contiguous, as they both discuss the actual topic whereas the second paragraph discusses the article itself. Whether the order should go 2-1-3 or 1-3-2 is up to you. Also, if the second paragraph discusses the format of the diagrams, shouldn't the accompanying image also be such a diagram?
 * 1-3-2 looks good to me. The reason the figure 2 image is in the article has nothing to do with the paragraph it is attached to, that is just a convenient place to park it.  The article benefits from a photograph (as opposed to a diagram) of a DE filter.  The lede image is good in that it provides a variety of filter structures, but it is busy with a lot of non-filter circuitry as well.  This image shows a filter by itself.  Such an image is best early in the article, but the lede position is already taken, and besides, the alternative of placing it in the parallel-coupled section is not viable because it has no room for another image.  Sp in ni  ng  Spark  23:05, 18 June 2010 (UTC)


 * "On the other hand, antenna structure dimensions" I'm thinking "antenna" would be a good thing to wikilink, but there are so many choices for articles to link to! Your call.
 * antenna (radio).  Sp in ni ng  Spark  23:05, 18 June 2010 (UTC)


 * "...can be chosen by design to function, to a first approximation, like lumped inductors..." Yikes, the use of "to a first approximation" makes this sentence way more technical than it needs to be, especially since this is the only instance of "first approximation" in this entire article. How about "...can be chosen by design to function in roughly the same manner as lumped inductors..." ?
 * Don't like "roughly", too imprecise. We could have "...can be chosen by design, over a prescribed frequency range, to have approximately the frequency function of a lumped inductor..." but that is hardly an improvement.  Sp in ni  ng  Spark  18:03, 18 June 2010 (UTC)
 * How about "...can be chosen by design to serve as approximations for lumped inductors..."? --Cryptic C62 · Talk 19:51, 18 June 2010 (UTC)
 * That's good.  Sp in ni ng  Spark  23:05, 18 June 2010 (UTC)


 * "The development of distributed element filters was spurred on by the wartime need of the military for radar and electronic counter measures during World War II." I suggest replacing "by the wartime need of the military" with "by the military need" or "by the military's need". The word "wartime" is very redundant given that the sentence ends with "during World War II".
 * Sounds good to me.  Sp in ni ng  Spark  23:05, 18 June 2010 (UTC)


 * "Development of distributed element filters was underway in the years before World War II." There is a bit of ambiguity with the phrase "the years before World War II". Some readers might interpret this as meaning the late 1930s, some (milhis buffs) might interpret it reaching all the way back to the end of World War I in 1918. Also, "was underway" is a useless phrase. Saying when something "was underway" gives no indication of when it started.
 * I don't really see the problem here. The opening sentence is a general statement that is then expanded on by the following sentences.  The nature of the work is stated and the papers referenced.  The probable birth is clearly identified as 1927, so no one should be thinking 1918.  The real drive to develop this technology, however, was during WWII which is why the opening sentence references the war.  Sp in ni  ng  Spark  23:23, 20 June 2010 (UTC)


 * "The introduction of printed planar technologies greatly" Should "printed planar technologies" link to Printed circuit board or am I mistaken?
 * That's a difficult one to answer, I would say no - it is linked already earlier in the article where it is made clear that pcb technology can be used to make planar DE circuits. That is a better link as pcb is linked directly, linking "printed planar technologies" has the potential to confuse the reader as the pcb article only  discusses these briefly.  Sp in ni  ng  Spark  23:23, 20 June 2010 (UTC)


 * "He published a set of transformations known as Kuroda's identities in 1955, but his work was written in Japanese and it took several years to reflect his ideas in the English-language literature" How about "He published a set of transformations known as Kuroda's identities in 1955, but his work was written in Japanese and it took several years for his ideas to enter the mainstream English-language literature." instead?
 * I agree except that "mainstream" should be dropped, implying as it does that one has to publish in English to be mainstream. The current wording of this sentence is due to Materialscientist - it may be worth checking his original edit summary to see if he had a particular reason for stating it that way.  Sp in ni  ng  Spark  23:23, 20 June 2010 (UTC)
 * I disagree that the use of "mainstream" implies that one must publish in English to be mainstream. What I intended to convey was that within the English-language literature, there is a mainstream which it took a while for Kuroda's work to break into. However, rather than fussing around with word order, perhaps both issue can be sidestepped with a different rewording altogether: "He published a set of transformations known as Kuroda's identities in 1955, but his work was written in Japanese and it was several years before his ideas were incorporated into the English-language literature." Better? --Cryptic C62 · Talk 15:31, 22 June 2010 (UTC)
 * Fine by me.  Sp in ni ng  Spark  21:25, 22 June 2010 (UTC)


