Chapter 9

dc-ac Inverters
Outlines
• Introduction
• Basic Block Diagram of Dc-Ac Inverter
• Basic Half-Bridge Inverter Circuits
– Resistive Load
– Inductive-Resistive Load
• Full-Bridge Inverter
– Approximate Analysis
– Generalized Analysis
• Harmonic Reduction
• Pulse-Width Modulation
– Equal-Pulse (Uniform) PWM
– Sinusoidal PWM
• Three-Phase Inverters
• Current-Source Inverters
 Introduction
• Power electronic circuits that produce variable-frequency ac outputs voltages from
dc sources.
• Depending on whether the source is dc or ac, power electronics circuits with ac
output voltages are referred to as dc-ac inverters or ac-ac cycloconverters.
• If the output voltage frequency is different from the source frequency , the
converter is called an ac voltage controller.
• Dc-to-ac inverters (also known as static inverters) use fixed dc sources to produce
symmetrical ac output voltages at fixed or variable frequency or magnitude.
• Inverting circuits are used to deliver power from a dc source to a passive or active
ac load employing conventional SCRs or gate-driven semiconductor devices such
as GTOs, IGBTs, and MOSFETs.
• Today’s inverters can operate in wide ranges of regulated output voltage and
frequency with reduced harmonics.
• Dc-to-ac inverters are used in applications where the only source available is a
fixed dc source and the system requires an ac load such as an uninterruptible
power supply (UPS).
• Applications where dc-to-ac inverters are used include aircraft power supplies,
variable-speed ac motor drives, and lagging or leading var generation.
 Basic Block Diagram of dc-ac Inverters
• Figure 9.1 shows a typical block diagram of a power electronic circuit utilizing a dc-to-ac
inverter with input and output filters used to smooth the output ac signal.
• The feedback circuit is used to sense the output voltage and compare it with a sinusoidal
reference signal as shown in Fig. 9.1.
• The control objective is producing a controllable ac output from an uncontrollable dc voltage
source.
• The load is broadly classified as either passive or active. If the load consists of impedance
only (i.e. passive), then its time domain response is determined by the nature of the load and
cannot be controlled externally.

Fig 9.1 Block

diagram of a
typical power
electronic
circuit with
dc-to-ac
inverter.
 Voltage-and Current-Source Inverters
• Inverters are divided into either Voltage-Source Inverters (VSI) or Current-Source Inverters
(CSI).
• The dc source in VSI is a fixed voltage such as battery, fuel cells, solar cells, dc generators
or rectified dc sources.
• In VSI, the output voltage, vo , is a function of the inverter operation and the load io
current, is a function of the nature of the load, and the dc input, V dc, is a constant input
voltage.

Fig 9.2 Block diagram

representations for (a) voltage
source inverter and (b) current
source inverter.
 Inverter Configurations
Figure 9.3 (a), (b) and (c) show, respectively, three possible single-phase inverter
arrangements: bi-phase, half-bridge, and full-bridge. The bi-phase inverter, also known as a
push-pull inverter, is drawn in two different ways in Fig. 9.3(a).

Fig 9.3 Single-phase inverter arrangements. (a) Biphase inverter. (b) Half-bridge inverter. (c)
Full-bridge inverter
 Output Voltage Control
Output voltage is controlled by varying the dc source voltage, by either controlling the dc
input by using a dc-dc converter as shown in Fig. 9.4(a), or by using an ac-dc phase control
converter as shown in Fig. 9.4(b).

Fig 9.4 Controlling the dc input using (a) dc-dc converter or (b) ac-dc phase-controlled
rectifier
 Basic Half-Bridge Inverter Circuit
Resistive Load:
• Half-bridge voltage source inverter circuit under a resistive load as shown in Fig. 9.5(a).
• Switching waveforms for S1, S2 and the resultant output voltage are shown in Fig. 9.5(b).
S1 and S2 are switched on and off alternatively at a 50% duty cycle.

Fig 9.5 (a) Half-bridge inverter under resistive load. (b) Switching and output voltage
waveforms
 Basic Half-Bridge Inverter Circuit
• In practice the above circuit does not require two equal dc voltage sources as shown in Fig.
9.5(a).
• Instead, large splitting capacitors, C , are used to produce two equal dc voltage sources as
shown in Fig. 9.6.
• The two capacitors, C, are equal and very large, so that RC  is much larger than half the
switching period. io
• The current from the source, Vdc, equals one-half of the load current, .
• In steady state, the average capacitor currents are zero, hence, C are used to block the dc
i
component of o .

Fig 9.6 Half-bridge inverter circuit with large splitting capacitors

 Example 9.1
Sketch the current and voltage waveforms for iin, io, v S 1 ,andin
v S 2 the circuit shown in Fig. 9.6
for =0 and 0. Assume that the switching waveform for S1 and S2 are shown in Fig. 9.7.
Determine the average output voltage in terms of V dc and  when the inverter operates in the
steady state.
Solution
Mode 1 when S1 is ON and S2 is OFF, then the output voltage and current equations are
given by,
vo  Vdc
vo Vdc
io  
R R
assumed that the capacitors are equal, then the load current, io, splits equally, i.e.
1 V
ic1   io   dc
2 2R
V
ic 2  iin  ic1  dc
2R
The voltages across the switches are,
v S1  0
v S 2  vo  Vdc  2Vdc
Mode 2 starts when S1 and S2 are OFF during the short interval ,
v 0
o

io  0
iC1  iC 2  iin  0
Vdc
vS 1  vS 2 
2
 Example 9.1
Mode 3 starts when S2 is ON and S1 is OFF, which yields the following equations:
Vdc
vo  
2
vo V
io    dc
R R
i V
ic1  o  dc
2 2R
V
ic 2  ic1   dc
2R
V
iin  ic1  dc
2R
vS 1  Vdc  vo  2Vdc
vS 2  0

Mode 4 is similar to Mode 2 since both switches are open.

The average output voltage is given by:
T
1 2
T 0
vo2, rms  vo (t )dt

1  
  2 

  Vdc d    (Vdc )d  
2
vo , rms  (9.1)
2      

Vdc
vo2, rms  
2

Notice that when   0 , the rms value of the output is V dc/2

 Example 9.1
Figure 9.7(b) and 9.7(c) show the currents and voltages for =0 and 0, respectively. The
waveform is symmetric, the average output voltage will always be zero.

