Lecture 56 - Power Amplifier Design - II

IIT Roorkee July 201
61 followers
Mar 4, 2024
Concepts Covered: Classification of Power Amplifiers, Class A, B and C Power Amplifiers, High Efficiency Power Amplifiers, Class D, E and F Power Amplifiers.
[Music]
[Music]
uh welcome to RF trans design course we
are discussing the power amp design and
we are on module 9 so regarding power
amplifier General consideration we
already have seen now we will go for the
classification of the power amplifiers
now classification of power amplifier
mostly done based on how the uh
transistor or moset is
conducting because the uh amount of
conduction that happens in the
transistor that much uh efficiency will
be degraded So based on the how we are
conducting the transistor and how much
power that is delivered to the load uh
Power Amplifier has been classified so
the first classification is a class A
amplifier so in class a power amplifier
what will happen that uh this uh mosat
or any transistor
will remain in saturation or active
region for entire cycle of the
input so if this is uh if we try to
consider this complete
uh complete RLC Network as a lossless
matching
Network where the supply is coming from
vdd and if we try to plot IDs and VDS uh
IDs is given by this green C green curve
and and uh VDS is given by this red
curve so we can write a drain voltage
VDS because there is vdd which is a
supply voltage through the inductor plus
VP sin Omega T so that is a sinusoidal
uh output that we are getting with the
peak voltage
VP so here Peak voltage is also
important in considering the efficiency
of the power amplifier now given the
load resistance RL we we have RL which
is a 50 ohm the drain current we can
write IDs equal to IDC plus DC current
plus VP by RL sin Omega T if we assume
the lossless matching Network so this
matching Network which we are designing
to transform the impedance which we are
looking from this side to 50 ohm if we
consider it it as a loss less the
efficiency we can write P out by P in so
P out by P in so if we consider the P
out uh and relation of RMS V RMS voltage
with the VP to Peak voltage that P out
is given by VP s by 2 RL divide by the
input power which is coming from the
supply voltage that is IDC total current
IDC into vdd
now if we want to go for the maximum
efficiency which is given by VP s by 2
RL uh the IDC current we can write VP by
RL into vdd so uh RL RL will be
cancelled out here and we'll get VP V VP
one of the VP will also be canceled out
so VP by 2 vdd now here the important
point is how your uh swing voltage swing
at the output it is at what level if VP
voltage swing is a maximum equal to vdd
then this class A amplifier can reach a
50% efficiency So based on this VP we
can have a three kind of uh uh
efficiency for class A
amplifier so that that is divided in
three part efficiency in different
scaling scenario so the first is the
supply voltage and bias current remain
at a level necessary for full output
power mean Supply voltage and bias
current remain at the level of levels
necessary for full output power and only
the input signal swing is reduced so if
the if we consider only the input uh
signal swing which is changing then we
uh as the input signal swing will change
our VP is also changing so we can write
VP squ by 2 RL that is uh uh output
power divide by the DC power vdd squ by
RL so we can write equal to VP s by 2
vdd square so this is the efficiency one
so if we uh if we plot this efficiency
one with with respect to the uh voltage
swing which is starting from 0 to vdd so
so here VP is a voltage SP and this is a
efficiency so this uh if we consider
input swing as a scaling parameter then
efficiency one has
a uh Behavior which looks like this
because it has a square behavior that is
VP Square by 2 vdd squ now the second
scenario is if the supply voltage
remains unchanged means we are are
keeping Supply voltage fixed but the
bias current is reduced in proportional
to the output voltage swing so it is the
same equation if uh that we have seen if
this bias current is changing with the
output voltage swing VP then we can
write a VP by 2 vdd so this efficiency
to what will happen that VP as VP is
changing your uh efficiency is also
linearly changing so this is a l
function of the uh VP how it is changing
and if we have a both Supply voltage and
bias current which are reduced in
proportional to the output voltage swing
so in this which is very important that
you required a supply voltage also
varying with the swing and it is very
difficult to design the amplifier in
which you need to scale the supply
voltage then only it will reach the 50%
efficiency so which is a constant means
if you want a
constant uh efficiency for class A
amplifier then your supply voltage also
should change with the swing so this is
the three scenario that is important to
understand based on the how the scaling
of the uh input uh swing or the bias
current or uh Supply voltage and the
current scaling is happening so while
