Ch7.

Steady-state errors

Steady-state error 1
 Learning Objectives
• Find the steady-state error for a unity feedback
system (Sections 7.1–7.2)
• Specify a system’s steady-state error performance
(Section 7.3)
• Design the gain of a closed-loop system to meet a
steady-state error specification (Section 7.4)
• Find the steady-state error for disturbance inputs
(Section 7.5)
• Find the steady-state error for nonunity feedback
systems (Section 7.6)
• Find the steady-state error sensitivity to parameter
changes (Section 7.7)Steady-state error 2
 Steady-state Error
• Control systems have the following three important
performance specifications:
– Transient Response
– Stability
– Steady-state error
•Steady-state error is the difference between the input and
the output for a prescribed test input as
•We hope that tends to either zero or a small number.
Note: steady-state error analysis is only meaningful for
stable systems. If a system is unstable, then =.

Steady-state error 3
 Steady-state Error (cont.)
• We are interested in the
steady-state error of a
system in response to
three different inputs:
step function, ramp
function, and parabola
function, which
correspond to real-world
control inputs for
constant position,
constant velocity, and
constant acceleration,
respectively. Steady-state error 4
Steady-state Error (cont.)

Steady-state error 5
Steady-state Error Examples of Step Input

Steady-state error 6
Steady-state Error Examples of Ramp Input

Steady-state error 7
Steady-state Error of Unity Feedback System
• Consider the following system with step input and pure gain K in
the forward path. Do you think the steady-state error can possibly
be zero?

• The steady-state error cannot possibly be zero. If it’s zero, then the
output c(t) is also zero.

• In order to reduce the steady-state error, the best we can do is to

increase the gain.
Steady-state error 8
Steady-state Error of Unity Feedback System
• Now, consider the following system with step input and an
integrator in the forward path. Do you think the steady-state error
is zero now?

• The steady-state error is now zero: when c(t) increases, e(t)=r(t)-c(t)

will decrease until the error reaches zero. At this time, the output
c(t) still has a non-zero value because an integrator can have a
constant output without input.
• A motor can be considered as an integrator: when a voltage is
applied to it, the shaft will turn; when the voltage is removed, the
motor will stop and remain at its present output position. So we can
have an angular displacement output without input.
Steady-state error 9
Steady-state Error of Unity Feedback System
• When analyzing steady-state errors, it is more convenient to
represent a closed-loop system using the following form, in
which the error is shown as the output:

• Because E(s)=R(s)-C(s) and C(s)=R(s)T(s), we can eliminate C(s)

and write E(s) as:
E(s)=R(s)[1-T(s)]
Using the above equation, we could get by first finding and then
solve for its inverse Laplace transform. However, there exists an
alternative, which is much simpler in most cases.

Steady-state error 10
Steady-state Error of Unity Feedback System
• The alternative uses the final value theorem of Laplace Transform:
.
Let’s see how it is derived for functions f(t) and F(s). We first start
with the differentiation theorem:

As

It yields:

Note: the final value theorem only works when (i) E(s) has poles
only on the left-half plane and at most, one pole at the origin and
(ii) T(s) is stable. Otherwise, the result is meaningless.
Steady-state error 11
Steady-state Error of Unity Feedback System
• Example: Find the steady-state error when input is unit step
function and
Solution: We first obtain E(s):

Since T(s) is stable, E(s) has two poles on the left-half plane and
one pole at the origin. Therefore, we can apply the final value
theorem and obtain:

Steady-state error 12
Steady-state Error of Unity Feedback System
• We now evaluate the steady-state error when the open-loop
transfer function G(s) is given.

From E(s)=R(s)-C(s) and C(s)=E(s)G(s), we can once again

eliminate C(s) and write:
.
After we verify that the closed-loop system is stable, we can apply
the final value theorem to obtain the steady-state error:

Next, we discuss the cases that the input is step, ramp, and parabolic.

Steady-state error 13
Steady-state Error of Unity Feedback System
• Step input: R(s)=1/s. We have
a.k.a position
constant Kp
We call the term the DC gain of the forward transfer function. In
order to have zero steady-state error, we must have:
=,
which means that G(s) must have the following form with

In another word, G(s) must have at least one pole at the origin; o.w.,
the steady-state error won’t be zero, e.g., when n=0:
Question: what does poles at the origin imply?
Answer: It implies that in order to achieve zero stead-state error,
the forward path must have at least one pure integrator.
Steady-state error 14
Steady-state Error of Unity Feedback System
• Ramp input: R(s)=1/s2. We have

a.k.a velocity
constant Kv

To have zero steady-state error, we must have:

which means in the form of G(s) below:

i.e., the system must have two pure integrations in the forward path.
If only one exists, Kv and the steady-state error are finite. If none
exists, then Kv=0 and steady-state error is infinity.
Steady-state error 15
Steady-state Error of Unity Feedback System
• Parabolic input: R(s)=1/s3. We have
a.k.a
acceleration
constant Ka
To have zero steady-state error, we must have:

which means in the form of G(s) below:

i.e., the system must have three pure integrations in the forward
path. If only two exist, Ka and the steady-state error are finite. If none
or only one exists, then Ka=0 and steady-state error is infinity.
Steady-state error 16
Steady-state Error of Unity Feedback System
• Example: Find the steady-state errors for inputs of 5u(t), 5tu(t),
and 5t2u(t) to the system shown below.

