[Octave-bug-tracker] [bug #43586] tf function outputs wrong results

[Lukas Reichlin]

Follow-up Comment #1, bug #43586 (project octave):

The tf function accepts any string as transfer function variable, even
'1/(s+1)':


octave:2> s = tf ('1/(s+1)')

Transfer function 's' from input 'u1' to output ...

 y1:  1/(s+1)

Continuous-time model.
octave:3> sys = (s - 1)/(s^2 + 2*s + 1)

Transfer function 'sys' from input 'u1' to output ...

             1/(s+1) - 1       
 y1:  -------------------------
      1/(s+1)^2 + 2 1/(s+1) + 1

Continuous-time model.
octave:4> s = tf ('string')

Transfer function 's' from input 'u1' to output ...

 y1:  string

Continuous-time model.
octave:5> sys = (s - 1)/(s^2 + 2*s + 1)

Transfer function 'sys' from input 'u1' to output ...

            string - 1       
 y1:  -----------------------
      string^2 + 2 string + 1

Continuous-time model.
octave:6> 


This can be useful, but also very misleading. As you can see below, the result
of your system c is always computed correctly:


octave:9> s = tf ('string')

Transfer function 's' from input 'u1' to output ...

 y1:  string

Continuous-time model.
octave:10> s(1)
ans =  0 + 1i
octave:11> s = tf ('s')

Transfer function 's' from input 'u1' to output ...

 y1:  s

Continuous-time model.
octave:12> s(1)
ans =  0 + 1i
octave:13> 


One could argue that only 's' and 'z' (and maybe 'p' and 'q' for time domain)
should be accepted by tf, though.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?43586>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/