Control and command

Gado TCHABODE
Rev 1.1 2023-2024
1
 1. Generalilties
1.1 Definitions
Automation : Set of processes designed at reducing or eliminating human intervention in production
processes. A system is said to be automated when it can autonomously manage a pre-established work cycle
which is made up of sequences or stages.
➢ Combinational automated system (combinational logic),
➢ Sequential automated system (sequential logic, programmable logic controller),

Control : Control is a set of mathematical theories and reasoning techniques that relate to decision-making
and system control. The aim of automatic control is to develop the control of the system under consideration.
There are different possibilities to carry out this command. We distinguish :
➢ Open loop controlled systems
➢ Closed-loop controlled system (controlled or regulated system).

In this course we will only be interested in the last two types of command

2 Gado TCHABODE
 1. Generalities
1.1 Definitions
System or process: It is a (complex) set organized for a fixed purpose or set of
physical - chemical processes in evolution. This set has inputs (managed or disturbance) and outputs. It
is made up of an adjustment system and the system to be controlled.

Signal : Physical quantity generated by a device or translated by a sensor (temperature, flow rate, etc.).
A quantity liable to change in value, it is associated with the physical quantity that it represents at a
given moment in a system.

Example: 𝑇(12) = 150°𝐶(temperature at 12s equals 150°C)

3 Gado TCHABODE
 1. Generalilties
1.1 Definitions
Input signal : independent of the system, it can be managed (setpoint, order) or undergone (disturbances).

✓ Setpoint (instruction), order : Set a goal for the system, for example set a temperature at 37°C or set
the trajectory of an airplane.
✓ Disturbances : A disturbance is a signal which tends to modify
the output value of the system. Their effect interfere the proper
functioning of the system. To recall this inconvenient character
for the user, we usually call them disturbances. (random variable
whose origin is unknown)

Output signal : depend on system and input signal. It is a signal that

we must control.

4 Gado TCHABODE
 1. Generalilties
1.1 Definitions
Command : Action likely to change the state of the system to be commanded. It is developed according to
orders (or instructions).
Example: Setting the temperature of an oven

In this functional diagram the large rectangle represents the physical system. The lines carrying arrows
represent signals, while the "boxes" represent relations between these signals (physical laws). The arrows
indicate that the phenomenon is generally not reversible.

5 Gado TCHABODE
 1. Generalilties
1.2 Structure of a servo system
An open-loop control system is symbolically represented by a diagram, which translates the existence of a
mathematical relationship between the input and the output.
Definition : A system is said to be open loop when there is no information on the variable to be controlled.

Example: Setting the temperature of the furnace The setting of the temperature of the furnace is carried out
by an outside person (from the control room), he therefore has no information on the quantity to be regulated
( Ts ).

Advantage : Stable system.

Disadvantages : Blind system,

no correction (sensitive to
disturbances)

6 Gado TCHABODE
 1. Generalilties
1.2 Structure of a servo system
The relationship between input and output is changed by disturbance inputs. Because of these
disturbances, when a certain value of the command is displayed, it is not certain in practice that the
output will have the desired value.

Servo-control is a controlled system whose output must follow the setpoint as faithfully as possible (variable
setpoint). Ex: trajectory tracking by a missile

The oven temperature is adjusted by an external

person (from the control room) with, in this case,
information on the quantity to be adjusted ( Ts ).

The temperature adjustment acts on the

adjustment device (the valve) according to the
difference between the desired value and the
actual value.
7 Gado TCHABODE
 1. Generalilties
1.2 Structure of a servo system

8 Gado TCHABODE
 1. Generalilties
1.2 Structure of a servo system

Advantages : precise, the speed of the system can be adjusted, possibility of correction, and much less
sensitive to disturbances.
Cons : can be unstable

1.3 Classification of systems

➢ Linear / Nonlinear : Most real systems are nonlinear. However, if the domain of
variations of the variables is small, it is always possible to linearize the system inside this
domain. For linear systems the principle of superposition applies.
➢ Invariant / Varying over time:
➢ Continuous / Discrete : In a continuous system the variables are continuous functions of time t. In a
discrete (sampled) system one or more variables are known only at
discrete instants of time.
➢ Single input - Single output / Multiple input - Multiple output :

9 Gado TCHABODE
 2. Study of linear systems
2.1 Complex numbers (reminder)

➢ Definition

➢ Properties of complex numbers

➢ Representation of a complex number

2.2 Typical input

➢ Step (unit)
It is a zero function for 𝑡 < 0, equal to a
constant for 𝑡 ≥ 0. Its representation is
given in the figure below. We denote it
by 𝑒(𝑡) = 𝐴. 𝑢(𝑡) , with u(t) the unit step

10 Gado TCHABODE
 2. Study of linear systems
2.2 Typical input
The step response is the response of a system following a step input. The system is initially at rest, 𝑒 𝑡 = 0
for 𝑡 < 0 the system is said to start from rest. The figure below illustrates the step response (response to a
step type input) of a system.

➢ Speed

➢ Damping

➢ Accuracy

➢ Exceeding

11 Gado TCHABODE
 2. Study of linear systems
2.2 Typical input
➢ Ramp

It is a zero function for t < 0, of constant slope A for 𝑡 ≥ 0 .

Its graphical representation is given in the figure on the
right. We notice it 𝑒(𝑡) = 𝐴 ∙ 𝑡 ∙ 𝑢(𝑡)

The system is initially at rest, 𝑒(𝑡) = 0 for 𝑡 < 0 the system is said to start from rest. The figure
below illustrates a system's response to a ramp type signal.

12 Gado TCHABODE
 2. Study of linear systems
2.2 Typical input
➢ Impulse

It is the limit of a null function outside a small interval of

time, but in this interval it takes values large enough for
its effect to be noticeable.

It is noted: 𝑒 𝑡 = 𝐴 ∙ 𝛿(𝑡) is the Dirac function or unit impulse. It is a vertical line at

the abscissa t = 0.