designing we need to uh make sure that
we we have a sufficient voltage swing in
this class A amplifier Class A amplifier
has one of the advantage that it is a
linear uh and uh because of its
linearity it is one of the choice in the
uh in using the in power amplifier
design so this is one of the the uh
widely used uh Class A amplifier power
amplifier because of its uh linearity
however the efficiency is not good so in
order to have a better efficiency there
are other class that is a Class B power
amplifi so he here that uh uh this LCM
means uh LC matching that is uh same
whatever is drawn here it is mentioned
as LC matching and uh this is uh uh
input that is supplied and we have IDs
current and VDS
voltage now here what what is happening
that instead of making this transistor
on the way we have done in class A
amplifier for entire cycle if we keep
only class B amplifier on in the
positive cycle
so if we try to make this transistor on
only in positive cycle only the IDS
current will flow and then there is a
due to the matching you will get a VDS
voltage which is a sinusoidal with the
with the peak voltage VP so ID is
current we can write in terms of ID s
equal to IP by Pi + IP by 2 sin Omega T
and it will go on we can write this uh
only the uh positive
cycle here when the fundamental
component of the current reges its peak
IP by two so this is fundamental uh
component sin Omega n t so the total
current saying from the load is VP by RL
that is VP by RL assuming the Pak swing
uh of uh Pak swing of uh this amplifier
is VP at the drain so we can write IP by
2 which is equal to VP by
RL so efficiency is equal to P out by P
in so output power is given by VP s by 2
RL uh as we have uh discussed in the
class A amplifier and the input power is
given by IP by pi so this is a DC term
DC current IP by pi into vdd so uh the
this Pi will go up
and we will have a PK by 4 into VP by
vdd so that uh here
IP this IP is replaced with the 2 VP by
RL so we can write it down uh
VP VP
s by 2
RL divide by
IP by pi so 1 by
pi
into
2 VP by RL into
V DD so one of the VP will be cancel
out then RL will be cancel out we will
have a pi by 4 this Pi will go up Pi by
4 VP by vdd
and if our PI to pick voltage VP will
reach to the
vdd then we will have a efficiency which
is a p by 4 so we will have a more than
uh 50% efficiency uh around uh greater
than 70 if we try to use a Class B power
amplifier now here what we are doing is
we are reducing the conduction of the
conducting transistor so that conduction
is also denoted as a conduction angle so
as we reduce the conduction angle uh
Theta for uh the transistor which is uh
driving the load that will reduce the uh
amount of power consumed by the
transistor so the maximum power will be
transferred to the load and efficiency
will
increase now now there is another
amplifier that is a Class C amplifier
now in class C amplifier we uh again
further reduce the conduction angle
instead of uh uh only the positive cycle
if we try to reduce the conduction angle
only for few uh cycle means half of the
cycle then as drawn in uh this figure
here vgs is drawn and there is IDs
current
so it is conducting only for this cycle
we we need to bias the transistor such
that it will conduct only for small
amount of uh period so if we uh here
assuming a conduction duration which is
a tow less than t by2 uh in the
previously it was a t by2 more than this
uh if it is where T is a one cycle we
can see that the device turns on and off
at a two points t on by off equal to T
by 4 uh plus or minus t by 2 we can
write a t which is equal to T cos cos
inverse vth so this vth threshold
voltage by VP voltage P by pi uh where
the conduction angle Theta which is
given by so this conduction is happening
uh only in the duration where cos
inverse we threshold by VT and it is
given by pi by T into T now during the
conduction the drain current IDs is
given by IP into- cos Theta + sin Theta
and here time T is between t on and off
so only this
period so for T is equal to uh if we
consider T is equal
to T is equal to T on and off which is
this
uh IDs is equal to 0 and IDs is equal to
IP by uh Pi so I for other period it is
zero and uh if we consider only this
much period which is a t by 2 IDs which
is equal to I IP by pi sin Theta minus
Theta cos theta plus uh Theta - sin 2
Theta by 2 this is the fundamental ter
sin Omega not C so to get the swing VP
at the RL during the conduction we must
have a IP which is equal to P into VP by
rlid by Theta - sin 2 Theta so this is a
as we have seen in the class B uh this
is a fundamental tone we need to
consider so IP which is given by pi into
VP by RL by Theta - sin 2 Theta by 2
because uh that current is only uh the
IDS is only provided during this period
and it is divided by this term now we
write P out by P in P out nothing but VP
Square by 2 RL it is same and p e IP by
pi into IP by pi into this uh sin Theta
minus Theta cos Theta into vdd that is
this DC term IP by pi sin Theta th minus
Theta cos Theta vdd and uh now we can