Fig 9.7 (a) Switching waveforms for Example 9.1. (b) Current and voltage waveforms for 
= 0 (c) Current and voltage waveforms for   0.
 Inductive-Resistive load:
Figure 9.8 (a) shows a half-bridge inverter under an inductive-resistive load, with its
equivalent circuit and the output waveforms are shown in Fig. 9.8(b) and (c), respectively.

(a) (b) (c)

Fig 9.8 (a) Half-bridge inverter with inductive-resistive load. (b) Equivalent circuit. (c)
Steady-state waveforms.
 MOSFET Implementation
• With S1 and S2 switched complementarily, each at a 50% duty cycle at a switching
frequency f , vin (t ) is defined:
 Vdc 0  t  T 2
vin (t )    (9.2)
 Vdc T 2  t  T

• The half-bridge inverter circuits shown in Fig. 9.9 with S 1 and S2 implemented by
IGBTs.
• In the steady state and its inductor current waveform is shown in Fig. 9.8(c) for , 0  t  t1
the inductor current is negative. At t  t1 ,the current flows through the transistor Q 1, as
shown in Fig. 9.8(c). At t  T 2 ,when S 2 is turned ON, since t  Tthe
2 current
t1 direction is
positive, the flyback diode, D2, turns ON until when Q2 starts conducting.

Fig 9.9 MOSFET implementation for S1 and S2 in the half-bridge inverter of Fig 9.8(a)
 Exact Analysis
In steady state,
i L (0)   i L T 2 
i L (0)  i L T 
During the first interval ( 0  t  T / 2 ) when S1 is on and S2 is off, v in (t )  Vdc , resulting in
the following equation for i L ( t ):
di L
L  RiL  Vdc (9.3)
dt
If the inductor initial value equals I L (0) , the solution for i L (t ) is given by:
Vdc  t  Vdc
iL (t )  ( I L (0)  )e  (9.4)
R R
L

R
Since I L (T / 2)   I L (0) , then the initial condition at t=0 is constant and given by:
T
Vdc 1  e 2
I L (0)  
R 1  e T 2 (9.5)
The second half cycle for t  T 2 produces the following expression for i L (t ) with the initial
condition at t  T 2 equaling - I L (0) .
V V
i (t )  ( I (0)  )e   
L L
dc  t T 2  dc
(9.6)
R R
The average power delivered to the load can be obtain from the following relation,
T
1
T 0
Po , ave  iL (t )vo (t )dt

T 2
2Vdc   Vdc  t  Vdc 

T    I
0
L (0) 
R 
e 
R 
 dt (9.7)
Where I L (0) is given by Eq. (9.5)
The time at which i L (t ) becomes zero, t  t,1 is obtained by setting i L (t )in Eq. (9.4) to zero at ,
t  t 1to yield, 2
t1   ln T
(9.8)
1 e 2
 Average Transistor and Diode currents
Obtain quantitatively the expressions for the diode and transistor currents, we represent the
load voltage and current by their fundamental components as shown in Fig. 9.10(a).
v o (t ) of and i L (t ) be given by
v o1 (t )  Vo1 sin t iL1 (t )  I o1 sin( t   )

where Vo1  2Vdc and I o1 and  are the peak current and the phase angle

2Vdc  L 
 and Z  R  L 
2
I o1  ,   tan 1  2

Z  R 

Fig 9.10 (a) Output voltage and current waveforms. (b) Average transistor and diode current
waveforms as a function of .
 Inductive Load Analysis
Using the fundamental component expression, the rms value of the diode and transistor
currents are given by:

1 2
2 0
I D ,rms  i L1 (t )dt

1
I D , rms 
2 I
2
o1 sin 2  td t (9.12)
0

I o21  sin 2  I o1  sin  cos 


4   2    2 

2
 

1
I Q ,rms  
2 
I o21 sin 2 tdt

I o21 I
(9.13)
 sin  cos
 (  cos sin    )  o1 1  
4 2  

and the average values of the diode and transistor currents are given by,

1
2 0
I D ,ave  I o1 sin tdt

I
 o1 (1  cos  )
(9.14)
2

I
I
 o1 (1  cos  )
(9.15)
Q, ave 2
 Inductive Load Analysis

The average output power delivered to the load is given by:

Po,ave  Vo1.rms I o1,rms cos (9.16)
Vo1 2V dc
Vo1, rms  
2 
I o1
I o1, rms 
2

Equation (9.16) yields the following expression for Po ,ave ,

2Vdc2
Po , ave  2 cos
 Z (9.17)
The ripple voltage in the dc-to-ac inverter is defined by,
Vo ,ripple  Vo2,rms  Vo2,ave (9.18)
The input ripple current expression in the dc-to-ac inverter is defined by:
I in ,ripple  I o2,rms  I o2,ave
(9.19)
I in , ripple  I in, rms 1 
8 o2 (9.20)
2

Example 9.2
Consider the half-bridge inverter of Fig. 9.9 with the following circuit components:
f  400 Hz L  40mH . V dc  408V R  8
(a) Derive the exact expression for iL (t.).
(b) Derive the expression for the fundamental component of iL
(c) Determine the average diode and transistor currents.
(d) Determine the average power delivered to the load.
(e) Determine the ripple current.

Solution
(a) The exact solution for iL (t ) is derived before, and given again in Eq. (9.21).
 Vdc   t  Vdc
 I L (0)  e  0t T 2
 R  R (9.21)
iL (t )  
  I (0)  Vdc  e t T 2  V
 dc T 2t T
  L R 

R


Since i L (T 2)   I L (0)

Vdc 1  e T 2 L
I L (0 )    (9.22)
R 1  e T 2 R

(b) To calculate the fundamental component of i L (t ) , we first determine the a 1 and b1

coefficients:2 T 2 T 
I L/ 1 (t ) 
T  0
  i L (t ) cos tdt   i L (t ) cos tdt  (9.23a)
T 2 
2 
T 2 T
I L1 (t )     i L (t ) sin tdt   i L (t ) sin tdt 
//
 (9.23b)
T  0 T 2 
 Example 9.2
The fundamental component of i L (t ) is given by,
i L1 (t )  I L/ 1 (t ) cos t  I L//1 (t ) sin t
i L1 (t )   0.57 cos(2513.3t )  57.8 A
(9.24)
The rms value of i L (t ) :

