Lecture 35 - Stability and Amplifier Design - IV

IIT Roorkee July 201
61 followers
Mar 1, 2024
Concepts Covered: Unilateral FOM, LNA Design
[Music]
[Music]
uh welcome to RF transer design course
uh we were discussing about the
amplifier design uh we also discussed
the stability in last class we'll go
further and uh in last class what we
have seen that if we want to design
amplifier for the maximum gain we have
considered the designing the amplifier
for maximum gain we have also derived
multiple uh gain formulas which are
important for the amplifier design it is
used it is useful for low noise
amplifier as well as power amp amplifier
so it is a common for both so that is
why we have a try to make it a common
for both the uh Amplified design now we
will see further that if we have a
specific gain and how can we go for the
amplifier design in last class we have
considered two case one is a unilateral
device or bilateral device unilateral
that means there is no feedback uh out
between output to input and in bilateral
you have a feedback from output to input
so in this case if S12 is equal to 0
that is considered as a unilateral but
if S12 is not is equal to0 then it is
called a bilateral so there is one
figure of Merit that is called
unilateral figure of Merit let's say you
have find out the S parameter of two
Port networks you have a two Port
network available
and that two Port Network consist a
multiple transistor as in last class we
discussed it can be one transistor two
transistor it can be differential
amplifier you can make it as a two Port
amplifier two Port Network now in the
two Port Network you know the S
parameter and there is a
S12 which is not equal to zero so if
your S12 which is not equal to Z still
you can consider it as a unilateral by
choosing this unilateral figure of Merit
so for many transistors or many
transistor network uh S12 is a small
enough to be ignored and uh when you
ignore the S12 uh we have already seen
that your uh methodology will be simpler
and you can design it faster and the
device can be assumed to be unilateral
so let it be a bilateral device but if
there is a S12 is very small then you
can consider it as a unilateral so the
error in the transducer
gain the error in the transducer gain
caused by the approximating S12 is is
equal to Z is given by the ratio it is
also called a unilateral figure of Merit
U that is 1 by 1 + U
² uh GT less than GT T by GT u 1 1 by 1
- u s so here U is defined as a
unilateral figure of Merit and the this
will provide you the range at which your
gain will differ if you have considered
the uh your transistor or transistor
Network as a unilateral or bilateral so
that difference in the gain if it is in
the range of plus or minus 2 or5 DB you
can consider it is you you can consider
this unilateral figure of Merit based on
the S parameter you can find out U and
you can consider your transistor as a
unilateral not a bilateral so this is
very important figure of Merit to find
out that you can consider unilateral or
bilateral
device so let's move to the uh how how
we can design the uh amplifier for
various gains
so that is called expression for GS and
GL so GS is a gain at the source side T
is a gain at the L load side so if you
have a unilateral case uh or bilateral
case uh you can write a GS uh from the
equation of GT uh we can write a GS that
is a 1 - GMA squ by 1 - S11 GMA s square
and your GL l
is equal to 1 - GMA l s / 1 - S2 GMA l
s so if you have a match condition if
your gamma s is a match conjugately and
Gamma L is mat conjugately with the
s22 this is the input reflection
coefficient of the transistor and s22 is
the output reflection coefficient of the
transistor so if it is matched
conjugately the with the S11 your gamma
s you can write GS Max that is 1 by 1 -
S11 whole squ and similarly you can
write a GL Max which is equal to 1 by 1
- s22 s so your GS so this is a maximum
gain at the source side you will get and
this is a maximum gain at the load side
you will get and you already know the G
node that is a S21 Square so GS now
let's say if you don't want to design
your amplifier for the maximum gain you
have a certain requirement of the gain
so that value GS so there is a
normalized gain Factor GS and GL can be
uh derived and what is that maximum
means gain that is required gain that is
a GS divide by the maximum gain that is
a GS Max so if you if you derive the
divide the GS by GS Max you will get 1
minus GMA s squ by 1 - S11 GMA s whole
Square 1 - S11 squ similarly for load
side at the load side you want a gain
which is a GL by GL Max so where GL is
your desired gain and you can find the
GL Max Bas based on the S gamma L which
is a match with the s22
conjugate so GL by GL max if you write
it down you will get a 1 - GMA l s / 1 -
