Talk:Feedback

Source supporting lead sentence

The lead sentence reads:

"Feedback occurs when outputs of a system are "fed back" as inputs as part of a chain of cause-and-effect that forms a circuit or loop.[2]

Source

[2] United States Naval Academy, "Elements of Feedback Control", Chapter 3, Fundamentals of Naval Weapons Systems (accessed August 29, 2014)

Presumably this source is meant to support the 'chain of cause-and-effect' claim. However, this idea does not show up in this source. The word 'chain' and the words 'cause' and 'effect' do not appear in this source. An alternative source is:

Andrew Ford (2010). "Chapter 9: Information feedback and causal loop diagrams". Modeling the Environment. Island Press. pp. 99 ff. ISBN 9781610914253.

This source states:

"This chapter describes causal loop diagrams to portray the information feedback at work in a system. The word causal refers to cause-and-effect relationships. The word loop refers to a closed chain of cause and effect that creates the feedback."

A bit disconcerting for the lede is that Ford restricts the 'chain of cause-and-effect' to a particular type of feedback, namely information feedback, and suggests that the 'chain of cause-and-effect' is not actually the feedback (which is the information, of course) but the mechanism that creates the feedback.

My suggestion is that the Ford source replace the present source that has nothing to do with the lead sentence, and that some thought be given as to what is the feedback and what is the feedback mechanism.1 2 3 4 5 Brews ohare (talk) 15:19, 30 September 2014 (UTC)

In this connection, CW Kirkwood suggests (§1.3 Feedback and causal loop diagrams in this source) that "Feedback is defined as the transmission and return of information (Richardson & Pugh, 1981)" This source is G. P. Richardson and A. L. Pugh III (1981). Introduction to System Dynamics Modeling with DYNAMO. Cambridge, Massachusetts: Productivity Press. Brews ohare (talk) 16:06, 30 September 2014 (UTC)

This definition is in widespread use. Brews ohare (talk) 16:09, 30 September 2014 (UTC)

That's Norbert Wiener's 1965 definition from his cybernetics work. It's not bad, but also not very used in engineering. Dicklyon (talk) 16:41, 30 September 2014 (UTC)

This feedback definition is not the main point. The main point is replacing the source. The next point is distinguishing between feedback and the way feedback is arranged for, the feedback mechanism. Brews ohare (talk) 21:53, 30 September 2014 (UTC)

I've replaced the Naval Academy source with Ford for the reasons given above, there being no objections raised here. Brews ohare (talk) 12:13, 1 October 2014 (UTC)

I've interchanged the first two sentences of the lede because the second sentence provides the definition of feedback and the first sentence describes its implementation. The reverse order is more logical and clearer. Both sentences are now solidly sourced. Brews ohare (talk) 13:07, 1 October 2014 (UTC)

It's not a particularly useful or clarifying definition. I agree with the revert of it. The figure was also not so useful; what source motivated that one? And what source motivates distinguishing between feedback and the way feedback is arranged for, the feedback mechanism? Is distinguishing these better than claiming they are synonyms? Or is "mechanism" essentially a meaningless term in typical uses? Dicklyon (talk) 17:49, 1 October 2014 (UTC)

Change in lede

The lede now reads:

Feedback occurs when outputs of a system are "fed back" as inputs as part of a chain of cause-and-effect that forms a circuit or loop.¹ The system can then be said to "feed back" into itself.

¹Andrew Ford (2010). "Chapter 9: Information feedback and causal loop diagrams". Modeling the Environment. Island Press. pp. 99 ff. ISBN 9781610914253. This chapter describes causal loop diagrams to portray the information feedback at work in a system. The word causal refers to cause-and-effect relationships. The word loop refers to a closed chain of cause and effect that creates the feedback.

This version has replaced the previous version:

Feedback is the return of information about a system or process that may effect a change in the process, for example, the regulation or optimization of performance.^1,2 Feedback occurs when outputs of a system are "fed back" as inputs as part of a chain of cause-and-effect that forms a circuit or loop.³ The system can then be said to "feed back" into itself.