1 T 2 2 T 
I L ,rms    i L (t )dt   i L (t )dt 
2 I L ,rms  64.1 A
T  0 T 2 
(c) The average diode current is given by Eq. (9.14) as,
I  L 
I D ,ave  o1 1  cos    tan 1    85.45
o

2  R 
2Vdc 2  408
I o1    2.58 A Z  R 2  L  100.85 
2
 Z   100.85
Substitute the above values in ID,ave, then ID,ave turns to be
is equal to,
2.58
I Q ,ave I D ,ave   
1  cos 85.45 o  0.53 A
2
I o1
I Q ,ave  1  cos  0.63 A
2
(d) The average power is given in Eq. (9.17)
Pave.  Vo1,rms I o1, rms cos

2Vdc2
 cos 
2 Z
2(408) 2
Pave.  cos(85.45)  26.5W
 2 100.85 
 Example 9.3
Draw the output voltage and v swaveforms
1 for the center tap bi-phase inverter shown in Fig.
9.11. Assume S1 and S2 are bi-directional switches and are switched at a 50% duty cycle. It
is used in a low input voltage application to reduce losses, since the current only flows half-
period in a section of the transformer (the transformer is not fully utilized). The two modes
of operations are shown in Fig. 9.11(b) and (c), the waveforms are shown in Fig. 9.11(d).

Solution:
The equivalent circuit for Mode 1 when switch S1 is on is shown in Fig. 9.11(b). The output
voltage is given by,
vo n2

Vdc n1
n2
vo  Vdc
n1

Figure 9.11(c) shows the equivalent circuit for Mode 2

vo n
 2
Vdc n1
n2
vo   Vdc
n1
 Example 9.3

The waveforms for vo are shown in Fig. 9.11(d)

(a)

(b) (c)

(d)

Fig 9.11 (a) Center-tap biphase inverter for Example 9.3. (b) Mode 1. (c) Mode 2. (d)
Voltage waveforms
 Full-Bridge Inverters
• Figure 9.12 shows the full-bridge circuit configuration for a voltage-source inverter
under resistive load.
• If S1, S3 and S2, S4 are switched ON and OFF at a 50% duty cycle as shown in Fig.
9.13 (a), the output voltage, shown in Fig.9.13 (b), is a symmetrical square wave whose
fundamental rms value is controlled only by varying Vdc.

Fig 9.12 Full-bridge inverter under a Fig 9.13 (a) Switching sequence
purely resistive load. for full-bridge voltage-source
inverter at 50% duty cycle. (b)
Output voltage waveform.
 Full-Bridge Inverters
The fundamental value of is given by:
Vdc (9.25)
v o1 (t )  sin(t )

Vdc
The rms value is .
2
The fundamental component of the output voltage, vo1 (t )is given by,
1 
v o1 (t )  Vdc  sin t    (9.26)
2 
1 
And the rms value is given by Vdc 
4 2

Fig 9.14 (a) Switching sequence with 

phase shift. (b) Output voltage. (c)
Fundamental component for vo(t).
 Modes of Operation
It can be seen from the switching sequence of Fig. 9.14(a) that in the steady state, there
exist four modes of operation as shown in Fig. 9.16.

1 
Vdc 
4 2

Fig 9.16 Modes of operation.

 Possible Switch Implementation

Under an inductive load, for the circuit to work using SCRs, a diode must be added in
parallel with each SCR as shown in Fig. 9.15.

Fig 9.15 Possible switch implementation.

Example 9.4: Consider the resistive load full-bridge voltage source inverter shown in Fig.
9.12 with the following circuit parameters: Vdc = 150V, R = 12  and fs = 60 Hz. Sketch the
waveforms for vo & determine
iin the average power delivered to the load0 for the two switching
sequences shown in Fig. 9.13(a) and Fig. 9.14(a), with.   10

Solution
For the switch sequence shown in Figure 9.13(a), vo and io are symmetric and given by,
Vdc ,0  t  T 2
vo  
Vdc ,T 2  t  T
 Vdc
 R ,0  t  T 2
io  
 Vdc ,T 2  t  T
 R

The average output power is given by,

T
1
Po ,ave   vo io dt
T 0
For the switch sequence shown in Fig. 9.14(a), the average output power as given by:
Vo2,rms
Po ,ave 
R
where the rms value is expressed as,
1 
Vo , rms  Vdc 
4 2 0
The resultant average output power for   10 is given by,
 2 
Vdc2 1  
   Po ,ave  1666.67 watt .
Po, ave 
R
 Full-Bridge with R-L Load
• Practical loads do not consist of a simple resistor with a unity power factor, but rather
have some sort of an inductance. Figure 9.17(a) shows a full-bridge inverter under an
inductive-resistive load.
• If the switches are operating at a 50% duty cycle with a two-state output, then the
current and voltage waveforms are shown in Fig. 9.17(b). The analysis of this inverter is
similar to the half-bridge voltage source inverter discussed earlier.
• The rms values for and of Fig. 9.17(b) v o are i based on a 50% square wave output
o
given by,
I o2,rms  I12,rms  I 22,rms  .......  I n2,rms (9.27a)
V 2 V (9.27b)
o , rms dc

In
Where I n,rms  2 ,
and I n is the peak current of the nth harmonic of i o (t ).

Fig 9.17 (a) Full-bridge inverter under R-L load.

 Full-Bridge Inverters
Fig 9.17 (b) Waveforms under 50% duty cycle. (c) Fundamental component of the inductor
current.

