Poweer Electroniccs

Prof. B.
B G. Fernan
ndes
Dep
partment of Electrical E
Engineeringg

India
an Institute of Technoloogy, Bombaay
Leccture No - 9

Hello, in
n my last class we discussed the
t
operatioon of pow
wer MOSFE
ET, metal ooxide
semicond
ductor field effect transistor. It is a very
v
fast devvice, much ffaster than a BJT. There are 3
terminalss similar to our
o BJT. Th
hey are drain
n, gate and soource. Gate is insulatedd from the soource.
There is a SiO2 layer which is an
a insulator. So, the inpuut impedancce or the imppedance bettween
gate and source is veery high and it is mainly
y a capacitor . So, how dooes a currentt flow from drain
to sourcee? When thee applied vo
oltage betweeen gate to source is hiigher than w
what is know
wn as
threshold
d value, N ch
hannel is form
med and wh
hich connectss the drain too source.
(Refer Sllide Time: 2:15)

See in this figure, wh

hen the appliied voltage is
i higher thaan the threshold value, N
N, a small channel
ucting channel is formed
d below the silicon
s
layer . So, electroons starts flow
wing from thhis N
is, condu
plus layeer to N minuss layer in thiis fashion orr the current starts from ddrain to sourrce.

(Refer Sllide Time: 2:44)

So, in the conduction

n mode, thee on state vo
oltage drop aacross the M
MOS is slighhtly higher thhan a
BJT. So, in the lineaar zone, in th
he sense, in the ohmic zzone, I can rrepresent M
MOS as a resistive
element. Just see thee characteristtics, they aree almost lineear. So, for the same coollector curreent or
urrent, see th
he characteriistics of a BJT and a MO
OS. See, I hhave plotted these
for the saame drain cu
characterristics for VDS
or
V
co
ollector
curr
rent
or
drain
current,
I
k
kept
IB constaant. Here, gaate to
D
CE,
source vo
oltage is held
d constant. See
S the charaacteristics, thhis is for a M
MOS, this is for BJT.
So, the eq
quivalent ressistance in th
his case is much
m
smallerr compared tto a MOS. S
So, on state ppower
loss in a MOS is sliightly higheer than a BJT
T and this iis one of thee drawbackss of a MOS
SFET.
Another point that I discussed was
w that the capacitance
c
between gatte and sourcce remains allmost
constant whereas, thee capacitancce between gate
g and draiin varies with
th VDS. So, thhis capacitannce is
w. In other words,
w
when MOSFET i s on, at that time VDS is low, capaciitance
high wheen VDS is low
between the gate and
d the drain iss high and th
his capacitannce decreasees as the volltage betweeen the
drain and
d source incrreases.
A BJT iss a minority carrier devicce, whereas, a MOSFET
T is a majoritty carrier deevice. The cuurrent
flow in the
t device is
i only due to the electtrons. It hass a positive temperaturee coefficiennt and
thereforee paralleling is easy. About the safe operating aarea in a BJT
T, there are 4 zones, whhereas
in a MOS
SFET there are
a only 3 zo
ones.

(Refer Sllide Time: 5:37)

See heree, ID limit, maximum

m
power
p
dissip
pation limit and CD is the VDS lim
mit. There is no
secondarry break dow
wn limit as in
n the case off BJT.
(Refer Sllide Time: 5:55)

Now, lett us see the variation of VGS, is a function of charge carried to the gate. O
OA is
correspon
nding to thee charging of the input capacitance
c
Ci under fulll VDS. See, VGS is veryy low,
less than
n the thresho
old value theen there is no
o flow of cuurrent. Whenn there is noo flow of cuurrent,
VDS is veery high. So, OA corresponding to charging
c
of input capacitance Ci unnder full VDSS. VDS
is high, th
herefore, gatte to drain caapacitance iss very low.

