HW5, MAE 171. Spring 2005. UCI
Nasser Abbasi
Calculate and plot the unit-step response at the sampling instances for the
case when
First find the open loop
Hence
Hence the closed loop transfer function is, noting that
For
we get from (1)
Now since the input is a unit step, then
Hence the output
Hence
From tables, the inverse Z transform of
for
Hence
And
Hence The inverse Z transform of
is
This is a plot of
Calculate the system unit step response for the analog system. Plot the response on the same graph as the result of (a).
The closed loop transfer function is
Hence
Hence
Here is a graph which shows the analog and discrete responses
Let
and
.
Calculate the unit step response and plot these on the same plot
From equation (1) in part(a), we have
for
Now since the input is a unit step, then
Hence the output
Hence
From tables, the inverse Z transform of
Hence
And
Hence The inverse Z transform of
is
The following is the plot comparing discrete system response for
and
with the pure analog system.
Notice the effect of changing
on the discrete response. For larger
the system had a larger overshoot and was less damped. As sampling rate
increased the response exhibited over damping.
Consider the sampeld-data systems with the following char. equation
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
This table summarizes the results.
system | ![]() |
![]() |
|||||||
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
||||
![]() |
![]() |
![]() |
N/A | N/A | ![]() |
N/A | N/A | ||
![]() |
![]() |
![]() |
N/A | N/A | ![]() |
N/A | N/A | ||
![]() |
![]() |
![]() |
N/A | N/A | ![]() |
N/A | N/A | ||
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
||
![]() |
![]() |
![]() |
N/A | N/A | ![]() |
N/A | N/A | ||
![]() |
![]() |
N/A | N/A | N/A | N/A | N/A | N/A | ||
![]() |
![]() |
N/A | N/A | N/A | N/A | N/A | N/A | ||
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
From the table above we observe that the
(rise time) and
(settling time) varies with
only. While
is not affected.
As
becomes larger (slower sampling), the response will take longer to rise and to
settle. The ratio of the change in
is the same ratio by which the response will take longer to settle for these
systems. To classify the responses, this table and the plots show the
classification. This Next plot is the transient response characteristics of
the systems for
system | ![]() |
||
![]() |
![]() |
stable | |
![]() |
![]() |
stable | |
![]() |
![]() |
stable | |
![]() |
![]() |
stable,exponential damped sinusoidal | |
![]() |
![]() |
stable,damped | |
![]() |
![]() |
marginal unstable. | |
![]() |
![]() |
unstable, exponential | |
![]() |
![]() |
stable, exponential damped sinusoidal |
What can you determine about the system natural response characteristic for
each system for
?
For second order system we can determine the transient response characteristic as follows.
From the location of the discrete transfer function poles, we determine
using the three equations below. And in turn from
we can determine, using the simplified equations, the rise time
,
the maximum overshoot
and the settling time
In addition we can find the time constant
from knowing
Let a Z system pole be expressed as
,
then these are the equations to use: (see book page 216 for the derivation)
We can now determine
for the system response to a unit step.
For
use
For
(since this is a unit step response, hence the final value is 1), we get
For
use
where
For a first order discrete system, the response will not exhibit oscillations
if the root is positive. So for a first order system with positive root we can
only say something about
,
the time constant (which is the time required for response to reach
of its final value.)
Assume root of Z system is at
but since this is a real positive root, then
,
and
and since
,
we get that
where
here is the magnitude of
Hence from
We see that
hence since
then we get
Now, For a first order discrete system, the response will exhibit oscillations if the root is negative.
Assume root of Z system is at
but since this is a real negative root, then
,
and
and since
,
we get that
where
here is the magnitude of
and hence since we are taking the natural log of a negative value, then
is a complex value say
,
from which we find
and
and then we can find the rest of the transient response properties since
and
we find
from
Now we can start to solve the problems
(1)
Since is a first order system. root
hence from
we get
(2)
Since is a first order system. root
hence from
we get
Notice that for a small change in the location of the root
(
change), the system now responds
times as fast.
(3)
Since is a first order system. root
hence from
we get
Notice that for a small change in the location of the root
(
change), the system now responds
times as fast.
(4)
Since is a first order system. since root
hence transient response will oscillate.
To evaluate
since
Let
Hence
From (1) we
get
Hence
Hence
and
Hence now we can obtain the transient response characteristics.
(5),
Solution is:
Since there are no complex poles, and both poles positive, then system will
not oscillate even though it is a second order system. Hence we know that
and
and
in this case since we will use
here as the time to go from
to
of the final value.
using
there
are 2 time constants since there are 2 roots, we get
and
We pick the larger time constant since that is associated with the dominant pole.
hence
(6)
poles are at
This is a marginally unstable system. For an input of
,
we see that the pole of the input equals the pole of the system at
,
hence resonance will occur. So not applicable to determine the transient
characteristic.
(7)
,
Solution is:
We have a pole outside the unit circle. Hence this is unstable system. So not applicable to determine the transient characteristic.
(8)
,
Solution is:
Here the pole is
where
and
rad
Redo part(a) for
(1)
Since is a first order system. root
hence from
we get
(2)
Since is a first order system. root
hence from
we get
(3)
Since is a first order system. root
hence from
we get
(4)
As from part (a), we get
Hence
and
Hence now we can obtain the transient response characteristics.
(5),
Solution is:
using
there
are 2 time constants since there are 2 roots, we get
and
We pick the larger time constant since that is associated with the dominant pole.
hence
(6)
poles are at
This is a marginally unstable system. For an input of
,
we see that the pole of the input equals the pole of the system at
,
hence resonance will occur. So not applicable to determine the transient
characteristic.
(7)
,
Solution is:
We have a pole outside the unit circle. Hence this is unstable system. So not applicable to determine the transient characteristic.
(8)
,
Solution is:
Here the pole is
where
and
rad
From the table above we observe that the
(rise time) and
(settling time) varies with
only. While
is not affected.
The following system has parameters
Find
and
for the sample period
Find open loop
then the closed loop
then find the zeros of the characteristic equation, and from these find
and
using the relations
So
Characteristic equation
,
Solution is:
Hence
and
rad
Hence
redo part (a) with
So
Characteristic equation
,
Solution is:
,
Solution is:
Hence
and
rad
Hence
repeat part(a) for analog system.
Hence
Hence
Give table listing the three parameters as function of sampling frequency
.
State the result of decreasing the sampling frequency on the parameters.
Solution summary table
T sec | ![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
analog | ![]() |
![]() |
![]() |
We see as the sampling frequency
becomes smaller, then the damping ratio decreases, this means the response
will become more oscillatory in the transient stage with larger overshoot.