(b) (c)
 Full-Bridge Inverters

Output voltage is a square wave with a 50% duty cycle, its Fourier series
4Vdc  sin 3t sin 5t sin nt 
vo (t )   sin t   . . .  (9.28)
  3 5 n 
Therefore,i o (t )is given by,
 
4Vdc  sin t sin 3t sin 5t sin nt 
io (t )    . . . (9.29)
  3 R  3L  5 R  (5L)
2 2 2
n R  nL  
2
 R  L 
2 2 2 2

rms value for the nth current component,

2 2Vdc
I n ,rms 
n Z n
where,
Z n  R 2  (nL) 2

we assume that the major part of the average output power is delivered at the fundamental
frequency, then Po ,ava
8Vdc 2 R
Po , ave  cos 
 2
R   L 
2 2
(9.30)
 Approximate Analysis
An approximate solution for the load current assuming that L R  T 2 .
io ,1  I o1 sin  t    (9.31)
where, 1  L
  tan
R
Vo1 4Vdc
I o1  and Vo1 
 L 2  R 2 

The average power delivered to the load,

Pave  I o ,1rmsVo ,1rms cos
I o ,1Vo ,1
 cos
2
8Vdc2 R
Pave  cos  (9.32)
 L   R 2
2
2

where I o1,rms is the rms of the fundamental component of the inductor current.
 Generalized analysis
• Figure 9.18(a) shows a typical output voltage under  control that is produced using the
switching sequence of Fig. 9.14(a).
• The equivalent circuit for the single-phase bridge inverter is shown in Fig. 9.18(b).

(a) (b)

Fig 9.18 (a) Output voltage using Fig 9.19 Generalized load
switching sequence given in Fig representation.
9.14(a). (b) Equivalent circuit for
the full-bridge inverter.
 Generalized analysis
The Fourier analysis representation for v o (t ) is given by the following equation,

v o (t )  V n sin n t (9.33)
where,
n 1, 3, 5

4Vdc (9.34)
Vn  cos n
n
and i o (t ) is obtained from the following equation,

i o (t )  I n sin( n t   n ) (9.35)
n 1, 3, 5
where,
Vn 4Vdc (9.36a)
In   cos n
Zn n Z n

 n  Z n
(9.36b)
where Z n is the magnitude of the nth harmonic impedance, and Z L is the angle of the nth
harmonic impedance. The overall rms output voltage is given by,
2
Vo , rms  V dc 1 
 (9.37)
In terms of the nth harmonics,
4Vdc 1 
Von2 ,rms 
2 n 1,3,5
n 2
cos 2 n (9.38)
rms value of the current for each harmonic

I on, rms  I
n 1, 3, 5...
2
n , rms (9.39)
 Generalized analysis

rms for the nth harmonic

4Vdc
I n ,rms  cos n (9.40)
2 n Z n

rms of the fundamental output current

4Vdc
I o1,rms  cos  (9.41)
2 Z 1
where,
Z 1  R 2   o L 
2

The total average power delivered to the load resistance is given by,
T
1
Po ,ave   io vo dt
T 0 (9.42)
 Generalized analysis

The voltage and current’s THD are given in Eqs. (9.43a) and (9.43b), respectively.
2
 Vo ,rms 
THDv    1

 Vo1,rms 
(9.43a)
 1 
 
2 cos  2 

2
I 
THDi   o ,rms   1
I
 o1, rms  (9.43b)
2
  Z1 cos  
   1
 n 1,3,5... n Z n cos n 
 

For the nth harmonic, the average power of Eq. (9.42) is given by,
Pon, ave.  I on, rmsVon , rms cos  n
the total average power

Po, ave.  I
n 1, 3, 5...
on , rms Von, rms cos  n (9.44)
 Example 9.5
Consider the full-bridge inverter whose equivalent circuit is represented in Fig. 9.19 with
the four different loads shown in Fig. 9.20 with R=8, L=30mH, C=147F, fo=60Hz, and
Vdc=120V.
(a)    Determine the rms fori o and vo for the 1st, 3rd, 5th and 7th harmonics.
(b)   Determine the total average power delivered to the load for each of the above harmonics.
(c)    Determine the output current and voltage total harmonic distortion.
Solution
To determine the rms for io and vo for the 1st,3rd, 5th and 7th harmonics, we use the nth
harmonic given by,
4V
v n ,o  dc cosn  sin nt (9.45)
n
4Vdc
I n ,rms  cos n
2n Z n

Fig 9.20 Various types of loads.

 Example 9.5
i) For the resistive load:
2 2Vdc Vn , rms
Vn ,rms  cosn  I n, rms  Z n  R  8  =/6.
n 0
n Zn
2 2120   V1,rms V1,rms
V1,rms  cos   93.56V I 1,rms   11 .69 A
 6 Z1 R
2 2120   V3, rms V3, rms
V3,rms  cos   0V I 3, rms   0 A
3 2 Z3 R
2 2120  5  V5,rms V5, rms
V5,rms  cos    18.71V I 5,rms    2.34 A
5  6  Z5 R

2 2120  7  V7 ,rms V7 ,rms

V7 ,rms  cos   13.37 V I 7 , rms     1.67 A
7  6  Z7 R

Using the first four harmonics, the approximate rms values for vo and io are given by,
Vo ,rms  V1,2rms  V32,rms  V52,rms  V72,rms  96.34V

I o ,rms  I 12,rms  I 32,rms  I 52,rms  I 72,rms  12.04 A

The exact rms values are given by

2 2Vdc  
Vo ,rms  cos    93.56 V
 6
V
I o, rms  o, rms  11.69 A
Zn
 Example 9.5

ii) For the series RL Load:

The rms value for the output voltage is the same as the one above, since the output voltage is
independent of the load.
Vn ,rms  nL 
I n ,rms  Z n  R 2  nL   n  tan 1  
2
Zn  R 
V1,rms
I 1,rms   6.75 A
R 2  L 
2

V3,rms
I 3,rms  0 A
R  3L 
2 2

V5,rms
I 5,rms    0.33 A
R 2  5L 
2

V7 ,rms
I 7 ,rms   0.168 A
 7L 
2 2
R
 Example 9.5

iii) For the series RLC Load:

The rms value for the output voltage is the same as in (i), since the output voltage is
independent of the load.
 1 
 1 
2
 n L  
2
Z n  R   nL    n  tan 1  nC 
 nC   R 
 
V1,rms  
I 1,rms   9.76 A
2
 1 
R 2   L  
 C 
V3,rms
I 3,rms  0A
2
 1 
R 2   3L  
 3C 
V5,rms
I 5,rms   0.35 A
2
 1 
R 2   5L  
 5C 
V7 ,rms
I 7 ,rms    0.17 A
2
 1 
R 2   7L  
 7C 
 Example 9.5

iv) For the parallel RLC Load:

The rms value for the output voltage is the same as the one shown above, since the output
voltage is independent of the load.
1
Zn   n 2 2 LRC  R 
2 2  n   tan 1  
 1   n  LC  1 
2 2
      n L 
R  nL 
V1,rms
I 1,rms   723.68 A
2 2
 1   1  LC  1 
2 2
    
