Lecture 55 - Power Amplifier Design - I

IIT Roorkee July 201
61 followers
Mar 4, 2024
Concepts Covered: Power Amplifier Specifications, General Considerations, Single-ended Power Amplifiers, Differential Power Amplifiers.
[Music]
[Music]
welcome to RF transo design course in uh
last few classes we have discussed about
the oscillator now uh one of the
important uh design component for RF
transer is a power amplifier so we'll
try to discuss the power amplifier
design in upcoming
lectures so we already have seen while
discussing the transceiver architecture
that how the power amplifier is
important for transmitter design so as
as we have discussed for receiver Lono
is Amplified design similarly for
transmitter power Amplified design is
important you can use all the
methodology that we have used to design
a power
amplifier uh that we have used to design
a lone noise amplifier however there are
certain parameters which are important
for a power amplifier design so if we
want to see the the power amplifier
specification there are various
specification that has been list out
here uh that one of the specification
important for power amplifier is a gain
and gain flatness what it means that for
all entire bandwidth of the operation
the gain should be flat because we are
going to give the modulated signal to
the power
amplifier so for that you should have a
gain with the flatness in for entire
band second is a power
consumption uh Power Amplifier is a
power hungry component and it is
increasing the power uh it required
higher power for the higher efficiency
so power consumption is also important
parameter third is the
linearity as we have discussed uh during
the initial lectures that there are
certain modulation scheme which has a uh
amplitude modulation which involve the
amplitude modulation and if we want to
recover the transmitted signal then the
power amplifier should be linear and we
already discussed the ofdm uh modulation
and for that uh Peak to average power
ratio is also important so power
amplifier linearity is important it
Signal Peak to meain ratio is also
important because uh for because of the
modulation scheme that we are
using then uh important parameter is a
bandwidth you should have a sufficient
bandwidth in your uh Power Amplifier so
that uh because we are transmitting high
band with signal to the antenna through
the antenna and that uh power amplifier
also should ensure the sufficient
bandwidth uh frequency band for transmit
and receive there are different
frequency band that uh we are using
because high power uh signal will feed
through to the receiver so uh frequency
band is also important what band we are
working then how much power is it
delivered uh to the antenna is also
important for power
amplifier and uh another important
parameter is the permissible inband
emission and permissible outut of band
emission because uh we we have a uh
Power Amplifier which is going to
transmit the signal in the inband
however some of the uh frequency
component of other harmonics will also
generate Auto band emission so that is
also important that how much power is
permissible in the band or out of band
that we should know prior to the design
of the power
amplifier stability over the uh voltage
standing W ratio so stability is also
important parameter for power amplifier
because uh if it's go to unstability
then there is a possibility that the
device or circuit that we are using
might burn because very high power is
involved ability to transmit into
unknown and varying load
efficiency uh so it should have
a uh because the power amplifier is
connected to the antenna so that antenna
might have a different load uh condition
and your uh Power Amplifier should work
it should have ability to transmit into
unknown and varying load
efficiency uh minimizing any loss in the
form of heat and
noise that is also important for uh
Power amplifier and we should find the
minimum size as much as possible uh so
while uh finding out the size of the
device or size of the power amplifier we
we we can have a minimum size then it is
a
better so these are the uh parameters
which are uh not the there are many uh
more parameters that is also possible
but these are the parameters that I
think is important for power amplifier
design and uh no doubt efficiency is
also part of this
so it should have a sufficient
efficiency for and entire frequency
range and uh so that the power loss DC
power loss will be
minimized so mostly for power amplifier
linearity and uh this efficiency are
more targeted parameter and other will
be the uh part of the
same so for this presentation uh if we
see
if if I will
outline the we will try to see the basic
uh class of the amplifier Class A Class
B Class C and uh if we try to improve
the high efficiency of our amplifier
what are the different kind of
efficiency that we will see uh Class A
if we want to improve the efficiency
Class E and class
F there are certain example we can try
to uh uh address that is a cast code
power amplifier positive feedback power
amplifier uh Power Amplifier with power
combining and uh polar modulation can
also be used in uh Power
amplifiers various linearization
techniques because linearity is
important so two things that we mostly
target linearity and efficiency so on
efficiency there there are on basis of
efficiency there are various classes has
