Rectifiers Calculations
Rectifiers Calculations
Aims
To explain the operation of, and analyse the waveforms in, 3-pulse and 6-pulse diode
rectifiers.
To examine and analyse the effect of non-zero source inductance on the operation of rectifier
circuits.
1.0 Introduction
The 3-phase rectifier forms the interface between the great majority of motor drive systems
and the AC utility, and also between many other pieces of power supply equipment and the
utility, for example supplies for welding and induction heating applications. However,
increasingly strict legislation on the harmonic content of the currents that may be drawn from
the utility is forcing many manufacturers to seek rectifiers that offer higher performance than
the basic 6-pulse circuit, such as 12-pulse systems or active rectifiers. An understanding of
the operation and characteristics of 3-phase rectifiers is therefore not only important for
power electronics engineers, but also for those who specify and purchase rectifier equipment,
and those who design and operate utility systems.
In the following sections we will first look at the 3-pulse rectifier, and then show how the 6-
pulse system is derived from this simple circuit. We will use Fourier techniques to examine
the input currents. Then we will look at one of the main second-order effects in rectifier
systems, namely the influence of non-zero AC source inductance.
The 3-pulse rectifier in Figure 1 is assumed to be supplying a resistive load in series with a
filter inductor, the inductor being large enough to ensure that the load current is continuous
and ripple-free. The circuit is a half-wave rectifier, with each supply line, A, B and C, being
connected through a single diode to the top of the load. The neutral wire forms the return
path. The circuit waveforms follow directly from the assumption of a continuous and smooth
load current:
• One of the three diodes must always be in conduction to provide a path for the load
current.
• Only one diode may conduct at once, since otherwise two of the supply lines would have
the same voltage, and this only occurs momentarily as the voltages cross each other.
• The diode that conducts at any instant is determined by the supply line with the largest
positive voltage, since this voltage will forward bias the diode in that line and reverse bias
the other two diodes.
Therefore the diodes conduct in sequence for 2π/3 radians or 120°, and the rectifier output
voltage vD is the maximum value of the three line-to-neutral voltages. Since the vD waveform
repeats every 2π/3 radians, or three times each utility cycle, it is known as a 3-pulse rectifier,
and the output voltage ripple frequency is three times the utility frequency.
1 5π / 6 3 3VLN
average v D =
2π / 3 ∫ π / 6
VLN sin θ dθ =
2π
(1)
Because the circuit only draws a uni-directional current from each supply line, and also uses
the neutral wire for the return current, it is very rarely used in practice, however it is a simple
circuit to understand, and is therefore useful in helping to analyse more complex circuits.
By reversing the diodes in the 3-pulse rectifier, and also swapping over the load connection, a
slightly different 3-pulse circuit is formed in which the rectifier now operates from the
negative halves of the supply voltages, but continues to provide a positive voltage to the load,
Figure 2. The vD waveform has an identical form to that in the positive-half-cycle circuit,
except that the ripple voltage is phase shifted by π/3 radians or 60°. The average of vD will
be the same as in the positive-half-cycle circuit, eq(1). The line currents now consist of a
series of negative current pulses, each pulse width again being 2π/3 radians.
Connecting two 3-pulse circuits, one operating on the positive half cycle and the other on the
negative half cycle, in a back-to-back configuration, Figure 3, forms the 6-pulse rectifier.
Since the average values of vD1 and vD2 are the same, then, the currents in the two load
resistors will also be the same, providing that the resistors have equal values, say R. Under
these conditions, the line currents will consist of 120° pulses of +Io as DA1, DB1 and DC1
conduct, and also 120° pulses of -Io as DA2, DB2 and DC2 conduct. The negative pulses will
flow in between the positive pulses, making the line currents quasi-squarewaves.
Furthermore, since the two resistors carry equal currents, the neutral current will be zero, and
the neutral connection may be removed without affecting the circuit operation. This leaves us
with the standard 3-wire, 6-pulse rectifier, Figure 4, in which the load is simply drawn as a
constant current element, representing the highly inductive nature of the load.
Figure 5 shows the detail of the two 3-pulse output voltages vD1 and vD2 and the total output
voltage vDD= vD1+vD2. Since the ripple components in vD1 and vD2 are phase-shifted by π/3
radians, the resultant ripple in vDD is reduced in amplitude and has a repetition interval of π/3
radians, giving 6 pulses per period of the input waveform. The average of vDD is simply twice
the average value of the 3-pulse waveforms and from eq(1):
3 3V LN 3 3V LN 3V LL
average v DD = 2 = = (2)
2π π π
where VLL is the amplitude or peak value of the line-to-line supply voltage.
For the purposes of analysing the quasi-squarewave input current, iA, the waveform is
redrawn in Figure 6 with the leading edge of the waveform aligned with the origin.