 * "There are many component forms used to construct distributed element filters, but all have the common property of causing a discontinuity on the transmission line." I'm not sure of the exact meaning of "discontinuity" in this context, which I suspect will be the case for many other readers. At first I thought it meant a physical break in the wiring, but after reading the rest of the paragraph, that doesn't seem right. Reading transmission line doesn't help either, as that article doesn't use the word "discontinuity" even once!
 * The sense continuity is being used here is in the mathematical sense of a continuous function. The transmission parameters are normally continuous functions of distance along the transmission line.  A discontinuity in the transmission parameters is caused by a sudden change in line geometry (this can be a narrow physical gap, but also a step change in conductor width will do it, or attaching a component at that point).  There is a step change in the transmission parameters at the point of the discontinuity.  Such functions are said to be piecewise continuous.  I have endeavoured to avoid introducing unnecessary mathematical concepts into this article and it would also be sidetracking the article somewhat to give lengthy explanations of DE theory, there are other articles for doing that.  Suggestions how this could be clarified while keeping the focus on filters are welcome.  Sp in ni  ng  Spark  18:03, 18 June 2010 (UTC)
 * Again, how about a footnote explaining what you just explained to me? I would also be happy simply linking the term, but the only articles available are Continuous function and Classification of discontinuities, neither of which are particularly helpful. --Cryptic C62 · Talk 21:53, 20 June 2010 (UTC)
 * I'll work on some footnotes later in the week.  Sp in ni ng  Spark  23:23, 20 June 2010 (UTC)
 * Oh yes, just remembered, I wrote an article on this last year:Reflections of signals on conducting lines, will that do?  Sp in ni ng  Spark  23:30, 20 June 2010 (UTC)
 * I don't fully understand the material that's being linked to, but if you're confident that it's the best way to explain a discontinuity, then I'm all for it. I've gone ahead and added the link. --Cryptic C62 · Talk 15:31, 22 June 2010 (UTC)


 * "Early stripline directly coupled resonator filters were end-coupled," Very awkward wording. I expected "stripline" to be the subject of the sentence and started hunting for the verb after "directly" and got very confused. Is "stripline" being used as an adjective here? Should "directly coupled" be interpreted as a single adjective ("directly-coupled")? --Cryptic C62 · Talk 15:31, 22 June 2010 (UTC)
 * Arguably, all three words should be hyphenated "directly-coupled-resonator", cf this very relevant paper.  Sp in ni ng  Spark  21:25, 22 June 2010 (UTC)


 * "by telecommunications companies to provide the "backbone" of their networks." Direct quotations should be followed by a citation. Did this come from Huurdeman or from some other source?
 * Perhaps I should have used italics instead of quotes. Or possibly wikilink backbone although sadly, the author of that article has written exclusively about computer networks even though the usage of backbone in telecommunications is much more general than just IT.  Sp in ni  ng  Spark  21:25, 22 June 2010 (UTC)


 * "...inductance and capacitance (and resistance, but that rarely features in filter designs) are distributed along the length of conductors, and the two..." It seems contradictory to have a parenthetical comment (intended to give "extra" details) in the lead (intended to give a broad overview of the topic). Unless you can think of a way to work it into the sentence without adding a huge amount of clunky commas, I suggest dropping the parenthetical altogether. Also, this seems somewhat contradictory with the opening sentence. Perhaps the opening sentence could be reworded to "in which capacitance, inductance and sometimes resistance" to avoid implying that all three are equally important.
 * Decline to do this. All three are equally important and resistance absolutely has to be mentioned.  The filter designer does not want it: it is parasitic to his design as they say in the terminology, but he is going to get it anyway.  It does not feature in the design in the sense that the designer has deliberately introduced capacitance and inductance, but resistance has crept aboard uninvited like a ship's rat.  I might also mention that there are some rare filter designs that do deliberately incorporate resistance, see this patent by Hendrik Bode for instance.  The lede should be readable by anybody, as reviewers never cease to remind me, and we should not assume prior technical knowledge.  Equally, we should not assume that the reader does not know what the basic three passive elements are and feel we are free to leave one out thinking it will not be missed.  In fact, a reader who has come to this article probably will know this, and almost everybody has at least heard of resistance and knows that it is an element of electrical circuits.  To not mention resistance will leave the reader asking what has happened to it.  I will, of course, welcome rewording to make it less clunky.   Sp in ni  ng  Spark  21:58, 17 June 2010 (UTC)
 * Very well explained. The question now is how to make it less clunky. I think the phrase "but that rarely features in filter designs" does not adequately explain what you've just told me, nor would it even be possible to concisely explain the matter without cluttering up the lead. Why not just "...inductance, capacitance and resistance are distributed..."? --Cryptic C62 · Talk 01:51, 18 June 2010 (UTC)
 * Because the lede has already opened with "capacitance, resistance, inductance". It then goes on to say LE filters usually consist of capacitors and inductors.  The secondary point is that resistance is not usually "designed in" with filter circuits.  To not say this leaves the reader wondering where it has gone in this story.  Sp in ni  ng  Spark  18:03, 18 June 2010 (UTC)
 * Two options that have not been explored: First is to use a footnote, which would allow you to explain the matter in more sufficient detail without clogging the lead. The second would be to simply add another sentence after the end of this one. The sentence that currently follows it, ("There is no precise frequency above...") doesn't seem closely related to this sentence, so the flow of the paragraph wouldn't be lost. --Cryptic C62 · Talk 19:51, 18 June 2010 (UTC)
 * I have tried restating this in a way which I think addresses your point.  Sp in ni ng  Spark  11:31, 26 June 2010 (UTC)


