Chapter 4

Inverters
Dewei(David) Xu
Dept. of Electrical & Computer Engineering
Ryerson University

Content
• Topologies
– Single-phase inverters
– Three-phase inverters
• Gating patterns
– Square wave operation
– Pulse width modulation (PWM)
– Selected harmonic elimination (SHE)
– Delta modulation
• Microprocessor control of inverters
1

1
4.1 Introduction
• Inverters:
– Input: dc voltage, fixed magnitude
– Output: ac voltage, adjustable frequency and
adjustable magnitude
• Main feature:
– dc-to-ac conversion
– Solid state device, no rotating machines.

4.1 Introduction
• Types
– Voltage source inverter (VSI): voltage source dc power supply.
Output voltage is defined, based on the gating pattern. Output
current is dependent on the load

Widely used
in industry

– Current source inverter (CSI): current source dc power supply.

Output current is defined, based on the gating pattern. Output
voltage is dependent on the load

2
 4.2 Square wave operation
• Single-phase inverter
– Circuit diagram

T1~T4 ─IGBTs,
P
switching devices + T1 D1 T3 D3
D1~D4 ─Freewheeling g1 g3
diodes io + vo
C A − B
C ─dc filter capacitor Vd L
(VSI) T4 D4 R T2 D2
g4 g2
Vd ─dc link voltage
− N

4.2 Square wave operation

Period I:
– waveforms vg1=vg2>0 → T1, T2 on, current path:
vg1, vg2 T
Vd+→T1→Load→T2→Vd-
t vo=Vd
vg3, vg4
Period II:
t vg3=vg4>0 → But io>0 → D3, D4 on,
vAN
energy stored in L is releasing to Vd,
Vd
t current path:
vBN Vd-→D4→Load→D3→Vd+
Vd Period III:
t
vAB vg3=vg4>0 → But io<0 → T3, T4 on,
Vd current path:
t Vd+→T3→Load→T4→Vd-
Vd Period IV:
io
vg1=vg2>0 → But io<0 → D1, D2 on,
t
energy stored in L is releasing to Vd,
Period I II III IV current path:
T1, T2 on T3, T4 on Vd-→D2→Load→D2→Vd+
D3, D4 on D1, D2 on 5

3
4.2 Square wave operation
– Analysis I on ,rms nωL
• rms value of output voltage R
1 T 2
T ∫0
Vo,rms = vo dt = Vd
Von ,rms
• Fourier series output voltage
∞
4Vd 4V ⎛ 1 1 ⎞ Harmonic
vo (t ) = ∑ sin nωt = d ⎜ sin ωt + sin 3ωt + sin 5ωt + ... ⎟
n =1, 3,5... n π π ⎝ 3 5 ⎠ Equivalent circuit
• rms value of fundamental component
2π
Vo1,rms = d
4V
2 = 0.9Vd ω = 2πf =
π T
• Fourier series of output current
∞ ∞
4Vd
io (t ) = ∑ 2 I on ,rms sin (nωt − θ n ) = ∑ sin (nωt − θ n )
nπ R + (nωL )
2 2
n =1, 3, 5... n =1, 3, 5...

use harmonic equivalent circuit

Von ,rms nωL
I on ,rms = and θ n = tan −1
R + (nωL ) R
2 2

4.2 Square wave operation

– Post analysis
• Question 1: How to change output frequency
• Answer: Adjust the period T of the gate signal
• Question 2: How to change Vo,rms ?
• Answer: Adjust dc link voltage Vd
• Question 3: Is io pure AC? no dc component in
practical?
• Answer: No. usually a capacitor is used to isolate the dc
component.

4
4.2 Example
– RLC load
P
+ T1 D1 T3 D3
g1 g3
+ vo io −
C A B
Vd L
T4 D4 R C T2 D2
g4 g2
Load
− N

Example
R = 10Ω, L = 31.5mH, C = 112µF, Vd = 220V, f o = 60Hz
Find : a) Fourier series of io (up to 9th order harmonic)
b) I o1,rms , c) THD, d) Pload , and e) I d

4.2 Example
a) Use harmonic equivalent circuit
X L = jnωL = jn 2πfL = jn × 2π × 60 × 31.5 ×10 −3 = j11.87 nΩ
1 1 1 1 23.68
XC = =−j =−j × =−j Ω
jnωC n × 2πfC n 2π × 60 ×112 ×10 −6 n
2
⎛ 23.68 ⎞
Z n = R 2 + ( X L + X C ) = 100 + ⎜11.87 n −
2
⎟
⎝ n ⎠
X + XC 11.87n − 23.68 n
θ n = tan −1 L = tan −1 = tan −1 (1.187 n − 2.368 n )
I on ,rms jnωL R − j (nωC )
R 10 −1
∞
4Vd
vo = ∑ sin nωt
n =1, 3, 5... nπ
Von ,rms
= 280.1sin ωt + 93.4 sin 3ωt + 56.02 sin 5ωt
+ 40.02 sin 7ωt + 31.12 sin 9ωt + ...
Harmonic
4Vd Equivalent circuit
I on , peak =
nπZ n
io = 18.1sin(ωt + 40.7°) + 3.17 sin(3ωt − 70.2°) + 1.0 sin(5ωt − 79.6°)
+ 0.5 sin(7ωt − 82.9°) + 0.3 sin(9ωt − 84.5°) + ... 9

