Talk:Analogue filter/GA1

GA Review
The edit link for this section can be used to add comments to the review.''

This article deserves GA status. It is well written, neutral, stable and well referenced with in-line citations (thus verifiable). The article is focused on its specific area, and the change in title during the review was necessary to accentuate that. There were minor problems with style, references and figures. Most of that was fixed in the review process. I am certain minor problems with the prose remain and would encourage the author to seriously brush it up if ever attempting FA nomination. I stress that the no-template reference formatting is well acceptable for this historical article, but again, FAC might reject that. GA comments are listed below. Materialscientist (talk) 07:13, 31 July 2009 (UTC)


 * "Overview" starts with "There are three main stages in the history" already suggesting the article is about the history. Please make you mind. If you focus on history, rename the article. If not, you'll need to rewrite the text to reflect the current status. I feel the former solution is more natural here.
 * I have renamed this "historical overview" which might make it more acceptable. I don't really mind this article being renamed, but have been reluctant to do so for the following reason - the article started off as a stub for "analogue filter"; renaming it will leave us without an analogue filter article except for a redirect to this very article, which seems a bit circuitous to me.  It is certainly true that the article has a strong historical bias (and a further criticisism could be it has a strong passive technology bias) but other key areas are at least mentioned.


 * I am not satisfied with the lead because: (i) it could be a bit friendlier to a non-specialist (ii) it silently assumes low frequencies. IMO, in high-freq. range, passive analogue filters are still a workhorse (e.g. in Europe, TV + internet signal is sent through one cable and is passively filtered for every consumer depending on subscription). If (i) appears unspecific, I'll add more on that.
 * Tried to address this with an additional first para in the lede and other rewording.


 * IMO, elastance is a very rare term (I mean in engineering, Google counts reflect here the definition, not its use), and I would not mention it at all. I would also not use the word "analogous" given the article is on the analogue filters.
 * I fully agree that elastance is a very rare term. Elastance is used, mainly, because it appears in Cauer's matrix representation and positive-real theorem.  It's use is so convenient here, and the use of capacitance is so awckward (it would be necessary to carefully distinguish between the meaning of inverse capacitance-matrix and an inverse-capacitance matrix) that I really think it should stay.  It is a very easy concept to grasp, even for the non-specialist (perhaps easier because they have no preconceptions), and it is also used by Norton as corresponding to a mechanical quantity, and there it would be necessary to re-interpret Norton in terms of the inverse mechanical quantity, which would not only now cause the same problem in  mechanics (compliance is very rare compared to stiffness) but is also a touch of OR.  I have reworded to make it clear that elastance is a rare term.
 * Remove analogous - done


 * The caption of File:Norton mechanical filter.png is unclear (there are 2 circuits).
 * Done


 * Q-factor needs a bit of explanation. Materialscientist (talk) 07:50, 22 July 2009 (UTC)
 * Added footnote.  Sp in ni ng  Spark  19:38, 23 July 2009 (UTC)

Second thoughts: (i) I shall referee the article whatever its name, for me its 1st two lines are clear enough and I shall not ask extending into what does not belong here, but. Sooner or later some engineer will jump this article to add, e.g., manufacturing, sales, modern designs and pictures, etc, etc. To protect from that, I would give the article a clear name and not worry now about absence of "Analogue filter" article - so many articles are missing on WP.
 * As you are the second person to complain about the title I guess I should think about compromising. I could live with passive analogue filter.  No real need to specify "linear" as that is almost taken for granted with passive filters and no need to put "history" in the title, although the article has a strong historical bias, it is not solely historical.  Would you agree with that? I would prefer that it was left open for others to expand the article, I do not see a need to protect it from that, but if that interferes with GA review then so be it.  Sp in ni  ng  Spark  16:44, 24 July 2009 (UTC)
 * I wouldn't stop there and add "history" or/and "development" to the title because the article does not reflect modern days (few existing comments are not enough). Materialscientist (talk) 05:36, 25 July 2009 (UTC)
 * Now renamed.  Sp in ni ng  Spark  20:48, 25 July 2009 (UTC)