 * "The increasing miniaturisation of electronics means that circuits which were once large compared to λ no longer are." This (quite wrongly) implies that an individual circuit will physically shrink over time. Perhaps "The increasing miniaturisation of electronics means that, unlike their older counterparts, modern circuit elements are generally not large compared to λ." ?
 * I sincerely doubt that anyone will put that reading into the sentence, but I have no objection to a rewording in principle. Your suggestion, however, is confusing components/elements/circuits and it is also incorrect that "modern circuits are generally not large compared to λ."  There is a trend, but there is still a point (implementation technology dependant) beyond which DE is still required.  Sp in ni  ng  Spark  23:05, 18 June 2010 (UTC)
 * My suggestion was merely an attempt to make it clearer that it is not the individual circuits themselves that are shrinking, but instead the designs for the circuits. Since I'm quite evidently not an expert on the subject, do you have any ideas for how to make this distinction clearer? Also, it would be nice to rid the sentence of the awkward ending "no longer are." --Cryptic C62 · Talk 15:31, 22 June 2010 (UTC)
 * Reworded to get rid of the ugliness. The idea that this could be misread as circuits physically shrinking is pedantic nonsense.  "The increasing streamlining of automobiles" would be understood by no one as meaning the "sit up and beg" radiator grille of ones Rolls-Royce Silver Shadow has slowly been worn away by air resistance over the years.  Sp in ni  ng  Spark  12:24, 26 June 2010 (UTC)


 * "Such applications are part of large capital investment programs. However, mass-production..." The use of "however" should set up a contrast between two ideas or two time periods. I fail to see such a contrast between these sentences.
 * The contrast was supposed to be between the initial applications where cost was not the overriding issue (the major costs of setting up a telecommunications network are getting all that kit to the top of obscure hills in inaccessible places and building towers to house it all. A piddling little electronics gadget does not really enter the economics, even if it is not made in the most cost effective way) and equipment for the consumer market where shaving fractions of a penny from unit costs can be significant for profitability because of the high volumes.  Sp in ni  ng  Spark  21:25, 22 June 2010 (UTC)
 * Then I think "are" should be changed to "were" in the first sentence, yes? --Cryptic C62 · Talk 13:18, 24 June 2010 (UTC)
 * ok  Sp in ni ng  Spark  18:45, 24 June 2010 (UTC)


 * "(often quarter wavelength)" I'm thinking that this can be replaced with "(0.25λ)" to make it consistent with the General comments section.
 * it can be, although some reviewers might prefer it in words. I would write it as "(often λ/4)"; stating as a fraction is more normal for mutiples of λ/4 and make the function clearer; "often" is still needed because other lengths can have a useful function as well.  Sp in ni  ng  Spark  18:45, 24 June 2010 (UTC)


 * "These multiple element filters can implement any of the filter families..." I'm confused. The lead explicitly states that all filter classes can be implemented using a distributed element approach. This is the only instance of "multiple element filters" in the entire article, leading me to believe that "these multiple element filters" refers to the "hybrid mixture" in the previous sentence.
 * You are confused because element is not being used with the same meaning here. This is bad, let me think of a better way of writing this.  I guess component will have to be used but I don't entirely like that because stripline circuits are made as a single piece, not from components joined together.  Sp in ni  ng  Spark  18:45, 24 June 2010 (UTC)
 * I've fixed this at the expense of introducing some more precise mathematical terminology - but this was probably needed in the article anyway.  Sp in ni ng  Spark  14:28, 26 June 2010 (UTC)