5
4.2 Example
b) I o1,rms = 18.1 = 12.8A
2
c) I o,rms ≈ I o21,rms + I o23,rms + I o25,rms + I o27,rms + I o29,rms
2 2 2 2 2
⎛ 18.1 ⎞ ⎛ 3.17 ⎞ ⎛ 1.0 ⎞ ⎛ 0.5 ⎞ ⎛ 0.3 ⎞
= ⎜ ⎟ +⎜ ⎟ +⎜ ⎟ +⎜ ⎟ +⎜ ⎟
⎝ 2 ⎠ ⎝ 2 ⎠ ⎝ 2⎠ ⎝ 2⎠ ⎝ 2⎠
= 13.02A
I o2,rms − I o21,rms 13.022 − 12.82
THD = = = 18.6%
I o1,rms 12.8
d) Pload = I 2
o ,rms R = 13.022 × 10 = 1.695kW
∞
or : Pload = ∑V
n =1, 3,5...
I
on ,rms on , rms cos θ n ≈ 1.69 kW

e) Lossless devices (ideal device)

Pin = Pload = 1.695kW
Pload 1695
Pin = Vd I d ⇒ I d = = = 7.7 A
Vd 220
10

4.2 Square wave operation

• Three-phase inverter
– Circuit diagram

T1~T6 ─IGBTs, P
switching devices + T1 D1 T3 D3 T5 D5
D1~D6 ─Freewheeling g1 g3 g5 R
iA
diodes A iB
C B n
C ─dc filter capacitor Vd C
iC
(VSI) g4 g6 g2
Vd ─dc link voltage − N T4 D4 T6 D6 T2 D2
R ─three-phase
resistive load

11

6
 4.2 Square wave operation
– Waveforms, gate signals
vg1 T
60° 60° 60° Note:
t (1) v and v are complimentary
vg4 60° 60° 60° g1 g4
t vg3 and vg6 are complimentary
vAN
Vd
t vg5 and vg2 are complimentary
vg3
(2) VAN is controlled by vg1. They
vg6 t have the same wave shape

vBN t VBN is controlled by vg3. They

120°
Vd have the same wave shape
vg5 t
VCN is controlled by vg5. They
vg2 t have the same wave shape

t
vCN 120°
Vd
t
12

4.2 Square wave operation

– Waveforms, line-to-line voltages
vAN
Vd Note:
t
vBN (1) VAN leads VBN 120º
120°
Vd VBN leads VCN 120º
t
vCN 120° VCN leads VAN 120º
Vd
t (2) V leads V 120º
AB BC
vAB
Vd VBC leads VCA 120º
t
60° Vd VCA leads VAB 120º
vBC (3) VAN lags VAB 30º
Vd
t VBN lags VBC 30º
Vd
vCA VCN lags VCA 30º
Vd
t
Vd Vd
13

7
 4.2 Square wave operation
– Switching sequence
vg1
Note:
T1 on T1 on
vg2 t Period I: T5, T6 and T1 on
T2 on
vg3 t Period II: T6, T1 and T2 on
T3 on
vg4 t Period III: T1, T2 and T3 on
T4 on
vg5 t Period IV: T2, T3 and T4 on
T5 on
vg6 t Period V: T3, T4 and T5 on
T6 on T6 on Period VI: T4, T5 and T6 on
t
I II III IV V VI I II III IV

14

4.2 Square wave operation

– Waveforms, phase voltage on load
Period I: T5, T6 and T1 on Period III: T1, T2 and T3 on Period V: T3, T4 and T5 on
A R A R B R
+ C R + B R + C R
Vd n Vd n Vd n
− B R − C R − A R
vAn = vCn = V , vBn = − V
1
3 d
2
3 d vAn = vBn = V , vCn = − V
1
3 d
2
3 d vBn = vCn = V , vAn = − 23 Vd
1
3 d

Period II: T6, T1 and T2 on Period IV: T2, T3 and T4 on Period VI: T4, T5 and T6 on
A R B R C R
+ C R + C R + A R
Vd n Vd n Vd n
− B R − A R − B R
vAn = 23 Vd , vBn = vCn = − 13Vd vBn = 23 Vd , vAn = vCn = − 13Vd vCn = 23 Vd , vAn = vBn = − 13Vd
15