write VP by RL uh we replace uh
IP with this IP so when we replace IP
with this IP this uh Pi will be
cancelled out and this Theta minus sin 2
Theta term will go up and this will be
down and this is VP will be canceled out
RL will be cancelled out so VP by 2 vdd
here if uh VP is equal to
vdd then this is a efficiency which is
depend upon the how much conduction
angle Theta now conduction angle Theta
is depend upon the V threshold by V pick
so uh what is a uh at what period of
time it is switching on based on that we
can have a uh efficiency that we can
Define so ideally if we try to plot the
efficiency versus the angle
Theta for a 0° we can get a maximum 100%
efficiency by using this equation but uh
we as this conduction
angle will uh increase because uh that
will reduce the efficiency Class B has a
79%
efficiency where it is conducting for
180 uh Dee and Class A which is for 360
it is a 50% duty cycle uh 50%
efficiency there is a tradeoff here also
has been observed that if the conduction
angle which is approaching to zero the
output power that P out will also uh
reducing uh to the very uh less so there
is a tradeoff that maximum output power
that we get and the efficiency that we
get versus the Theta which is given in
this another plot so we need to choose
the uh if we want to achieve the higher
efficiency the amount of power output
power will reduce so while designing we
need to consider uh this also that what
is the output power that we are going to
reach for the different conduction
Angle now uh there is a another class uh
in all the class that we have seen class
ABC which is a widely used uh class
where the conduction we have uh defined
the conduction how the conduction is
happening now in uh another uh high
efficiency power
amplifiers what what we try to do is we
uh try to use uh two features one is we
try to modu the impedance for different
uh means different harmonics we try to
modulate the impedance with the
different harmonics and we also try to
make the instead of using a transistor
as a uh voltage uh uh current control
voltage device so where where voltage we
are applying a voltage and its output
current is changing
uh that instead of that we try to use a
moset as a switch so uh if we Tred to
use a mosat as a switch then amount of
uh time that uh transistor is on will be
reduced and the maximum power will be
transferred now in order to when there
is a uh digital switching we are doing
we need to make sure that our output uh
means the output waveform is should be
uh shaped properly uh so that the sun
soidal output will uh you will get so in
that case the impedance matching is very
important so if we have a class a stage
let's say uh and uh if we suppose the
matching network is design such that uh
its input impedance is low at the
fundamental and high at the second
harmonics so what what we try to do is
we try to make a a the fundamental
harmonics uh which is passed so when
there is a fundamental harmonics will
come uh at that time the impedence is a
low uh that J One impedance is low and
for other harmonics the impedance will
high
so we we need to properly Design This
impedance matching Network so there is a
sign Omega t uh if uh for second
harmonic 2 Omega + 90 the if we have a 3
sin 2 Omega not + 90 and then if we have
a combination of these two uh signal it
will something look like this so it is
more most likely the what we are trying
to do is uh trying to make it more
digital instead of a uh analog uh
[Music]
the way we have done in the class A
amplifier so here the average power
consumed by the output transistor
decreases and the efficiency increases
the that is that is why this be this
kind of methods will be used to improve
the efficiency of the power amplifier
there is a
another class which is defined as a
Class D
d means digital also where uh if we try
to uh this is a basic inverter
configuration where uh uh during the
positive cycle it will charge and
negative it will discharge so uh it will
uh mostly VDS SN means idsn will look
like this and vsn is more like a switch
so uh and this is the inverter now uh
this uh if we try to redraw this as a
switch where it is on and uh uh it is f
charging and discharging Behavior we can
write write a load voltage Peak current
we load equal to 2 by pi because we are
considering the switching uh and we have
seen that when we are considering the
square wave the uh DC is a 2 Pi Pi so
that is we load 2 Pi Pi vdd into RL that
is uh uh total uh voltage vdd into
RL divide by uh what we are doing is uh
the amount of voltage that is coming at
this node so the if we use a voltage
division rule that RL divide by RL uh
plus r switch plus RL will give the load
voltage V
load and current I load which is G by 2
by pi vdd divide by the total resistance
R switch RL by plus RL small RL plus RL
so to efficiency we can write the P out
by P in equal to V load into I load by
two uh IDC into vdd so if we replace V
load and I load in this equation what we
