Switch-Mode DC-AC Inverter

Four quadrants of operation

The most widely used control technique in power

electronics

Basic principles of PWM

Similar response to different shape of impulse
input

Application of the equal-area theorem

The question then becomes how to change the duty

cycle with a sinusoidal rule. The following figure
illustrates one of the methods, which is named as
sinusoidal PWM (SPWM).

A list of PWM techniques

Triangular-wave sampling
Natural sampling
Uniform sampling
Calculation
Calculation based on equal-area criterion
Selective harmonics elimination
Hysteresis band control
Space Vector Modulation (SVM, or SVPWM)
Random PWM

Some major PWM techniques

Natural sampling
Uniform sampling
Selective harmonics elimination
Some practical issues
Synchronous modulation and asynchronous
modulation
Harmonics in the PWM inverter output voltages
Ways to improve DC input voltage utilization and
reduce switching frequency
Connection of multiple PWM inverters

Pulse-Width Modulated VSI

PWM Methods
Single Pulse-width Modulation
Multiple/Uniform Pulse Width Modulation
sine PWM (SPWM)
Modified sine PWM
Harmonic Elimination Technique (SHE)
Minimum ripple current PWM
Hysteresis (Bang-bang)
Space Vector PWM

Control circuit waveforms for a square wave PWM inverter

a)Comparator input voltages.
b)Comparator output voltage and pole voltage
mf = fc/fo and p = mf /2 = pulses/half cycle
MI = Am/Ac, MI is amplitude modulation index, Ac peak carrier amplitude
and Am is peak amplitude of modulating wave
Where, mf is frequency modulation index, fc carrier freq., and fo output freq.

Voltage waveforms for a 3ph square wave PWM inverter

a),b),c) comparator input voltages;
d),e),f) pole voltages; g) pole voltages; h) line to neutral voltage.

Voltage waveforms for a 3ph square wave PWM inverter when the carrier
wave is shifted by one quarter-cycle .
a),b),c) comparator input voltages;
d),e),f) pole voltages; g) pole voltages;

Harmonic content of the

square wave PWM
voltage as a function of
the modulation index.
a)Harmonic amplitude
relative to maximum
fundamental amplitude.
b)Harmonic amplitude
relative to actual
fundamental amplitude.

In synchronized PWM the frequency of the triangle signal is an integral multiple of

that of the reference signal. Therefore, the generated PWM signal is identical in
every cycle of the reference signal. This ensures a stable voltage output where the
triangle signal has low frequencies in order to reduce the switching loss of the
power transistors.
Asynchronous PWM doesnt ensure the relation between the frequencies of both
signals. The method is simple but causes different voltage forms in different cycles.
However, if the triangle frequency is much higher than the reference frequency, this
influence is negligible.

Triangular-wave natural sampling

Uni-polar PWM in single-phase VSI

Uni-polar sampling is used to realize unipolar PWM.

Symmetric and Asymmetric PWM

In symmetric PWM, the positive (or negative) pulse of every PWM cycle
is located in the middle of the cycle period, while in the asymmetric PWM,
the pulses are usually aligned to the start or end of the PWM cycle.
Practically, asymmetric methods are relatively easier to realise, but
symmetric methods evoke fewer harmonics. Therefore, symmetric PWM
should be used when possible.

Symmetric and Asymmetric PWM

Dead time

The insertion of the dead time in every PWM cycle distorts the output voltages.
In accurate motor control, this negative effect will be compensated by
prolonging some of the pulses.

Triangular-wave uniform sampling

Easier to
realize by
computer
control
Modulation
factor

Bi-polar PWM in single-phase VSI

Bi-polar sampling is used

to realize bi-polar PWM.

In 3-phase VSI

Three-phase
bridge inverter
can only realize bipolar PWM
therefore should
be controlled
by bipolar
sampling.

Ways to improve utilization of DC input voltage

and reduce switching frequency
Use trapezoidal waveform as modulating signal instead of sinusoidal

Use 3k order harmonics bias

in the modulating signal

Reference signal of third-harmonic PWM

Sixty-degree PWM
The sixty-degree PWM is an extension of third-harmonic PWM. It
is also implemented in the same manner as SPWM. It is based
on the consideration that not only third harmonic, but also all
non-even triplen harmonics are filtered out by the delta
connected motor windings. Adding all these harmonics with the
fundamental together, a function with flat segments is obtained
as shown in the figure. The period of the flat part covers 60
signal phase
The modulation index of this method also reaches 1.