8
4.2 Square wave operation
– Waveforms, phase voltage on load
vAn
2
3 dV Note:
1
V
3 d
t VAn lags VAB 30º
VBn lags VBC 30º
vBn VCn lags VCA 30º
t

vCn
t

I II III IV V VI I II III IV

16

4.2 Square wave operation

– Analysis, line-to-line voltage
• rms value
1 T 2 2
T ∫0
VLL,rms = vo dt = Vd = 0.816Vd
3
• Fourier series
∞
4Vd nπ nπ ⎛ nπ ⎞
v AB (t ) = ∑ sin sin sin ⎜ nωt + ⎟
n =1, 3, 5... nπ 2 3 ⎝ 6 ⎠
2 3Vd ⎡ ⎛ π⎞ 1 ⎛ 5π ⎞ 1 ⎛ 7π ⎞ ⎤
= sin⎜ ωt + ⎟ − sin ⎜ 5ωt + ⎟ − sin⎜ 7ωt + ⎟ + ...⎥
π ⎢⎣ ⎝ 6⎠ 5 ⎝ 6 ⎠ 7 ⎝ 6 ⎠ ⎦
• rms value of fundamental component
2 3Vd 6
VAB1,rms = 2= Vd = 0.78Vd
π π
• triple - order harmonics
1 4Vd 3π 1 4Vd 9π
VAB 3,rms = sin = 0, VAB 9,rms = sin =0
2 3π 3 2 9π 3
• Other harmonics
1 4Vd 5π
VAB 5,rms = sin = 0.156Vd
2 5π 3 17

9
4.2 Square wave operation
– Analysis, phase voltage
• rms value

1 ⎡⎛ Vd ⎞ π ⎛ 2Vd ⎞ π ⎛ Vd ⎞ π ⎤
2 2 2
1 T 2 V
V p ,rms =
T ∫
0
vo2dt = ⎢⎜ ⎟ +⎜ ⎟ +⎜ ⎟ ⎥ =
π ⎣⎢⎝ 3 ⎠ 3 ⎝ 3 ⎠ 3 ⎝ 3 ⎠ 3 ⎥⎦ 9
Vd = 0.471Vd = LL,rms
3
• Fourier series
∞
8Vd nπ nπ nπ 2V ⎛ 1 1 ⎞
v An (t ) = ∑ sin sin cos sin nωt = d ⎜ sin ωt + sin 5ωt + sin 7ωt + ... ⎟
n =1, 3, 5... 3nπ 2 3 6 π ⎝ 5 7 ⎠
• rms value of fundamental component
2Vd 2
VAn1,rms = 2= Vd = 0.45Vd
π π
• triple - order harmonics
VAn 3,rms = 0, V An 9 ,rms = 0
• Other harmonics
1 2Vd
VAn 5,rms = = 0.09Vd
2 5π
18

4.2 Square wave operation

– Analysis, harmonic equivalent circuit
V An ,n ,rms I jn ω L R
An , rms

V Bn , n , rms I Bn , rms jn ω L R
n n
VCn ,n ,rms I Cn , rms jn ω L R

Harmonic Equivalent circuit

• Fourier series of i A
∞
4Vd nπ nπ
iA = ∑ sin sin sin( nωt − θ n )
3nπ R + (nωL ) 2 3
2 2
n =1, 3,5 , 7...

nωL
where θ n = tan −1
R
19

10
 4.2 Square wave operation
– Load neutral voltage

+ T3 D3 T5 D5
Note: In practical
Vd C T1 D1
design, the neutral of g1 g3 g5 R
capacitors is not 2 iA
A

Ground
− iB
grounded because of B n
+ iC
the grounded three- C +
phase power supply. Vd g6
g4 g2 vn
2 C T D4 T6 D6 T2 D2 −
− 4

Neutral point grounded three-phase inverter

vn: voltage of load neutral respect to ground
vA: voltage of node A respect to ground
vB: voltage of node B respect to ground
vC: voltage of node C respect to ground
20

4.2 Square wave operation

– Load neutral voltage • frequency :
vg1 3 f1 , where f1 is fundamental frequency
• magnitude
t
vg4 1
peak - to - peak value : ± Vd
t 6
vA • harmonic component :
Vd
2
t Triplen - order harmonics of v A
Vd
2 2Vd ⎛ 1 1
vA = ⎜ sin ωt + sin 3ωt + sin 5ωt
vAn
2
V
3 d π ⎝ 3 5
V
1
3 d 1 1
+ sin 7ωt + sin 9ωt... ⎟
⎞
t 7 9 ⎠
2Vd ⎛ 1 1 ⎞
vn v An = ⎜ sin ωt + sin 5ωt + sin 7ωt + ... ⎟
V
1
6 d π ⎝ 5 7 ⎠
2Vd ⎛ 1 1 1 ⎞
vn = ⎜ sin 3ωt + sin 9ωt + sin 15ωt + ... ⎟
− 16 Vd π ⎝3 9 15 ⎠

