Community Tip - When posting, your subject should be specific and summarize your question. Here are some additional tips on asking a great question. X
Hi,
I am trying to plot nyquist diagram of my transfer function. I dont know what could be wrong. There should be a nice curved diagram, and I get such a mess. Can anyone help please? MathCAD file is attached. Thank you.
Are you sure about the "1 - e^..." in the denominator of y1(w)? Shouldn't it be simply 1 +/- s(w)*T ?? Or rather "e^..." without the "1 - " ??
Yes, I am sure of that. I tried it also with your suggestion, but I didnt get the "nyquist curve".
This denominator 1 - exp(-j*w*T) is getting zero(!) for every w which is an integer multiple of pi/16 (=2*pi/T) and your y1(w) is going high up to infinity that way, regardlesss of W(w). This happens quite often in the w range of 1 to 1000 and you have seen this effect already in the 3D-plots of your previous posts (the spikes). Because you set w to integer values you never get an w which makes the denominator to zero, but obviously you can't expect a smooth curve that way.
PS: Watch your plot for w:=0,0.0001 .. 1 (or 10, or 100) and look at the axis scales
It's been a while since controls class, so I'm a bit rusty. I take it that W is your system (out/in)? So in that case, shouldn't you be looking at the real and imaginary parts of W? Is this close to what you're looking for?
Have you considered using a log axis:
I am affraid that it is not what I was expected, but Thank you.
BTW: How can I close discussion? Just by checking correct answer? Thx
Did you look at Rogers post? Looks plausible to me.
You can't close a discussion literally. You may chose a post as the correct answer or chose "assume as answered" to indicate you don't need any help on the topic anymore.
Or you may comment on the various posts which tried to help telling why you think the correct output should be different. If you ask a question in any forum people usually will try to help, so you should come back yourself and show up. Your comments also may help others in the future who search the forum archive for a similar problem.
As your problem seems not to be solveed so far in your opinion you may either state that you are not interested in a solution anymore (cause the work is already overdue, e.g.) or provide additional information. Where are those equations coming from and how were they derived. What output do you want to duplicate, what kind of curve do you know is the correct one, etc.
Good luck!