The response of a system to an impulse-like input is called an impulse response . It makes it possible to
characterize the behavior of the system in the face of a shock or a parasite.
13 Gado TCHABODE
 2. Study of linear systems
2.2 Typical input
➢ Sinusoid or harmonic input

It is an input which is a sinusoidal function

of time. The figure beside represents its evolution
over time. Mathematically it is given by the equation
𝑒 𝑡 = 𝐸𝑚 sin(𝜔𝑡 + 𝜃)
The steady-state response of a linear system
𝐸𝑚 : 𝑙 ′ 𝑎𝑚𝑝𝑙𝑖𝑡𝑢𝑑𝑒
𝜔ൗ : 𝑙𝑎 𝑓𝑟é𝑞𝑢𝑒𝑛𝑐𝑒 𝐻𝑧 (process) to a sinusoidal excitation of amplitude 𝐸𝑚
2𝜋 is a sinusoid of the same frequency of amplitude 𝑆𝑚
𝜔: 𝑙𝑎 𝑝𝑢𝑙𝑠𝑎𝑡𝑖𝑜𝑛 𝑜𝑢 𝑓𝑟é𝑞𝑢𝑒𝑛𝑐𝑒 𝑎𝑛𝑔𝑢𝑙𝑎𝑖𝑟𝑒 (𝑟𝑎𝑑ൗ𝑠)
out of phase on the excitation by an angle 𝜑
The characterization of the system
includes the determination of the
𝑆
amplitude ratio 𝑚 and the phase
𝐸𝑚
𝜑 for a sequence of pulsations𝜔.

14 Gado TCHABODE
 2. Study of linear systems
2.2 Typical input
➢ Sinusoid or harmonic input

The study is carried out in steady state. The results of these measurements
can be graphically represented in 3 different forms

Bode Nyquist Black

15 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
The Laplace transformation allows the determination of the response of a linear system to
zero excitation for t < 0 and of any form for t ≥ 0. The variable is denoted p. It is a complex variable
such that 𝑝 = 𝑎 + 𝑗𝜔. We then obtain complex functions of the variable p, example F (p).
We denote by:
▪ f(t) is a function of time such that f(t)=0 for t < 0
▪ p: complex variable
▪ L: Laplace transform notation
▪ F(p): Laplace transform of f(t)
➢ Definition
To any null time function f(t) for t < 0 (so-called causal function: the effect never precedes the cause) we
match a function F(p) of the complex variable p, such that:
+∞
𝐹 𝑃 =𝐿 𝑓 𝑡 =න 𝑓(𝑡) ∙ 𝑒 −𝑝𝑡 𝑑𝑡
0

16 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ LAPLACE transform of the step
∞
0, 𝑡 < 0 −𝑝𝑡
𝐴 −𝑝𝑡 ∞ 𝐴
𝑓 𝑡 =ቊ ℒ 𝑓(𝑡) = 𝐹 𝑝 = න 𝐴 ∙ 𝑒 𝑑𝑡 = − [𝑒 ]0 =
𝐴, 𝑡 ≥ 0 0 𝑝 𝑝
➢ LAPLACE transform of the ramp
∞ ∞
0, 𝑡 < 0 −𝑝𝑡
𝐴 −𝑝𝑡 ∞
𝐴 −𝑝𝑡
𝐴 𝑒 −𝑝𝑡 ∞
𝑓 𝑡 =ቊ ℒ 𝑓(𝑡) = 𝐹 𝑝 = න 𝐴𝑡 ∙ 𝑒 𝑑𝑡 = [−𝑡𝑒 ]0 + න 𝑒 𝑑𝑡 = [− ]0
𝐴𝑡, 𝑡 ≥ 0 0 𝑝 𝑝 0 𝑝 𝑝
𝐴
ℒ 𝑓(𝑡) = 𝐹 𝑝 = 2
𝑝
➢ LAPLACE transform of the sinusoid
0, 𝑡 < 0 After two integrations by parts, it is easy to show that
𝑓 𝑡 =ቊ
𝐴 ∙ 𝑠𝑖𝑛(𝜔𝑡), 𝑡 ≥ 0
𝐴∙𝜔
ℒ 𝑓(𝑡) = 𝐹 𝑝 = 2
17
𝑝 + 𝜔2 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Properties of the LAPLACE transform

✓ Linear Superposition: If 𝑓(𝑡) and 𝑔(𝑡) have Laplace Transforms then:

ℒ [𝐴𝑓(𝑡)+ 𝐵𝑔(𝑡)]= 𝐴 ℒ[𝑓(𝑡)] + 𝐵 ℒ[𝑔(𝑡)] = 𝐴𝐹(𝑝) + 𝐵𝐺(𝑝), A and B are constants .
Example: Either 𝑓 𝑡 = 𝑒 −𝑡 − 𝑒 −2𝑡 ; ℒ[𝑓(𝑡)] = 𝐹(𝑝) = ℒ [ 𝑒 −𝑡 − 𝑒 −2𝑡 ] = ℒ [𝑒 −𝑡 ] − ℒ [𝑒 −2𝑡 ]
𝑡 1
✓ Integration: The Laplace transform of the integral of f(t) is given byℒ ‫׬‬0 𝑓 𝜏 𝑑𝜏 = ℒ[𝑓 𝑡 ]
𝑝

𝑑𝑓(𝑡)
✓ Derivative: The Laplace transform of the derivative of f(t) is given byℒ = 𝑝ℒ 𝑓 𝑡 −𝑓 0 .
𝑑𝑡
If 𝑓(0) = 0 then𝐹 𝑝 = 𝑝ℒ 𝑓 𝑡
𝑑(𝑠𝑖𝑛 𝜔𝑡 𝜔
Example :ℒ = 𝑝 ∙ ℒ 𝑠𝑖𝑛 𝜔𝑡 =𝑝∙
𝑑𝑡 𝑝2 +𝜔2

18 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Table of usual LAPLACE transforms

19 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Properties of the LAPLACE transform

✓ Translation : If a function is zero for t < 0, i.e. 𝑓(𝑡) ∙ 𝑢(𝑡), has a Laplace transform, the transform of the
function 𝑓(𝑡 - 𝜏) (delay of τ , shift of τ towards the positive times) is deduced by multiplying by 𝑒 −𝑝𝑡 . The
factor 𝑒 −𝑝𝑡 is called the lag operator.
Example :ℒ[𝑓(𝑡 − 𝜏) ∙ 𝑢(𝑡 − 𝜏)] = 𝑒 −𝑝𝜏 ℒ[𝑓 𝑡 ]