21

11
4.2 Square wave operation
– Load neutral voltage
VAn ,rms I jn ωL R
An , rms

VBn ,rms I Bn ,rms jn ωL R

n
VCn ,rms I Cn ,rms jn ωL R

Harmonic Equivalent circuit

•Triple-order harmonic:
voltages are in phase. But no current path between node n and ground. Then the
voltage will be applied to the load neutral
•Other harmonics:
voltages are not in phase (120º lagging or leading). The three harmonic voltages
are three-phase balanced voltages, then load neutral voltage is zero.
22

4.2 Square wave operation

– Post analysis
• Question 1: How to change output frequency
• Answer: Adjust the period T of the gate signal
• Question 2: How to change VLL,rms ?
• Answer: Adjust dc link voltage Vd
– Feature:
• Has low order harmonic, 5th, 7th, 13th, 15th
• Voltage THD is high
– How to reduce the voltage THD? Use low pass filter (LPF)
• Suitable for high power application due to very low
switching frequency
• Usually, the multiple converters are connected with
phase shift transformer to reduce the harmonics.
23

12
 4.3 Project 3
• Objective: use microprocessor to generate IGBT gate
signals for three-phase inverter
– Variable frequency:10-70Hz
– Square wave operation (with a fixed dc voltage → ac
output voltage is fixed)
• Gate signals: three gate signals are to be generated.
Other three signals are obtained by logic inverters.
– vg1 and vg4 are complimentary
– vg3 and vg6 are complimentary
– vg5 and vg2 are complimentary

24

4.2 Square wave operation

• Gate signal generation
use port A to generate the gate signals
vg1 port A
T1 on T1 on 7 6 5 4 3 2 1 1
t
vg3
T3 on vg1 vg3 vg5
t
vg5
T5 on
t
Period
Interrupt request :
generated by OC1
101 100 110 010 011 001 101 100 110
5 4 3
25

13
4.2 Square wave operation
• Interrupt Service Routine
Task
– Determine with gating pattern (1,2,3,…6) should
be used
– Send the gate signals (101, 100, …) to port A
– Determine the time interval (“period”) for the next
interrupt
– Clear flag and return

26

4.2 Square wave operation

• How to determine the “period”
1 1 106
Period = sec = f E −cycle [# of E - clocck ] = [# of E - clocck ]
6 f1 6 f1 6 f1
For example :
106 106
f1 = 20Hz → Period = = = 8333[# of E - clocck ]
6 f1 6 × 20
where f1 is fundamental frequency of output

27

14
4.2 Square wave operation
• How to change frequency
– Change period according to frequency
– Due to the slow uP, use look-up table to calculate
division
Table FDB 16667 (10Hz)
FDB 15152 (11Hz)
FDB 13889 (12Hz)

FDB 8333 (20Hz)

– You are required to develop the full table until

70Hz
28

4.3 Pulse Width Modulation

• Purpose of using PWM
– To make inverter output voltage adjustable
– To make inverter output frequency adjustable
– To eliminate low order harmonics
• Method
– Change the pulse width according to the
modulating waveform (sinusoidal, trapezoidal, et
al)
– Carrier waveform: triangular wave

29

15
 4.3 Pulse Width Modulation
• Principle
– Half bridge inverter

T1, T2 ─IGBTs,
switching devices + T1 D1
Vd C
D1, D2 ─Freewheeling g1
diodes 2 R
− A
C ─dc filter capacitors O
(VSI) +
Vd
Vd ─dc link voltage g2
2 C T2 D2
R ─load −

30

4.3 Pulse Width Modulation

– Waveforms, gate signals
Note:
Vˆm Vˆc (1) vg1 and vg2 are complimentary
ωt
(2) vm is modulating wave (sine)
vc is carrier wave (triangular)
vg1 (3) When vm≥vc →vg1=logic “1”
ωt vm<vc →vg1=logic “0”
vg 2
(4) When vg1=logic “1”→T1
ωt on→vAO=+Vd/2
v AO When vg1=logic “0”→T2
Vd on→vAO=-Vd/2
2
ωt
Vd
2
v AO1

31

16
4.3 Pulse Width Modulation
– Terminology
• Amplitude modulation index (ratio) • Frequency modulation index (ratio)
Vˆ mf =
fc
ma = m
Vˆ c
fm
where Vˆm is peak value of vm where f m is frequency of vm
f c is frequency of vc
Vˆ is peak value of v
c c
usually m f > 9
usually 0 < ma ≤ 1

• How to find fundamental component?

use Fourier analysis
1 2π 1 2π
an =
π ∫
0
v AO cos nωtdωt and bn =
π ∫
0
v AO sin nωtdωt

• Fundamental freqency?
Fundamental frequency = f m
32

4.3 Pulse Width Modulation

– Spectrum • Given : ma = 0.8, m f = 15, f m = 60 Hz
Find spectrum of v AO
VˆAO ,h
0 .8 0.818
0.5Vd

0.22 0.22

h
1 13 15 17 27 29 31 33

Assume Vd = 100V, VAO,1 = ?