(ii) One of my early GAs was rejected only because I couldn't make it understood by everyone. This is the main problem here. The text is flowing nicely, I like it, but it is overloaded with unexplained jargon. Wikilinks do help, but whenever you can add a few explanatory words, please do. I'm adding myself too, but clumsily, and you have to check that I didn't mess up something (BTW, delinking was a semi-automatic AWB job - I only saw overlinking, but couldn't see well where to remove a wikilink and where to leave).
 * Your edits are good, except for one, the wikilinking of antimetric filter which I have undone for now. This is really strange, it is a redirect to a section of an article I wrote a long time ago which discusses antimetric filters briefly in relation to image parameter design, but is not really relevant here.  The entry I put in Wiktionary antimetric is actually more useful, if that is an acceptable link.  Failing that, can deal with it as a footnote.  Or could create a stub at antimetric filter, but it won't be much more than a definition and I need to find out what the editor was trying to achieve with that redirect first.  Sp in ni  ng  Spark  20:55, 24 July 2009 (UTC)

*:"Antimetric filter", and all other technical jargon should be explained. For vocabulary terms (e.g. elastance) a footnote will do. For wider notions (e.g. RC filter) a separate article is better, stub is fine, it will get expanded some day. Materialscientist (talk) 05:36, 25 July 2009 (UTC)
 * "all the transformed networks are equivalents" - should it be "equivalent" ?
 * "...of the original network." are the unsaid words there. Article now says them.  Sp in ni ng  Spark  22:10, 25 July 2009 (UTC)
 * Terms to explain:
 * "earth return phantom circuit"
 * open-circuit and short-circuit impedances (good wikilink will do, but I haven't seen one)
 * "impedance constructed of a finite number of meshes" (Note 12) - there is space there to provide an easier explanation.
 * Done - but I am surprised I am the first on Wikipedia to explain open and short-circuit impedances. I would have expected that at two-port network, and it should definitely be at Z-parameters and Y-parameters.
 * I have now created an article at Antimetric (electrical networks). "Filter" does not appear in the title of this new article since it is a subject of the wider topic of network analysis and not specific to filters.  RC filter can go in as a wikilink, it redirects to RC circuit.  I don't know what a footnote for elastance should say, the text already states that it is D=1/C and that is all there is to say.  Sp in ni  ng  Spark  20:48, 25 July 2009 (UTC)
 * Actually, RC filter does not appear anywhere in the text so there is nothing to link.  Sp in ni ng  Spark  22:10, 25 July 2009 (UTC)
 * Unfortunately, you'll have to bear with my unclear formulations. Examples of linking "elastance" and "RC filters" were abstract, not referring to this article. I have delinked elastance because that wikilink was awkward (notion too simple, and my screen jumps on that link).