20 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Inverse LAPLACE transform
To go from the LAPLACE domain to the time domain, the inverse LAPLACE transform is calculated. We
call 𝒇 𝒕 the original of 𝑭 𝒑 and 𝑭 𝒑 the image of 𝒇 𝒕 .
One determines 𝑓 𝑡 thanks to the table of the usual transforms and by making a decomposition in simple
elements if that is necessary.
4
Example: Determine the original of 𝐹 𝑝 =
2𝑝+1

application exercise

21 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ LAPLACE transform of a differential equation
We only consider the case of a linear differential equation with constant coefficients. Consider two
time-dependent quantities x(t) and y(t) linked by the following differential equation:
𝑑𝑦 𝑑2 𝑦
𝑏0 𝑦 + 𝑏1 ∙ + 𝑏2 ∙ 2 = 𝑎0 𝑥
𝑑𝑡 𝑑𝑡
𝑑𝑦 𝑑2 𝑦
ℒ[𝑏0 𝑦 + 𝑏1 ∙ + 𝑏2 ∙ 2 ] = ℒ[𝑎0 𝑥]
𝑑𝑡 𝑑𝑡
𝑑𝑦 𝑑2 𝑦
ℒ[𝑏0 𝑦] + 𝑏1 ∙ ℒ[ ] + 𝑏2 ∙ ℒ[ 2 ] = ℒ[𝑎0 𝑥]
𝑑𝑡 𝑑𝑡
𝑑𝑦
𝑏0 𝑌 𝑝 + 𝑏1 ∙ 𝑝𝑌 𝑝 − 𝑦 0 + 𝑏2 ∙ [𝑝ℒ − 𝑦(0)]
ሶ = 𝑎0 𝑋(𝑝)]
𝑑𝑡

22 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform

The transformation of the differential equation gives F(p). Knowing x(t), the tables of transforms make it
possible to calculate X(p). We must then find y(t) from Y(p). This operation constitutes the inverse Laplace
transform. The inverse transformation is performed by using the table of transforms:

▪ Directly, if the function is in such a form that each part has an entry in the table
of transforms;
▪ Indirectly, by decomposition into simple elements then use of the table
The inverse Laplace transform of a rational fraction in p is obtained using the table of Laplace transforms
after decomposition of the fraction into simple elements and application of the properties (linearity).

𝑁 𝑝
𝑆 𝑝 = →decomposition into simple elements →then we deduce s 𝑡
𝐷(𝑝)
6
Example: Find f(t) knowing that 𝐹 𝑝 =
(𝑝 + 1)(𝑝 + 3)
23 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Transfer function
Consider any linear system with an input e(t) and an output s(t). We assume that it is governed by a
differential equation of degree n:

If we apply the Laplace transformation to both sides of this equation, while assuming
the different initial conditions are zero, we get:

This rational fraction of two polynomials of the complex variable p is called the transfer function of the
system and commonly denoted:

24 Gado TCHABODE
 2. Study of linear systems
2.3 Mathematical tools: LAPLACE transform
➢ Transfer function
As this function is a rational fraction of two polynomials in p, it is possible to factor these
two polynomials in the field of complexes. We obtain :

• The roots 𝑧𝑖 that cancel the numerator are called the zeros of the transfer function.
• The roots 𝑝𝑖 that cancel its denominator are the poles of the transfer function.
These parameters can be complex or real. We will see later in this course that the sign or the belonging to the
set of reals of these poles or zeros play very important roles in the study of systems.

The complex transfer function is obtained by replacing 𝒑 by 𝒋𝝎

𝑵 𝑗𝜔
𝐺 𝑗𝜔 =
𝑫 𝑗𝜔
25 Gado TCHABODE
 3. Study of servo systems
3.1 Transfer function

26 Gado TCHABODE
 3. Study of servo systems
3.1 Transfer function
Consider the block diagram opposite
We define :
The system is assumed to be linear
𝑌(𝑝)
➢ The forward path transfer function 𝐺 𝑝 = (differential equation with constant
𝜀(𝑝) coefficients)
𝑅(𝑝)
➢ The feedback path transfer function 𝐻 𝑝 =
𝑌(𝑝)
𝑅(𝑝)
➢ The open loop transfer function (OLTF) 𝐺 𝑝 ∙ 𝐻(𝑝) =
𝜀(𝑝)
𝑌(𝑝)
➢ The servo system transfer function or CLTF 𝑇 𝑝 = =
𝑋(𝑝)
𝜀 𝑝 1
➢ Transfer function gap 𝐹𝑇𝜀 𝑝 = =
𝑋 𝑝 1 + 𝐺(𝑝) ∙ 𝐻(𝑝)
➢ The characteristic equation of the 1 + 𝐺 𝑝 ∙ 𝐻 𝑝 = 0(FTBF denominator = 0)
system

27 Gado TCHABODE
 3. Study of servo systems
3.1 Transfer function
Special case of unit feedback
We define :
The system is assumed to be linear
𝑌(𝑝)
➢ The forward path transfer function 𝐺 𝑝 = (differential equation with constant
𝜀(𝑝) coefficients)
𝑅(𝑝)
➢ The feedback path transfer function 1=
𝑌(𝑝)
𝑅(𝑝)
➢ The open loop transfer function (OLTF) 𝐺 𝑝 =
𝜀(𝑝)
𝑌(𝑝)
➢ The servo system transfer function or CLTF 𝑇 𝑝 = =
𝑋(𝑝)
𝜀 𝑝 1
➢ Transfer function gap 𝐹𝑇𝜀 𝑝 = =
𝑋 𝑝 1 + 𝐺(𝑝)
➢ The characteristic equation becomes 1 + 𝐺 𝑝 = 0(FTBF denominator = 0)

28 Gado TCHABODE
 3. Study of servo systems
3.1 Transfer function
➢ Reduction of a block diagram to a unit feedback system
When the system is not unitary feedback, it is possible to reduce it to a unitary feedback system. In reality the
sensor does not exactly measure the output to compare it to the input. It must be taken into account that the
signal Y (p) passes through a certain function H(p). So 𝜀 = 𝑋 - 𝐻 ∙ Y and not X - Y.

In general H(p) is the transfer function of the sensor. In this case r(t) represents the measure of y(t). It is
necessary
to choose the sensor well so that its transfer function H(p) ≈ 1 in the bandwidth of the system.