VˆAO ,1
= 0.8 → VˆAO ,1 = 0.8 × 0.5Vd = 0.8 × 0.5 × 100 = 40V
0.5Vd
Vˆ 40
VAO,1 = AO ,1 = = 28.28V
2 2 33

17
4.3 Pulse Width Modulation
– Features
• No low order harmonics: in the example, no 3rd, 5th, 7th,
9th, 11th order harmonics
• High order harmonics can be easily filtered out

f ↑→ X L ↑→ Blocking harmonic current

→ X C ↓→ Bypassing harmonic current

34

4.3 Pulse Width Modulation

– How to determine harmonic components
• Use table 8-1 on text page 207
• Note:
– mf ≥9, if mf <9, table 8-1 may not be used
– Based on half bridge inverter
h ma 0.2 0.4 0.6 0.8 1.0
1st 0.2 0.4 0.6 0.8 1.0
mf 1.242 1.15 1.006 0.818 0.601
mf ±2 0.016 0.061 0.131 0.22 0.318 VˆAO ,h
mf ±4 … … … … … 0.5Vd
2mf ±1 …. … …. … …
2mf ±3 … … … … …
2mf ±5 … … … … …

35

18
4.3 Pulse Width Modulation
– Conclusions
• The fundamental component is proportional to
amplitude modulation index (ma)
• The fundamental frequency is determine by the
frequency of modulating waveform (vm)
– Questions
• Q: How to change inverter output frequency?
• A: Change the frequency of vm (modulating frequency)
• Q: How to change inverter output voltage?
• A: Change the amplitude modulation index.

36

4.3 Pulse Width Modulation

– Example
• Given : Half bridge inverter, Vd = 300V, ma = 0.8, m f = 39, f m = 47 Hz
Find : (1) VAO1,rms = ?
(2) Dominant harmonics, m f − 2, m f , m f + 2 (rms value)

Solution :
Check m f = 39 > 9 → Table 8 - 1 can be used
(1) fundamental component at ma = 0.8
VˆAO ,1 0.8 × 0.5 × 300
= 0.8 → VAO ,1 = = 84.85V
0.5Vd 2

37

19
 4.3 Pulse Width Modulation
– Example
• Given : Half bridge inverter, Vd = 300V, ma = 0.8, m f = 39, f m = 47 Hz
Find : (1) VAO1,rms = ?
(2) Dominant harmonics, m f − 2, m f , m f + 2 (rms value)

VˆAO ,37 0.22 × 0.5 × 300

(2) m f − 2 = 37 th , = 0.22 → VAO ,37 = = 23.3
0.5Vd 2
VˆAO ,39 0.818 × 0.5 × 300
m f = 39 th , = 0.818 → VAO ,39 = = 86.7
0.5Vd 2
Vˆ 0.22 × 0.5 × 300
m f + 2 = 41st , AO ,41 = 0.22 → VAO ,41 = = 23.3
0.5Vd 2
2m f − 1 = 77 th , VAO ,77 = 33.3
2m f + 1 = 79 th , VAO ,79 = 33.3

38

4.3 Pulse Width Modulation

– Over modulation
Vˆm Vˆc Vˆm
• Define : if ma = >1
ωt Vˆc
→ over modulation
• Waveform :
vg1
if ma >> 1,
ωt v AO become a squarewave
vg 2

ωt
v AO No switching
Vd
2
ωt
Vd
2
No switching
39

20
4.3 Pulse Width Modulation
– Fundamental versus ma
VˆAO ,1
0.5Vd
4
π
1.0

Linear Over modulation Square wave

ma
1.0 3.24

40

4.3 Pulse Width Modulation

– Drawback of over modulation
• Low order harmonic components start to appear
Example : ma = 2.5, m f = 15
VˆAO ,h
0.5Vd
1 .0

h
1 3 5 7 9 11 13 15 17 19

Due to over mf
modulation

41

21
 4.4 Single-phase full bridge inverter
with PWM operation
• Circuit diagram

P
+ T1 D1 T3 D3
g1 g3

C A Load B
Vd
T4 D4 T2 D2
g4 g2
− N

42

4.4 Single-phase full bridge inverter

with PWM operation
• Waveforms, bipolar PWM
Note:

Vˆc (1) vg1 and vg4 are complimentary

Vˆm
ωt signals
vg2 and vg3 are complimentary
signals
vg1 , vg 2 (2) vm≥vc →vg1=vg2=logic “1”
ωt vm<vc →vg1=vg2=logic “0”
vg 3 , v g 4
(3) When vg1=vg2=logic “1”→T1 and
ωt T2 on→vAB=Vd
v AB When vg1=vg2=logic “0”→T3 and
Vd T4 on→vAB=-Vd
ωt
Vd
v AB1

43

22
4.4 Single-phase full bridge inverter
with PWM operation
• How to determine harmonic components
– Use table 8-1 on text page 207
– Note:
Table 8-1 is designed for half bridge inverter
Half bridge inverter: VAO,p-p=Vd
Full bridge inverter: VAB,p-p=2Vd
(Coefficients in Table 8-1) ×2 for full bridge inverter
• Given : Full bridge inverter, bipolar PWM, Vd = 300V, ma = 0.8,
m f = 39, f m = 47 Hz
Find : (1) VAB1,rms = ?
(2) Dominant harmonics, m f − 2, m f , m f + 2 (rms value)

44

4.4 Single-phase full bridge inverter

with PWM operation
• Example, bipolar PWM
Solution :
(1) Check m f = 39 > 9 → Table 8 - 1 can be used
fundamental component at ma = 0.8
VˆAB ,1 0.8 × 0.5 × 300 × 2
= 0.8 × 2 → VAB ,1 = = 169.7V
0.5Vd 2
Vˆ 0.22 × 0.5 × 300 × 2
(2) m f − 2 = 37 th , AB ,37 = 0.22 × 2 → VAB ,37 = = 46.7
0.5Vd 2
Vˆ 0.818 × 0.5 × 300 × 2
m f = 39 th , AB ,39 = 0.818 × 2 → VAB ,39 = = 173.5
0.5Vd 2
Vˆ 0.22 × 0.5 × 300 × 2
m f + 2 = 41st , AB , 41 = 0.22 × 2 → VAB , 41 = = 46.7
0.5Vd 2

45

23
4.4 Single-phase full bridge inverter
with PWM operation
• Waveforms, unipolar PWM
Note:
vm − vm (1)vm≥vc →vg1=logic “1”,vg4=logic “0”
ωt vm<vc →vg1=logic “0”,vg4=logic “1”
-vm≥vc →vg3=logic “1”,vg2=logic “0”

vg 1 -vm<vc →vg3=logic “0”,vg2=logic “1”

(2) vg1=vg3=logic “1”→T1 and T3
ωt on→vAB=0
vg 3
vg1=vg3=logic “0”→T4 and T2
ωt on→vAB=0
v AB
Vd vg1=logic “1” and vg3=logic
ωt “0”→T1 and T2 on→vAB=Vd
Vd vg1=logic “0” and vg3=logic
v AB1
“1”→T4 and T3 on→vAB=-Vd
46

4.4 Single-phase full bridge inverter

with PWM operation
• Spectrum for unipolar PWM

VˆAB ,h
0 .8
Vd

h
1 mf 2mf
2mf -1 2mf +1
Bipolar PWM
Dominant harmonics: mf , mf ±2, mf ±4
Unipolar PWM
Dominant harmonics: 2mf ±1, 2mf ±3
47

24
 4.4 Single-phase full bridge inverter
with PWM operation
• Example, unipolar PWM
• Given : Full bridge inverter, unipolar PWM, Vd = 300V, ma = 0.8,
m f = 38, f m = 47 Hz
Find : (1) VAB1,rms = ?
(2) Dominant harmonics, 2m f ± 1, 2m f ± 3 (rms value)
(3) THD (consider 4 dominant harmonics only)

Solution :
(1) Check m f = 38 > 9 → Table 8 - 1 can be used
fundamental component at ma = 0.8
VˆAB ,1 0.8 × 0.5 × 300 × 2
= 0.8 × 2 → VAB ,1 = = 169.7V
0.5Vd 2

48

4.4 Single-phase full bridge inverter

with PWM operation
• Example, unipolar PWM
(2) Four dominant harmonics, 2m f − 3 = 73rd (3431Hz), 2m f − 1 = 75th (3525Hz ),
2m f + 1 = 77 th (3619Hz), 2m f + 3 = 79th (3713Hz )
VˆAB ,73 0.139 × 0.5 × 300 × 2
2m f − 3 = 73rd , = 0.139 × 2 → V AB ,73 = = 29.5
0.5Vd 2
Vˆ 0.314 × 0.5 × 300 × 2
2m f − 1 = 75th , AB ,75 = 0.314 × 2 → VAB ,75 = = 66.6
0.5Vd 2
Vˆ 0.314 × 0.5 × 300 × 2
2m f + 1 = 77 th , AB ,77 = 0.314 × 2 → V AB ,77 = = 66.6
0.5Vd 2
Vˆ 0.139 × 0.5 × 300 × 2
2m f + 3 = 79 rd , AB ,79 = 0.139 × 2 → VAB ,79 = = 29.5
0.5Vd 2
(3) VAB ,rms = V AB
2
,1 + VAB , 73 + V AB , 75 + VAB , 77 + V AB , 79 = 169.7 + 29.5 + 66.6 + 66.6 + 29.5 = 198.5 V
2 2 2 2 2 2 2 2 2