*Ref. 52 (Transactions of the American Institute of Electrical Engineers) is incomplete.
 * I will come back to this one, first I am going to revert the editor who converted all the refs to templates. This one has totally the wrong ISSN number inserted and has lost some other essential information and I can no longer locate where I found it online (google books can find the relevant text string but there is not preview so I cannot discover the author and article title in the journal).  I really do not like templates and it was actually against the guidelines to make this change.  I was not goign to bother to revert, but as it is now clear the job has not been done reliably I will.  I will do this later today before addressing the rest of your comments.
 * Why? Templates are widely endorsed as they allow many auto operations (formatting and checking). Just recover the information, not the formatting! Even if a bot is misbehaving, he can always be tricked by tweaking the fields. Materialscientist (talk) 10:53, 25 July 2009 (UTC)
 * Too late, I had already done it before getting around to reading your comment. The reason I do not like cite templates is they add to the "text clutter" and are a major obstacle to newbie editors because they don't understand them.  Sp in ni  ng  Spark  22:10, 25 July 2009 (UTC)
 * My experience is that with old refs, "no-tags" style is indeed better, but with new ones not - I have stopped typing full refs quite some time ago, only and let the bot fill it up. As a WP cleaner, I see so many mistyped refs, which I sometimes can't recover. Materialscientist (talk) 04:41, 26 July 2009 (UTC)
 * I think we are going to have to agree to disagree over this one, this is not the place to have a general debate. I will just say that what I have done is in accordance with the guidelines, but I am not going to do battle over this and insist on doing it my way.  Sp in ni  ng  Spark  13:09, 26 July 2009 (UTC)
 * "The concept of a port also played a part in the development of the theory, and proved to be a more useful idea than network poles (the terminals of the circuit connecting to the outside world)". I am getting confused here as I thought "pole" means singularity all through the text and port is reserved for "2x terminal" (don't know though which term sounds better to non-specialist, port or terminal).
 * You are right, there are two meanings of pole in this article which needs to be made clear. "Pole" is used extensively in the literature with the meaning of "external connection point" especially amongst the mathematicians and theoretical workers (and often with the other meaning - opposite of "zero" - in the same paper).  A 2-pole can be considered synonymous with a one-port, but more than this they start to diverge.  A simple "T" circuit is often used as 2-port network in filters &C, but as a load on a 3-phase generator, the same network must be considered as a 3-pole star topology.  No pair of poles have the same current going in and out which is the definition of a port.  Some network theorems are only valid if ports are being considered (Bartlett's bisection theorem for instance) whereas others (eg Kirchoff's current law) apply in the more general case.  It is therefore important to distinguish them.  I might write an article one day - but not today!  Sp in ni  ng  Spark 
 * My scientific experience tells me that even the "established" literature is often messed up regarding terminology. I am not happy with two "poles" in this article and do suggest using consistently "ports" or "terminals" for connections and poles for singularities; my intuition tells me "conventions" don't need to be followed in this specific case.
 * My references definitely favour pole, but a quick search of the literature shows that "4-terminal network" is sufficiently established terminology to be able to use it on Wikipedia.  Sp in ni ng  Spark  13:09, 26 July 2009 (UTC)
 * we need a note or wikilink explaining "2-pole" (at first occurrence of "pole").
 * This is done, but I will change it to 2-terminal as above, maybe leaving the explanation of pole in the note.  Sp in ni ng  Spark  13:09, 26 July 2009 (UTC)
 * "and the distributed elements are called the primary line constants." - is this piece necessary ?
 * Don't know whether its necessary (is anything?) but it is certainly going to be useful for understanding line elements when there is eventually an article there.  Sp in ni ng  Spark  13:09, 26 July 2009 (UTC)


 * I would add a small note to one of the figure captions on similarity of a filter drawing with a ladder.
 * Done.
 * A minor issue, consider unifying elliptical and elliptic.
 * Done.  Sp in ni ng  Spark  00:11, 29 July 2009 (UTC)


 * Wherever simple illustration of the described filters are available, please add, this would make the text much easier to understand. There are images available for "resonance" and other sections, but you can choose much better than I. Also maybe a drawing of one of Campbell's filters, File:Zobel original drawing.PNG, etc.Materialscientist (talk) 09:57, 24 July 2009 (UTC)
 * I will produce some illustrations if you are willing to allow me time to do so, but the image you have linked is not a Campbell filter, it is a matching network along the lines of an m-derived filter before Zobel started calling them such.  Sp in ni ng  Spark  10:30, 25 July 2009 (UTC)
 * No rush. Quality is primary here, and figures will add much to that. I'm not sure I'll have ample time for review during the week. Materialscientist (talk) 04:41, 26 July 2009 (UTC)
 * I have added an original sketch by Campbell but now find I am at loss for what else would be useful to this article. I am open to suggestions but here is my thinking: later filters all often use the same ladder topology so additional diagrams would be no different from Campbell's filter - other than it could be a clean svg and the component values would be different, but this is best done with maths and in the individual articles, not here in a general article.  Also, with later filters the use of ladder topology is a matter of choice, a point I have explained in the annotation to the new diagram.  Putting in more ladder diagrams (or any diagrams for that matter) would give the false impression that that topology was associated with that particular design methodology when it is not.  The canonical forms of Cauer etc are really only of interest theoretically and giving examples is merely adding confusion.  Other practical realisation topologies are not really discussed in the article and to add them is starting to get into the level of detailed design which is not really suitable for a broad-brush article.  Sp in ni  ng  Spark  15:51, 26 July 2009 (UTC)
 * Maybe one from these (perhaps File:Leidenflaska.png) for the Leiden jar?
 * I quite like the Oudin coil one for the "electrical resonance" section, but I would not want to explain Oudin or Tesla coils, it is going too far off-topic for this article:- just wikilink to Oudin coil.  Sp in ni ng  Spark  18:46, 28 July 2009 (UTC)