 * "The generic term stripline in modern usage usually refers to the form then known as triplate" Let me just check to see if I'm reading this correctly: Barrett invented something which he called "stripline". That particular technology received competition from another technology called "triplate". Today, the term "stripline" no longer refers to Barrett's technology but to triplate instead. Correct?
 * Not quite as clear cut as that. Stripline has become a generic term in the same way as hoover refers to vacuum cleaners and xerox machine means a photocopier regardless of who the manufacturer really is.  Most of the time triplate would be meant in modern usage but it still could be applied to Barrett's original form, or some other variant.  Sp in ni  ng  Spark  23:23, 20 June 2010 (UTC)
 * I see. Such subtleties are often difficult to accurately convey in an article. How about Replacing the aforementioned sentence with the more general "In modern literature, stripline has become a generic term to describe several variations of the technology"?
 * While that is true, I would still like to say that stripline unqualified most often means the triplate form. Some authors use stripline and triplate interchangeably, some will use a qualifying adjective where distinction is necessary, such as air stripline, triplate stripline or dual stripline for instance.  Sp in ni  ng  Spark  21:25, 22 June 2010 (UTC)
 * If triplate is the more common meaning of the modern term stripline, does that mean that triplate has become the more popular format? If so, how about "Though triplate has since become the more popular format, stripline is often used as a genericized trademark." or something like that? --Cryptic C62 · Talk 01:17, 28 June 2010 (UTC)
 * That would not really be accurate. If anything, microstrip is now the most popular format, and although a planar strip, the term stripline is rarely, if ever, applied to it.  It is more usual that stripline and microstrip are counterposed, both terms being used generically.  Unless you have a source that states x, y, or z is the most popular format I would be very uncomfortable making any such statement.  It might be true that triplate is the most popular form where the medium above and below the strip are identical (this is what distinguishes it from microstrip which is air dielectric above the strip) and I can verify that was true in the time-period I was involved in design in low-power applications.  But it may not be now, and high-power applications usually require an air-dielectric format.  I appreciate you would like to say something more definite, but as I say, I would not want to make any sweeping statements about popularity without references.  Sp in ni  ng  Spark  17:02, 1 July 2010 (UTC)


 * "The simplest structure that can be implemented is a step in the characteristic impedance of the line" Suggest adding "in a distributed element filter" after "implemented", unless you meant that an impedance step is literally the single simplest structure ever in the history of the universe.
 * I'm not really convinced that is necessary given the context of the article, but it is not objectionable.  Sp in ni ng  Spark  18:45, 24 June 2010 (UTC)
 * Actually, now that I think about it, there is an ambiguity here that may need correction: does this sentence mean the specific "the simplest structure that can be implemented in a distributed element filter" or the more general "simplest structure that can be implemented in the distributed element model"? --Cryptic C62 · Talk 01:17, 28 June 2010 (UTC)
 * Saying that of "distributed element model" would not make sense, the model is for modelling an implementation, not for implementing a structure. You could generalize the statement to "distributed element circuit", but is not really any ambiguity here.  In any case, simplest is not being used in any formal technical sense; it is being used in a loose and woolly sense and I would be unable to give you a firm definition if challenged.  Sp in ni  ng  Spark  17:56, 1 July 2010 (UTC)


 * "behave as shunt LC (anti-)resonators" Very awkward construction, and the meaning isn't entirely clear. Suggest replacing with "behave as shunt LC resonators or antiresonators", assuming that that is the correct interpretation.
 * No, resonator or antiresonator is not meant. The reason for the clunky construction is that antiresonator is a bit of a rare term, although it does exist in a large number of sources.  In lumped-element circuits it is more common to refer to series resonators and parallel resonators (which are resonant and anti-resonant respectively).  In DE circuits the series and parallel classifications have no meaning, unless we are discussing the LE equivalent of a DE circuit.  This was my attempt at using the precise antiresonator while preserving the resonator terminology familiar to those who understand LE circuits.  On second thoughts using just antiresonator would be simpler and clearer.  Sp in ni  ng  Spark  17:56, 1 July 2010 (UTC)


 * "and consist of a coil of wire, on a former and core" What is a "former"? The article former is of no help.
 * The meaning of former is bobbin and the meaning of core is magnetic core.  Sp in ni ng  Spark  17:56, 1 July 2010 (UTC)


 * The following two sentences seem to contradict each other: "While open-circuit stubs are easier to manufacture in planar technologies, they have the drawback that it is impossible to create an ideal open circuit (see figure 4(b)), often leading to a preference for short-circuit stubs (one can always be used in place of the other by adding or subtracting λ/4 to or from the length)" and "Coupled lines tend to be preferred in planar technologies, where they are easy to implement, whereas stubs tend to be preferred elsewhere." If open-circuit stubs are easy to manufacture in planar technologies, then how can it be that coupled lines are preferred in planar technologies because they are easy to implement?
 * There is no contradiction; coupled lines are preferred over stubs especially for higher order (which means more complex) filters if all other considerations are equal. Where stubs are being used (and there can be good reasons for using stubs) the open-circuit version is easier to manufacture, but has less accuracy than the short-circuit version.  The example in the lede image shows these forms: in the full monty wide-band high-order filters, the designer has used parallel coupled resonators.  For the simple bias filters where only one resonator is used, s/he has chosen to use a stub (parallel coupled lines would not actually work at all in this application because dc is not passed by such a structure).  In both cases s/he can't be asked to bother with the complication of providing short-circuits in the filter (again, in the case of the bias filter this is ruled out in any case by the need to pass dc) and has settled for the less accurate open-circuit version.  Sp in ni  ng  Spark  17:56, 1 July 2010 (UTC)
 * I think the most pertinent piece of information which you just explained to me is that coupled lines are preferred for higher-order filters, as this is not made clear in the article. --Cryptic C62 · Talk 00:06, 4 July 2010 (UTC)
 * I think the point you are missing is that the first sentence you highlight is comparing different types of stub in a paragraph discussing stubs. The second sentence is comparing stubs as a whole to coupled lines as a whole in a completely different section of the article.  Sp in ni  ng  Spark  09:47, 4 July 2010 (UTC)