29 Gado TCHABODE
 3. Study of servo systems
3.2 Block diagrams reduction
After construction of a functional block diagram according to the physical equations, it may be useful to
modify its structure. It is for example possible to simplify a group of blocks by a single block, to highlight
a measurable quantity or to disentangle intertwined loops. To be very general,
the blocks will be designated by G1 to G4 which can depend on a variable p. We give here the main
useful rules, without claiming that they are exhaustive. We can do :
▪ A combination of blocks
▪ A displacement of the summers
▪ Moving branch points
➢ Association of blocks
There are 5 basic rules
𝑦1
𝑢 = 𝑢1 𝑦1 = 𝑢1 𝑦2 = 𝑦 𝐺1 (𝑝)
𝐺1 (𝑝) 𝐺2 (𝑝) 𝑢 𝑦 = 𝑦1 +𝑦2
𝑢 𝑦
𝐺3 𝑝 = 𝐺1 (𝑝) ∙ 𝐺2 (𝑝) 𝐺2 (𝑝)
𝑦2
30 Gado T C HABODE
 3. Study of servo systems
3.2 Transformation of block diagrams
➢ Association of blocks

𝑢 + 𝑦
𝐺1 (𝑝) 𝑢 + 𝑦
𝐺1 (𝑝) 𝐺2 (𝑝)
−
−
𝐺2 (𝑝)

𝑢 𝐺1 (𝑝) 𝑦 𝑢 𝐺1 (𝑝) ∙ 𝐺2 (𝑝) 𝑦

𝐺3 𝑝 = 𝐺3 𝑝 =
1 + 𝐺1 (𝑝) ∙ 𝐺2 (𝑝) 1 + 𝐺1 (𝑝) ∙ 𝐺2 (𝑝)

𝑢 + 𝑢 + 𝑦
𝐺1 (𝑝) 𝐺2 (𝑝)𝐺1 (𝑝) 𝐺4 (𝑝)
− −
𝐺2 (𝑝)
1
31 Gado T C HABODE
𝐺4 𝑝 =
𝐺2 (𝑝)
 3. Study of servo systems
3.2 Transformation of block diagrams
➢ Movement of summers
For summers, we can cite three basic rules. It is possible to permute adders in cascade or replace them by an
equivalent adder . This property follows from the commutativity of addition. You may also need to move an
adder upstream or downstream of a block.

32 Gado T C HABODE
 3. Study of servo systems
3.2 Transformation of block diagrams
➢ Movement of summers

𝑈 + 𝑌 𝑈 + 𝑌
𝐺1 (𝑝) 𝐺1 (𝑝)
± 𝑋 ±
𝒀 𝒑 = 𝑼 𝒑 𝑮(𝒑) ±𝑿(𝒑) 1 𝑋
𝐺1 (𝑝)

𝑈1 + 𝑌
𝐺(𝑝) 𝑈1 + 𝑌
𝐺(𝑝)
𝑈2 ±
𝒀 𝒑 = [𝑼𝟏 𝒑 ± 𝑼𝟐 (𝒑)]𝑮(𝒑) 𝑈2 ±
𝐺(𝑝)

33 Gado T C HABODE
 3. Study of servo systems
3.2 Transformation of block diagrams
➢ Move branch point

𝑈 𝑌 𝑈 𝑌
𝐺1 (𝑝) 𝐺1 (𝑝)
𝑌
𝑌
𝐺1 (𝑝)

34 Gado T C HABODE
 3. Study of servo systems
3.3 First-order system
𝑑𝑠(𝑡)
A first-order system is governed by the differential equation 𝜏 + 𝑠 𝑡 = 𝐻0 𝑒(𝑡)
𝑑𝑡
The TL makes it possible to obtain the transfer function. For a first order we obtain the following transfer
function: 𝑆(𝑝) 𝐻0
𝐻 𝑝 = =
𝐸(𝑝) 1 + 𝜏𝑝
➢ Step response
1 1 𝐻0
𝐸 𝑝 = 𝑆 𝑝 =𝐻 𝑝 ∙𝐸 𝑝 = ∙
𝑝 𝑝 1 + 𝜏𝑝
▪ When 𝑡 = 𝜏 we get𝑠 𝜏 = 𝐻0 1 − 𝑒 −1 = 0,63 ∙ 𝐻0
▪ When 𝑡 → ∞, we get 𝑠(∞)= 𝐻0 . 𝐻0 is the final value
▪ The equation at the coordinate point 𝑡 = 0+, 𝑠 0+ = 0 has
the form 𝑦 = 𝑎𝑡 + 𝑏
➢ Frequency response
𝐻0
By replacing 𝑝 with 𝑗𝜔, 𝐻 𝑗𝜔 =
1 + 𝑗𝜔𝜏
𝐻0
▪ Its modulus 𝐻(𝑗𝜔) =
1+(𝜏𝜔)2
35
▪ His argument 𝜑 = arg 𝐻 𝑗𝜔 = 𝑡𝑎𝑛−1 (𝑗𝜔)
 3. Study of servo systems
3.3 First-order system
➢ Nyquist curve
It is the locus of the endpoint of the complex number 𝐻(𝑗𝜔) as 𝜔 varies from 0 to ∞. Let the harmonic
response be plotted in the complex plane such that:
▪ 𝑿 = 𝑹𝒆 𝑯 𝒋𝝎 𝑿 = 𝑹𝒆 𝑯 𝒋𝝎 = ? ? ?
▪ 𝒀 = 𝑰𝒎 𝑯 𝒋𝝎 𝒀 = 𝑰𝒎 𝑯 𝒋𝝎 = ? ? ?
We can write 𝑯 𝒋𝝎 = 𝑿 + 𝒋𝒀

Nyquist transfer locus is graduated and oriented in the

direction of increasing 𝜔
▪ For 𝜔 → 0, 𝐴 𝐻0 ; 0
1 𝐻 −𝐻
▪ For 𝜔 = ,B 0 ; 0
𝜏 2 2
▪ For 𝜔 → ∞, 𝐶 0; 0

36 Gado T C HABODE
 3. Study of servo systems
3.3 First-order system
➢ Black–Nichols curve
This is the plane such that on the abscissa: Arg [𝐻(𝑗𝜔)] (in degrees) and on the ordinate: 20 ∙ log| 𝐻(𝑗𝜔)| (in
dB). We calculate :

▪ The phase shift 𝜑 = 𝐴𝑟𝑔 𝐻 𝑗𝜔 = 𝑡𝑎𝑛−1 𝜔𝜏

𝐻0
▪ The linear gain𝐴 = 2
1+(𝜔𝜏)
𝐻0
▪ The logarithmic gain𝐴𝑑𝐵 = 20 log 𝐴 = 20 log =
1+(𝜔𝜏)2

