Lecture 4 [DR.

MUHAMMAD NAEEM ARBAB]

Lecture # 4
Block Diagrams
Generally it is more convenient to describe a function by its pictorial representation. For example
the space remaining and used in a computer hard-disk can be easily recognized by a pi-chart
display rather than numerical data. In order to describe and understand a physical process it is
more convenient to provide a pictorial representation of different stages through which input is
synthesized. Control systems consist of multiple subsystems, each can be represented by its
single individual transfer function in a block. Since the response of a single transfer function can
be calculated, the response of multiple subsystems can be calculated on the basis of a single
block transfer function as shown in Figure (1) with a terminating (inward) arrow to the block
represents an input signal R(s) and originating (outward) arrow from the block represents the
output quantity C(s).

Figure 1: Block Diagram Representation

Mathematically the output C(s) is R(s) multiplied by the transfer function G(s) that is: C(s) =
R(s)G(s). A feed-forward or a forward block G(s) transmit signal from input to output direction
whereas; a feedback block H(s) transmit signal from output to input direction. Multiple
subsystems are represented in two ways: as block diagrams and as signal-flow graphs. Although
neither representation is limited to a particular analysis and design technique, block diagrams are
usually used for s-domain analysis and design, and signal-flow graphs for state-space analysis
and design.

Summing Junction
When multiple subsystems are interconnected, a few more schematic elements must be added to
the block diagram. These additional elements are summing junctions and takeoff points as shown
in Fig (7.2).

Figure 2: Summing Junction and Takeoff Point

1 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

The characteristic of the summing junction is that the output signal, C(s), is the algebraic sum of
the input signals. A takeoff point, distributes the input signal, R(s) to numerous points.

Cascade Form
Blocks are said to be connected in cascade if the output of the preceding block appears as
an input to the following block. Thus a cascade arrangement of blocks is viewed as blocks in
series (one after the other). Figure (3a) shows an example of cascaded subsystems in which three
subsystems with their transfer functions G1(s), G2(s) and G3(s) are connected in cascade so that
the equivalent representation is a single block as shown in Figure (3b) with an overall transfer
function, G(s) as:

C ( s)
G( s)   G1 (s)G2 (s)G3 (s) 1
R( s )

Figure 3: Blocks in Cascade

Thus the overall transfer function of blocks connected in cascade is the product of the transfer
function of the individual subsystem block.

Parallel Form
Blocks are said to be connected in parallel if they derive their inputs from a common
source, whereas the output of each block is different depending on their transfer function. Thus a
summing junction at the output is necessary. Figure (4a) shows an example of arrangement of
three subsystems with their transfer functions G1(s), G2(s) and G3(s) connected in parallel.
Knowing the output of each subsystem, we can find the equivalent transfer function. From
Figure (4b), at the summing junction the output of each block Y1(s), Y2(s) and Y3(s) are added so
that:

C(s)  Y1 (s)  Y2 (s)  Y3 (s)

C (s)   R(s)G1 (s)  R(s)G2 (s)  R(s)G3 (s)

C ( s)
Or G( s)   G1 (s)  G2 (s)  G3 (s) 2
R( s )

2 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

Figure 4: Subsystem Blocks in Parallel

Thus when subsystems are in parallel, the combined transfer function placed in a single block is
the sum of transfer functions of individual subsystems.

Feedback Form
All automatic control systems incorporate feedback to modify the input to suit the
requirements at the output. Automatic control systems are therefore referred to as feedback
control systems. A typical feedback system is shown in Figure (5) in which the forward transfer
function G(s) is obtained from the cascade combination of the controller and the plant, whereas
the transfer function H(s) represent the transfer function of cascaded feedback element.

Figure 5: Simplification of a Feedback Control System

3 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

At the summing junction: E(s)  R(s)  C(s)H (s)

1
Or C ( s)  R( s )  C ( s ) H ( s )
G( s)

Simplifying and rearranging yields the overall closed-loop transfer function as:

C ( s) G( s)
T ( s)   3
R( s ) 1  G ( s ) H ( s )

Important Point: Thus to obtain a single block representation and overall transfer function, the
first step is to simplify those blocks which are directly connected in either cascade or in parallel.
The next step is to consider the loop. In a “loop within a loop” situation, the inner loop simplified
first followed by the outer loop. The arrows direction must be considered in order to identify
feed-forward and feedback blocks.
Example 1: Simplify the given block diagram as shown in Figure (6) to a single block and
obtain its transfer function.

Figure 6
Solution: The first step in the simplification is that all those single blocks, which are connected
directly in series (cascade) or in parallel are simplified. In Figure (6), however, there is
apparently no such situation of blocks directly in parallel or cascade. The next step is to consider
the feedback loops. We have a “loop within a loop” situation; the innermost loop is simplified
first followed by the outer loops in sequence. The arrows direction must be considered in order
to identify feed-forward and feedback blocks. In Figure (6), the loop containing G2(s) and G4(s)
are simplified to a single block G5(s) as:

G2 (s)
G5 (s) 
1  G2 (s)G4 (s)

The resultant system is shown in Figure (7a). It can be seen that the block G1(s) is now in
cascade (series) with block G5(s), whose transfer functions can be multiplied to obtain the overall
transfer function G6(s) as:

4 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

G6 (s)  G1 (s)G5 (s)

The resultant block diagram with block G6(s) is shown in Figure (7b).

(a)

(b)

(c)
Figure 7: Step-by-step Block Diagram Reduction
In Figure (7b) the blocks G3(s) and G6(s) form a parallel combination, so that their transfer
functions are added to result in G7(s) that is:

G7 (s)  G3 (s)  G6 (s)  G3 (s)  G1 (s)G5 (s)

G2 (s)
Or G7 (s)  G3 (s)  G1 (s)
1  G2 (s)G4 (s)

The resultant system is shown in Fig (7c), which a simplified canonical form.
G7 (s)
T ( s) 
1  H (s)G7 (s)

5 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

Substituting the values of the G7(s):

 G2 (s) 
G3 (s)  G1 (s)
1  G2 (s)G4 (s) 
T ( s) 
 G2 (s) 
1  H (s)G3 (s)  G1 (s)
 1  G2 (s)G4 (s) 

G3 (s)[1  G2 (s)G4 (s)]  G1 (s)G2 (s)

T ( s) 
1  G2 (s)G4 (s)  H (s)[G3 (s)(1  G2 (s)G4 (s))  G1 (s)G2 (s)]

G3 (s)  G2 (s)G3 (s)G4 (s)  G1 (s)G2 (s)

T ( s) 
1  G2 (s)G4 (s)  G1 (s)G2 (s) H (s)  G3 (s) H (s)  G2 (s)G3 (s)G4 (s) H (s)

Converting to Unity Feedback System

In numerous control system studies, it is usually convenient to represent a canonical form
as an equivalent unity feedback system. This can greatly simplify control system problems. In
unity feedback systems, the forward block has a transfer function G(s), whereas the feedback
path has a transfer function of unity [H(s) = 1], as:

C ( s) G( s)
T ( s)   4
R( s ) 1  G ( s )

The first step is to add and subtract 1 from the feedback element, leaving +1 as being
independent block while combining –1 with the feedback block transfer function. The second
step is to decompose the feedback block in two parallel parts whose input C(s) and terminates at
the same summing point in the same form (positive or negative) as that of the given system. One
portion is a block whose function is +1 and the other block is a modified function in which –1
has been combined with feedback transfer function in the first step. The separation thus forms
two loops, an inner and an outer loop. The third step is to consider the inner loop for obtaining
the closed-loop transfer function, leaving the +1 feedback block as it is which forms a unity
feedback system.
Example 2: Consider a canonical form of negative feedback control system with the feedback
element s as shown in Figure (8a).

(a)

6 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

(b)

(c)
Figure 8: Converting to Unity Feedback System
The first step and second step is illustrated in Figure (8b). The third step is to simplify the inner
loop leaving the outer loop with 1 as shown in Figure (8c). Thus the transfer function of the
system shown in Figure (8c) is:

1 /( 2s  1) 1 1/ 2
T ( s)   
1  1 /( 2s  1) 2s  2 s  1

The transfer function of Figure (8a) is:

1 /( s  2) 1 1/ 2
T ( s)   
1  ( s) /( s  2) 2s  2 s  1

The two are the same.

Signal-Flow Graphs (SGF)

First developed by S. J. Mason, the SFG shows the flow of information between different
subsystems from the input to the output. SFG are an alternative to block diagrams and are much
simpler to understand. SFG consists only of branches (transmittance or path gain) and nodes
(junctions). A branch links two nodes and may be forward or feedback. A forward branch is
recognized by arrow from input side to output, whereas the feedback branch is recognized by
arrow from output side to input. A node performs the function of addition of the signals on all
incoming branches and the transmission of the total node signal (the sum of all incoming signals)
to all outgoing branches. The following three types of nodes are of particular interest in SFG.

7 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

Source nodes (independent nodes): These represent independent variables and have only
outgoing branches; e.g. input nodes.
Sink nodes (dependent nodes): These represent dependent variables and have only incoming
branches; e.g. output node.
Mixed nodes (general nodes): These have both incoming and outgoing branches.
Path: A path is defined as connected sequence of branches with their arrows in the same
direction. Paths in SFG are recognized as forward and feedback. A forward path between two
nodes is one that follows the arrows of successive branches from input to output and in which a
node is traversed only once. On the contrary, a feedback path between two nodes is one that
follows the arrows of successive branches from output to input and in which a node is traversed
only once.

Mason's Rule
The relationship between an input variable and output variable of a SFG is the net gain or
transmittance between the input and output nodes and represents the overall gain or transfer
function of the system. Mason’s rule allows us to reduce SFG of a block diagram to single
closed-loop transfer function T(s) that requires the application of the Mason’s gain formula,
which is:

T ( s) 
Pi i

P11  P2  2  ...
5
 

Where:   1   La   Lb LC   Ld Le L f  ... 6

And L a = algebraic sum of all the possible loop gain

L L
b C = algebraic sum of product of set of two non-touching loop gain

L L L
d e f = algebraic sum of product of set of three non-touching loop gain

Systems can have numerous loops, so that there is a possibility of sets of two, three or
more non-touching loops, whose product is then included in Eq (6). Δ1, Δ2,… are the cofactors of
their forward paths P1, P2,… respectively. For example, Δ1 is the determinant of the remaining
sub-graph when the forward path which produces Pi is removed. In other words Δ1 includes the
number of leftover loops when a path P1 is eliminated. The leftover loops are apparent from the
resultant sub-graph obtained by removing a forward path. Thus Δi is equal to unity when the
forward path touches all the loops in the SFG (that is the forward path and loops have common
nodes) or when the sub-graph contains no loop.
Application of Mason’s rule to obtain the overall transfer function requires two steps. In
the first step, the given block diagram is converted to SFG, which if possible can be simplified
by elimination or absorption of node(s). The second step is the application of Mason’s gain

8 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

formula. Before applying Mason’s gain formula, the components of Mason’s formula must be
evaluated.

Points to remember:
1. The transmittance of each forward path is obtained by multiplying the transmittance
along that path with each node traversed once.
2. Loop is identified by a contour following the direction of arrows so as to return at the
same node from where we started with nodes in the loop-path traversed once.
3. Loops are touching if they have a node in common, whereas; loops are non-touching if
they do not have any node in common.
4. Sometimes in SFG, we can come across self-loop. Self-loop is independent loop with
having a single node from which the path originate and terminate.
Example 3: Represent the block diagram of Figure (9) by its SFG.

Figure 9: Block Diagram for Example (7.4)

Solution: The step-by-step approach to the formation of SFG is illustrated in Figure (10).

Figure 10: Formation of SFG

9 Control System Engineering
Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

As a first step, all possible nodes are laid down representing each summing junction and takeoff
points as shown in Figure (10a). The shaded nodes represent summing junctions.
In the second step, the nodes are joined by connecting lines with arrows to indicate the direction
of flow of signal, and the transfer function of the respective block representing the transmittance
is also indicated. In case there is no transmittance, it is indicated by 1 as in Figure (10b). In the
third step the SFG is simplified by absorbing all possible single input and single output nodes.
The final SFG of the given block diagram of Figure (9) is shown in Figure (11).

Figure 11: SFG of Block Diagram in Example (3)

Example 4: Given the SGF in Figure (12), identify the forward path(s) and loops as components
of Mason’s formula.

Figure 12

Solution: The main components of the Mason’s gain formula are the transmittance of all the
forward paths and transmittance of all the individual loops. The components for Mason’s
formula with reference to Figure (12) are as follows:
1. Independent forward paths:
P1  G1G2 G3G4 G5G7

P2  G1G2 G3G4 G6 G7

10 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

2. Individual loops:
L1  G2 H1

L2  G4 H 2

L3  G4 G5 H 3

L4  G4 G6 H 3

3. Product sets of two non-touching loops:

L1L2: G2 H1 and G4 H 2 [ G2 H1 G4 H 2 ]

L1L3: G2 H1 and G4 G5 H 3 [ G2 H1 G4 G5 H 3 ]

L1L4: G2 H1 and G4 G6 H 3 [ G2 H1 G4 G6 H 3 ]

4. There are no three non-touching loops.

There are two forward paths; there exist Δ1 and Δ2, which must also be determined for the use in
Mason’s formula. For example, Δ1 is obtained by referring to the SFG of Figure (12) in which
the removed forward path P1 is indicated by dotted lines and the nodes that are traversed by this
forward path are shown shaded. It is thus illustrated in Figure (13) that no loops with solid lines
remain and therefore loses the definition of loop. In the expression for Δ all the loops containing
Gs and Hs are put to zero, resulting in Δ1 = 1.

Figure 13: SGF when the Forward Path P1 is removed

The same procedure can be applied for finding Δ2. The application of Mason’s rule is illustrated
in the following examples.

11 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

Example 5: Using Mason’s gain formula, obtain the transfer function of a system whose signal
flow graph is given in Figure (14).

Figure 14: SFG of the System with Self-Loop

Solution: The second step is applicable that is components required for the use of Mason’s gain
formula, which are as follows:

Forward paths: P1  G1G2 G3G4

P2  G1G5

Individual loops: L1  G1 H1

L2  G3 H 2

L3  G1G2 G3 H 3

L4  H 4 (Self-loop)

Product sets of two non-touching loops:

L1L2 = G1G3 H1 H 2

L1 L4  G1 H1 H 4

L2 L4  G3 H 2 H 4

L3 L4  G1G2 G3 H 3 H 4

There is one combination of three non-touching loops, whose product set is:

L1 L2 L4  G1G3 H1 H 2 H 4

  1  G1 H 1  G3 H 2  G1G2 G3 H 3  H 4  G1G3 H 1 H 2  G1 H 1 H 4  G3 H 2 H 4
Therefore:
 G1G2 G3 H 3 H 4  G1G3 H 1 H 2 H 4

12 Control System Engineering

Lecture 4 [DR. MUHAMMAD NAEEM ARBAB]

Since all the loops touches the forward path P1, therefore: 1  1 .On the other hand the loop
formed by the combination of G3 and H2 does not touch the forward path P2, so that:
1  1  G3 H 2 . Using:

T ( s) 
P i i

P11  P2  2
 
G1G2 G3G4  G1G5 (1  G3 H 2 )
T ( s) 

It must be noted that the product set of two and three non-touching loops formed in this example
does not contain Gs and Hs that are common in the expressions of the individual loop gains from
which the product set is formed. For example, since L1 and L2 contain entirely different Hs and
Gs, so they are non-touching. Similarly in the product set formed by L1, L2 and L4, of the three
non-touching loops there are no Gs and Hs that are common in their expressions, but this cannot
be regarded a general case.

13 Control System Engineering