 * "The purpose of incorporating resonators is to improve the stopband rejection." What does "rejection" mean in this context? There doesn't appear to be an article about the subject.
 * There is an article, band rejection, but a not terribly informative stub. I leave it to you to decide if it is worth a wikilink.  The meaning of rejection is attenuation and a key measure of this in most modern filter design methods is insertion loss.  In filter designs which do not deliberately incorporate resistive elements (the vast majority) the rejected part of the signal is not dissipated within the filter, but rather, is reflected back towards the source - it is literally "rejected" by the doorman.  Sp in ni  ng  Spark  20:11, 8 July 2010 (UTC)


 * "Two radial stubs in parallel (one either side of the main line) is called a butterfly stub" It is grammatically incorrect to use a plural noun and the singular "is". I suggest prepending the sentence with "A pair of" or a similar phrase. The same is true of the sentence that follows this one, in which case "A group of" might work well.
 * I recuse myself from pronouncing on questions of grammar. There is no problem with your suggestions from a technical point of view.  Sp in ni  ng  Spark  20:48, 8 July 2010 (UTC)


 * "The purpose of this is to prevent coupling between adjacent stubs which detracts from the filter performance." What does "performance" mean in this context? The word is used later on in the article where it is clear that it refers to insertion loss. Perhaps a similar construction can be added here.
 * The exact effect depends on the precise design details of the filter and the exact amount of coupling between each resonator. The most likely undesirable effect is a reduction in stop-band rejection.  "Ripples" or "dips" can also be caused in the passband.  All the source book says on this is "such coupling can alter the filter's frequency response in undesired ways" which is a fair summary of these effects.  It is also worth noting that coupling between resonators can produce useful effects, notably to increase the steepness of the filter "skirt" at the expense of some loss of rejection (which can be made up for with additional resonators).  This is not much done in these kind of filters but was once a popular technique in mechanical filters and I have explored this idea in more depth in that article.  Sp in ni  ng  Spark  20:48, 8 July 2010 (UTC)
 * So... do you think there any way to concisely clarify the meaning of "performance" within the article? If it's a broad concept here, then there's not really anything we can do about it, but if it can be replaced by a more informative word, that would help. --Cryptic C62 · Talk 18:50, 10 July 2010 (UTC)
 * We could say something like "...detracts from the filter performance by altering the frequency response." All the effects I noted above are changes to the frequency response, both the undesirable and desirable ones.  This is about all that can be said from the existing source, any more depth will require new sources.  Frequency response can also be read as including phase and delay effects as well as amplitude, which are also changed by any coupling, but it is usually amplitude that is of primary concern.  I am not keen to go too deeply into this area; to explain it properly would probably need a few worked examples and lots of math markup making it a completely different article, and the issue of coupling between resonators is not, in any case, unique to DE filters. Saying the response is changed does clearly demarcate which kind of parameters are being affected.  Sp in ni  ng  Spark  21:13, 10 July 2010 (UTC)


 * "Filters using stubs can clearly be made band-pass" I'm not a fan of phrasing that includes "clearly" or "obviously", as it implies that the reader is stupid for not having already realized the fact. I suggest replacing with "easily" or perhaps "readily".
 * The point being made is that stubs naturally form band-pass filters because they are resonators and this sentence is to be read together with the preceding one to get this meaning. "Stubs are resonators and can form band-pass filters; numerous other..." is more direct and should overcome your objection.  Sp in ni  ng  Spark  21:13, 10 July 2010 (UTC)


 * "The minimum width of gaps, like the minimum width of tracks" Eh? What does "tracks" refer to?
 * Track, also called trace, is the flat printed conductor on a printed circuit board. There used to be a stub article at signal trace, but it has been changed to a redirect to printed circuit board.  Sp in ni  ng  Spark  21:13, 10 July 2010 (UTC)