R  L 
V3,rms
I 3,rms   0A
2 2
 1   3  LC  1 
2 2
    
R  3L 
V5,rms
I 5,rms    64.97 A
2 2
 1   5  LC  1 
2 2
    
R  5L 
V7 ,rms
I 7 ,rms    33.80 A
2 2
 1   7  LC  1 
2 2
    
R  7L 
 Example 9.5

(b) Power calculations are obtained from the following equation:

Pon ,ave  I on , rmsVon ,rms cos n
i) For the resistive load:
cos n For
1 n=1,3,5…

Po1,ave  (93.56)(11 .69)  1118 .04watt

Po 3,ave  0 watt

Po 5, ave  ( 18.71)(2.34)  43.78 watt

Po 7 ,ave  (13.37)(1.67 )  22.33 watt

Po ,ave.  I
n 1, 3, 5...
on , rms Von,rms cos n  1184.15watt

ii) For the series RL Load:

Po1, ave  (93.56)(6.75) cos(54.72 )  364.52watt
Po 3,ave  0 watt
Po 5,ave  (18.71)(0.33) cos(81.93)  0.867 watt
Po 7 ,ave  (13.37)(0.168) cos(84.22 )  0.226watt

Po ,ave.  I
n 1, 3, 5...
on , rms Von,rms cos n  365.61watt
 Example 9.5

iii) For the series RLC load:

Po1,ave  (93.56)(9.76) cos( 40.091)  698.48 watt

Po 3,ave  0 watt
Po 5,ave  (18.71)(0.35) cos(81.36)  0.98watt
Po 7 ,ave  (13.37)(0.17) cos(84.037 )  0.24watt

Po ,ave.  I
n 1, 3, 5...
on , rms Von,rms cos n  699.69watt

iv) For the parallel RLC load:

Po1,ave  (93.56)(723.69) cos(14.78)  65465.11watt

Po 3,ave  0 watt

Po 5,ave  (18.17) (64.97 ) cos64.274  512.42watt

Po 7 ,ave  (13.37) (  33.80) cos71.579 142.92watt


Po ,ave.  I
n 1, 3, 5...
on , rms Von,rms cos n  66120.45watt

(c) The total harmonic distortion can be obtained from Eq. (9.43) given by,
2
 I o ,rms 
THDi     1  0.25

 I o1,rms 
 Example 9.6
Consider the active load in a bridge inverter that consists of an R-L load and an ac sinusoidal
voltage source as shown in Fig. 9.21 with the load voltage vo as shown. Obtain the
expression for the fundamental load current and the average power delivered to the load for the
following circuit parameters: v ac  100 sin 2 100t  30
,Vo dc =180V, L=42mH, R=0.5 and =15o,
fo = 60Hz.
Solution:
(a) To obtain the expression for the fundamental load current we can use the following
equations:
 
 V V
I o1  o1 ac
R  jL
 4V 4  180
Vo1  dc cos  cos(15o )  221.370 o V
 

Vac 100 30 o
 221.370o  100300
I o1   9.08  108.54o A
(0.5  j 2 (60))(42 10 3 )

The total power delivered to the load is given by the following equation:
1  
PT  I o1 Vo1
2

1
PT  (221.370o )(9.08  108.54o )  1005.02  108.54o W
2
 Harmonic Reduction
• Harmonic reduction includes the elimination and cancellation of certain harmonics of the
output voltage.  

• Unlike in dc-to-dc converters, harmonics in the output waveforms are very significant.

• In dc-to-dc inverters, the reduction or cancellation of the output harmonics is done

actively by controlling the switching technique of the inverter.

• Unlike ac-to-dc conversion, harmonic filtering in dc-to-ac is harder since it will affect the
attenuation and/or the phase shift of the fundamental component.

• The harmonics that are present in the inverter’s output voltage are high for many
practical applications.

• By controlling the width or the number of pulses, it will be shown that certain harmonic
contents can be removed without the need to use complex harmonic filtering circuits.
 Harmonics
It is possible to cancel certain harmonics by simply selecting the duration of the pulse in the
half-cycle of the output voltage. Peak component of the nth harmonic for Fig. 9.18(a) is given
by,
4V (9.46)
V  dc cos 
n
n
Considering the 3rd harmonic, we have,
4Vdc (9.47)
V3  cos 3
3

To cancel the 3rd harmonics, set   , this results with the cancellation of all the harmonics
6
of the order of 3n.
 
The nth harmonic for Fig.9.22 is given by,

4Vdc
Von  1  cos n1  cos n 2  (9.48)
n
For example to eliminate the 3rd and 5th harmonics, we make Von  0at two frequencies as follows,
1  cos 3 1  cos 3 2  0 1=17.8o and 2=38o
1  cos 5 1  cos 5 2  0

Fig 9.22 Two-angle shift

control of inverter output
voltage.
 Example 9.7
Consider two cascaded push-pull inverters as shown in Fig. 9.23(a), the switching
waveforms for S1-S4 are shown in Fig. 9.23(b). Sketch the waveforms for the outputs vo,
v v
, o1 and o 2 .
Solution:
The waveforms for the output voltage waveforms are shown in Fig. 9.23(c).

The fundamental component of vo is given by,

2V 
vo1 (t )  dc cos sin( t   )
 2
where,
   t1

Fig 9.23 (a) Two push-pull inverters for Example 9.7. (b) Typical switching waveforms.
 Harmonic Analysis
The harmonics contents present in the output of the inverter could be significant. Recall
that the nth component for a square wave output   0 is 4Vdc . For   0 ,the
n
expression for vo (t ) is given by,
vo (t )  V
n 1, 3, 5,...
n sin n t (9.49)
where,
4Vdc
Vn  cos n
n
The fundamental output voltage component is given by,
4Vdc
v o1  cos  sin  t (9.50)

 Harmonics
Figure 9.24(a), (b) and (c) show the resultant output voltage where harmonics up to the 9 th
 
are included for   0,   6 and   3 , respectively.

(a) (b)

Fig 9.24 The first nine output harmonics: (a)  = 0. (b)  = /6.
 Harmonics

The sketch for the magnitude of the harmonics of Fig. 9.24, as a function of , is shown in
Fig. 9.25.

