c2d with 'impulse' method

According to the z-transform theory, the z-transform of the step unit u(t) is z/(z-1), regardless the sampling time. Aldo the s-transform of u(t) is 1/s. I want to obtain in Matlab z/(z-1) using the fucntion c2d for a sampling time T=0.1. This is my code:

T=0.1.

g_s = tf([1],[1 0]);

g_z = c2d (g_s, T, 'impulse');

However, the result for g_z is: 0.1*z/(z-1). May anyone explain why the factor 0.1 (the sampling time) appears in g_z? Which command should I use to obtain z/(z-1)? I would appreciate any help.

PS: I also observed that when I plotted the impulse response of g_z using the command impulse(g_z), the plot was right.

Answer 1

[Paul]

I don't think you can use c2d() and get z/(z-1) with a sample time T = 0.1.

In, short, the factor of 0.1 is needed to have the gain of g_s and the gain of g_z match in the frequency domain. You can see this with:

bode(g_s,g_z)

impulse(g_z) gives the "right" plot because the Control System Toolbox command impulse(g_z) does not return the inverse z-transform of g_z, or the unit pulse response. Rather, it returns the ouput in response to a unit pulse scaled by 1/T. So the output of impulse(g_z) is really the inverse z-transform of 1/T*g_z = 1/T * T * z / (z-1) = z/(z-1) which are the samples of the continuous time unit step.

[Buddy]

Ok. Thanks Paul. Your explanation is clear.