 * "The design consists of a series of λ/2 resonators side by side but coupling over only λ/4 to each of the resonators either side in a staggered line as shown in figure 9." Is there a word missing between "resonators" and "either side"? I'm not really sure how to read this sentence.
 * Messed with this sentence, changing resonators either side to neighbouring resonators among other things. --catslash (talk) 17:21, 13 July 2010 (UTC)


 * "The "U" shapes so formed give rise to the name hairpin filter." There is a mathematical symbol "∩" in the Insert Symbol library, perhaps it could be added next to "U": "The "U" and "∩" shapes so formed give..." Also, would you object to linking "hairpin" to either hairpin turn or bobby pin?
 * I would say no to the inappropriate use of mathematical symbols. Linking to bobby pin is fine, and is a more direct explanation than hairpin turn.  Sp in ni  ng  Spark  23:46, 15 July 2010 (UTC)


 * "which takes up more real estate on the board which can be at a premium" Very unencyclopedic phrasing. How about "which can take up an undesirably large amount of space on the board" ?
 * Real estate is a commonly used metaphor for this I think (there are two occurrences in the article). Try typing "PCB rea..." into Google and it will auto-complete it. Otherwise, area is both snappier and more precise than amount of space. --catslash (talk) 17:01, 13 July 2010 (UTC)


 * "the short-circuit provides a mechanical fixing point for the line" What does "fixing" mean in this context? When I first read it, I thought it meant "repairing", but that doesn't really fit into the sentence.
 * You must be American. I was about to change this to attachment point, but found there were four occurrences of fix in the article. I don't believe you could read all these as mend, but I'm English, so could you check these yourself? Thanks. --catslash (talk) 17:01, 13 July 2010 (UTC)


 * "The end which is short-circuited alternates on alternate line sections." The repetitive use of "alternate" is both awkward and unclear in meaning.
 * Suggestion: "The end which is short-circuited alternates on each line section."  Sp in ni ng  Spark  21:48, 21 July 2010 (UTC)


 * "The spacing requirement between lines is not so onerous as the parallel line structure; so higher fractional bandwidths can be achieved, and Q values down to 1.4 are possible." Are the spaces between the lines large or small? "Onerous" is an odd word choice that does not correspond to any particular size.
 * It is narrow gaps (or lines) that is an onerous requirement. This should be clear from the previous discussion of the printing limitations.  Sp in ni  ng  Spark  21:48, 21 July 2010 (UTC)


 * I think the Interdigital filter section would benefit from a comb-line filter diagram. Perhaps the existing Figure 11 could be made smaller and augmented with such a diagram.
 * Don't really agree with that. The number of diagrams in the article is already quite difficult to fit in without causing sandwiching issues.  There are many different designs and variants that diagrams could be created for, but it is not necessary to show every singly one in an overview article.  The comb-line is similar enough to other depicted forms that it can be handled in prose.  Perhaps articles on specific types will be created in the future at which time more diagrams would be appropriate.   Sp in ni  ng  Spark  21:48, 21 July 2010 (UTC)


 * "The lines can be either rods or bars" What is the difference?
 * Suggestion: "The lines can be either rods or rectangular bars"  Sp in ni ng  Spark  21:48, 21 July 2010 (UTC)
 * I suppose that this implies that "rods" are necessarily circular? If so, I think adding "circular" into the sentence would be helpful. I'm curious about something: Is this only true for interdigital filters, or is it just a general statement? --Cryptic C62 · Talk 14:22, 23 July 2010 (UTC)
 * That was exactly the meaning, circular and rectangular cross-sections. However, it is not an important point and there is no need to make a meal of it.  Such constructions are not limited to interdigital filters, but they are particularly suited to interdigital and combline filters which can be assembled in a compact box.  A stub filter, for instance would have "bits" sticking out all over the place.  While PCB construction has become pervasive, it is not suitable for every application.  High power military radar, for instance, is likely to turn the average printed circuit into a charred and useless mess.  Hence constructions with solid conductors still find uses.  Sp in ni  ng  Spark  20:10, 23 July 2010 (UTC)


 * "This topology has both low-pass and band-pass characteristics." Does "band-pass" have a distinct meaning from "bandpass", or is this just an inconsistency?
 * Inconsistency. It should probably be standardised on "band-pass" for consistency with other bandforms which are also hyphenated.  Note that this is different from "passband" which should not be hyphenated.  Sp in ni  ng  Spark  20:10, 23 July 2010 (UTC)


 * "Genuine high-pass filters are difficult, if not impossible, to implement in distributed elements." I'm thinking this should either say "in distributed element models" or "with distributed elements."
 * Not sure what problem you are trying to address here, but the first suggestion is logically wrong, a model is not an implementation, it models an implementation. The second suggestion is ok.  Sp in ni  ng  Spark  20:10, 23 July 2010 (UTC)