been decided defined class AB BC and uh
on the linear linearization technique
there are various linearization
techniques like feed forward cartisian
feedback pre-distortion polar modulation
out phasing and doar TPS so these are
the various linearization techniques
we'll try to
address so uh before going into the
class of the PA some of the issues uh
while designing on cheap power amplifier
that we will discuss the first
issue uh or general
consideration that there is a tradeoff
between the output power and the voltage
swing so what it says that if I want to
transmit one wat of
power if I want to transmit one water of
power
and if that one wat of power if we try
to convert into dbm it will be 30
dbm for 50 ohm
load so for this this much power you
require a supply voltage PE to Peck
voltage should be around 20
volt and your load current load Peak
current
load Peak current is
around 200
milliamp so this uh if we try to use the
common source amplifier with RL load or
if we have a common source amplifier
buas through the current source and this
is a RL load then what what you required
is your pick to pick voltage requirement
is too
high so in order to deliver that much
power now that what will happen that
that voltage is across the uh that is a
drain voltage which is a 20 volt and
very high so one solution we can think
that we can put a inductor here and that
inductor will increase the swing two
times at here that is a 2 vdd still you
your M1 transistor is experiencing a 20
volt
and what is happening that uh in order
to deliver that much power you required
a very high voltage swing so there is a
tread of that uh because uh in the mosad
we have a limited volt voltage swing due
to the uh limitation in the vdd so this
output power and voltage always has
a
limited
uh it has a limited voltage swing and
you want to deliver the high output
power so if we want to deliver so here
the same thing has been written whatever
I discuss if the load is realized as a
inductor that drain uh ac voltage
exceeds
vdd it it will even reach to 2 vdd but
the maximum drain Source voltage
experienced by M1 is still 20 volt so uh
in certain literature it has been proved
that uh the product of breakdown voltage
and ft of silicon devices around 200
gahz into
volt 200 GHz volt so that uh we limit
the voltage swing and output power so
what is the
solution so the solution is
incorporating the interposing a matching
Network so in order to reduce the peak
voltage experienced by the output
transistor a matching
network is interpose between the PA and
the
load so uh what will happen this Network
transform the load resistance to a lower
value that is
RT uh so this RL at this point will be
transferred to the RT which is a low
resistance so that the smaller voltage
swing still deliver the required output
power so you do not need a very high
voltage swing so we'll try to take one
example for that the
above PA if it want to deliver one watt
of power RL with Supply voltage of 1
volt because here Supply voltage is 1
volt what what we need to do is find the
value of
RT so here the peak to Peck voltage
swing VPP at the drain of M1 is
approximately equal to 2 volt because uh
it is Pi to Peck voltage swing that we
are thinking so your P out which is
equal to half VP to Peak by 2 squ into 1
by
RT now what we want to deliver is a 1
volt and we pick to Peck is also fix so
what is the value of RT 2x 2 is 1 so RT
is nothing but 1 by 2 so we need to have
a
half ohm means. 5 ohm resistance at this
place if you try to see 0.5 ohm
resistance and that uh RL is a 50 ohm so
if we want to convert 05 to uh
50 that should experience and if the
voltage swing is uh double then your
requirement is a 1 by 10 ratio of the
Transformer so
what will happen the matching Network
must therefore transform RL down by a
factor of 100 and figure above shows an
example where is a lossless Transformer
having a turn ratio 110 converts a 2v p
to pick swing at the drain of M1 to 20
volt P top swing across the RL so the
here the requirement is a matching that
Network we should have a matching
Network which can also help us to uh
transform the uh lower voltage uh swing
to the higher voltage swing however the
amount of current that is experienced by
the transistor is always high and the
onchip transistor should also withstand
that much
current so now the next point that
effect of the high current now that much
High current if you want to uh if your
transistor want to uh draw the that much
large current then it's width of the
transistor let's say for most but the W
is very high and that width if you try
to increase the weight if we increasing
the width the capacitance that this is a
output stage of the power
amplifier that has higher
capacitance which the previous stage you
need a previous stage because if there
is a mixer is directly connected to the
output stage or the
PA it cannot drive that much High
capacitance so what you need is a you
need to introduce a driver Stage in
between so this driver Stage has has
been introduced in between so that it
can sufficiently drive that much High uh
capacitance now the another issue that
is arising in the power
amplifier is package
parasitics
so if we consider the bond wire at the
source side and so in common source
amplifier
this is a source which is connected to
the ground but it in the package it will
