This is the talk page for discussing Signal-flow graph and anything related to its purposes and tasks. This is not a forum for general discussion of the subject of the article.
This article is within the scope of WikiProject Systems, which collaborates on articles related to systems and systems science.SystemsWikipedia:WikiProject SystemsTemplate:WikiProject SystemsSystems
This seems to be a lot of algebra signifying nothing. At best it shows that you can change the weights on some of the branches and still have the same overall gain. The relationship of the SFG to anything is not defined. Even if you can represent asymptotic gain by a SFG, there is no useful reason to do so. You can represent a voltage divider by a SFG, but it would not bring any insight to the analysis.Constant314 (talk) 17:52, 31 December 2014 (UTC)[reply]
The reason that a 2 port's admittance parameters are an unrevealing example is because it should have been transformed to the equivalent scattering parameters. See Mason's 1953 paper, fig 35 page 1196. IIRC a lossless transmission line can't be represented using an admittance matrix, although there's a way to fudge it that's good enough if the application isn't too demanding. Scattering parameters are a relatively direct application of signal flow graphs with more intuitively meaningful results, e.g. S21 is the input to output voltage gain/attenuation of a 2 port. — Preceding unsigned comment added by PolychromePlatypus (talk • contribs) 19:40, 9 April 2019 (UTC)[reply]
It was actually example 2 Ideal negative feedback amplifier, figure 3. It was retained. I still think it is a bunch of algebra signifying nothing.
If I have the equation V = kθ there are two possible graphs: an arrow pointing from θ to V with a weight of k, or an arrow pointing from V to θ with a weight of 1/k. If I have other knowledge such as V is the velocity of my car and θ is the deflection of the speedometer, then I know that the arrow should point from V to θ. The SFG has more information than the equation. I do not have access to the reference, but assuming that the reference is correct, there is probably some requirement that the equations be in some sort of canonical form.Constant314 (talk) 19:47, 31 December 2014 (UTC)[reply]
A signal flow graph should model the actual physical process. Note that removing the bezel from a speedometer to directly manipulate the indicator needle has no discernible effect on the forward velocity of the vehicle. (Feel free to insert an analagous humorous reference to a fuel or oil treatment or to the application of flame decals) Conversely altering the forward ground velocity of the vehicle by any available means has a direct and proportional effect on the position of the speedometer's indicator needle. (note that because a modern instrument cluster doesn't have a speedometer the system also has to be powered to permit the microprocessor controlled stepper motor with gear reduction to simulates a speedometer) ... also there is no requirement that a signal flow graph be expressed in a canonical form. A signal flow graph can often be reduced Isee Mason, 1953) and sometimes its pleasant not to cover the page with boxes and arrows, but the validity of the analysis doesn't depend on reducing it.PolychromePlatypus 20:06, 9 April 2019 (UTC) — Preceding unsigned comment added by PolychromePlatypus (talk • contribs)
The issue was settled to everyone's satisfaction long ago. The article talk page is not a forum for discussing the topic. It is a forum for discussing improvements in the article. If you think there is something to improve in the article, its best to just start a new topic rather than responding to long dormant discussions. Constant314 (talk) 21:39, 9 April 2019 (UTC)[reply]
I think the section on causality can be reduced to a couple of sentences in the introduction. Since a system of simultaneous linear equations can be made equivalent to a SFG, the SFG inherets the ability of linear equations to represent both causal and acausal; the is no controversy. Constant314 (talk) 21:38, 17 February 2015 (UTC)[reply]
I favour keeping the extended critisism of SFG causality, since causality is a major point of differentiation from bond graphs (in which causality is explicit)Pierre5018 (talk) 12:18, 18 February 2015 (UTC)[reply]
Brews, I have no objection to a clear statement; I just don't think it needs a section. Pierre, I don't see anything that could be criticized. I think bringing in bond graphs is extraneous to this article, but a single sentence would be sufficient to say something like "The SFG may represent both causal and acausal relationships, unlike the bond graph in which causality is explicit."
Constant: I am with Pierre on this one for a different reason. That reason is that Kuo (a very influential text), Mason himself, Paynter, and Willems stress causality, which means confusion reigns without a clear disclaimer that SFG has no necessary connection to causality. Brews ohare (talk) 16:02, 18 February 2015 (UTC)[reply]
Yes. The SFG is a mathematical device, and just as Newton's law F=ma could be interpreted as saying force causes acceleration, it can be interpreted as saying acceleration is an indication of the presence of a force, or interpreted as how much acceleration corresponds to how much force. The math is neutral on the causal interpretation. Brews ohare (talk) 16:36, 19 February 2015 (UTC)[reply]
Brews ohare is right, see Shannon's or Mason's original works. Example: (a) Before path inversion(b) After path inversion The first graph corresponds to the equation , the second graph to the equation . I suggest we remove the whole philosophy discussion, and replace it by the simple statement that sfgs do not contain causal information, but can under some circumstances be drawn in a way such that the way they are drawn does convey it. This is then not a mathematical property of the graph, but a meta-property of the way to draw the equations system. I think it is sufficient to cite Mason56 on this. Hanspi (talk) 06:00, 24 September 2020 (UTC)[reply]
I agree. The section is not remotely encyclopedic. It reads like a conversation. It could probably be reduced to one sentence in another section. The arrow does not imply causality. Constant314 (talk) 06:27, 24 September 2020 (UTC)[reply]
The re-write I suggest is to massively abbreviate "Systematic reduction to sources and sinks" by giving one reference to reduction rules and obe example, rewrite "Basic components" to the much shorter description originally used by Mason, and changing "Solving linear equations" to a correspondence between the matrix notation of a linear equation systemm instead of the Kronecker delta notation. My estimate is that the section will then be a third of the length. Opinions?Hanspi (talk) 06:23, 24 September 2020 (UTC)[reply]
Yes, I agree with greatly reducing the ""Systematic reduction to sources and sinks" section. And the section should probably be renamed "simplification."Constant314 (talk) 06:34, 24 September 2020 (UTC)[reply]
I think most people would write the content in their sandbox and then transfer it here. You have your own sandbox which you access by clicking the "Sandbox" link in the upper right corner of the page. Once you have it like you want it, you could invite us to view your sandbox; but in this case I'm going to suggest that you be bold WP:BB and simply replace the sections that we have discussed recently here. Some other editor may disagree and revert it. Don't let that bother you. It happens frequently. It means that we come back here to the talk page and find a consensus. Constant314 (talk) 15:16, 24 September 2020 (UTC)[reply]
