Talk:Control Systems/Feedback Loops

From Wikibooks, the open-content textbooks collection

Open Loop vs Closed Loop: There are some signs wrong in this section - the junction shown in the block diagram is a summing one, and so the transfer function should have a negative in it. (Feedback is typically negative, and so the form of the equation is what is usually seen, so maybe the block diagram is what needs changing)

I also thought the open loop would just be KGp(s) not KGp(s)Gb(s) a shown but I don't know how authors are using these representations in later parts of this book and so leave it to someone else to figure out what the correction should be.

Thanks for the comment. You're right that maybe this diagram is a little confusing. However, even if the sign is "+" at the adder, the gain of the feedback path Gb(s) could still be negative. Using a "+" sign is just a general form, and changing the sign of the gain term can give both positive and negative feedback. Negative feedback is much more common in control systems, but positive feedback has uses in other types of system modeling. The focus here is obviously on controls, but having a little bit of generality is never a bad thing. We could leave an explanatory note here to clarify the situation, you can do it if you want or I can add it later.

I've always read that the open loop gain was KGp(s)Gb(s), and that's how it's used throughout the rest of the text. Again, if there are alternatives we should probably leave an explanatory note about it. Do you want to do it or should I? --Whiteknight (Page) (Talk) 13:17, 2 December 2008 (UTC)

[edit] Elevator Example

In the Elevator Example the following text appears:

"Where we are measuring in units of "floors". At time t0, the error signal is:

e(t0) = x(t1) − y(t1) = 5 − 1 = 4

Which means that the elevator needs to travel upwards 4 more floors. At time t1, when the elevator is at the second floor, the error signal is:

e(t1) = x(t1) − y(t1) = 5 − 2 = 3"

I think that there must be an error here, shouldn't e(t0)=x(t0)-y(t0), rather than e(t0) = x(t1) − y(t1)? I'm probably failing my controls class right now, so I didn't want to edit the page, just thought I'd bring this point up.