So, Ci, th
he input capaacitance whiich is equal to
t CGS plus CGD, now CGGD is very low
w, therefore,, Ci is
approxim
mately equal to CGS, a cap
pacitance beetween gate aand source.
(Refer Sllide Time: 7:34)

Now, AB
B, What is AB?
A AB is it correspondss to VDS decrrease from thhe supply vooltage to VDSS (ON).
Once, thee applied vo
oltage betweeen the gate and the souurce is higheer than the thhreshold, floow of
electronss starts. So, current staarts flowing
g from drainn to source, VDS falls. So, that reegion
correspon
nds to AB, see AB. VGS,
voltage
between
b
gatte and sourcce remains constant and the
G
charge su
upplied is used to vary the voltage across CGD , gate to draain. Now, I told you, seee the
voltage is falling from the full voltage, full rated
r
voltage to DS on. So, as the vvoltage falls from
the full voltage
v
to thee on state vo
oltage, CGD goes
g
on increeasing.
s
is used
u
to chargge the capaccitor and thiss BC correspponds
So, whattever the chaarge that is supplied
to the inp
put capacitan
nce charge when
w
the dev
vice is on. N
Now, what is the input ccapacitance w
when
the devicce is on? It is
i CGS plus CGD (ON). Du
uring on, CGGD is high. S
So, this is coorrespondingg to a
device on
n

(Refer Sllide Time: 9:37)

Similarly
y, the chang
ge in voltagee with the ch
harge or durring turn offf, E to F foor the removval of
excess ch
harge, F to H is corressponding to discharge oof CGD. Youu are turningg off the deevice.
Voltage across the MOS is increasing. In other w
words, VDS is increasiing. So, thhis is
nding to thee discharge of
o CGD and HJ is corressponding to the dischargge during cuurrent
correspon
fall, CGS discharge. See,
S these wave forms we
w require w
while designinng or while studying thee turn
on and tu
urn off of MOS.
M
So, I will just show
w you how poower MOS IIRF640 is m
mounted on a heat
sink.
(Refer Sllide Time: 11:11)

See, the parameters. This data I got it from their site. ID at 25 degreees, continuuous drain cuurrent
VGS is 10 volts, max
ximum is 18
8 amperes, gate to sourrce is plus oor minus 200 volts. See here,
RDS(ON) iss 0.18 ohm, it is mention
ned here very
y clearly.
(Refer Sllide Time: 11:42)

o source threeshold voltaage, see, minnimum is 2 vvolts,

The otheer voltage orr the parameeters, gate to
maximum
m is 4 volts. Gate to sourrce charge QGS, 13 nanoo coulombs, tturn on delaay, that is 14 nano
seconds, rise time 51
1 nano secon
nds, turn offf delay 45 naano seconds,, fall time 366 nano seconnds. I
will expllain to you what exactly
y this mean
ns. They are almost sim
milar to BJT. See, just nnow I
showed you
y the varriation of VGS
with
Q,
harge.
See
total
gate
ch
the variationn and this iis the
G
variation
n of the capaccitance.

(Refer Sllide Time: 13

3:18)

Now, thee switching characteristiics, if you know

k
how exxactly VGS vvaries with Q
Q, understannding
the switcching characcteristics beccomes very simple. Seee, the variattion of VGS is shown heere. I
have not shown you
u the variatio
on of drain current
c
becaause it depennds on the eexternal circcuit. I
s
state current here.
have justt shown the steady
Gate to source
s
voltag
ge, it starts from
f
zero, iss increasing.. Till it is the threshold vvalue, there is no
conductio
on, VDS is sttill high. All the charge or
o all the currrent that is ssupplied is uused to chargge the
gate to so
ource capaciitance. This region, whaatever the chharge that is supplied, it is used to chharge
the capaccitor between
n gate to draain.
VGS remains constan
nt here and VDS falls to
o a very low
w value andd when VDS is very low
w, the
capacitan
nce between
n gate and drain is very high and thhis is the saame, the variation is sam
me as
what we seen before. See, after sometime
s
deevice is turnned off. Sam
me, there is a delay time, after
sometimee current staarts falling an
nd this is the fall time, t f. Again I have not show
wn the rise iin the
drain voltage becausse it depend
ds on the ex
xternal circuuit. So, thiss wave form
m is same aas the
previous one. See herre, on and th
his is off.
hat are the major
m
differen
nces between
n a BJT and a MOSFET
T? So, let mee summarize. BJT
Now, wh
is a curreent controlleed device, MOSFET
M
is a voltage conntrolled deviice. BJT is a minority carrier
device, th
herefore, it has
h a negativ
ve resistancee coefficient,, therefore, pparalleling iss difficult.
What abo
out MOS? It
I is a majorrity carrier device,
d
has a positive reesistant coeff
fficient, thereefore,
parallelin
ng is easy. BJT has a secondary breakdown
b
bbecause it iis a minoritty carrier deevice,
negative resistance coefficient.
c
MOSFET
M
is a majority ccarrier devicce. So, theree is no seconndary
breakdow
wn. On state power loss in a BJT is low that is Vc sat into Ic is low. Pow
wer MOS, itt is ID
squared into RDS (ONN) is higher than the on
n state lossees of a BJT. Turn off ttime of a B
BJT is
generally
y higher becaause of the storage
s
time. MOSFET iis a very fastt device.

(Refer Sllide Time: 17

7:25)

So see, here
h are the major
m
differeences between a BJT annd a MOSFE
ET that I havve summarizzed. It
looks like, exact for on state volttage drop orr on state po wer loss, M
MOSFET has an edge oveer the
BJT.
7:41)
(Refer Sllide Time: 17

So, BJT has an exceellent on statte characteriistics. The pproblem withh that is a ccurrent contrrolled
device. So,
S if you compare
c
BJT
T and a MO
OSFET, theyy almost coomplement eeach other. Input
power requirement fo
or a MOSFE
ET is very lo
ow, BJT has higher. It is a current coontrolled devvice. I
a
PT, ppower transiistor.
told you that a small PT is requirred to drive another

Now, can
n we combine this both
h the devicess, the advanntages of botth the devicces and com
me out
with the new
n device, what is kno
own as IGBT
T. See here, iinsulated gatte bipolar traansistor, IGB
BT. It
was deveeloped in 1983 by Jayant Baliga, verry popular ddevice in pow
wer electroniics, very poppular.
Prior to the
t advent off IGBT, BJT
T was used. It
I was very ppopular, mayy be till the 11990s or so, BJT
is very popular. As IGBT started
d coming to the market, the populariity of the BJJT started cooming
down. So
o, what is th
his IGBT? In
nsulated gatee: almost sam
me as the M
MOS. There is a SiO2 layyer, bi
polar tran
nsistor, output is a BJT.
(Refer Sllide Time: 20
0:07)

So, see this

t
structuree, almost thee same as th
he MOS herre, almost thhe same. Yoou just addedd a P
layer, a P plus layer and that forrms the drain
n. So this laayer, above tthis same ass MOS, you have
just addeed p plus lay
yer, that form
ms the drain
n. Operationn at the inpuut stage is thhe same wheen the
applied voltage
v
betw
ween the gatte and sourcce is higherr than the thhreshold value, N channnel is
formed, N channel iss formed here.
So, electrrons starts fllowing form
m N plus to N minus regiion. When thhe positive vvoltage is appplied
to the drrain and the source betw
ween drain and the souurce, this junnction is forrward biasedd and
thereforee, holes are injected
i
into
o N minus zone.
z
See, thhis was not tthere in the MOS. MOS
S, the
flow of current
c
only due to flow of electron. Whereas, heere you havee here input stage similaar to a
MOS, flo
ow of electro
ons from N plus
p to N min
nus. Similarrly, holes froom P plus to N minus.

(Refer Sllide Time: 21:59)

So, whatt happens? Some

S
electro
ons recombin
ned with hooles and rem
maining holes are collectted at
the sourcces. Now, having
h
this P plus layeer, what havve we achieeved? Now, this junctioon in
principall can block a negative vo
oltage. I willl explain it sometime laater, what exxactly we neeed to
do for IG
GBT to block
k a negative voltage.
Now, theere is a P plu
us and N plu
us. So, in prrincipal, it caan block a nnegative voltage. What iis the
magnitud
de of negativ
ve voltage it can block? I will tell yoou some tim
me later. See,, you can obbserve
one moree thing, see, there was an
n anti paralleel body diodde in a MOS
S, this is this diode. Now
w, that
diode is absent
a
in IGB
BT. But then
n almost all the IGBT, thhey do have an inverse ddiode.
Now, thiis inverse diiode is built in which iss optimized tto match thee IGBT swiitching operaation,
almost siimilar to a MOS.
M
MOS had a body
y diode but tthen we disaabled it and separate a ddiode
which waas optimize to match thee MOS charaacteristics, thhe switchingg characterisstics was buiild in,
whereas, IGBT does not have thee body diodee. In the lateer half of the course, we will find thaat this
diode is required.
r
ost of the IGB
BTs, they do
d have an innverse diodee and this is bbuilt in, wheereas,
So in thee market, mo
in a MOS
S, it was inherently preseent. So, that is the differrence.

10

(Refer Sllide Time: 24

4:25)

Now, com
ming to the VI characteristics, how do they loo k? They loook similar too BJT. Only thing
is, contro
ol parameterr is gate to source
s
voltag
ge, almost ssimilar to a B
BJT, a steepp rise. So, dduring
conductio
on RDS is low
wer, RDS (ON)) is lower, I told
t
you. Seee here, I justt redrawn thhem, drain cuurrent
verses VDS
BT.
D for a MOS, for an IGB
(Refer Sllide Time: 25
5:06)

for IGBT
T is less com
mpared to MOS.
M
What iis the reasonn? The reasoon for havingg low
So, RDS (ON)
(
RDS (ON) in the case of IGBT, I will tell yo
ou some tim
me later. Onne more thinng that had to be
d from the strructure, see here,
h
we had
d a drain whhich is P pluss and N plus and N minuus.
observed

11

(Refer Sllide Time: 25

5:50)

Do we need to have these 2 N laayers? Whatt if there is nno N plus laayer? What w
will happen?? The
operation
n of IGBT will
w not get effected
e
becaause there is N minus layyer always tthere. But thhen, it
will mak
ke a significaant differencce in reversee voltage bloockage capaability of thee IGBT. If P plus
and N pllus are preseent, like, wh
hat I showed
d you in thee beginning, both are heeavily dopedd and
when it is reverse biaased or in otther words, when
w
I am cconnecting nnegative term
minal to drainn, the
positive to
t source, J1 should block the negativ
ve voltage.
Now sincce, both of th
hem are heaavily doped, negative bloocking voltaage capabilityy of this sorrt of a
IGBT is low. The IG
GBT which has
h P plus N plus N minnus layer is aalso known aas punch thrrough
IGBT.

12

(Refer Sllide Time: 27

7:45)

See here, N plus lay

yer between P plus and N minus driift layer is nnot essentiall for operation of
IGBT. Iff both are preesent, in fact, if N plus layer
l
and N minus layerr, both of theem are theree, it is
known as the punch through IGB
BT. So therrefore, the puunch througgh IGBT cannnot block a very
mall negativee voltage beccause both P plus and N plus
high negative voltagee. It can blocck a very sm
w
form J1 junction, itt has a very low reverse voltage bloccking capability.
which aree adjacent, which
So, punch
h through IG
GBT has a lo
ow negative voltage bloccking capabiility. So therrefore, this iss also
known as
a non symm
metrical IGB
BT, positivee voltage bloocking capaability whichh is not equual to
negative voltage bloccking capabiility. Whereaas, a non punnch throughh IGBT, see tthe structuree, it is
P plus an
nd N minus,, non punch through, P plus and N minus struccture. Now, this junctionn can
block a high
h
negativ
ve voltage beecause we have
h
a lightlly doped N jjunction or N layer. Soo, non
punch th
hrough IGB
BT can blocck a high negative
n
vooltage. Therrefore, it is also know
wn as
symmetrical IGBT.

13

(Refer Sllide Time: 29

9:38)

So, the sttructure of punch

p
throug
gh, PT stand
ds for punch through, loooks like this.. See here, P plus
N plus N minus and again, now if I have onlly P plus, thiis is known as non puncch through IG
GBT.
This can
n block the negative vo
oltage. This junction caan block neggative voltaage, whereass this
junction block can block
b
a verry small neg
gative voltagge. Hence, this is asym
mmetrical orr non
symmetrical, this is symmetrical IGBT. Theere are otheer differencees also in puunch throughh and
non puncch through IG
GBTs. I willl tell you so
ome time lateer.
Now, com
ming to the turn
t
on and turn
t
off of IG
GBT, the turrn on of IGB
BT is approxximately the same
as that off a power MOS.
M
The ratte at which you
y charge tthe input cappacitance or the rate at w
which
you charrge the gate to source caapacitance, determine
d
thhe turn on tim
me. How abbout the turnn off?
MOS, th
here is only the majority
y carriers, th
here are no minority caarriers. How about in IG
GBT?
There waas the MOS
S operation and
a there is or the flow
w of electronns from N pplus to N m
minus.
There is also a flow of holes fro
om P plus to
o N minus. So, there arre majority ccarriers, therre are
minority carriers also. So, in thee turn off of the IGBT,, there are 2 distinct inttervals. Whaat are
they? I will
w show you
u.

14

(Refer Sllide Time: 31:55)

See this figure, the voltage

v
variaation of gate to source, ssame as thatt of a power MOS. The drain
current starts
s
falling very rapidlly because at
a this point the channell that is form
med betweeen the
drain and
d the source has disappeaared.
So, MOS
SFET blockss quickly. I told
t
you MO
OSFET is a vvery fast devvice, the mooment the vooltage
between the gate an
nd the source is less thaan the thres hold, the chhannel disapppears. MOS
SFET
blocks quickly,
q
ID drops.
d
This current,
c
MO
OSFET curreent has droppped but then there aree still
minority carriers in N minus lay
yer. There arre minority ccarriers in N minus layeer. They willl take
w take a lon
nger time to recombine.
or they will
So thereffore, ID, the drain curren
nt now fallss relatively sslowly. See the slope, tthere was a rapid
flow of drain
d
current. This is th
he MOSFET current andd this is the BJT currentt. Because oof the
minority carriers in N minus lay
yer and this current fallss gradually w
what is som
mething know
wn as
the tail current, almo
ost all the miinority carriers. Say, BJJT that was tthe tail curreent, GTO is a tail
current, even
e
IGBT as
a a tail current. Why? because
b
of thhe minority ccarriers, the holes in N m
minus
layer.
w see how
w to reduce the tail currrent, somettime later. S
So, I am not shown you the
So, we will
voltage variation,
v
VDS.
ncreasing somewhere at this point. S
So, during taail current peeriod,
D It starts in
even in the GTO or a BJT also, there
t
was a high turn off power loss. Because, thhe voltage aacross
drain to source
s
as cerrtain a reason
nably a high
h value.
o losses wiill be high. How
H do you reduce this turn off lossses or how ddo you reducce the
So, turn off
tail curreent period orr how do you
u make this tail current tto fall rapidlly? What didd we do in G
GTO?
We used anode shortt structure, th
hat N plus laayer in P layyer which forrm the anodde. By havingg that
not block a negative voltage but thhen with N pplus layer, iit accelerateed the
structure, GTO cann
recombin
nation of the charges in N minus layeer. This I tolld you whilee discussing a GTO.

15

So, if I compare
c
a no
on punch thrrough IGBT
T and a puncch through IG
GBT, a puncch through IIGBT
has N pllus layer. So
o, compared
d to a non punch
p
througgh IGBT, a punch throough IGBT hhas a
smaller taail current, almost
a
similaar to a GTO there.
6:08)
(Refer Sllide Time: 36

unch throughh. So, this IG

GBT has a sm
maller tail cuurrent
See here,, we have N plus and N minus in pu
time. No
ow, it may not
n block a negative
n
volltage. So, thhis is the priice that you pay. Now, what
about saffe operating area for IGB
BT?
6:50)
(Refer Sllide Time: 36

16

It is the same
s
as thatt of MOSFE
ET. There arre only 3 zonnes: ID, TJ mmax and CD llimit. There is no
secondarry break dow
wn zone as in
i the case of
o IGBT. Soo, IGBT is a device which has almoost all
the good qualities off BJT as welll as a MOS, except the iincrease in tuurn of time. Now, becauuse of
the minority carriers,, turn off is not
n as fast ass a MOS. Soo, that is abouut the IGBT.
(Refer Sllide Time: 00
0:37:51)

I will jusst show you a module, a 75 amperee, 1200 voltts dual IGBT
T. Dual, I m
mean, there are 2
IGBTs and
a this is a diode conn
nected acrosss it. The sw
witching timee of this diode is compaarable
with the IGBT itself.. This is not a body diod
de, unlike in a MOS, rem
member. So, these 3 term
minals
are correspond to 1 2 3 and thesee 2, 1 and 2,, 1 and 2, thee control terrminals, gatee and source, gate
and sourcce. A common heat sink
k can mount on,
o a 75 amppere, 1200 vvolts IGBT.
BT does requ
uire a driverr circuit and that driver ccircuit shoulld have the pprotection ccircuit
This IGB
also. Say
y, 1 way to protect
p
a BJT
T from overr current prootection, I diiscussed in m
my last classs. So,
driver cirrcuit also sh
hould have, in
i a BJT, it should able to supply a high currennt during staarting.
So, you require
r
a sep
parate driver circuit to co
ontrol the IG
GBT.
mart power modules.
m
W
What are theyy? That moduule has the ppower
There aree something known as sm
module, the power devices
d
plus the driver circuit
c
plus tthe protectioon circuit, evverything is built
in. You do
d not need to use the discrete
d
com
mponents to make a drivver circuit annd the proteection
circuit. Everything
E
iss built in, in the smart po
ower modulees. The proteections like, over temperrature
protection, over currrent protectiion, over vo
oltage prote ction, all thhese protectiions are buiilt in,
inside the module. So,
S just imag
gine or just think
t
the redduction in thhe size that is possible uusing
wer moduless.
smart pow

17

(Refer Sllide Time: 41:10)

Smart po
ower modulee or intelligen
nt power mo
odule, these are the poweer terminals and these arre the
control teerminals, con
ntrol pins.
(Refer Sllide Time: 41:26)

And, thiss is for the heat

h sink mod
dule. What this
t intelligeent power paack consists?? See, I will show
you.

18

(Refer Sllide Time: 41:37)

See, there are 3 legs, 1, 2 and 3. Each leg con

nsists of 2 IG
GBTs and 2 diodes. Eacch IGBT ratiing is
75 amperre, 1200 volts plus the driver
d
circuitt for all 6 IG
GBTs, the prrotection cirrcuit, over cuurrent
protection, over tem
mperature pro
otection, oveer voltage pprotection. Juust see the reduction inn size
that can be
b achieved by using thee smart pow
wer module. IIt has 3 suchh modules. T
These are juust the
power module.
m
It req
quires additional driver circuit and tthe protectioon circuit. Evverything is built
in this po
ower pack.
So, thesee power mod
dules, poweer packs, inteelligent pow
wer packs arre increasinggly being ussed in
modern power
p
electronic equipm
ments and thaat has resulteed in a signifficant reducttion in size oof the
power eleectronic equ
uipment. So, one of the reasons for thhe progress iin power eleectronics is ddue to
the progrress in poweer semi con
nductor techn
nology. It iss the heart aand soul of power electtronic
equipmen
nt. So, with that I have come
c
to an end of powerr semiconducctors chapterr.
What aree the variouss power sem
miconductor devices
d
that we discusseed so for? Diodes, there are 2
types wh
hich are bein
ng used in po
ower electro
onic equipmeent. One is rrectifier diodde, other one is a
fast reco
overy diode; Rectifier diodes,
d
wherre the frequuency of opperation is loow and for high
frequency
y application
n fast recoveery diodes arre being usedd.
It is an uncontrolled
u
d switch, a SCR and TRIAC, semii controlled switches. W
Why they caan be
turned on
n by applyin
ng a control signal?
s
At th
he gate, haviing turned onn the devicee, you cannot turn
it off by applying a control
c
signaal at the gatee. SCR has aalmost all thhe qualities oof an ideal sw
witch
except th
hat it cannot be turned offf using a gaate and it cann block a neggative voltagge. A track w
which
is nothing
g but or whiich is functio
onally equal to 2 thyristoors connecteed back to baack which iss used
in fan reg
gulators.
G
gate turrn off thyristtor. So, it is possible to tturn off a GT
TO by applyying a
Then we studied a GTO,
control siignal at the gate
g though the gain or the
t current, tthe gate currrent requiredd to turn off GTO
is high. A GTO is a high
h
current device. Theere are again 2 types we studied. A ssymmetrical GTO
19

which can block both positive as well as negative voltage but then it has a longer tail current
duration. To fasten the turn off process, we modified the structure what is known as the anode
short structure. It cannot block the negative voltage. Then we studied BJT, bipolar junction
transistor. It can block only positive voltage, it cannot block negative voltage because base
emitter junction is heavily doped.
It is a current controlled device. The problem with power transistor is gain is low. But then, as
the voltage rating increases and more over these transistors are operated in saturation, gain still
falls. So therefore, we require a small power transistor to drive another high current power
transistor. It has an excellent on state characteristics. Vc sat into Ic is very small. Then we
discussed a MOSFET, ideal gate characteristics, in the sense, input power at the gate is
approximately 0. Input impedance is very high, very fast device and majority carrier device,
positive resistance coefficient, paralleling is very easy, no secondary break down but has one
limitation.
What is that? On state power loss is high. Then we studied an IGBT. It is the device, has
characteristics of MOS as well as a BJT. Input is a MOS, the power stage is a BJT. Gate power
requirement is very small. But then, the turn off time of an IGBT is slightly higher compared to
that of a MOS. That is because there are both majority as well as minority carriers.
So, this turn off time can be reduced by using a punch through IGBT. But then it cannot block a
negative voltage. In non punch through IGBT, it can block a negative voltage. So, there are very
few devices, it can block the negative voltage; a SCR, a symmetrical GTO and IGBT. Other
devices cannot block a negative voltage. So, that is about power semiconductor devices.
From next class onwards, I will discuss the various power electronic circuits. The power
semiconductor devices are used as switches and in our entire analysis, we will assume that these
devices are ideal. We need to know the 2 basic laws to understand power electronics. They are;
Kirchhoffs current law and Kirchhoffs voltage law.
So, we will start the course assuming that you all know these 2 laws. It is going to be very
simple. I will keep the mathematical content to a very low value. I will use the graphical
approach, mathematical approach is very important. So, if there is a circuit, you can always write
a differential equation and you can solve. I will not take that route.
I will take the graphical approach, I will draw the wave form and I will use the basic properties
limit of the various passive components, I will use the basic that is L and C. Average voltage
across inductor should be 0, the average value of the current flowing through the capacitor is 0 at
steady state. These are concepts that we will be using. So with that, I will conclude my todays
lecture. From next lecture, I will start the power electronic circuits.
Thank you.

20