▪ For𝜔 → 0, 𝜑 = 0°And𝐴𝑑𝐵 = 20 ∙ log(𝐻0 )

1
▪ For𝜔 = , 𝜑 = −45°And𝐴𝑑𝐵 = 20 ∙ log 𝐻0 − 10 ∙ log 2
𝜏

▪ For 𝜔 → ∞, 𝜑 = −90°and𝐴𝑑𝐵 → −∞

37 Gado T C HABODE
 3. Study of servo systems
3.3 First-order system
➢ Bode curve
The representation in the Bode plane consists of 2 curves:
▪ the gain curve: on the abscissa 𝜔 and on the ordinate (20 ∙ 𝑙𝑜𝑔(|𝐻(𝑗𝜔)|) in dB).
▪ the phase curve: in abscissa 𝜔 and in ordinate (A𝑟𝑔(|𝐻(𝑗𝜔)|) in degree).
We use a logarithmic scale of the pulsations:
𝜔2
▪ We call a decade an interval of pulsation such that [𝜔1 ; 𝜔2 ] such that = 10
𝜔1
𝜔2
▪ We call an octave a pulsation interval such that 𝜔1 ; 𝜔2 such that =2
𝜔1
The asymptotic gain curve makes it possible to obtain a summary shape of the gain curve:
1
Whether𝜔 ≪ , ⟹ 𝜔2 𝜏 2 ≪ 1.We obtain 𝐴𝑑𝐵 = 20𝑙𝑜𝑔𝐻0 − 10𝑙𝑜𝑔1 ≈ 20𝑙𝑜𝑔𝐻0
𝜏
1
Whether𝜔 ≫ , ⟹ 𝜔2 𝜏 2 ≫ 1.We obtain 𝐴𝑑𝐵 = 20𝑙𝑜𝑔𝐻0 − 10𝑙𝑜𝑔𝜔2 − 10𝑙𝑜𝑔𝜏 2 ≈ 20𝑙𝑜𝑔𝐻0 − 20𝑙𝑜𝑔𝜔 − 20𝑙𝑜𝑔𝜏
𝜏

38 Gado T C HABODE
 3. Study of servo systems
3.3 First-order system
➢ Bode curve
Actual gain curve

Real phase curve

39 Gado T C HABODE
 3. Study of servo systems
3.4 Second-order system

1 𝑑 2 𝑠(𝑡) 2𝜁 𝑑𝑠(𝑡)
A second-order system is governed by the differential equation + + 𝑠(𝑡) = 𝐻0 𝑒(𝑡)
𝜔𝑛2 𝑑𝑡 2 𝜔𝑛 𝑑𝑡
The TL of this differential equation makes it possible to obtain
the transfer function: 𝑆(𝑝)
𝐻 𝑝 = = ???
𝐸(𝑝)
The poles of H(p) are given by the equation: 𝑝2 + 2𝜁𝜔𝑛 𝑝 + 𝜔𝑛2 = 0

The reduced discriminant is given by:𝜔𝑛2 𝜁 2 − 𝜔𝑛2 = 𝜔𝑛2 (𝜁 2 − 1)

𝜻 = 𝟏, critical case
2
𝐻0 𝜔𝑛 𝐻 1
We have a double root 𝑝0 = −𝜔𝑛 and 𝐻 𝑝 = = (1+𝜏0𝑝)2 with 𝜏0 = −
(𝑝−𝑝0 )2 0 𝑝0

40 Gado T C HABODE
 3. Study of servo systems
3.4 Second-order system
𝜻 > 𝟏, this is the aperiodic regime
We have two negative real 𝑝1 = −𝜁𝜔𝑛 + 𝜔𝑛 𝜁 2 − 1
ቐ
poles 𝑝2 = −𝜁𝜔𝑛 − 𝜔𝑛 𝜁 2 − 1

1 1
with : 𝑝1 𝑝2 = 𝜔𝑛2 , 𝑝1 + 𝑝2 = −2𝜁𝜔𝑛 , 𝜏1 = − ;𝜏2 = −
𝑝1 𝑝2

𝜻 < 𝟏, this is the oscillatory regime

𝑝1 = −𝜁𝜔𝑛 + 𝑗𝜔𝑛 1 − 𝜁 2
We have two complex conjugate poles ቐ
𝑝2 = −𝜁𝜔𝑛 − 𝑗𝜔𝑛 1 − 𝜁 2

We pose 𝜔𝑝 = 𝜔𝑛 1 − 𝜁 2 the so-called

pulsation of the damped oscillations
41 Gado T C HABODE
 3. Study of servo systems
3.4 Second-order system
➢ Step Response We obtain s(t) by the inverse Laplace transform
1 𝐻0
𝑆 𝑝 =𝐸 𝑝 ∙𝐻 𝑝 = ∙
𝑝 1 + 2𝜁 𝑝 + 1 𝑝2
𝜔𝑛 𝜔𝑛2

Depending on the value of 𝜁 we obtain

different index responses

We pose 𝜔𝑝 = 𝜔𝑛 1 − 𝜁 2 the so-called

pulsation of the damped oscillations
42 Gado T C HABODE
 3. Study of servo systems
3.4 Second-order system
➢ Step Response
It can be seen that for 𝜁< 1 there is an overshoot. This overshoot is greater as the value of 𝜁 decreases. In
control, we are interested in the first overshoot, which is the most
important. On the other hand, it can be shown on an index response of a second order that the tangent at
the origin is zero. We show that the value of the first overshoot is obtained by the formula:

We define the value of the first overrun D1 as the quantity:

The table opposite gives an overview of the value of the

1st overrun according to 𝜁

We pose 𝜔𝑝 = 𝜔𝑛 1 − 𝜁 2 the so-called

pulsation of the damped oscillations
43 Gado T C HABODE
 3. Study of servo systems
3.4 Second-order system
➢ Frequency response
Replacing p by 𝑗𝜔 in H(p), we get: 𝐻 𝑗𝜔 = ? ? ?
We then calculate:
𝐻0 𝜔2 2𝑗𝜁
𝐴 = 𝐻 𝑗𝜔 = And 𝜑 = 𝐴𝑟𝑔 𝐻0 − 𝐴𝑟𝑔 1− 2 +
𝜔𝑛 𝜔𝑛
𝜔 2 2 2𝜁𝜔 2
(1 − 2 ) +( 𝜔 )
𝜔𝑛 𝑛