Fig 9.24 (continued) (c)  = /3. Fig 9.25 Magnitude of the output harmonics
as a function of .
 Harmonics
 
The total harmonic distortion for v o (t ) is given by,
2
 Vo ,rms 
THDV    1

 Vo1,rms  (9.51)
Vo2,rms  Vo21,rms  Vo22,rms .....  Von2 ,rms
 V
n 1, 3, 5..
2
on , rms

where,
Von
Von2 , rms 
2
4Vdc
Von  cos n
n
Substitute Eqs. (9.52) and (9.53) into Eq. (9.51) to yield,
  Von2 ,rm 
 n 1,3,5.. 
THDV   2  1
 Vo1,rms 
 
  Von2 ,rm 
 n 3,5,7.. 
THDV   
 Vo1,rms
2
 Fig 9.26 Plots of THD
  as a function of .
1 1
THDV 
cos 

n  3, 5... n 2
cos 2 n

The plots of the THD for   0 to    2 are shown in Fig.9.26.

 Pulse Width Modulation
• Figure 9.27 shows the simplified block diagram representation for a single-phase
switching mode inverter.
• The outputvo (t ) shows different types of possible output waveforms that can be
produced depending on the Pulse Width Modulation (PWM) control technique
employed.

Fig 9.27 Simplified block diagram of single-phase switching-mode inverter.

 Pulse Width Modulation
• Since varying the width of these pulses within the half-cycle carries out the control,
the process is appropriately known as Pulse Width Modulation (PWM). The PWM
method may be grouped into classes:
• The PWM method may be grouped into classes:

1) Non-sinusoidal PWM in which all pulses have the same width and are
normally modulated equally to control the output voltage as shown in Fig. 9.28.
The widths of these pulses are adjusted equally to control the output voltage.

Fig 9.28 Typical waveform for equal-pulse PWM technique.

 Sinusoidal PWM

2) Sinusoidal Pulse-Width-Modulation that allows the pulse width to be modulated

sinusoidaly, i.e. the width of each pulse is proportional to the instantaneous value of a
reference sinusoid whose frequency equals the fundamental’s as shown in Fig. 9.29.

Fig 9.29 Typical waveforms for sinusoidal PWM technique.

 PWM Definition
Reference voltage vref . (t ) is square and sinusoidal waveforms for the equal pulse and
sinusoidal PWM. It is important that first we define some terms
vtri . t  :
Repetitive triangular waveform (also known as a carrier signal).
VP ,tri . :
The peak value of the triangular waveform.
TS , f S : The period and the frequency of the triangular waveform (also known as a
carrier or switching frequency).
v ref . : Reference signal that can be either a square or a sinusoidal waveform (also
known as a control signal).
VP ,ref . :The peak value of the reference signal.
To , f o : The desired inverter output period and output frequency, which are equal to the
period and frequency of the reference or control signal.
ma : Inverter amplitude modulation index.
mf : Inverter frequency modulation index.
k: Number of pulses per half cycle.

The amplitude and frequency modulation indices are defined as follows:

V P ,ref .
ma  (9.55)
V P ,tri.

fs (9.56)
mf 
fo
 Equal Pulse (Uniform) PWM
The Equal Pulse PWM technique, known also as a single pulse PWM control. Generating
the equal and multiple pulses is acheived by comparing a square wave reference voltage
waveform v ref (t )to a triangular control (carrier) voltage waveform vcont.. (t )

(a) (b)

Fig 9.30 Examples of equal pulses. (a) One-pulse output. (b) Two-pulse output.
 can be expressed in terms of VP ,tri. and VP ,ref . as follows:
  VP ,ref .  
   1  1  ma  (9.57)
2  V P ,tri.  2
• Frequency of the control signal, f tri. , is twice the frequency of the reference signal, f ref . .The
number of pulses, k, is equal to the number of periods of the control signal per half a period of
the reference signal, i.e. k is the number of switching periods, Ts , in the To 2 period,
1 fS
k (9.58)
2 fo
• In terms of the frequency modulation index,m f , k may be expressed by the following relation:
1
k mf (9.59)
2
Notice that k 1 is a special case where . f S fo  2

The output frequency is equal to the frequency of the reference signal, i.e. f o  f ref . and the
switching frequency is equal to the frequency of the carrier or triangle signal, f S  f tri. .
 Example for k =7 and mf = 14
The maximum width of each pulse occurs when ma  1 and is given by,
T
t width ,max  o (9.60)
2k
the maximum conduction angle width of each pulse is given by,

 width,max   o t width,mox  (9.61)
k

(c) (d)

Fig 9.30 (continued) Examples of equal pulses. (c) Three-pulse output. (d) Seven-pulse
output.
A general expression for the ith pulse-width in a given k-pulse output in terms of i, k, and
. Referring to Fig.m9.31,
a the start of the ith pulse is,
TS  VP ,ref . 
ti  (i  1)TS  1   (9.62)
2  VP ,tri . 
1
Substituting for fS  from Eq. (9.58) into Eq. (9.62), t i becomes,
TS
To T (9.63)
ti  (i  1)  o 1  ma 
2k 4k
In terms of the starting angle,  i   o t i , of the i th pulse,
  ma 1 
i  i  2  2  (9.64)
k  
for the two-pulse waveform, k  2 , and m=0.5, the angles at which the pulses start are
given by,
  1 1    1 1  9
1  1   3  3  
2  4 2  8 2  2 4  8

  1 1  5   1 1  13
2  2    4  4   
2  4 2 8 2  2 4 8

because of the symmetry, 3 = 1 +  and 4 = 2 + .

The Output Voltage
The average output voltage over a period of T s,
Vo ,ave.  maVdc 0  ma  1 (9.66)
The rms value for the pulse
th
i
 i  width
1 2
Vo ,rms 
2  2kV

dc dt
(9.67)
i

k
Vo ,rms  Vdc  width

ma
From Eq. (9.65)  width   , hence,
k
k m a
Vo ,rms  Vdc
 k
 Vdc ma (9.68)

The rms of the output voltage is a function of the modulation index.