1 2π
An = ∫0 v(θ )e − jnθ dθ (3)
2π
then the complex Fourier coefficients of the quasi-squarewave input current may calculated
as:
1 2π / 3 1 5π / 3
An = ∫0 I 0 e − jnθ dθ + ∫π (− I 0 )e − jnθ dθ (5)
2π 2π
2π / 3 5π / 3 ⎤
I 0 ⎡ e − jnθ e − jnθ
An = ⎢ − ⎥
2π ⎢ − jn 0 − jn π ⎥
⎣ ⎦ (6)
I
(
= 0 1 − e − jnπ − e − jn 2π / 3 + e − jn5π / 3
2 jnπ
)
simplifying
An =
I0
2 jnπ
(
1 − e − jnπ − e − jn 2π / 3 (1 − e − jnπ ) )
( )( )
I jnπ
(7)
= 0 1− e −
1 − e − jn 2π / 3
2 jnπ
2 for n odd
0 for n even
and we note that sin(nπ/3)=0 for n=3,6,9…, therefore, in addition to the fundamental, the
harmonics present in the quasi-squarewave are of order 6k±1, where k=1,2,3…
From eq(8), the fundamental of the waveform is obtained by setting n=1, the fundamental is
therefore:
2 3I 0
2 A1 ∠A1 = ∠ − π / 3 radians (9)
π
Recalling the fundamental definition for the rms value, IRMS of a current i(θ), having a period
2π:
2π
∫ 0 [i(θ )] dθ
1
I RMS = 2
(10)
2π
then the rms value of the quasi-squarewave rectifier input current in Figure 6 is:
2π / 3 2 5π / 3 2π / 3 2
I RMS =
∫0 I 0 dθ +∫
π
(− I 0 ) 2 dθ
=
∫0 I 0 dθ
= 2 / 3I 0 (11)
2π π
Active Power
Power Factor = = k d kθ (12)
Apparent Power
I 1− RMS
kd = kθ = cos φ (13)
I RMS
I1 is the fundamental component of the current, and φ is the angle of the fundamental current
component with respect to the voltage.
Looking at the waveform in Figure 3, then compared to the analysed waveform in Figure 6,
iA is delayed by π/6, therefore the total phase of the fundamental of iA in Figure 3 is –
(π/3+π/6)= -π/2. That is the expression for the fundamental is
2 3I 0 / π cos(θ − π / 2) = 2 3I 0 sin θ , and we conclude that the fundamental of IA is in
phase with vAN, giving φ=0, therefore kθ=1. Using eqs(9) and (11)
kd =
I (1)
=
(2 )
3I 0 / π / 2
=
3
(14)
I RMS 2 / 3I 0 π
In considering the basic rectifier circuits, we have assumed a perfect AC source, that is a
source from which it is possible to draw any amount of current without affecting the source
voltage. However, to provide an isolated DC output a rectifier would normally be supplied
from the AC system though a transformer. A transformer could also be used on the AC side
to change the output voltage of the rectifier, for example to produce a low DC voltage supply
from the high voltage mains.
In these circumstances the leakage inductance of the transformer will appear as an impedance
in series with the rectifier input, and this can significantly affect the rectifier operation. The
reason for this is that the ideal input currents in the circuits we have examined are stepped
waveforms, have points of infinite di/dt, and such waveforms cannot be drawn through an
inductive impedance. In addition to a transformer, the supply cables themselves will also add
inductance into the source
In the following sections we will see how the ideal operation of a 3-pulse rectifier is changed
by the presence of non-zero source inductance. The results for the 3-pulse circuit can be
easily extended to other more complex systems.
Figure 8 shows the 3-pulse circuit re-drawn with a small inductor of value LS included in
series with each of the supply lines. The three inductors represent the leakage inductance of
the supply transformer and any other inductance present in the source. Since the three line
inductors are assumed to be small, they will only affect the circuit operation in the region of
the commutation instants, that is when the constant load current transfers between any two
diodes.
To look at the detailed operation of the circuit, we will examine the commutation of current
from DA to DB, which is initiated at the instant when vAN exceeds vBN. This is taken as the t=0
origin in the expanded waveforms of Figure 9. The presence of the source inductance will
prevent an instantaneous transfer of current from DA to DB and instead the current will
transfer gradually. During this commutation transient both diodes DA and DB will be in
conduction, giving rise to the term ‘overlap’, and the equivalent circuit during the
commutation process is shown in Figure 10.
The analysis of the commutation transient assumes that the line-to-line voltage vBA drives a
current iOL around the loop shown in Figure 10. This current will be superimposed on top of
the current distribution that exists at time t=0, so for the overlap period we can write
i A = i A (0) − iOL = I 0 − iOL and i B = i B (0) + iOL = iOL .
The current iOL is zero at t=0 and gradually rises to the load current level I0, at which time iA
will be zero, DA turns off, and iB will be equal to I0. The commutation process is then
complete.
During the commutation or overlap transient the load current is constant, therefore
diOL
v BA = 2 LS (16)
dt
and the voltage across each source inductance in the direction indicated in Figure 10 is
diOL v BA
LS = (17)
dt 2
The voltage vBA/2, shown as the second waveform in Figure 9, is responsible for changing
the current in each inductor; driving iB from zero to I0, and reducing iA from I0 to 0. The
currents iA and iB are shown as the third set of waveforms in Figure 9.
During the overlap transient the rectifier output voltage, vD, is reduced by an amount equal to
the voltage dropped across LS, the output voltage is therefore
v BA (v − v AN ) (v AN + v BN )
v D = v BN − = v BN − BN = (18)
2 2 2
So immediately prior to t=0, vD is equal to vAN, during the commutation transient vD is the
instantaneous average of the A and B line-to-neutral voltages, (vAN+vBN)/2, and once the
current in DA has fallen to zero, vD becomes vBN. The form of vD is indicated in the top
waveform of Figure 9.
Analytic expressions for the instantaneous line currents during overlap may be derived from
eq(17), since during overlap iB=iOL therefore
1 t
iB =
2 LS ∫v
0
BA dt (19)
From Figure 9 we can see that with respect to the origin at t=0, v BA = VLL sin ωt , and since
i B (0) = 0 , then integrating eq(19)
V V
i B = LL (1 − cos ωt ) and i A = I 0 − i B = I 0 − LL (1 − cos ωt ) (20)
2ωLS 2ωLS
Eq(20) gives the detailed expressions for iA and iB during overlap and we can determine the
duration of the commutation transient, denoted TOL in Figure 9, by putting iA =0 or iB=I0,
giving
During the overlap transient the rectifier output voltage, vD, is reduced from the ideal value
by the voltage dropped across one of the input inductors, LS, equal to vBA/2, eq(18), and the
lost volt-seconds, shown shaded in Figure 9, will result in a reduction in the average output
voltage. Assuming that the average output voltage under ideal conditions, with no source
inductance, is VD , and that due to overlap the average output voltage falls by VOL , then the
average output voltage in the presence of overlap, VD′ , may be written as
These voltages are depicted in Figure 11, where the lower plot shows the voltage pulses that
are lost from the rectifier output, and the average value, VOL , of the pulses of lost voltage.
By integrating eq(17) over the commutation interval, we can obtain an expression for the
volt-time area of each pulse of lost voltage
I0 TOL v BA
∫ 0
LS diOL = ∫
0 2
dt = volt-time area of each lost voltage pulse (23)
Since there are three commutations per utility cycle, the total volt-time area of the lost
voltage pulses in each cycle is 3LS I 0 . Then, by dividing by the utility period, we can obtain
an expression for the reduction in average rectifier output voltage due to overlap
3LS I 0 3ωLS I 0
Reduction in average output voltage VOL = = (25)
2π / ω 2π
3VLL 3ωLS
VD′ = − I0 (26)
2π 2π
Eq(26) shows that the average rectifier output voltage may be characterised by a simple
Thevenin equivalent form in which there is a no load output voltage of 3VLL/(2π) in series
with an output resistance of 3ωLS/(2π). This is illustrated by the equivalent circuit in Figure
12. It is important to note however that there is no power dissipation or power loss associated
with the rectifier output resistance. The output resistance simply reflects the fact that volt-
seconds are lost in transferring the rectifier current, and stored energy, between the line
inductors.
Since we have seen that a 6-pulse rectifier consists of two 3-pulse circuits operating in series
across a common load, then the equations derived for overlap in the 3-pulse circuit may be
readily applied to the 6-pulse system. An overlap transient will occur as the current transfers
between each of the three top diodes and each of the three lower diodes.
Since the output voltage of a 6-pulse circuit is formed by two 3-pluse circuits, we can obtain
the average output voltage expression for a 6-pulse system by simply doubling eq(26).
Therefore the average output voltage of a 6-pulse rectifier with source inductance is
3VLL 3ωLS
VD′ = − I0 (27)
π π
Figure 7
MSc in Electrical Energy Conversion Systems 14
EEEN60192 Principles of Power Electronic Converters
AJF 21-11-08
3-Pulse Rectifier with Source Inductance
LS DA
A
LS DB
B
LS DC L
C vD O VD
A
D
Figure 8
vBA
vD 2
1 (v + v (
2 AN BN
vBN
t=o
vAN
vBA/2
iA iB
IO
t=o t=TOL
Figure 9
LS
iA
vAN
Io Io
vBA i OL
iB
vBN
LS
Figure 10
Rectifier Output Voltage Showing ‘Lost’ Volts Due to Overlap
VD
vD when Ls = 0
'
VD
vD with Ls = 0
VOL
Figure 11
MSc in Electrical Energy Conversion Systems 17
EEEN60192 Principles of Power Electronic Converters
AJF 21-11-08
Equivalent Circuit for 3-Pulse Rectifier Including Overlap
s
Re =
2π
+
3VLL
VD = '
VD
2π
Figure 12