• Solution: We first verify that the closed-loop system is stable

(details omitted).

Steady-state error 17
Steady-state Error of Unity Feedback System
• Example: Find the steady-state errors for inputs of 5u(t), 5tu(t),
and 5t2u(t) to the system shown below.

Solution: We first verify that the closed-loop system is stable.

Steady-state error 18
The System Type of Unity Feedback Systems
• We define system type to be the value of n in the denominator.
It’s also the number of pure integrations in the forward path.

Steady-state error 19
Interpreting Steady-state Error Specification
• Just as rise time, settling time, %overshoot, etc., are used as for
the specifications of a control system’s transient response, the
position constant, velocity constant, and acceleration constant can
be used for the stead-state error specification of a control system.
• Example: if a control system has specification Kv=1000, what can
we conclude?
Solution:
1. The system is stable
2. The system is type 1 (because only type 1 systems have finite
velocity constant).
3. The ramp input is the test signal (because Kv is specified).
4. The steady-state error between the input ramp and output
ramp is 1/1000.
Steady-state error 20
Meeting a Steady-state Error Specification
• Example: given the control system below, find the value of K so that
there is 10% error in the steady state.

Hint: What type of system is it? What kind of input do we have in this
problem?
Solution: It’s a type 1 system and there is a constant steady-state
error of 0.1. Therefore, the input must be a ramp. We have:

Finally, we use the Routh-Hurwitz criterion to verify that it’s stable.

Question: What if K=672 yields undesirable transient response?
This question will be answered in Ch.9 when we discuss PID controller.
Steady-state error 21
 Steady-state Error Under Disturbances
• A well-designed feedback control system should compensate
for undesirable disturbances. Consider the system below:

We can represent c(s) in the following two forms:

After we eliminate C(s), we get:

Error caused by the input Error caused by the disturbances

Steady-state error 22
Steady-state Error Under Disturbances
• Suppose the system is stable, i.e., the roots of 1+G1(s)G2(s) are
all in the left-half plane, we can apply the final value theorem
to obtain the steady-state error:

where
Error contributed by R(s).
This is what we have discussed.

Error contributed by the

disturbances. This is what we
will focus next.
Steady-state error 23
Steady-state Error Under Disturbances

• To analyze we assume D(s) is unit step. Then, we obtain:

The error can be reduced if we increase the DC gain of G1(s)

or decrease the DC gain of G2(s). This can also be seen by
examining the diagram below:

Note that this is identical to the original

diagram except that R(s) is now assumed
to be zero.
Steady-state error 24
Steady-state Error Under Disturbances
• Example: Find the steady-state error component due to a step
disturbance for the system below.

Solution: Let’s first use this diagram to derive the diagram

from the previous slide.
Then, we can solve the error
caused by the disturbance:

Steady-state error 25
 Steady-state Error for Nonunity Feedback
Systems
• Our discussion thus far has been restricted to unity feedback
systems. Can we still figure out the steady-state error for
nonunity feedback control systems, which are very common?
• The answer is “Yes”. We just need to transform the nonunity
feedback system to a unity feedback system.
• Let’s consider the general feedback system below. Our first
step is to push the input transducer G1(s) to the right of the
summing junction.

Steady-state error 26
 Steady-state Error for Nonunity Feedback
Systems (cont.)
Next, we manually create a
unity feedback loop:

Finally, we simply the feedback loop:

This is a unity feedback

Then, we combine the two system!
parallel branches in the middle:
Steady-state error 27
 Sensitivity
• In a nutshell, sensitivity is the degree to which
changes in system parameters affect system
performance:

Steady-state error 28
 Sensitivity (cont.)
• Example: Given the system below, calculate the sensitivity of the
closed-loop transfer function to changes in the parameter a. How
would you reduce the sensitivity?

Solution: We first obtain the closed-loop transfer function:

Using the sensitivity formula,

function of s

Can be reduced
by increasing K
Steady-state error 29
 Sensitivity (cont.)
• Example: In the previous system, find the sensitivity of the
steady-state error to changes in parameter K and parameter a
with ramp inputs.

Solution: The steady-state error is:

The sensitivity of to changers in parameter a is:

The sensitivity of to changers in parameter K is:

Steady-state error 30