have a bond wire inductance this is a
bond
wire
inductance now this bondware
inductance and there is a
ground and there are uh the there are
inductance let's say one inductor
here that is a load and when we are
connecting this to the vdd at the pad
there is another inductor here so this
inductance are parasitics and what will
happen that that will also add the loss
so uh our efficiency uh will reduce by
due to this parasitics also and this
matching network if we also try to if we
want to make a onchip that will ALS also
at too much loss because on cheap
inductor we already have seen inductor
and Transformers are lossy so in many
case the uh all the matching network uh
needs to be designed of cheap so that
the loss uh will be
reduced and that will try to give the
high power so that is why and this Bond
wire when we consider very high current
this all the parasitics is adding too
much loss and that is why we we can make
a matching network of Chip and this
other stages on chip Active Components
on chip so that our losses are reduced
at high
currents now another important parameter
in the
power amplifier is efficiency so there
are two kinds of efficiency that we can
Define one is a drain efficiency or
collector
efficiency for bipolar transistor
because there are many application many
uh trans power amplifier has been
designed in a silicon germanium byos
process which is a bipolar seos and
widely used for a power amplifier design
because its ft is high and the power
handling capability is also high so the
efficiency D efficiency is given by the
uh loss uh PL means power loss into the
in the uh circuit divide by the power
supply now so PL is a power uh average
power delivered to the load divided by
the power supplied it's a average power
drawn from the supply voltage so the if
average power delivered to the load is
equal to the average power that we are
drawing for
Supply then our efficiency is a maximum
so if we want to have a efficient power
amplifier then your circuit should
consume as loss as
minimum uh Power as
possible and there is a no any power
amplifier in which circuit is consuming
no power there is no
possibility so that is why we cannot get
100% efficient uh efficient power
amplifier but it has a some efficiency
that we can Target to
achieve there is another
uh uh important efficiency parameter
that is the power added
efficiency so uh if you have a power
amplifier and there is input power p in
and output power
PL there is a supply voltage means power
supply it is a p
Supply
then what we can get is a power
deficiency PL L output
power minus P because in last example we
have seen if some of the power is
provided by the uh driver Stitch that is
a p in so p l minus p in by P Supply so
that is how much power p in already
has uh giving the power how much power
power amplifier is adding to the uh
actual power that is received so that is
why it is called a power added
efficiency that PL minus P because the
uh actual power which is added is given
by PL minus P divide by P Supply where p
in is the average input
power so for common source stage in many
case for low frequency your power added
efficiency is is equal
to uh drain efficiency at uh low or
moderate frequency because its input
impedence is capacit and hence average
input power is zero but as we are
increasing the frequency high frequency
your input Port uh input Port impedance
has a real part in
Jing and that draws some
power so in that casee you uh need to
consider the power added efficiency and
for stand alone for stand alone mixer uh
power added efficiency is less than uh
the actual drain efficiency because you
need to match input impedance with the p
ohm and when you are matching your input
is also uh drawing the power so this is
a PL L minus P by P Supply
so uh this is uh about the efficiency
and its parameter now we will try to see
the uh various
configuration before going to the
various uh uh Power amplifiers class
another important parameter for power
amplifier is a
linearity so how we can characterize the
power
amplifier so in uh Power
Amplifier uh has a more important so
power amplifier non nonlinearity leads
to two
effects uh means because of the
modulation uh there is a means when
there is a nonlinearity in the power
amplifier there is a high adjacent
Channel power which results in a
spectral regrowth and we already have
seen how the spectral regrowth happens
when we discuss the uh basics of the
communication and another is a amplitude
modulation so these
two uh will because of this amplitude
compression and the uh spectral regrowth
introduced by the power amplifier
nonlinearity we our power amplifier
should be as linear as
possible and power amplifier linearity
is directly related to the modulation
scheme that we are
interested so
nonlinearity of power
amplifier uh if we are using amplitude
modulation it will distort our signal
and it will be very difficult to recover
the same signal which is a
transmitted so what it's requir is that
we if we try to do the circuit
simulation with the modulated signal if
we apply and that will uh increase the
time timeline of the uh designing of the
power amplifier so in that case your
circuit simulation will take a longer
time with the modulated signal so what
is a way we can uh also ensure the
linearity and we make sure that that it
uh uh it will also perform better for
