1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Homework Help: Digital Control Systems

  1. Aug 6, 2015 #1
    < Mentor Note -- thread moved to HH from the technical engineering forums, so no HH Template is shown >

    Please see question no. 1 in the attachment.
    If i take the inverse Laplace transform or the Z transform, how can i actually get the value of sampling period 'T'.
    or is there any other way to solve this?

    Attached Files:

    Last edited by a moderator: Aug 6, 2015
  2. jcsd
  3. Aug 6, 2015 #2
    Ok i took the inverse Laplace transform.
    I got

    20sin(1.3t) /2.7e^(t/2)
    Is the frequency of the wave (2pi f)=1.3 ?

    Or do i have to consider the exponential term as well?
  4. Aug 6, 2015 #3


    User Avatar
    Gold Member

    Do you know what the nyquist theorem is?
  5. Aug 6, 2015 #4
    Well what i am trying to do is,
    Finding out the frequency of the continuous signal.
    Two times this frequency is the minimum sampling frequency,and inverse of that is the maximum time period.
    Please correct me if i am wrong.
    Thank you.
  6. Aug 6, 2015 #5


    User Avatar
    Gold Member

    Almost. You need to take into account the frequency of your transient, as well as your steady state signal. . The initial function you are given is your function in the frequency domain. If you get a bode plot, or look at the poles, you should get an idea of what the frequencies will be.
  7. Aug 6, 2015 #6

    rude man

    User Avatar
    Homework Helper
    Gold Member

    Is F(s) a transfer function or a response to a delta (impulse) or step input? I'm guessing it's the former even though f(t) looks like a response. But if it's a response you'd have to know what the input is, remove it from F(s), then proceed as below. (You could assume a delta function input of course, in which case the response and transfer functions would be the same.)

    In either case there is no transient to consider. You have a low-pass filter with either two real poles or one complex-conjugate pair (hint: which is it?). Draw the Bode frequecy plot, then use the Nyquist theorem to come up with the min. sampling rate by picking the rolloff frequency off the plot. (NOTE: that rate will be an approximation. Theoretically, any finite frequency is passed by the network to some extent so an infinitely high sampling rate would be required for 100% accuracy in restoring to the time domain, but you have to cut off at some point in reality.)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted