# IGBT Modules - Technologies, Driver and Application (Second Edition) - page 157

145
example, a
dt
di
C
feedback can be used from the load circuit to the gate circuit.
Particulars on this canbe found in chapter
The IGBT turn-off losses P
off,I
are ascertained or calculated as follows: By measuring
the collector current and the collector-emitter voltage with an oscilloscope and later
multiplying the curves the peak power loss can be determined. By then integrating this
power over time the turn-off loss energy E
off
is calculated. The integration bounds are
set here at 10% of U
CE
and 2% of I
C
. Finally the IGBT switching losses P
off
,I are
calculatedbymultiplyingE
off
with the switching frequency f
sw
.
Alternatively the switching losses P
off,I
may be calculated according to the datasheet
value of E
off
(I
nom
U
nom
,T
vj
) considering the application parameters. Usually the value
E
off
(I
nom
, U
nom
,T
vj
) is shown in the datasheet of the IGBT for a specific current I
nom
and
voltage U
nom
. Also a particular voltage slop
dt
du
CE
is specified which the manufacturer
defines. Matching this to the real application is done by linear adaptation. This approach
is sufficiently accurate for adeviation from thenominal values by
±
20%.
nom
DC
nom
vj
nom nom off
sw
I,on
U
U
I
)T, U, I
f1 P
π
=
Eq. 3.28
By adding the turn-on and turn-off losses the total IGBT switching lossesP
sw,I
are:
(
)
nom
DC
nom
vj
nom nom off
vj
nom nom on
sw
I,off
I,on
I,sw
U
U
I
)T, U, I
)T, U, I(E f1 P P P
+
π
= + =
Eq. 3.29
3.5.3 Gate chargeandMiller effect
Similar as a capacitor the IGBT gate shows a dependency of the charge on the voltage
which is determinedby the capacitance.
UCQ
⋅ =
Eq. 3.30
Whilst a capacitor shows a linear dependency of the chargeon the voltage, because the
capacitance is constant, this is not so with an IGBT as depicted in
Here the
gate chargeQ
G
is scaled to the final value at point E and plotted over the resulting gate
voltageU
GE
. The behaviour can then be divided into four sections which are detailed as
follows.
At the timeA the gate is in accumulationmode (chapte
. During the sectionA-B
the capacitance C
GE
is charged and the voltage U
GE
rises according to
In the real application the time t
A-B
is determined by the gate resistor (both external
10
Valid for sinusoidal current.
11
Valid for sinusoidal current.
12
The Miller effect is named after the American physicist John Milton Miller (1882 – 1962), who was the first to
describe theeffect in 1919.
1...,147,148,149,150,151,152,153,154,155,156 158,159,160,161,162,163,164,165,166,167,...548