[SciPy-user] Small problem with scipy.signal

[Stefan van der Walt]

Hi Clovis

On Tue, Dec 11, 2007 at 05:09:34PM -0200, Clovis Goldemberg wrote:
>    I am having some problems with scipy.signal
>    (specifically with scipy.signal.step) which can be
>    explained by the following text,
>    which was edited with some comments.
>    Configuration is also given in the text.

[...]

>    >>> num=[1.0]
>    >>> den=[1.0,0.0]
>    >>> ss.step((num,den))
>    #####################################
>    # This fails!
>    # The transfer function would be "1/s"
>    # which is a simple integrator.
>    #####################################

Looks like the system is converted to state space.  After that,
somewhere along the line A is inverted, which causes the problem.  I
think a person can get away with only inverting [I-A], but I'll have
to poke around the source a bit to be sure.

Regards
Stéfan