44 Gado T C HABODE
 4. Stability of servo systems
4.1 Introduction
A system is said to be stable if it moves away from its equilibrium position and tends to return to it, and
unstable if it moves further away.
The closed-loop stability of an automatic control system is an imperative condition.
Consider a transfer function system T(p) excited by a Dirac pulse (thus moved away from its
equilibrium position). This signal has the notable advantage of considerably reducing the calculations since
x(t) = δ(t) then X(p) = 1 and therefore: Y (p) = T (p)X(p) = T (p ). The impulse response of a system is
represented by its transfer function. Let T (p) be its transfer function:

The simple element decomposition of T(p) is of the form

Then we apply the inverse Laplace transformation The first

term for the real poles, the second for the imaginary poles.
45 Gado T C HABODE
 4. Stability of servo systems
4.1 Introduction
▪ if the real parts are all negative (ci and aj ), then the transient response of the system is
composed of damped exponentials, the response tends to zero for 𝑡 → ∞, the system returns to its
equilibrium position, the system is stable .
▪ if one of the real poles (ci) is positive, the system is unstable: the system is of
exponential divergent type.
▪ if one of the complex poles has a positive real part ( aj ), the system is unstable the system is of the
divergent oscillatory type

The fundamental condition of stability of a

continuous linear closed-loop system is that the
poles of its transfer function have their real parts
negative.

46 Gado T C HABODE
 4. Stability of servo systems
4.2 Algebraic criterion of a looped system
Consider the looped system in the figure below. Its closed-loop transfer function (FTBF)
is written:

According to the previous theorem, this looped system is stable if the poles of its FTBF have a
strictly negative real part. Now, these poles are the roots of the equation:

This equation is also called the characteristic equation. It is a polynomial of degree n.

➢ Routh 's algebraic criterion

Routh 's rule is used to determine the sign of the roots of a
polynomial. For this we construct Routh 's table , using the
coefficients of the characteristic equation. To fill the table of Routh
one arranges so that the coefficient bn is positive. The table is filled
until the coefficients are exhausted.
47 Gado T C HABODE
 4. Stability of servo systems
4.2 Algebraic criterion of a looped system
➢ Routh 's algebraic criterion
The coefficients (c1, c2, d1, d2, e1;…q1) are calculated
as follows:

condition of stability.
The system is stable if:
▪ all the coefficients of the characteristic equation exist;
▪ all the terms in the first column of Routh 's tableau have the same sign,
otherwise the number of roots with positive real part is equal to the
number of sign changes. If one of the coefficients is zero, the system
exhibits sustained oscillations.
48 Gado T C HABODE
 4. Stability of servo systems
4.2 Algebraic criterion of a looped system
➢ Routh 's algebraic criterion 1
With K=7 and 𝐺 𝑝 =
𝑝6 + 2𝑝5 + 4𝑝4 +6𝑝3 +5𝑝2 + 8𝑝 + 1

The characteristic equation is given by:

▪ There are 2 sign changes in the first column of the table

▪ 2 roots with positive real part.

Conclusion, the system is unstable

49 Gado T C HABODE
 4. Stability of servo systems
4.2 Algebraic criterion of a looped system
➢ Graphic criterion of stability of a looped system
The use of methods based on the frequency responses of the isochronous transfer functions allow a clearer
visualization of the influence of the parameters on the stability of the system.
The characteristic equation 1 + 𝐺𝐻 𝑝 = 0or 𝐺𝐻(𝑝) = −1. (-1;0) is called the critical point
We solve this equation by plotting in the complex Nyquist plane the graph of GH(𝑗𝜔) and
considering its passage through the critical point. We will study only one graphical method, the
reverse criterion or simplified
We will study only one graphical method, the
reverse criterion or simplified Nyquist criterion.

50 Gado T C HABODE
 4. Stability of servo systems
4.2 Algebraic criterion of a looped system
➢ Graphic criterion of stability of a looped system
Depending on the order of the transfer function in OL, which determines the asymptotic value of its argument
(===>number of quadrants where the Nyquist locus is located ), we can predict whether the stability of a
loop of regulation is conditional or not

We want to determine the stability as a function of K.

▪ If K < 1 the system is stable

▪ If K =1 limit of stability
▪ If K > 1 the system is unstable

51 Gado T C HABODE
 5. Accuracy of servo systems
5.1 Introduction
The role of the servo systems is to cause the output 𝑦(𝑡) to follow a law fixed by the setpoint input 𝑥(𝑡),
with the ideal ε(𝑡) = 𝑥(𝑡)-𝑦(𝑡) ≡ 0, ∀𝑡. In practice, what can happen?
• the input varies, the system operates as a follower and performs the servo function .
• the input is constant and a disturbance signal can be superimposed on the useful signal at
any point in the chain. Maintaining ε(𝑡) ≡ 0 despite this disturbance requires the
system to operate as a regulator
In fact, both sources of error are present simultaneously, but by virtue of the superposition theorem, we can
write that: ε(𝑡) = 𝜀𝑒(𝑡) + 𝜀𝑝(𝑡)
The error signal is written:

In general, the variation of the error signal is studied by

assuming either:
- the input signal X(p) = 0.
52 - the disturbance signal Pert(p) = 0 Gado T C HABODE
 5. Accuracy of servo systems
We then apply the principle of superposition. The variation of 𝜀(𝑡) over time, with respect to
input and disturbance signals, can often be broken down into two parts, the transient and
the steady state. We are therefore led to define two kinds of precision according to the regime:
▪ Steady State Static Accuracy
▪ Dynamic accuracy in transient state
5.2 Static accuracy
We study in this paragraph the permanent error, ie the difference which remains in
permanent mode. This error is also called static error. It is defined by the limiting value, if it
exists, of ε(t) when t tends to infinity.