the modulated signal so that is why
there is a two generic test which can be
used which is a inter modulation and
compression so uh inter modulation when
uh two t we will apply to the power
amplifier and then we try to see the
third order component so uh I'm three
component and we Define the
ip3 and we also Define the 1 DB
compression
point for the uh Power
Amplifier now when we see the two ton
analysis we uh we should ensure that the
output power level is 60b down because
when we are taking a differential output
it is uh its swing is going to have
experience a 6 DB high so we should uh
ensure that the power level is a 60b
down and uh this uh in power
amplifier uh we we should not use uh
amplifier at the 1 DB compression point
it should be used at the uh back uh back
of the power so that should be in the
linear range
so that it will not express the too much
compression so when we are uh try to
design the power amplifier we need to
ensure the linearity by in doing the
testing on inter modulation and the
compression now we also see there are a
two kind of uh Power uh we can get the
power amp alfier with a single-ended and
a
differential so if we have designed a
complete transmitter which is I and Q in
phase and quadrature component and we
have discussed while discussing the
transceiver architecture that uh inas
and quadrature will come from the
modulated signal then we do the up
conversion and if we use a power
amplifier
single-ended which is uh required for
the antenna what will happen that it
Wast half of the transmitted voltage
which is coming from the other part of
the up converter so what is the
advantage of single-ended PS which is
mentioned the antenna is typically
single-ended and single-handed RF
circuits are much simpler to test than
the their differential
counterpart what is a drawback drawback
number one that they waste half of the
transmitter voltage gain because they
sense only one output of the up
converter so what is the solution the
solution is we can introduce a
balloon we can introduce a ballon in
between up converter and the power
amplifier this Balon what it will do
differential single signal to the
single-ended signal it convert the
differential signal to a single-ended
signal and then it amplify
PR so this issue can be elevated by
interposing a balance between the up
converter and the PA but balance will
also introduce a
loss if you have a single ended PA we
will see more
issues if you have a single ended PA the
another drawback that this is a bond
wire inductance at the the between
external vdd and internal uh point at
the power amplifier and bond wire at the
source side so what will happen that
this uh will be because of this
inductor uh very large Transit current
that they pull from the supply to the
ground and that
lb1 there's a possibility that if this
inductor is in similar value to the LD
it will change the resonance or
impedance transformation
properties and there is a signal
feedback at the preceding
stages lb2 degenerate the output stage
and introduce a feedback so because it
is a single-ended the common mode effect
will also uh come at the power amplifier
and that we cannot
neglect uh in the singlehanded power
amplifier so what is a solution we can
use a differential power
amplifier now differential power
amplifier there's
a possibility to uh have the if we have
a differential power amplifier when we
take a differential output it will uh
help to uh reduce the leakage means from
the external
vdd so when we do the differential
realization less sensitive to lb1 and
lb2 so your circuit will not be that
much sensitive to this two Bond wire
inductance and the degeneration issue
also will be
relaxed so the your power amplifier must
still drive a single-ended antenna in
most of the cases we do not have a
differential so you need to introduce
the balance at the output stages after
the power amplifier so
if uh there is a two ways you can either
use a power
amplifier uh B and before the uh Power
amplifier and Bon after the power
amplifier now Bon after the power
amplifier and before the amplifier it
will add the
loss so
if if if you take a example of
efficiency and loss due to the balance
design if your balance has a 1.5 DB
loss if your
Balon has a 1.5 DB loss and that Balon
if uh it is introduce before so there is
a two
case before
4 uh
pa uh
like what we have seen here if we use
this if we add before the PA this is a
PA and if it has a loss of 1.5
DB now if the preceding stage is adding
uh around 0 dbm
power 0
dbm
by up
converter then the loss in terms of
heat in the Balon
is3
M if Bon has a 1.5 DB insertion loss if
it is after the
PA and if your power amplifier is uh
around 1 Wat or uh 30
dbm
so by
PA Power output is 30 dbm then loss is
around
300
mwatt so again uh we if we try to use a
power amp B and after the power
amplifier there are too much loss so
either we uh it is better to use the
power amplifier uh Balon before the
PA if we use the Bon before the PA it uh
it will needs to be designed on cheap it
has a higher loss and after the PA it is
uh going to add uh too much uh loss in
terms of power so there is a tradeoff
based on the requirement and the output
power we can choose that we should use
before or after
uh we'll discuss more about the power
amplifier different class topology in
the next class thank
[Music]
you
[Music]
a