¹Christopher G. Morris, ed. (1992). Academic Press Dictionary of Science and Technology. Academic Press Inc. p. 812. ISBN 0122004000.

²Norbert Wiener (1965). Cybernetics Or Control and Communication in the Animal and the Machine (Reprint of 1948 ed.). MIT Press. p. 96. ISBN 9780262730099. ...the chain of the transmission and return of information: in what we shall from now on call the chain of feedback.

³Andrew Ford (2010). "Chapter 9: Information feedback and causal loop diagrams". Modeling the Environment. Island Press. pp. 99 ff. ISBN 9781610914253. This chapter describes causal loop diagrams to portray the information feedback at work in a system. The word causal refers to cause-and-effect relationships. The word loop refers to a closed chain of cause and effect that creates the feedback.

This change was made by GliderMaven with the comment removed unnecessary definition that is probably wrong, or at least incomplete and deceptive, an action supported by Dicklyon with the comment It's not a particularly useful or clarifying definition. These are rather odd evaluations, as the definition originates with Norbert Wiener as cited above, a particularly authoritative source, and additionally is found in Morris' dictionary also cited above, in Ford, and has wide currency.

In addition, the earlier version and its three sources emphasize feedback as an exchange of information, but also that for it to work we need a mechanism to insure consequences. There exist a variety of structures for putting feedback to work, but the various possible mechanisms can be described generally as "closed chains of cause and effect". Although the present lede retains Ford as a source, his identification of these two aspects, the information exchange and its activation,(also pointed out by Wiener and Morris) has been lost. Brews ohare (talk) 20:41, 1 October 2014 (UTC)

What you refer to as "the earlier version" is your recent new version, right? I think you got the order wrong. Anyway, Wiener's cybernetics definition is not very clarifying or useful, even though widely quoted. I'm not sure what the history of the ref on the lead sentence is. We should fix that. Usually we don't put refs in the lead, but summarize what's in the article. So yes maybe there's work to do. Dicklyon (talk) 01:08, 2 October 2014 (UTC)

Ah, I see you attributed that one to me, a while back, [1]. Not sure what I said to provoke that. Dicklyon (talk) 01:15, 2 October 2014 (UTC)

And here is where GliderMaven replaced it with something that to me at least seems more meaningful, sourced to a Navy thing. Dicklyon (talk) 01:27, 2 October 2014 (UTC)

Dick: You repeat that Wiener's definition from his well-regarded book is 'not very clarifying'. GliderMaven says it's probably wrong and misleading. Well, I feel in good company with Wiener, and suggest that the notion that feedback involves making use of an information exchange is perfectly clear and helpful, while the present lead sentence, in contrast, actually does not undertake to define feedback at all, but only tries to say under what circumstances it can happen - it 'occurs when...' That is like a so-called explanation of ebola as something that 'occurs when you visit Liberia'. The change of the source to Ford and the lack of relevance of the Navy source is very carefully explained in the thread above that provides reasons you have not taken into account. Brews ohare (talk) 01:47, 2 October 2014 (UTC)

I agree that the "occurs when" construct is not great. But Wiener's "Feedback is the return of information about a system or process that may effect a change in the process" is very unclear. What does "return" mean here? And "information" is also so general as to be uninformative. And "effect a change in the process" is very unclear; how do the examples of "regulation or optimization of performance" relate to feedback as we know it, which doesn't regulate or optimize or change the process, but just modifies an input? His cybernetic view was trying to take the usual feedback and generalize it to something so big as to be almost anything. It's not necessarily wrong, just not useful or clarifying, which is why you don't see anything like it in most engineering texts. What Wiener talks about is more what I'd call "learning" or "reinforcement" than the usual concept of feedback. Dicklyon (talk) 03:25, 2 October 2014 (UTC)

Perhaps Wiener's definition is clear to me because I already know what he is talking about. You do too. So perhaps you can reword his definition to be clearer? That would be better than abandoning ship for a non-definition about 'it occurs when...'.