Concept of sixty-degree PWM

Connection of multiple PWM

inverters

Purpose:
Expand output power rating
Reduce harmonics

PWM techniques with feedback control

Current hysteresis control

Voltage hysteresis control
Triangular-wave comparison (sampling)
with feedback control

I. Voltage Source Inverter (VSI)

A. Six-Step VSI
Six-Step

three-phase Voltage Source Inverter

Fig. 1 Three-phase voltage source inverter.

I. Voltage Source Inverter (VSI)

A. Six-Step VSI
Gating signals, switching sequence and line to negative voltages

Fig. 2 Waveforms of gating signals, switching sequence, line to negative voltages

for six-step voltage source inverter.

Voltage Source Inverter (VSI) Six-Step VSI 1800 operation

Switching Sequence:

561 (V1) 612 (V2) 123 (V3) 234 (V4) 345 (V5) 456 (V6) 561 (V1)

where, 561 means that S5, S6 and S1 are switched on

Fig. 3 Six inverter voltage vectors for six-step voltage source inverter.

3-ph Bridge
Inverter output
voltage waves
in square wave
(or Six Step)
mode

I. Voltage Source Inverter (VSI)

A. Six-Step VSI
Line to line voltages (Vab, Vbc, Vca) and line to neutral voltages (Van, Vbn, Vcn)
Line

to line voltages

Vab = VaN - VbN

Vbc = VbN - VcN

Vca = VcN - VaN
Phase voltages
V

= 2/3VaN - 1/3VbN - 1/3VcN

= -1/3VaN + 2/3VbN - 1/3VcN

= -1/3VaN - 1/3VbN + 2/3VcN

an

bn
cn

Fig. 4 Waveforms of line to neutral (phase) voltages and line to line voltages
for six-step voltage source inverter.

I. Voltage Source Inverter (VSI)

A. Six-Step VSI
Amplitude of line to line voltages (Vab, Vbc, Vca)
Fundamental Frequency Component (Vab)1

3 4 Vdc
6
(Vab )1 (rms)

Vdc 0.78Vdc

2 2
Harmonic

Frequency Components (Vab)h

: amplitudes of harmonics decrease inversely proportional to

their harmonic order

0.78
(Vab )h (rms)
Vdc
h

where, h 6n 1 (n 1, 2, 3,.....)

Pulse-Width Modulated VSI

Objective
Control

of PWM

of inverter output voltage

Reduction

of harmonics

Disadvantages of PWM
Increase of switching losses due to high PWM frequency
Reduction
EMI

of available voltage

problems due to high-order harmonics

 Pulse-Width

Modulation (PWM)

Fig. 5 Sinusoidal Pulse-width modulation.

Pulse-Width Modulated VSI

Inverter

output voltage

When

vcontrol > vtri, VA0 = Vdc/2

When

vcontrol < vtri, VA0 = -Vdc/2

Control

of inverter output voltage

PWM frequency is the same as the frequency of vtri
Amplitude is controlled by the peak value of vcontrol
Fundamental

frequency is controlled by the frequency of

vcontrol
Modulation Index (m)

vcontrol peak of (VA0 )1

m

,
vtri
Vdc / 2
where, (VA0 )1 : fundamental frequecny component of VA0

PWM METHODS Sine PWM

Amplitude modulation ratio (ma)
peak amplitude of vcontrol
peak
ma

amplitude of vtri

value of
Vdc / 2

where, (VA0 )1 : fundamental frequecny component of

(VA0 )1

VA0

Frequency modulation ratio/index (mf)

mf
mf

fs
, where, fs PWM frequency and f1 fundamental frequency
f1

should be an odd integer

if

mf is not an integer, there may exist subhamonics at output voltage

if

mf is not odd, DC component may exist and even harmonics are

present at output voltage
mf should be a multiple of 3 for three-phase PWM inverter

An

odd multiple of 3 and even harmonics are

suppressed

PWM METHODS
Sine PWM
Three-phase inverter

Fig. 6 Three-phase Sine PWM inverter.

PWM METHODS sine PWM

Three-phase sine PWM waveforms
vtri

vcontrol_A

vcontrol_B

vcontrol_C

Frequency

of vcontrol = f1

where, fs = PWM frequency

V B0

Frequency of vtri = fs

VA0

Frequency of vtri and vcontrol

vcontrol > vtri, VA0 = Vdc/2

When vcontrol < vtri, VA0 = -Vdc/2

where, VAB = VA0 VB0
VBC = VB0 VC0
VCA = VC0 VA0

V BC

When

output voltage

V CA

Inverter

VAB

V C0

f1 = Fundamental frequency

Fig. 7 Waveforms of three-phase sine PWM inverter.