 Harmonics of k equal pulses
th
Determine the harmonic components for the i pulse shown in Figure 9.32.
Since it is an odd, function only the odd harmonics exist in the bn coefficients. The harmonics of
th
the output voltage due to the i pulse acting alone is given by,
V  i  width  i  width 

Von,i  dc   cos n t dt   cos nt dt 
   i  i  
(9.69)
2V  n          
 dc sin width  sin  n i  width    sin n    i  width  
n  2    2    2  

For the total k pulses, Von is given by,

k
2Vdc  n          
Von   sin width  sin n i  width    sin n    i  width   n  1,3,5......
i 1 n  2    2    2  
(9.70)

In terms of k,  i ,  width and ma , can be expressed as,

  1  (9.71)
 i   i    width where i  1, 2, ......k
k  2 2
 (9.72)
 width  ma
k

Fig 9.32 Symmetric al

representation of ith pulse for a
given inverter output.
The nth harmonic component for the k-pulse output voltage maybe expressed as:
4Vdc m   k  
Von  sin  a n  sin  i  1 2n  (9.73)
n  2k  i 1  k 
 
To verify the above equation for a single pulse, we calculate the harmonic components for
k=1, i=1, and 1   , to yield

4Vdc  1  
Vn  sin  i  1 2  sin  1  1 2 n  (9.74)
n k  i 1  k 

Substituting for ma from Eq. (9.72), Eq. (9.74) becomes,

 4V    width  (9.75)
Vn   dc
 n
 sin  n
 

2 
for odd n

Equation (9.75) represents the harmonic components of the output voltage for k = 1 as a
function of the width  width .
Example 9.8
For a uniform PWM with a value of k=5, and a modulation index ma  0.2 , calculate the
output harmonic components up to the 15 th harmonic.
Solution:
Using Eq. (9.73), Table 9.1 shows the values of the first 15 harmonics. Figure 9.33 shows the
plot for the harmonic contents of Table 9.1. Figure 9.34 shows the harmonic ratios with
respect to the fundamentals for m=0.2 and k=1 to k=7 pulses per half cycle.

Fig 9.33 Harmonic contents for

k = 5 and ma = 0.2.
 Example 9.8 (cont’d)
From Table 9.1 for ma=0.2, as the number of pulses increases per half a cycle, the
magnitude of the lower harmonics (third, fifth, seventh) decreases with respect to the
fundamental component. The THD for this example is 223%! The higher the modulation
index, the lower the THD.

Fig 9.34 Harmonic contents for k = 1 to k = 7, and ma = 0.2.

 Sinusoidal PWM
Basic concept:
Recall that in PWM dc-dc converters the duty cycle is modulated between 0 and 1 in
order to regulate the dc output voltage. In the steady state, the duty cycle in PWM switch
mode converters is relatively constant and does not vary with time:

Vo  DVdc (9.76)
where D is the duty cycle representing the ratio between the on-time of the switch to
the switching period, and Vo is the average output voltage.

Fig 9.35 Simplified buck converter.

 Sinusoidal PWM
Basic PWM Concept
If the duty cycle, d (t ) , varies or is modulated according to a certain time function, with a
modulating frequency, f o , then it is possible to shape the output voltage waveform, vo , in
such a way that its average value over the modulating period synthesizes a sinusoidal
waveform.If the duty cycle is defined according to the following function,
d (t )  Ddc  Dmax sin  o t
(9.77)
where,
Ddc : dc duty cycle when no modulation exists.
Dmax : maximum modulation constant.
 o : frequency of modulation.

the output voltage , vo , is given by,

v o  d (t )Vdc
 Vdc Ddc  Vdc M max sin  ot (9.78)
For a buck converter, since the output voltage cannot be negative, then Dmax  Ddc as shown
in Fig. 9.36.

Fig 9.36 Example of Ddc

= 0.5, Dmax = 0.8Ddc, and
fo = fs/12.
 Sinusoidal PWM
Basic PWM Concept
As an example, if Dmax  0.8 Ddc , Ddc  0.5 , and f o  f s 12 , the duty cycle, , is given
by,
d (t )  0.5  0.4 sin( 2f oti )
where t i = 0,1,2,3,…,12.

Since the switching frequency is 12 times faster than the modulating frequency, fo,
then d (t ) is sampled 12 times between 0  t  To .
 Sinusoidal PWM

PWM Modulation Function

In the single pulse inverter with a two-state output of Fig. 9.13b, the modulation function is
simply unity.

vo  m(t ) Vdc m(t )  1 0t T 2 (9.79)

In the single pulse inverter with a tri-state output, the modulation function in a half cycle is
given by,
vo  m(t ) Vdc

0   t  

m(t )   1    t  T 2   (9.80)
0 t T 2

In a k-pulse inverter with a constant duration, the modulation function is given by.

vo  m(t ) Vdc  1 0  t  d (t ) (9.81)

m(t )  
0 d (t )  t  TS
 Sinusoidal PWM

Switching Schemes
• Depending on the switching sequence, the output voltage in PWM inverters can be either
bi-polar or unipolar. Fig. 9.37 shows a bi-polar output voltage in a PWM inverter. When
the reference sinusoidal signal is larger or smaller than the triangular wave, the output
equals,  Vdc or  Vdc .
• In the bipolar voltage switching, m f is an odd number with the same switching
frequency, f S .The output frequency, f o ,in the unipolar voltage switching, is twice
that of the frequency in the bipolar voltage switching ( m f is doubled).

Fig 9.37 Example of a bipolar PWM output waveform.

 Sinusoidal PWM
Unipolar Output Waveform

(a) (b)

(c)

Fig 9.38 Unipolar PWM output. (a) A positive sinusoidal reference to produce v o1 (b)
Positive sinusoidal reference to produce v o2. (c) The differential output vo = vo1 – vo2.
 Sinusoidal PWM

Signal Generation
Advanced digital and analog techniques exist in today’s inverters to generate the driving
signals that produce a sinusoidal PWM. Fig. 9.39 shows a comparator that compares a
triangular signal to a sinusoidal reference signal.

Fig 9.39 Simplified circuit showing how signals are generated in sinusoidal PWM inverters.
 Analysis of Sinusoidal PWM
• The output voltage signal in sinusoidal PWM can be obtained by comparing a control vcont
signal, against a sinusoidal reference signal, v ref , at the desired frequency.