will get RL by RL + RL + r switch so in
this your efficent efficiency is depend
upon the load resistance and uh uh the
inductor quality factor and the switch
resistance R
switch so uh this if we consider this uh
RL which is very high in this case and
if we try to reduce the r switch
resistance then uh this term RL divide
by RL and it will B efficiency is equal
to 1 so there is a possibility to get
more than 100% means more than 90 uh
percentage efficiency or around 100%
efficiency if if we Design This switch
with a low resistance and if we have a
better uh matching
Network we can also have a uh
differential amplifier configuration
which looks like here where uh this is
uh tuned uh uh circuit for impedance
matching and this is uh uh how we can uh
make this uh more look like a square L
only for positive uh for positive signal
and negative signal the differential
signal will be applied to the tune
circuit so this is called a class D
amplifier power
amplifier then there is a Class E
amplifier now what we need to do is in
class E stage we need to revisit the
output stage with the switching
transistor so in class E also what uh we
do is we use a uh transistor as a switch
the way we have used in the class D so
in class E stage are nonlinear
amplifiers that achieves efficiency
which is approaching 100 percentage
while delivering the full power and it
is a very
remarkable advantage over a Class C
circuits uh in switching power amplifier
uh it achieves the high efficiency if
this M1 transistor which is shown here
uh sustain a small voltage when it
carries a current means it should uh
have a very small voltage when it uh so
it is like a complete shot and M1
carries a small current when it is
finite voltage
me
means when it is on it does not it
should pass the current means it is kind
of Ideal switch and when it is off it it
should have a very minimal means it
should not have any current and the
transition time between the on and off
states are minimized so if if we have a
ideal switch kind of configuration then
uh we can have a uh better efficiency uh
though there there is a as shown in this
figure there is a certain on Resistance
and when it is off if there is some
lickage current uh so that uh we cannot
because of that we cannot achieve the
100% efficiency but uh it will be close
to the uh good efficiency that we
want so in class A amplifier the main
challenge of a Class D paa that we have
discussed is to minimize the voltage
across the switch when it is carrying
the current and the current through it
when there exist a voltage across it so
uh making it a more ideal switch that
was a one challenge when we have seen
the class D now uh this condition what
is called a fast switching uh which
which is very difficult at high
frequency and if the switch voltage and
current and which is if it will overlap
they uh they generally do in practice
due to the finite rise and fall time we
cannot have a ideal uh digital voltage
and ideal switch and the efficiency
degradation May substantially happen and
uh this can be avoided by proper design
of a matching Network so we need to
design the matching Network properly and
which adds in in degree of freedom to
ensure that there is no appreciable time
duration that voltage and current
simultaneously exist so uh that that is
uh that way we can improve the
efficiency now there is a class F power
amplifier so in class F power
amplifier what uh has to be done we'll
use the M1 transistor as a switch only
but the way we have improved the
efficiency and class amplifier by making
the uh input impedance looks low for the
uh fundamental frequency and high for
the other harmonics it is improving the
efficiency so same thing if we try to
make this imped impedance J uh one if
this try to uh make this input impedance
low for fundamental tone and high for
other harmonics then we can achieve the
higher efficiency uh in class F
amplifier so uh in a generic switching
State the load Network provides a high
termination impedance at the second or
third harmonics and the voltage waveform
across the switch exhibit a sharper
edges than a sinoid thereby reducing the
power loss in the transistor and this
such uh this kind of circuit is called a
class F stage and where VX which is
drawn instead of square it looks like a
sinusoidal uh waveform so this is what
the all the efficiency
that uh means this is about the
efficiency of power amplifiers so we
have seen all the class different class
of the power
amplifier Class A B C D E F and class CL
ABC which is defined based on the
conduction angle class DF is how
digitally we switch and reduce the
amount of power that is consumed by the
transistor and it improves the
efficiency uh in all the cas uh
impedance matching network is uh very
important for Designing and we have
given enough time to design the
impedance to understand the impedance
matching uh during the initial lectures
uh thank you we'll discuss about the uh
linearity about the linearity
Improvement in the next
[Music]
l
[Music]
a