V CA

V BC

VAB

V C0

V B0

VA0

PWM METHODS Modified sine PWM

Improves short comings of sine PWM techniques, while
retaining its merits.
For Y connected load Van = Vd/2 and Van1= 2Vd/ = .636Vd
Corresponding fundamental rms line to line voltages are
Vab1 = 0.61Vd for SPWM technique and 0.78Vd for 6 step
Implementation of modified sine PWM Techniques is more
complex.
Trapezoidal PWM technique
Harmonic injection sine PWM technique

Effect of Blanking
Time

Results in nonlinearity

Effect of Blanking Time

Voltage jump when the current reverses direction

Effect of Blanking Time

Effect on the output voltage

Programmed Harmonic Elimination (SHE)

5th and 7th harmonic elimination

Angles based on the desired output

 The general fourier series

v(t )
of the wave is given as
Where,

For quarter cycle a 0

n
symmetry

(an cos nt bn sin nt )

n 1

1 2
an v(t ) cos nt dt
0
1 2
bn v(t ) sin nt dt
0
and

4 2
bn 0 v(t ) sin nt dt

Assuming that the wave has unit amplitude i.e. v(t) = 1,

bn can be expanded as

bn [ 0 1 (1) sin nt dt 2 (1) sin nt dt 3 (1) sin nt dt

1
2

/2
k 1
. ... (1) sin nt dt ( 1) sin nt dt ]
k

k 1

Using the relation

1st and the last terms
are

1
sin nt dt n (cos n1 cos n 2 )
1
1
0 (1) sin nt dt n (1 cos n1 )
1
/2
k (1) sin nt dt n cos n k
2
1

Integrating the other components of the above Eqn and substituting

4
[1 2 ( cos n1 cos n 2 ....... cos n k )]
n
k
4

(1 2 (1) k cos n k
n
k 1

bn

Above eqn has k variables and needs k simultaneous eqns to solve their
values

Example: Say, we need to eliminate 5th & 7th harmonic, while

fundamental magnitude is to be controlled.
For a star connected load with isolated neutral, triplen harmonics
can be ignored. Here we need to solve three simultaneous eqns:
Fundamental:
5th Harmonic:
7th Harmonic:

4
b1
(1 2 cos 1 2 cos 2 2 cos 3 )

4
b5
(1 2 cos 51 2 cos 5 2 2 cos 5 3 ) 0
5
4
b7
(1 2 cos 71 2 cos 7 2 2 cos 7 3 ) 0
7

Above non linear transcendental eqns can be solved

numerically and 1, 2, 3 can be determined.

Programmed
Harmonic
Elimination
Method

Angles based on
the desired output

 angle Table for Vs from 93 to 100%

Vs

93
94
95
96
97
98
99
100
50

0
0
0
0
0
0
0
0
20.9

15.94
16.17
16.41
16.88
17.34
11.02
4.69
0
35.8

22.03
21.56
20.86
20.39
19.92
13.59
7.27
0
51.2

SHE typical waveform at 98% output

Minimum Ripple Current PWM

The harmonic loss in a m/c is
dictated by the rms ripple
current, therefore, rms ripple
current should be minimized
instead of individual
harmonics. The rms ripple
current in a m/c is given by :

2
I ripple I 52 I 72 I11
...

I52

I72

2
I11

. ..

Vn 2
1
(
)

2 n 5,7,11 nL

Where, I5, I7 .= rms harmonic currents

L is the effective leakage inductance of the machine per phase

I 5 , I 7 , ...... peak value of harmonic currents

n = order of harmonics and = fundamental frequency

Hysteresis (Bang-bang) PWM

Three-phase inverter for hysteresis Current Control

Three-phase inverter for hysteresis current control.

Hysteresis (Bang-bang) PWM

Hysteresis Current Controller

Hysteresis current controller at Phase a.

Tolerance-Band Current Control

Results in a variable frequency operation

Current
control
Block
Diagram

B. Hysteresis (Bang-bang) PWM

Characteristics of hysteresis Current Control

Advantages
Excellent
Low

dynamic response

cost and easy implementation

Drawbacks
Large

current ripple in steady-state

Variation of switching frequency
No intercommunication between each hysteresis controller
of three phases and hence no strategy to generate zerovoltage vectors. As a result, the switching frequency
increases at lower modulation index and the signal will leave
the hysteresis band whenever the zero vector is turned on.