 * "Konishi describes a wideband..." Who's this? If the person is notable enough to be mentioned, some introduction should be included to give the reader an idea of who he is. Alternatively, the passing reference could just be dropped altogether.
 * I have seen this idea debated before (that is, the idea that a source should not be mentioned by name unless they are notable) and afaik no definite conclusion has ever been reached. It is certainly not the case in scholarly papers.  Nor do they "introduce" the author first like a character in a novel.  If you can point to a policy or guideline that requires this then fair enough, but otherwise I will oppose such a change.  I think rather, that the boot is on the other foot, it is for the person who wants to remove Konishi to show that his filter is in widespread use by others and it is not therefore necessary to refer to him specifically.  Sp in ni  ng  Spark  20:10, 23 July 2010 (UTC)
 * I'm not interested in policy here, I'm interested in what's best for the reader. My concern is that dropping a name into an article like this leaves the reader thinking "Who is this person?" and then proceed to either try to find a relevant article or search through this article to see if they missed the first reference. My personal preference is to always have the first reference to a person include their full name and some indication of who they are / what they do, after which any other references to the person can use simply the surname. I realize that this is not the convention in scientific papers, and I also realize that most of the people who would find this article of interest would already be familiar with such conventions. However, this is not a scientific paper, and we must strive to make the article accessible to everyone. --Cryptic C62 · Talk 19:54, 25 July 2010 (UTC)
 * We are going to have to disagree on that, it is a perfectly normal construction. The place the reader should be looking is the reference where they will find Konishi's book and his journal publications aren't hard to find either.  If you are arguing that the refs are never looked at, there is no point in having them at all.  Sp in ni  ng  Spark  23:17, 25 July 2010 (UTC)
 * Another concern which I just realized: You said above that I should "show that his filter is in widespread use by others and it is not therefore necessary to refer to him specifically" If his filter is not in widespread use, why is mentioned in this article?
 * The bottom line is that I do not know whether Konishi's design is in widespread use, is exceptional, or simply run-of-the-mill. So I have followed the convention of naming sources: rather than have Wikipedia say that such-and-such design exists, I have had Konishi say it. Certainly radial stubs in general are fashionable at the moment and Konishi seems to have some status in the field but I don't feel confident enough to synthesise those two things into a general statement.  Sp in ni  ng  Spark  17:58, 27 July 2010 (UTC)


 * Not sure I like the change from "There are a great number of topologies of stub filters to choose from" to "There are several different topologies of stub filters". While "a great number" may be overstating the case "several" has gone the other way.  It has also lost the sense that the designer has a choice of forms which in some cases are identical in response.  Sp in ni  ng  Spark  20:10, 23 July 2010 (UTC)
 * How about "many" instead of "a great number"? My problem with the "to choose from" construction is that it reads like a how-to guide, implying that the reader is necessarily an electronics manufacturer. --Cryptic C62 · Talk 19:54, 25 July 2010 (UTC)
 * Oh for goodness sake, throwing "how to" around is a pedantic application of the guidelines. Of course its not a "how to".  There is a big difference between discussing design choices (perfectly valid thing for an article to do) and an instruction manual (which is what WP:NOTHOWTO is addressing).  I am ok with "many", but that is not addressing the point about design choice.  Please either find a form of words that continues to convey that meaning or reinstate the original sentence.  Sp in ni  ng  Spark  23:17, 25 July 2010 (UTC)
 * You mentioned above that the change has "lost the sense that the designer has a choice of forms which in some cases are identical in response". The notion that different topologies can produce the same response is not made clear in either variation, so how about we try to resolve that issue while also avoiding any how-to phrasing? If the issue I just highlighted is important to you, how about stating it bluntly: "There are many different topologies of stub filters, some of which produce identical responses." Alternatively, if you really want to emphasize the matter of choice, perhaps it would be better to mention the designer explicitly: "Designers have many different topologies of stub filters to choose from." The two could also be combined: "Designers have many different topologies of stub filters to choose from, some of which produce identical responses." --Cryptic C62 · Talk 02:12, 27 July 2010 (UTC)
 * All of your suggestions are good, the last one in particular says it exactly to my mind.  Sp in ni ng  Spark  17:58, 27 July 2010 (UTC)

Unresolved comments
Here are some comments on the article's prose intended to help make the article accessible and precise:


