Associated Reference Convention :

A current direction is chosen

entering each positively-referenced terminal.
vj
i1 v2
+ i1 i2 i2 ij
+ +
v1 D D D
i1 i j −1
− v1 v2
v1 v j −1
- -

Device Graph : DIGRAPH (Directed Graph)

1 2 j
1 2

1 j −1
Digraph of Op Amp

2 i2
e1 _
i3 3
e3
i1
e2 +
1
0

2
3

1
0
 A Circuit with 3 different digraphs

1. Choose 3 as datum for D

1 i2 i3 2 1 2
i1 D i4 3
+ + 2
+
v1 v2 v3
+
v4 ⇒ 1 4
- - - -
3 3

2. Choose 2 as datum for D

v2
1
+ - 2 1
2
2
i1
i2
D - i4
+
v1
i3 v3
+
v4
⇒ 1
3
4
- + - 3
3

3. Choose 1 as datum for D

v3 3
1 - + 2 1 2

i1 - D i4
i3
+
v1 v2 i2
+
v4
⇒ 1
2
4
- + - 3
3
• Circuits containing n-terminal
devices can have many distinct
digraphs, due to different (arbitrary)
choices of the datum terminal for
each n-terminal device.
• Although the KCL and KVL
equations associated with 2 different
digraphs of a given circuit are
different, they contain the same
information because each set of
equations can be derived from the
other.
Associated Reference Convention :
i1
+
v1
i1 i2 −
+ + i2
+
v1
−
D v2
−
v2
− D
2-port Device in
+
vn
−
n-port Device
Device Graph
1

1 2
2

n
 1 1 2 2 i4 i5 4
i2 i3
4
i1 + + + i6
+ + D +
v1 v2 v3 2-port v6
v4 device v5
− - - −
- -5 5
3 3

1 2 4

2 3 5 6
1
4 i7 = 0
3 5

Adding a wire connecting one node from

each separate component does not change
KVL or KCL equations.

{7} is a cut set ⇒ i7 = 0

 1 1 2 2 i4 i5 4
i2 i3
4
i1 + + + i6
+ + D +
v1 v2 v3 2-port v6
v4 device v5
− - - −
- -5 5
3 3

1 2 4

2 3 5 6
1 4

3 5

disconnected digraph

KCL at 2 : i3 + i4 = 0
KCL at 4 : i5 + i6 = 0
KVL around 2 3 2 : v4 − v3 = 0
KVL around 4 5 4 : v6 − v5 = 0
 1 1 2 2 i4 i5 4
i2 i3
4
i1 + + + i6
+ + D +
v1 v2 v3 2-port v6
v4 device v5
− - - −
- -5 5
3 3

1 2 4

2 3 5 6
1 i7 = ?
4
3 5

Adding a wire connecting one node from

each separate component does not change
KVL or KCL equations.
 1 1 2 2 i4 i5 4
i2 i3
4
i1 + + + i6
+ + D +
v1 v2 v3 2-port v6
v4 device v5
− - - −
- -5 5
3 3

1 2 4

2 3 5 6
1
4 i7 = 0
3 5

Adding a wire connecting one node from

each separate component does not change
KVL or KCL equations.

{7} is a cut set ⇒ i7 = 0

 1 1 2 2 i4 i5 4
i2 i3
4
i1 + + + i6
+ + D +
v1 v2 v3 2-port v6
v4 device v5
− - - −
- -5 5
3 3
1 2 4

2 3 5 6
1 4

3 5

HINGED DIGRAPH
1 2
Since nodes 3 and 5

2 3 are now the same node,

1 4 they can be combined into
3 one node, and the redrawn
6 5 digraph is called a
hinged graph.
4
 e1 1
i1
v1 + i2 +
− D1
v2
D2 −
− e2 − v3 +
v6 i3 3
D6 D3 e3
+ 2 i4
i5
+ D4 D5 +
v4 v5
− −

2
1
3 3
2
1
6 4 5
4
 Writing KCL Equations at Nodes
1
1 2
3
6 2 3
4 5
4
KCL at nodes
1 i1 + i2 – i6 = 0
2 – i1 – i3 + i4 =0
3 – i2 + i3 + i5 = 0
4 – i4 – i5 + i6 = 0
Incidence Matrix Aa
i1
1 1 1 0 0 0 -1 i2 0
2 -1 0 -1 1 0 0 i3 0
0 -1 1 0 1 0 i4 = 0
3
4 0 0 0 -1 -1 1 i5 0
i6
branch 1 branch 6
The 4 × 6 matrix just obtained is called the
incidence matrix Aa of G
 Incidence Matrix Aa

For an n-node b-branch connected digraph G

which does not contain self-loops the
incidence matrix Aa is specified as follows:
For i = 1, 2, …, n and k = 1, 2, …, b
+1 if branch k leaves node i
aik = –1 if branch k enters node i
0 if branch k does not touch node i

KCL in terms of Incidence Matrix

Aa i = 0
where i = (i1, i2, …, ib)T
is called the branch current vector.
The

Incidence Matrix Aa
of a digraph G is just a
connection table
giving instructions on
how to reconstruct
the digraph G.
 Writing KCL Equations at Nodes
1
1 2
3
6 2 3
4 5
4
KCL at nodes
i1 + i2 – i6 = 0 (1)
– i1 – i3 + i4 =0 (2)
– i2 + i3 + i5 = 0 (3)
– i4 – i5 + i6 = 0 (4)
Adding equations (1), (2), (3), and (4)

(i1 + i2 – i6 ) + (– i1 – i3 + i4 )
+ (– i2 + i3 + i5 ) + (– i4 – i5 + i6 ) = 0

∴ Since any one of these 4 equations can

be derived from the other 3 equations,
these 4 KCL equations are
not linear independent.
 Writing KCL Equations at Nodes
1
1 2
3
6 2 3
4 5
4
Incidence Matrix Aa
i1
1 1 1 0 0 0 -1 i2 0
2 -1 0 -1 1 0 0 i3 0
0 -1 1 0 1 0 i4 = 0
3
4 0 0 0 -1 -1 1 i5 0
i6
branch 1 branch 6
Since each column of Aa has exactly one “1”
and one “-1” entries, each column sums to zero.
This means the rows of Aa are not linearly
independent, or equivalently.

rank Aa < 4
Observation : The 4 KCL node
equations are not linearly independent.
Adding the left side of the 4 KCL node
equations, we obtain:
(i1 + i2 − i6 ) + (−i1 − i3 + i4 ) + (−i2 + i3 + i5 )
1 2 3

+(−i4 − i5 + i6 ) ≡ 0
4
This means we can derive any one of
these 4 equations from the other 3.
Example: Derive KCL equations at
node 4 :
Adding the first 3 node equations
gives:
(i1 + i2 − i6 ) + (−i1 − i3 + i4 ) + (−i2 + i3 + i5 )
1 2 3

= i4 + i5 − i6
4
By applying the various versions
of KCL, we can write many
different KCL equations for each
circuit. However, these equations
are usually not linearly
independent in the sense that each
equation can be derived by a
linear combination of the others.

How can we write a maximum set

of linearly-independent KCL
equations?
 Simplest Method
to write linearly-Independent
KCL Equations.

Given a connected circuit with

“n” nodes, choose an arbitrary
node as datum. Write a KCL
equation at each of the remaining
(n-1) nodes.