 * Same filter diagram is described as "Lord Kelvin's model" and then as a part of a Campbell's 1915 patent. Can this be clarified a bit in the captions (e.g., that the drawing is a modern-day illustration of a model described by Kelvin's equations)?
 * That is not correct. Firstly, the Kelvin model consists of RC elements and the Campbell filter is LC elements.  Secondly, all the line models are distributed elements whereas the Campbell filter is lumped elements.  I deliberately avoided annotating the diagrams with the usual infinitesimal symbols to avoid giving the impression that Kelvin was working in those terms which he was not (and Ohm certainly wasn't either).  However, since it has clearly confused you, it is taken for granted that other readers will be confused as well, so I will annotate the diagrams fully and explain in the caption.  Sp in ni  ng  Spark  18:46, 28 July 2009 (UTC)
 * I obviously missed the difference between those RC and LC drawings .. Could you fix "and the infinitesimal," in the Heaviside's picture? Materialscientist (talk) 00:55, 29 July 2009 (UTC)
 * Done, I assumed you wanted an expanded description.  Sp in ni ng  Spark 
 * Note 12: "A class of filters are a group of filters which are all described by the same class of mathematical functions" - "class of mathematical functions" needs clarification or rewriting.
 * "Function class" is a common enough term in mathematics and philosophy, but unfortunately there is no wiki article. The more general article class (set theory) would be an adequate link but better to add a note here I think.  The footnotes section seem to be getting quite large now.  Do you think it would help to separate out into its own section those items that are purely "definition of terms"?  I can imagine readers getting frustrated after repeatedly going to the footnotes thinking they will learn some interesting tidbit only to find a definition of something they already understand.  Sp in ni  ng  Spark  18:46, 28 July 2009 (UTC)
 * I think the present formatting is still Ok and that the number of notes is higher than it could be because of missing WP articles (c'est la vie). Materialscientist (talk) 00:55, 29 July 2009 (UTC)

Materialscientist (talk) 00:55, 29 July 2009 (UTC)
 * "in America", "the behavior" (several instances), "are all across the same line" and "further developed by several others" are somewhat unclear.
 * "in America". Done, I assume you meant this should read "in the US". although there is also a United States of Mexico, but no one thinks of that as being the United States any more than anyone would consider the term "America" to be anything other than the USA.  Sp in ni  ng  Spark 
 * "the behaviour". Not really sure what the problem is here.
 * The occurence in the "electrical resonance" section describes the specific behaviour being referred to in the immediately following sentence. I have tried to address this by changing the period to a colon.
 * Both occurences in "image filters" are referring to all behaviours of the filter in general (gain, transfer function, phase and delay etc). The whole paragraph is saying this in a nutshell: old method: take a specific circuit diagram, calculate the response, see if that is any use for filtering. New method: take a bunch of poles and zeroes, calculate the response (which remarkably can be done just as well as with a definite circuit), see if it is any use for filtering, convert the poles and zeroes to a choice of circuits, choose the most convenient one to build.  If what is in the article is not clearly saying that, please explain where it is unclear.
 * "are all across the same line". Done
 * "further developed by several others". I don't understand the difficulty here, the major players are named, this phrase lets the reader know that there were others working in the field without needing to name every minor player, with the consequent risks of both leaving someone out and boring the reader.  Sp in ni  ng  Spark 
 * I felt others does need a noun there (e.g. "other researchers"). As to behavior, I would let it go in "electrical resonance", but it should be clarified at 1st occurrence in "image filters". I see this term so often in science literature and dislike it as very confusing (the reader is left to guess which parameter(s) are implied - it is so easy to specify them) Materialscientist (talk) 00:55, 29 July 2009 (UTC)
 * Done  Sp in ni ng  Spark 
 * I would rewrite "Asking the two related questions of what specification of poles and zeroes are realisable as passive filters and what realisations are equivalent to each other led Cauer to develop network synthesis techniques"
 * I have changed "asking" to "pursuing" but am unclear what the difficulty is.  Sp in ni ng  Spark 
 * I always worry about numerous non-native readers and usually split up such sentences (or use commas to separate clauses). Materialscientist (talk) 00:55, 29 July 2009 (UTC)
 * Done  Sp in ni ng  Spark  16:01, 29 July 2009 (UTC)