 * "they have the drawback that it is impossible to create an ideal open circuit" Because I'm a physics nerd, I automatically assume that "ideal" has some special meaning, as is the case for ideal gas. What meaning is it intended to convey here?
 * Yes, exactly in the sense of an ideal gas which perfectly obeys some law of physics. Even more relevant, it is the same sense in which we talk of ideal voltage or current generators in electrical analysis.  Specifically here is meant a circuit element through which zero current flows regardless of the voltage across it.  Sp in ni  ng  Spark  17:56, 1 July 2010 (UTC)
 * Is it ever possible to create an ideal circuit? I only ask because ideal gasses do not exist, so I would assume that the same is true of ideal circuits. If such is the case, then the highlighted sentence doesn't really make any sense. If it is possible to create an ideal circuit, I think this sentence would benefit from a wikilink (I couldn't find any appropriate articles) or a brief explanation&mdash;the article should be accessible to everyone, not just to circuitry nerds. --Cryptic C62 · Talk 00:06, 4 July 2010 (UTC)
 * Ideal circuit elements do not in reality exist but it is perfectly valid to examine how well a real component corresponds to the ideal. Electrical analysis is entirely based on networks of ideal elements so it is not only unavoidable, but actually helpful, to discuss ideal elements here and there.  Even when it is desired to model a non-ideal component, this is done by adding additional ideal elements to its network representation - a real non-ideal resistor might be modelled, for instance, by adding a capacitor in parallel, and an inductor in series with it.  In the case of a printed line being left open-circuit, this is a non-ideal open-circuit because of the unavoidable dielectric effect of the board on which it is printed.  Consequently, this non-ideal open circuit is frequently modelled as shown in figure 4b as a parallel capacitor.  This is to be compared to a line terminated with a short-circuit achieved with a strap through a via hole to the ground plane beneath.  To be sure, this will not be an ideal element either, the strap must possess some resistance for instance.  However, it is close enough to the ideal to be treated as ideal for most design purposes.  The dielectric effect causing such a problem for open-circuits is not a problem here since anything in parallel with a short-circuit is still a short-circuit; the capacitive effect has been shorted out. I will try tweaking the wording though, to make this a bit clearer.  Sp in ni  ng  Spark  09:47, 4 July 2010 (UTC)

More to come. As you make changes in the article, please respond below individual concerns so I know which are done and which need further discussion. Thanks! --Cryptic C62 · Talk 00:52, 15 June 2010 (UTC)
 * As you have not struck this, does it mean that your edit summary of "final comments" is untrue?
 * It is quite true. I will review your responses tomorrow. --Cryptic C62 · Talk 04:40, 24 July 2010 (UTC)
 * The review is now complete. Distributed element filter has earned my stamp of approval:


 * Feel free to contact me if you want a review for any other articles you're working on. --Cryptic C62 · Talk 03:12, 28 July 2010 (UTC)

Great article
Nice job. I know electronics, but not microwave techniques, and I found the article very readable. Wizzy&hellip; &#9742; 16:14, 17 July 2010 (UTC)

Good work. Howver, I see a way to further improve the presentation. For example, the citation starts in the text with the number 4, but the reference list starts from 1. Also the reference section does not contain the entire informnation so that to be able to find that reference in on-line library. — Preceding unsigned comment added by 199.64.72.253 (talk) 12:54, 23 December 2014 (UTC)
 * The references are numbered automatically by the Wikimedia software and cannot be changed. They are numbered in the order they appear in the source text.  Refs #1, #2, and #3 are in the "block converter" sidebox.  In the source this is above the introductory text (to ensure that it appears to the side, and not beneath, the opening text) so those refs get numbered first and hence the first ref in the text is #4.  The full citation information is contained in the bibliography section below the references.  It is organised this way to avoid the full citation having to be entered each time the same reference is used.  Not all references are online, but for book references one can often find an online preview by clicking on the ISBN number.  This will take you to a page listing numerous book catalogues.  The most useful ones for previewing are Google Books and Amazon, both of which allow limited preview for selected books.  The most useful for finding a copy in a library is the WorldCat site. SpinningSpark 14:42, 23 December 2014 (UTC)

super duper long img alt
I recall a recent WT:FAC discussion suggesting that img alt text should be short short. This has the longest I've ever seen... clocking in at nearly 500 words. ♦ Lingzhi2 (talk) 15:50, 2 August 2019 (UTC)
 * I agree with you. At one time the standard was to write long alt text, almost an audio description of the image.  The current thinking is that it should be minimalist unless there is some information to impart that is not covered in the text or caption.  Nowadays I generally put in one word alts like "photograph" or "diagram"; alt text should always exist, even if only one word, to prevent screen readers from reading out the file name, which is generally less than useful. SpinningSpark 22:40, 2 August 2019 (UTC)
 * hey added a clear because on my laptop the first word of the paragraph after the image (i.e., " The") gets separated from the rest of the sentence and isolated on the same line as the image, before the image. As for alt text, I dunno how someone with a screen reader would respond to the 5 minute descriptions. ♦ Lingzhi2 (talk) 00:34, 3 August 2019 (UTC)
 * You must be using a very low res monitor, something like 800×600 px for that orphaning to happen. I've added a fix that should stop it happening.  As I already said, there's no objection from me if you want to remove the long alt texts. SpinningSpark 01:07, 3 August 2019 (UTC)