Frequency response function...

Hi!

I have an s domain transfer function H(s)
and trying to determine the response of the system
to the unit step function.

I tried this:

Y(s) = H(s) * X(s) = H(s) * Hc(s) = H(s)/s

(* is product, not convolution)

But the frequency response using freqs(Y(s)) is wrong.
(I bet it must be obvious to someone why). The
frequency spectrum should a have a value at DC (omega*j = 0)
but in this case it is undefined.

What is going wrong?

Thanks!
Atmapuri.

Hi!

Found the solution. The frequency response should be evaluted
at some s = alfa + wj where alfa is different from zero. By performing
IFFT on the sampled frequency response, the time series then has to
be compensated by multiplying it with e^(alfa*t) to obtain the true
step response.

Thanks!
Atmapuri

"Atmapuri" <janez.makovsek@usa.net> wrote in message
news:aKlTd.9507$F6.1836303@news.siol.net...
> Hi!
>
> I have an s domain transfer function H(s)
> and trying to determine the response of the system
> to the unit step function.
>
> I tried this:
>
> Y(s) = H(s) * X(s) = H(s) * Hc(s) = H(s)/s
>
> (* is product, not convolution)
>
> But the frequency response using freqs(Y(s)) is wrong.
> (I bet it must be obvious to someone why). The
> frequency spectrum should a have a value at DC (omega*j = 0)
> but in this case it is undefined.
>
> What is going wrong?
>
> Thanks!
> Atmapuri.
>
>

"Atmapuri" <janez.makovsek@usa.net> writes:

> The frequency response should be evaluted
> at some s = alfa + wj where alfa is different from zero. 

"What's it all about, Alfa?"
-- 
%  Randy Yates                  % "Remember the good old 1980's, when 
%% Fuquay-Varina, NC            %  things were so uncomplicated?"
%%% 919-577-9882                % 'Ticket To The Moon' 
%%%% <yates@ieee.org>           % *Time*, Electric Light Orchestra
http://home.earthlink.net/~yatescr