The

modulation process generates sub-harmonic components

Space Vector Modulation

PWM can be generated by analogue or digital control
techniques.
The advantages of digital control over analogue are:
Stability (no drift, offsets or aging effects)
Precision (noise immunity)
Flexibility (can be customized by changing software)
Even if done digitally, significant computing time is required,
as the PWM signals have to be calculated in real time. By
using Space Vector Modulation this calculation process is
simplified. As it is simplified, less computing time is required,
and therefore better performance can be obtained.

Space Vector PWM

Output voltages of three-phase inverter

where, upper transistors: S1, S3, S5

lower transistors: S4, S6, S2
switching variable vector: a, b, c
Three-phase power inverter.

3 Phase Voltage Source Inverter (VSI)

Comparison between SPWM and SVM

Sinusoidal Pulse Width
Modulation (SPWM)

Space Vector Modulation

(SVM)

Comparing high frequency

triangular carrier signal with 3
sinusoidal reference signals
(treated as separate identity)

Taking all 3 modulating signals

into account simultaneously within
a 2D reference frame (in d-q axis
or complex form)

Available DC supply voltage not

fully utilized

Increased utilization of DC supply

voltage, 15% more than SPWM

More Total Harmonic Distortion

Less Total Harmonic Distortion

Does not facilitate advanced

vector control implementation

Enables advanced vector control

implementation

3-ph Bridge
Inverter output
voltage waves
in square wave
(or Six Step)
mode

I. Voltage Source Inverter (VSI)

A. Six-Step VSI
Line to line voltages (Vab, Vbc, Vca) and line to neutral voltages (Van, Vbn, Vcn)
Line

to line voltages

Vab = VaN - VbN

Vbc = VbN - VcN

Vca = VcN - VaN
Phase voltages
V

= 2/3VaN - 1/3VbN - 1/3VcN

= -1/3VaN + 2/3VbN - 1/3VcN

= -1/3VaN - 1/3VbN + 2/3VcN

an

bn
cn

Fig. 4 Waveforms of line to neutral (phase) voltages and line to line voltages
for six-step voltage source inverter.

Space Vector PWM

Output voltages of three-phase inverter

S1 through S6 are the six power transistors that shape the output voltage

When an upper switch is turned on (i.e., a, b or c is 1), the

corresponding lower switch is turned off (i.e., a', b' or c' is 0)

Eight

possible combinations of on and off patterns for the three upper

transistors (S1, S3, S5)

Line to line voltage vector [Vab Vbc Vca]t

Vab
1 1 0
V V 0 1 1
dc
bc

Vca
1 0 1

a
b ,

c

where switching variable vector [a

Line to neutral (phase) voltage vector [Van Vbn Vcn]t

Van
2 1 1 a
1

1
2

1
b
dc
bn

3
Vcn
1 1 2 c

c]t

Space Vector PWM

Output voltages of three-phase inverter
The

eight inverter voltage vectors (V0 to V7)

Space Vector PWM

Output voltages of three-phase inverter
The

eight combinations, phase voltages and output line to line voltages

Space Vector PWM

Basic switching vectors and Sectors
6 active vectors (V1,V2, V3, V4, V5, V6)

Axes of a hexagonal
DC link voltage is supplied to
the load
Each sector (1 to 6): 60 degrees

2 zero vectors (V0, V7)

At origin
No voltage is supplied to the
load

Basic switching vectors and sectors.

Space Vector PWM

Comparison of Sine PWM and Space Vector PWM (2)
Space Vector PWM generates less harmonic distortion

in the output voltage or currents in comparison with sine PWM

Space

Vector PWM provides more efficient use of supply

voltage in comparison with sine PWM

Sine

PWM :
Locus of the reference vector is the inside of a circle with
radius of 1/2 Vdc

Space Vector PWM

: Locus of the reference vector is the inside of a circle with
radius of 1/3 Vdc
Voltage Utilization: Space Vector PWM = 2/3 times of Sine PWM

PWM METHODS Space Vector PWM

Comparison of Sine PWM and Space Vector PWM

Locus comparison of maximum linear control voltage

in Sine PWM and SV PWM.