The key point of Wiener's definition is the role of information. To elaborate using an example you are familiar with, error-controlled regulation, the information begins with the performance gap. This information is processed by the controller to figure out what correction is needed, the implications of the initial information. Then this combination of what the system status is and information about the nature of the system's operation is repackaged by the controller as information in the form of instructions to an effector that can modify the system. The conduit for this information in its various forms as it passes around the loop is the 'chain of cause and effect'.

Now, your take on things, that Wiener has unfocussed the content of his definition so it can encompass a ridiculously broad range of topics, could be seen instead as an abstraction that covers the various forms of feedback succinctly and can be illustrated for an audience not familiar with its ramifications by providing a few examples so its generality can be understood. That approach is used by these authors, for instance Lein. Brews ohare (talk) 04:41, 2 October 2014 (UTC)

Inadequacies of the present lede

The version of the lede installed by GliderMaven with Dicklyon's blessing has the demerits of containing no definition of 'feedback' (it states, not what it is in fact, but only circumstances attending its appearance - it 'occurs when...' ), and of ignoring the role of information pointed out by a seminal source (Wiener) and adopted by many subsequent authors. The present lede employs a source focused upon information (Ford) without benefiting from its teachings. This version removed by GliderMaven has none of these defects and is more completely sourced. Brews ohare (talk) 14:34, 2 October 2014 (UTC)

In making this change in the lede that de-emphasizes the importance of information, GliderMaven is reversing his previous recommendation for a generality of approach involving explicit reference to the conduit of information. Why? Brews ohare (talk) 16:16, 4 October 2014 (UTC)

Apparently the only changes to the lede are the addition of links to two seemingly opposing sources. Ford suggests that feedback is a causal loop, while Aström & Murray point out that in a simple feedback loop it is impossible to say whether part A is a cause acting on part B or vice versa as each affects the other. Probably Ford would agree that a "casual loop" is inherently ambiguous in this way, although he has no hesitation in calling part A a cause acting on part B, and ignores the fact that part B influenced the cause provided by part A via part B’s own action. For more complex systems, feedback may not be a loop of any kind, of course, as pointed out by Ashby, p. 54 who goes so far as to say the concept of feedback "becomes artificial and of little use when the interconnections between the parts become more complex". In any event the present substitution of some circumstances where feedback might occur (Feedback occurs when ...), instead of what the lede needs, namely a definition of what feedback actually is, has not been addressed. Brews ohare (talk) 18:21, 11 November 2014 (UTC)

Sections on flip-flops and on oscillators have not been connected to the topic of feedback

The section on Feedback#Oscillator and on Feedback#Flip-flop make no connection to the role of feedback in their operation. The oscillator subsection is little more than a repeat of the lede for Electronic oscillator, and likewise, the subsection on flip-flops is little more than a repetition of the lede for Flip-flop (electronics). Without clearly describing the nature of the feedback used in these applications and its role, preferably sourced and not just some WP editor's opinion, these subsections have no place in an article on feedback. In their present form, these subsections could be replaced by See also links at the end of the feedback article. Brews ohare (talk) 18:11, 11 November 2014 (UTC)

A possibly useful discussion for oscillators is provided by Kal "Sinusoidal oscillations can be produced using two different types of principle: negative resistance oscillators and feedback oscillators." and by Chattopadhyay "Both negative-resistance and feedback oscillators can be sinusoidal and relaxation types." Brews ohare (talk) 18:56, 11 November 2014 (UTC)

See also is for things that they are not the same topic. Flip flops are multivibrators and multivibrators are devices based in feedback theory- they're cross-linked feedback loops. Flip-flops are also extremely important, and incredibly widespread.