,rms − VAB ,1
2 2
V AB 198.52 − 169.72
THD = = = 0.61
VAB ,1 169.7
49

25
4.4 Single-phase full bridge inverter
with PWM operation
• Summary
– Table 8-1 can be used for
(1) Single-phase half bridge inverter
(2) Single-phase full bridge inverter with bipolar
PWM
Note: coefficient×2, dominant harmonics: mf, mf
±2, mf ±4
(3) Single-phase full bridge inverter with unipolar
PWM
Note: coefficient×2, dominant harmonics: 2mf ±1,
2mf ±3
50

4.5 Three-phase inverter with PWM

operation
• Circuit diagram

P
+ T1 D1 T3 D3 T5 D5
g1 g3 g5 R
iA
A iB
C B n
Vd iC
C
g4 g6 g2
− N T4 D4 T6 D6 T2 D2

51

26
4.5 Three-phase inverter with PWM
operation
• Waveforms Note:
(1) Modulating waves: three-phase sine
vmA vmB vmC
waves (vmA, vmB, vmC) with
adjustable amplitude and frequency
ωt (2) Carrier wave: triangular wave, fixed
amplitude, frequency may be
v AN adjusted, depends on applications
(3) vmA≥vc →T1 on→vAN=Vd
ωt
vBN vmA<vc →T4 on→vAN=0
vg1 and vg4 are complementary
ωt (4) vmB≥vc →T3 on→vBN=Vd
vCN vmB<vc →T6 on→vBN=0
ωt vg3 and vg6 are complementary
v AB v AB ,1 (5) vmC≥vc →T5 on→vCN=Vd
ωt vmC<vc →T2 on→vCN=0
vg5 and vg2 are complementary
52

4.5 Three-phase inverter with

PWM operation
• Spectrum

No low order harmonic

53

27
4.5 Three-phase inverter with PWM
operation
– How to determine harmonic components
• Use table 8-2 on text page 228
h ma 0.2 0.4 0.6 0.8 1.0
1st 0.122 0.245 0.367 0.49 0.612
rms value!
mf ±2 0.010 0.037 0.080 0.135 0.195
mf ±4 − − − 0.005 0.011 VLL ,h
2mf ±1 0.116 0.2 0.227 0.192 0.111 Vd
2mf ±5 − − − 0.008 0.020
3mf ±2 … … … … …
3mf ±4 … … … … …

54

4.5 Three-phase inverter with PWM

operation
• Example
• Given : Three - phase PWM inverter, Vd = 240V, ma = 1,
m f = 21, f m = 100Hz
Find : (1) rms value of the fundamental line - to - line voltage
(2) rms value of 4100Hz and 4300Hz harmonic voltage
Solution :
(1) Use table 8 - 2
fundamental component at ma = 1
VLL ,1
= 0.612 → VLL ,1 = 0.612 × 240 = 146.9V
Vd
4100 V
(2) 4100Hz : h = = 41 = 2m f − 1 → LL , 41 = 0.111 → VLL , 41 = 0.111× 240 = 26.4V
100 Vd
4300 V
4300Hz : h = = 43 = 2m f + 1 → LL , 43 = 0.111 → VLL , 43 = 0.111× 240 = 26.4V
100 Vd
55

28
 4.6 PWM Signal Generation
• Analog circuit Features:
• Simple circuit and low cost
Control Comparator • Sensitive to the noise
wave
+ Gate signal for
Carrier - top device
wave
Gate signal for
bottom device
Logic inverter

Synchronous PWM: Asynchronous PWM:

The triangular waveform is The triangular waveform is NOT
synchronized with the sine wave synchronized with the sine wave
Producing “sub-harmonics”: the
frequency of the harmonic is not
integer times of fundamental
frequency.
56

4.6 PWM Signal Generation

• Digital circuit In the interrupt service routine, CPU write the
data of sine wave to the compare register
Digital comparator
Compare Buffer
register register +
Processor
-
System bus

Counter Digital
control counter

Advantages: Triangular wave is generated by hardware

• Stable, very robust to the noise automatically.
• Easy to be controlled by CPU/DSP CPU can control the frequency and the
• Reduce the load of CPU/DSP magnitude.
Disadvantages: Every time when the triangular wave reaches
the maximum or minimum value, a interrupt
• Expensive hardware
will be generated.
• High frequency counter 57