Space Vector PWM

Realization of Space Vector PWM

Step 1. Determine Vd, Vq, Vref, and angle ()

Step 2. Determine time duration T1, T2, T0

Step 3. Determine the switching time of each transistor (S 1 to S6)

Graphs of 3 modulating voltages where reference voltage is shifted

from one sector to another

Space Vector PWM

Step 1. Determine Vd, Vq, Vref, and angle ()
Coordinate transformation: abc to dq

Vd Van Vbn cos60 Vcn cos60

Van

1
1
Vbn Vcn
2
2

Vq 0 Vbn cos30 Vcn cos30

Van

3
3
Vbn
Vcn
2
2

1
1 V
an
1

Vd 2
2
2
Vbn

3
3
Vq 3 0
Vcn

2
2
2

Voltage Space Vector and its

components in (d, q).

V ref Vd Vq
1

tan (

Vq
Vd

) s t 2f st

(where, f s fundamental frequency)

Space Vector PWM

Step 2. Determine time duration T1, T2, T0

Reference vector as a combination of adjacent vectors at sector 1.

Space Vector PWM

Step 2. Determine time duration T1, T2, T0
Switching time duration at Sector 1
Tz

V
0

T1

ref

V1dt
0

T1 T2

Tz

V dt V
2

T1

T1 T2

Tz V ref (T1 V1 T2 V 2 )
Tz V ref

cos ( )
1
cos ( / 3)
2
2

T1 Vdc T2 Vdc

3
3
sin ( )
0
sin ( / 3)

(where, 0 60)
sin ( / 3 )
sin ( / 3)
sin ( )
T2 Tz a
sin ( / 3)
T1 Tz a

1
T0 Tz (T1 T2 ), where, Tz

fs

and a

V ref

2
Vdc
3

Space Vector PWM

Step 2.

Determine
time duration
T1, T2, T0

Switching time
duration at any
Sector

T1

3 Tz V ref
n 1

sin

Vdc
3
3

3 Tz V ref
n

sin
Vdc
3

3 Tz V ref
n
n

sin cos cos sin

Vdc
3
3

T2

3 Tz V ref
n 1

sin

Vdc
3

3 Tz V ref
n 1
n 1

cos

sin

sin

cos

Vdc
3
3

where, n 1 through 6 (that is, Sector1 to 6)

T0 Tz T1 T2 ,
0 60

Algorithms of SVM

Algorithms of output signals based on sector 1

Output signal based on Symmetrical Sequence algorithm in sector 1

Space Vector PWM

based on symmetrical sequence
Step 3. Determine the switching time of each transistor (S 1 to S6) (1)

(a) Sector 1.

(b) Sector 2.

Space Vector PWM switching patterns at each sector.

Space Vector PWM

Step 3. Determine the switching time of each transistor (S 1 to S6) (2)

(c) Sector 3.

(d) Sector 4.

Space Vector PWM switching patterns at each sector.

Space Vector PWM

Step 3. Determine the switching time of each transistor (S1 to S6) (3)

(e) Sector 5.

(f) Sector 6.

Space Vector PWM switching patterns at each sector.

Space Vector PWM

Step 3. Determine the switching time of each transistor (S 1 to S6)
Table 1. Switching Time Table at Each Sector

 COMPARISON OF SPWM-SVM-SHE-HB PWM TECHNIQUES

SHE
NO CARRIER BASE - Number Of Notches Determine Switching
Frequency
DIFFICULT TO APPLY AT LOW FREQUENCY
OUTPUT MAY NOT BE HARMONICALLY OPTIMUM
MOST USEFUL WHEN SPECIFIC ORDER OF HARMONICS ARE
HARMFUL
EASY LINEARIZATION IN WHOLE MODULATION RANGE
DC LINK VOLTAGE RIPPLE INTRODUCES ADDITIONAL OUTPUT
RIPPLE
MICROCOMPUTER/DSP BASED LOOK-UP TABLE IMPLEMENTATION

HB
EASY OPERATION FROM ZERO FREQUENCY
NEEDS CLOSE LOOP CURRENT CONTROL
DC LINK RIPPLE IS COMPENSATED PERMITS LOWER CF IN DC LINK
FAST TRANSIENT RESPONSE
SWITCHING FREQUENCY VARIES
SMOOTH TRANSITION FROM UNDERMODULATION TO
OVERMODULATION

III. REFERENCES
[1] N. Mohan, W. P. Robbin, and T. Undeland, Power Electronics: Converters,
Applications, and Design, 2nd ed. New York: Wiley, 1995.
[2] B. K. Bose, Power Electronics and Variable Frequency Drives:Technology
and Applications. IEEE Press, 1997.
[3] H.W. van der Broeck, H.-C. Skudelny, and G.V. Stanke, Analysis and
realization of a pulsewidth modulator based on voltage space vectors,
IEEE Transactions on Industry Applications, vol.24, pp. 142-150, 1988.

35