• The control frequency f cont determines the number of pulses per half cycle of the output
voltage signal. The output frequency f o is determined by the reference frequency, f ref .The
amplitude modulation index, m a ,is defined again as the ratio between the sinusoidal
magnitude and the control signal magnitude.
V p ,ref
ma  (9.83)
V p ,cont
 Sinusoidal PWM

PWM Concepts

The duration of the pulses is proportional to the corresponding value of the sine wave at that
corresponding position. Then, the ratio of any pulse duration to its corresponding time
duration is constant as shown in Figure 9.40.
1  2  3 
    i  constant (9.84)
y1 y2 y3 yi

Fig 9.40 PWM figure illustrating the constant ratio Fig 9.41 High-frequency
between the width and height of a given pulse. sinusoidal PWM.
 Sinusoidal PWM

Approximation
The proportional variation of each pulse width with respect to the corresponding sine wave
amplitude could be seen by applying a triangular relationship as shown in Fig. 9.42.

Fig 9.42 Triangular approximation.

 Sinusoidal PWM

Output voltage Harmonics

• For a sinusoidal PWM, the width of each pulse varies according to its position. The
output voltage is obtained using a Fourier series transformation for vo ,

vo (t )   (a
n 1, 2,...
n cos nt  bn sin nt ) (9.85)
• Since the inverter output voltage is an odd function, only odd harmonics exist.
• The calculation of the output voltage harmonic components can be done using one
single pair of pulses as shown in Fig. 9.43.
1 2
V  v ( t ) sin(n t ) d ( t )
n, i   o
0 (9.86)
      
1 i wi i wi 
V    V sin( nwt ) d ( wt )   V sin( nwt )d ( wt ) 
n, i  dc dc
    
 i i 
• The nth component of vo is given by,
k  2V            
V    dc  sin  n wi  sin n   wi   sin n   wi     (9.88)
n    2   i 2   i 
i  1 n      
2


where,  i is the starting angle of the ith pulse, and   is the pulse width at the corresponding
angular position.
Fig 9.43 Single sinusoidal
PWM pair of pulses.
 Sinusoidal PWM

Approximating the Pulse Width, i

• Assume each pulse is located at the discrete value of  .i

• From the geometry of the triangle ABC in Fig. 9.44 we have,

h y  V p ,tri hx  V p ,ref sin  i To 
y o 
2k k

Fig 9.44 Approximated pulse width, i.e., e = 0.

 
 i , app   m a sin  i (9.91)
k
 Sinusoidal PWM

Exact expression for i

For the ith pulse with an angle i, the point of intersection between the control signal and
the t axis is 2i  1 2k ,0  as shown in Fig. 9.45.
The general expression for vcont (t ) having all the lines of negative slopes is given by:
mf   
vcont (t )   V p ,tri  t  2i  1  i  0,1,2,..., k (9.92)
  mf 
 

Fig 9.45 Intersections between vref and vcont.

 Sinusoidal PWM

Exact expression for i

The reference signal is
v ref (t )  V P ,ref sin(t ) (9.93)
Evaluating Eqs. (9.92) and (9.93) at  t  i , vcont ( i )  v ref ( i ), yields Eq. (9.94).
mf   
 V p ,tri  i  (2i  1)   V p , ref sin  i (9.94)
  mf 
 
Equation (9.94) can be re-written as ,
mf
ma sin  i    i  (2i  1) (9.95)

The value of  i can be found when solving Eq. (9.95) numerically.
 Sample PWM Pulses
• Consider the cases for a different number of pulses per half cycle as shown in Fig. 9.46.
• The exact width for each pulse is found by adding the approximated width with the
corresponding error e in each interval. The two-pulse case, we have e1  e2

Fig 9.46 Illustration of two, three, and four PWM pulses per half-cycle.
 Sinusoidal PWM
Expression for Exact Width
Calculation for any width can be done as shown in Figure 9.47.
v ref ( i/ )  v ref ( i )
tan   (9.96)
ei
From the triangle DEF:
V p ,tri VP ,tri 2k
tan     V P ,tri
 To o  1  (9.97)
  2k
 4k 
Equating Eqs. (9.96) and (9.97),e maybe expressed by,
   v ref ( i )  v ref ( i ) 
/

ei    (9.98)
 2k  V p ,tri 

V p ,ref 1
Substituting for vref (t )  V p ,ref sin(t ), ma  V p ,cont and k
2
mf in Eq. (9.98), e becomes
m a
ei 
2k
sin  i/  sin  i  (9.99)
From sin x  sin(  x) : , Eq. (9.99) becomes
 m 
ei   a sin  K 1i  sin  i 
 2k 
and since  w,i   w,i , app  ei :, go to the general expression for the exact width,
ma  1 
 w ,i   sin  i  sin  i  sin  k i 1 
k  2  (9.100)
 Sinusoidal PWM

Pair or Pulses

Fig 9.47 Symmetric pair of pulses about /2.

Example 9.9
Consider a full-bridge inverter with an R-L load with an output sinusoidal PWM voltage
with a reference frequency f o of 60 Hz, Vdc = 280V, ma = 0.6, and m f = 24.
(a) Find the carrier frequency (triangular wave)
(b) Find the number of pulses per half cycle
(c) Find the angles of intersection and the pulse widths in a half cycle
(d) Find the harmonic components
(e) Find the total harmonic distortion

Solution
(a) f s  m f  f o  24  60Hz  1.44kHz
(b) k= 24  12 pulses
(c) The2exact angles were numerically calculated using MathCad as shown in the table
below:
 Example 9.9 (cont’d)
(d) Harmonic components in volts:
V( 1 )  167.931
V( 11 )  0.019 V( 21)  19.909
V( 3 )  0.048
V( 13)  0.085 V( 23)  103.541
V( 5 )  0.127
V( 15)  0.021 V( 25)  103.74
V( 7 )  0.03
V( 17)  0.071 V( 27)  19.736
V( 9 )  0.118
V( 19)  1.058 V( 29)  0.898

These values are plotted in Fig. 9.48.

Fig 9.48 First 29 harmonics using exact analysis for Example 9.9.
 Example 9.9 (cont’d)
(e) The total harmonic distortion is approximated by
1
THD  2
(V212  V232  V252  V272  V292 )  0.89
V1

Fig. 9.50, shows the plot of the normalized magnitude of the harmonic components
using uniform pulse width modulation with a modulation index of 0.6 for 12 pulses per
half a period.

Fig 9.50 Uniform PWM for Example 9.9.