I don't think we're covering negative resistance oscillators here though; the article is about feedback. Feedback based oscillators are completely on-topic though.GliderMaven (talk) 20:49, 11 November 2014 (UTC)

Not the point. The point is to include a description of the role for feedback and its type in these devices, with sources. Brews ohare (talk) 21:55, 11 November 2014 (UTC)

Poorly written article

This article is poorly written and it's not concise. The introduction is lousy and it does not list positive feedback so I wonder why only negative feedback should be listed. The section on terminology implies two confusing definitions and later, in the electronic engineering section, a third definition based on direction is included. Later, in the negative feedback loops section, a definition based on phase is introduced. In my opinion, the definition in terms of direction and phase are confusing or even incorrect. Also, if multivibrators are a form of positive feedback, they should be listed under the positive feedback section. Oscillators are presented and my understanding is that they use positive feedback. This is not explained or mentioned at all. Lastly, as far as I know, latches use positive feedback. This is not state anywhere in the article. Nothing is mentioned about flip-flops either. Both oscillators and latches/flip-flops should be listed under their respective categories. This article needs a huge rework.

ICE77 (talk) 16:50, 14 June 2015 (UTC)

Response speed positive/negative feedback in Biology section

> A negative feedback loop is one that tends to slow down a process, whereas the positive feedback loop tends to accelerate it.

This sentence is capable of being completely misunderstood, seemingly supporting a common misunderstanding in certain fields of biology, namely that systems possessing a positive feedback tend to respond faster, and systems possessing a negative feedback tend to respond slower.

The problem arises because it is not stated at all as compared to what exactly a "negative feedback [..] tends to slow down a process". That is: given a process with a positive or negative feedback loop, how does the corresponding process without the feedback loop exactly looks like? This is very problematic because in a very common way to answer this question, rather the direct opposite of the statement above holds.

Example: Consider the linear system

$\displaystyle {\frac {d}{dt}}x(t)=(1-k)(u(t)-x(t)),x(0)=0$

where 0<k<1 corresponds to a positive feedback of "strength" k, k<0 to a negative feedback of "strength" -k, and k=0 to "no feedback". Consider the input u(t) switching from zero to one at t=0. The solution is $\displaystyle x(t)=1-e^{(1-k)t}$, and the time constant is $\displaystyle T={\frac {1}{1-k}}$. That is, a positive feedback (k>0) increases the time constant (makes the system slower), and a negative feedback (k<0) decreases the time constant (makes it faster).

I propose to completely delete the statement rather than keeping it in its current form. — Preceding unsigned comment added by Langmo (talk • contribs) 12:50, 22 February 2017 (UTC)

life sciences & engineered systems - what' the breakdown? 50/50? 90/10?

this article/ subject matter & its related concepts- are they more closely associated with life sciences or with engineered systems? can anyone estimate a proportion break-down, like a percentage that it pertains to each?

leaving this question unanswered will leave the article muddled and possibly confusing, depending on the initial assumptions of the reader. skakEL 19:41, 9 July 2020 (UTC)

Medical electronic 3330303

Op-amp with nagative positive feedback 2402:3A80:16E1:87D:0:3B:122A:2A01 (talk) 08:57, 22 February 2022 (UTC)

@ 159.192.43.62 (talk) 14:20, 13 January 2024 (UTC)

MANAGING 196.189.127.101 (talk) 03:32, 21 March 2024 (UTC)

Privacy confidentiality

I am Kedrian Hambrick I am American this is a privacy notice for Indonesia

I do not want to share my or financial credentials money property or assets.
They are strictly confidential.
Under the FTC state law the Consumer Federal Protection act the Civil rights act the United States Constitution and Civil rights laws.
No one has the right to use share trade link group dictate withhold supervise narrate access or alter my personal or financial credentials money property assets or credit. Including postal mail in home or telephone.  I am not religious.My personal and financial information is Strictly Confidential 
This means media TV radio news and news papers I have the right to my privacy do not reuse post share or broadcast my personal or financial information in any way.
Remoove all links withholdings restrictions grouping alternits return all financial credentials money  credit property or assets u have received I haven't authorized the use the share the trade or the grouping including religious groups and medical institutions.You have received them fraudulent.

Do not use share trade or sale my personal information

I am protected by the United States Constitution and Civil rights laws that protect all Americans 
Sincerely Kedrian Hambrick 2601:C4:CA00:8160:98F8:9824:ACE5:56AB (talk) 17:41, 29 September 2024 (UTC)