Applying the final value theorem, we can write that

𝜖 = lim 𝜀 𝑡 = lim 𝑝𝜀 𝑝
𝑡→∞ 𝑝→0
➢ Non-disturbance system with variable input
By setting 𝐺 𝑝 = 𝐺1 𝑝 ∙ 𝐺2 𝑝 = 𝐹𝑇𝐵𝑂(𝑝) we get

53 Gado T C HABODE
 5. Accuracy of servo systems
5.2 Static accuracy
➢ Non-disturbance system with variable input
G(p) is the open loop transfer function of the system. Thus, the deviation 𝜀 is related on the one hand to the
shape of the input signal and on the other hand to the shape of the FTBO. Depending on the input, we
distinguish the error from:
▪ position for a noted step input 𝜀𝑝
▪ dragging for a noted ramp input 𝜀𝑣
▪ of acceleration for a parabola type input denoted 𝜀𝑎
It is convenient for calculations to put the open loop transfer function in the form:
𝐾𝛼 1 + 𝑎1 𝑝 + 𝑎2 𝑝2 … + 𝑎𝑚 𝑝𝑚
𝐹𝑇𝐵𝑂 𝑝 = 𝛼 ∙
𝑝 1 + 𝑏1 𝑝 + 𝑏𝑎2 𝑝2 … + 𝑏𝑛 𝑝𝑛
𝜶 𝒊𝒔 𝒄𝒂𝒍𝒍𝒆𝒅 𝒕𝒉𝒆 𝒄𝒍𝒂𝒔𝒔 𝒐𝒕𝒉𝒆 𝒔𝒚𝒔𝒕𝒆𝒎
𝑁(𝑝) 𝐾1 𝑁(𝑝) 𝐾2 𝑁(𝑝)
𝐹𝑇𝐵𝑂 𝑝 = 𝐾0 𝐹𝑇𝐵𝑂 𝑝 = 𝐹𝑇𝐵𝑂 𝑝 =
𝐷(𝑝) 𝑝 𝐷(𝑝) 𝑝2 𝐷(𝑝)
Class 0 Class 1 Class 2
54 Gado T C HABODE
 5. Accuracy of servo systems
5.2 Static accuracy
➢ Non-disturbance system with variable input
The table below summarizes the errors according to the class of the FTBO:

55 Gado T C HABODE
 5. Accuracy of servo systems
5.2 Static accuracy
➢ System with disturbances only
We are now working at a constant input, therefore as a regulator. We are still studying the system of the
figure in the introduction where the transfer function of the chain of action is separated into two, namely G1
and G2. The disturbance signal produces its effect between G1(p) and G2(p). According to the equation (P51)
and taking into account only the disturbance, the error is written:

We notice that in the case of a unitary feedback system and for a zero input we have, 𝜀(𝑝) = -𝑌(𝑝). The most
common disturbance is of the step type, it is a permanent disturbance.

We write G1(p) and G2(p) in the general form

With

56 Gado T C HABODE
 5. Accuracy of servo systems
5.2 Static accuracy
➢ System disturbances only
We distinguish 3 cases:
𝑃0 𝐾2𝛼
▪ 𝐺1 (𝑝) and 𝐺2 (𝑝) do not have an integrator (α1 = 0 and α2 = 0) then 𝜀 = − . The error due
1+𝐾1𝛼 ∙𝐾2𝛼
to a constant disturbance is proportional to the amplitude 𝑃0 and inversely proportional to the gain of
the forward channel. As 𝐾2𝛼 is in the numerator, it is in our interest to take K1 as large as possible, at
the risk of destabilizing the system.
𝑃0
▪ 𝐺1 (𝑝) does not have an integrator and 𝐺2 (𝑝) has an integrator (α1 = 0 and α2 = 1) then 𝜀 = − .
𝐾1𝛼
The error is still a function of 𝑃0 but no longer a function of K2. It is therefore in our interest to take
K1 as large as possible, but this will still be done to the detriment of stability.
▪ 𝐺1 (𝑝) has an integrator and 𝐺2 (𝑝) does not have an integrator (α1 = 1 and α2 = 0) then 𝜀 =0. An
integrator placed upstream of a disturbance cancels its effects.

57 Gado T C HABODE
 5. Accuracy of servo systems
5.2 Static accuracy
➢ Conclusion
When the system is stressed by input variations and disturbances, the superposition theorem applies. It is
then observed that, except in the case of an integration placed furthest upstream
of the direct chain, it is imperative to increase the gain to obtain better accuracy.
As it is not possible to affect this quantity too much because of the stability, one solution consists in
placing an integrator upstream of the chain, therefore just behind the comparator: this is the place of the
corrector.

58 Gado T C HABODE
 5. Accuracy of servo systems
5.3 Dynamic Accuracy
Perfect static precision for a given input if the dynamic precision takes large values
for a long time, is not very useful. The figure below illustrates an example of a
step response where the input is changing faster than the system time constants.

59 Gado T C HABODE
 5. Accuracy of servo systems
5.3 Dynamic Accuracy
Dynamic accuracy is characterized by the step response of the control loop. Indeed, the
transient state is all the more precise as the response is fast and sufficiently damped. These two
objectives are contradictory. The speed of the step response is quantified by the rise time Tm
(time required to go from 10% to 90% of the final value) and by the response time Tr at 5% (time
required to reach the final value within ± 5%).
The damping of the step response is evaluated
by the first overshoot D1 or by the damping
coefficient 𝜁if the system can be assimilated to a
second oscillating order.
However, the two criteria (speed and damping) can be
combined
by a global criterion:

60 Gado T C HABODE
 5. Accuracy of servo systems
5.3 Dynamic Accuracy
However, the two criteria (speed and damping) can be combined by a global criterion:

5.4 Conclusion

objectives of the analysis of controlled systems are

to determine the following characteristics of the system:
▪ Stability and degree of stability,
▪ Accuracy and behavior in steady state,
▪ Damping and transient response.

61 Gado T C HABODE
 6. Controllers
6.1 Principle of correction
We saw in the previous chapter on the analysis of controlled systems, that to satisfy the specifications
of stability and precision, we are led to formulate certain conditions on the FTBO. In general, a
servo system must be robust enough to guarantee three levels of performance:
▪ Its stability,
▪ Good static accuracy,
▪ sufficient speed
The big problem is that these three criteria are contradictory: both precision and speed are linked to gain, but
too much gain can have a destabilizing effect. Correcting a controlled system means ensuring compatibility
between these contradictory criteria (accuracy-stability) and the corrector will be the
"intelligent" element that will be added to the initial system to ensure this compatibility. Its elaboration
constitutes the synthesis of the controlled system.
The most common forms which will be added to the proportional action, ie the gain in
open loop, are:
• the derivative action which will ensure a correct response time.
• the integral action which will ensure good static precision,
62 Gado T C HABODE
 6. Controllers
6.1 Principle of correction
These actions are performed as a rule (serial correction)
on the error signal 𝜀, the only way to
really observe the proximity or the remoteness of the
desired goal. It consists of introducing the corrector into
the direct chain behind the comparator.
6.2 Proportional control law
It can therefore be seen that it makes it possible to dose the command as a function of the distance from the
desired goal. The greater the difference, the greater the reaction must be. The effect on U can be felt in
several ways.
▪ If K is large, the correction is strong but not without danger (risk of pumping),
▪ If K is low, the correction is soft and slow, but there is no risk
If the gain K is average and constant, its effect is felt only if the
deviation 𝜀 is large; however, too great a difference risks saturating
the amplifier. We see that beyond 𝜀𝑚𝑎𝑥, the action U no longer
increases. The error range for which 𝑈 = 𝐾𝜀 holds is called the
63 Proportional Band (PB). Gado T C HABODE
 6. Controllers
6.2 Proportional control law
The properties of the BP are as follows:

▪ it is arranged symmetrically around the operating point (setpoint),

▪ it is expressed in % of the scale range of the input,
▪ the BP is always a positive value and the product 𝐾 ∙ 𝐵𝑃 always remains constant,
▪ for industrial regulators, the adjustment ranges usually extend between 0 and 200%.
▪ it is inversely proportional to the gain and is expressed by: BP% = 100 K . The following table gives the
correspondence of some values

64 Gado T C HABODE
 6. Controllers
6.2 Proportional control law
The proportional control law has one more flaw and not the least: if 𝜀 = 0, then 𝑈 = 0, the
process is no longer powered! The proportional control law must therefore be written: 𝑈 = 𝐾 ∙ 𝜀 + 𝑈0 ,
where 𝑈0 is a necessary residual voltage, but which will cause a static error. This concept
is not unknown to us: the static error decreases when the gain is increased, but it can be
significant if the latter is low. It is obvious that the shift of the operating point shifts the
phenomenon since we wrote that:

With 𝑊0 the amplitude of the input step

The proportional control is straightforward but its use is limited. It does not make it possible to eliminate the
static error and too large a gain risks destabilizing the process. Proportional action therefore proves
insufficient to regulate, on its own, the imperfections of a system.

65 Gado T C HABODE
 6. Controllers
6.3 Derived control law
It is possible to improve the operation of the system by taking account of variations in the output quantity
. Consider, for example, the cruise control of a car. At a given instant, this speed
is 5% lower than the set speed. What can this mean? It is certainly necessary to act on
the accelerator but not just anyhow because:
▪ the speed may be decreasing or increasing, which means that you have to act more or
less depending on the case,
▪ it is also necessary to take into account the speed of evolution of this variation: indeed a
rapid decrease in speed would require a vigorous reaction.
𝑑𝜀(𝑡)
The derivative of the error signal provides this information and we will create the term for this: 𝑢(𝑡) 𝑇𝑑
𝑑𝑡

It can be seen in the figure opposite that the derivative action is

proportional to the "slope" of the error signal. When this signal does not
vary (𝜀 = 0 or constant), then the adjustment quantity is zero, which
also explains why this action is never used alone, since it does not
correct a constant deviation. It is associated with proportional action.

66 Gado T C HABODE
 6. Controllers
6.3 Derived control law
➢ Derivative proportional corrector
𝑑𝜀
The control law is written:𝑢(𝑡) = 𝐾[𝜀 + 𝐾𝑑 ]
𝑑𝑡
We deduce the transfer function of the corrector: 𝐶 𝑝 = 𝐾[1 + 𝑝𝑇𝑑 ]. Since the pure PD corrector is not
feasible, the so-called phase advance or approximate derivative corrector assembly is preferred. In fact, we
will filter the derived action.
6.4 Integral control law
The proportional and derivative actions are therefore insufficient to obtain a zero error for a step input. As for
driving a car, instead of crushing the accelerator, it may be advantageous to ensure a more progressive U
control

67 Gado T C HABODE
 6. Controllers
6.4 Integral control law
If 𝜀 changes suddenly, the output U increases gradually, providing a smooth start to the process. Ti therefore
serves to manage the speed of correction and must be tuned to the dominant time constant τ which measures
the inertia of the process. "to be tuned" means to be of the same order of magnitude as τ. Indeed:
▪ if Ti is too small, signal U increases too quickly without giving the system time to react: the command has
no effect,
▪ conversely, if Ti is too large, this action increases the inertia of the system and this results in a start that is
much too soft.
The integral control law is therefore progressive. It is still said that she is persevering. As long as the
static error (𝜀 > 0 or 𝜀 < 0) exists, the integral action acts (positively or negatively) until it
cancels out. Still by analogy with driving a car, there is no point in driving in jerks: it is better to
gradually press the accelerator, let the car reach the desired speed, then
maintain it by keeping your foot at the same height.
𝑡
For a proportional integral, the control law is written:𝑢 𝑡 = 𝐾[𝜀 𝑡 + 𝐾𝑖 ‫׬‬0 𝜀 𝑥 𝑑𝑥]
1
We deduce the transfer function of the corrector:𝐶(𝑝) = 𝐾[𝜀 + ]
𝑝𝑇𝑖
68 Gado T C HABODE
 6. Controllers
6.4 PID corrector
The regulating system is called a regulator. The set point is generally designated by the letter W. It is
sometimes denoted SP (set point). The setpoint is a signal expressed in the same unit as the
measurement signal and M represents the measurement signal m(t). U is the process command , Y the output
(physical quantity) and 𝜀 the error signal. Finally Z designates all the secondary inputs disturbing the process
(disturbances).

69 Gado T C HABODE
 6. Controllers
6.4 PID corrector
➢ First series type
We see this kind of structure flourishing in the literature, so we will say a word about it in order to
demonstrate that it is not the natural structure of PIDs.
It is not a PID but a simple
proportional corrector!!!

➢ Second type series

This structure, if it leads to a PID, is not very convenient either since
the pure proportional gain is not regulated by a single potentiometer,
but also depends on the adjustment of the derivative and integral
actions. The integral and derivative actions are also modified by the
gain K.

70 Gado T C HABODE
 6. Controllers
6.4 PID corrector
➢ parallel type
It is the most moderate of the associations because the gain does not act on the integral and derivative
actions.

➢ mixed type

This mixed PID controller is the most energetic,

because the gain acts on the other two actions. As
its name suggests, the mixed PID controller is a
hybrid structure of the two previous ones whose use
is undoubtedly the most common.
71 Gado T C HABODE