29
4.6 PWM Signal Generation
• Software
Interrupt request

Digital comparator
OC1
+
Processor
-
System bus

TCNT Advantages:
• Software, less hardware
• Stable, very robust to the noise
• Easy to be controlled by CPU/DSP
Disadvantages:
TCNT: Free running counter • Increase the calculation load of
OC1: output compare register #1 CPU/DSP
58

4.6 PWM Signal Generation

• Software • use vm' to approximate
vm (sine) vm' (stair ) vm for simplicity
vc

• Ts − sample period

Ts
Ts
• Tp = [1 + ma sin θ k ]
2
Ts − Tp
Tg T p Tg Tg =
Interrupt 2
request
θk θ k +1 θ k +2
• θ k +1 = θ k + ∆θ
∆θ = ω ⋅ ∆t
59

30
4.6 PWM Signal Generation
• Software programming
– At t1, the number in free running counter (TCNT) matches
the number of the output compare register (OC1), an
interrupt request will be made.
– Interrupt service routine
• Determine the type of interval (gap, or pulse)
• Set the output pin to logic “0” for gap interval or logic “1” for pulse
interval.
• Calculate the time interval for following T period
• Clear the interrupt flag and return
– If the calculation in the interrupt service routine can not be
completed during Tg, the program will crash. What shall we
do?
60

4.7 Selected Harmonic Elimination

Techniques (SHE)
• Purpose:
– To eliminate a number of unwanted low order harmonics
– To control the fundamental output voltage
• Waveforms

61

31
4.7 Selected Harmonic Elimination
Techniques (SHE)
• Waveforms
– Three independent switching angles: α1, α2, and α3
Given α1, α2, and α3, the other switching angles are determined due
to symmetrical
– α1, α2, and α3 are used to
•Eliminate two harmonics (usually 5th and 7th), and
•To control the fundamental output voltage
– α1, α2, and α3 should be pre-calculated (offline
calculation)
Use numerical iteration methods to solve a set of nonlinear equations.

62

4.7 Selected Harmonic Elimination

Techniques (SHE)
• Calculations results
Example:
At 50% of maximum fundamental
voltage, α1=22º, α2=36º and α3=52º.
• Microprocessor programming
–Use look-up table
fundamental α1 α2 α3

1% 29.2º 30.3º 59.5º

2% 29.6º 30.8º 59.1º
3% … … …
4% … … …
… … … …
63

32
 4.8 Delta Modulation

• Block diagram Three-phase inverter

ia* + ∆ia g1
Controller
−

g4

ia

ia* , reference current (phase A)

ia , actual inverter output current (phase A)

∆ia , error signal (phase A)

64

4.8 Delta Modulation

• Waveforms Upper band limit (UBL)

Note:
ia*
ia* (1) Assume vg1=“1”→T1
ωt on→ia↑ until t1
(2) At t1, ia reaches the UBL
Lower band limit (LBL) →vg1=“0” →vg4=“1”→T4
on→ia↓ until t2
vg 1 (3) At t2, ia reaches the LBL
→vg1=“1” →vg4=“0”→T1
ωt
t1 t2 on→ia↑
As a result, the actual current ia will be kept within the
upper and lower band limits
65

33
4.8 Delta Modulation

• Q &A
– Q: How to adjust the frequency of ia?
– A: Adjust the frequency of reference current
– Q: How to adjust the magnitude of ia?
– A: Adjust the magnitude of reference current
– Q: What if the band width is reduces?
– A: ia will follow reference current more closely. But
switching frequency↑

66

4.8 Delta Modulation

• Summary
– If reference current is sine wave, actual current is
also sine wave on which some high order
harmonics are superimposed. High order
harmonics can be filtered out easily.
– No low order harmonics.
– Inverter output current can be accurately
controlled.

67

34
4.8 Delta Modulation

• Implementation

R6
R2
R1
ia* ∆ia R5
vg 4
ia
R3 R4
vg1

R1 = R2 = R3 = R4 = R
Band Width
−
2

Band Width
2
68

Summary
• Voltage source inverter, Topologies
– Single-phase inverters
• Half-bridge inverter
• Full-bridge (H-Bridge) inverter
– Three-phase inverters
• Gating patterns
– Square wave operation
– Pulse width modulation (PWM)
– Selected harmonic elimination (SHE)
– Delta modulation
• Harmonics
– Fourier series
– Harmonic components, table 8-1 and 8-2
– Dominant harmonic component
• Waveforms
– Gate signals
– Voltage waveforms
• Signal generation
– Analog circuit
– Digital circuit
– Microprocessor
69

35