s22 GMA L whole square into 1 - s22
Square so what is the maximum value this
is a normalized GS so this normalized GS
can have a value from 0 to up to one so
the all the Valu
are in the range of zero to maximum of
one similarly GL because you cannot get
more than GS Max so when your GS so here
GS is equal to GS
by GS Max so when your GS is equal to GS
Max you will have a GS equal to 1
similarly your when your GL
by G GL
Max when GL is equal to GL
Max it is equal to 1 so you'll get
maximum value of GL that is a normalized
gain Factor GS or GL that is a maximum
is a
one so if you try to uh make this
equation into
the form of the circle so you will get a
gamma s minus GS S11 conjugate 1 - 1 -
CS S11 squ equal to square otk of 1 - GS
1 - S11 squ 1 - 1 - GS S11 squ so here
what is important that you can represent
the same equation because you are
interested to find out the value of
gamma s and if you uh if you want to uh
find the value of gamma s and gamma L
which will give you the desired gain GS
and GL L that will come from this
equation and this equation is nothing
but a circle equation which is a center
is given by the Cs and radius is given
by the RS so this is nothing but gamma s
minus CS equal to
RS so you can have a constant gain
circles for the input section that is a
source side which is given by CS equal
to GS S11 conjugate by 1 - 1 - CS S11
squ so here if you put GS is equal to 1
this term will be zero then you will
have a center equal to means your Center
will be at S11 conjugate get only it is
at one dot and your radius is also zero
so this is at the source side you can
have a constant gain circles for the
output section that is at the load side
so the way we have written the equation
for gamma s you can also write equation
for the gamma L and this gamma L will
provide you the two uh equation that is
the center of the circle CL and the
radius of circle RL so this CL here now
is depend upon s22 conjugate so the
angle of your CL is depend upon the s22
conjugate so this is a GL s22 conjugate
by 1 - 1 - CL s22 s this RL is square
root of 1 - CL 1 - s22 squ 1 - 1 - CL
s22
squ so this is how you can also plot the
gain Circle so up to now if you uh go
back and see that how many circles that
we have studied we have studied the
input stability Circle we have studied
the output stability Circle to find the
stability at the input side to find the
stability at the output side now there
is other circles that are the gain
circles so there is a uh for different
gains you can have a different Circle so
while finding out the optimum uh gain
and stability parameter you should uh
draw the circle on the SMI chart you can
also use a uh simulator in order to draw
the
same so what are the attributes here the
angle of the circles are defined by S11
conjugate and s22
conjugate when GS or GL which is equal
to one that is a maximum gain I already
told the radius RL or RS is equal to
zero and the center reduces to S11
conjugate and and s22
conjugate the 0 DB Circle will always
pass through the center of the Smith
chart so if you uh if you want a zero DB
gain that will pass through the pass
through the center of the SMI chart so
in order to understand we will use the
same parameters that we have discussed
in the amplifier design and we'll take
one
example what what this example says we
have S parameter is given and you uh in
the last class what we have done we have
find out the maximum gain for maximum
gain we have find out the input matching
and output matching similarly here we
want a gain of 11
DB and frequency is a 4
GHz so if you want to go for the gain of
11 DB and frequency of 4 GHz you need to
plot the constant gain
circles which is GS is equal to 2 and 3
DB and ZL is equal to 0 and 1 DB and you
can get the
maximum uh gain of 11 DB either you can
use a two and Z combination of two and z
g note square is already given or you
can use a three and 1 uh three and zero
so the any combination is possible based
on the gain that you
want so let's uh this is the S parameter
given and here we have consider the S12
is equal to Z uh in order to uh make it
simple and here uh because this is a
unilateral case uh this s parameter for
3 4 and 5 GHz is given here make sure
that this s parameter is coming after
biasing the circuit and it is given by
the foundary for the particular bias
first you need to find the stability in
the last class also we have find
stability here K greater than 1 Delta
less than 1 and we need to find the
maximum gain so what is the maximum gain
GS so that is the GS Max which is 1 by 1
- S11 squ equal to 2.29 which is equal
to 3.6 DB what is the GL Max 1 by 1us
s22 Square 1.56 it is 1.9 DP and what is
the G not it is S21 squ equal to 6.25 =
to 8
DP so your G2 Max is given by 3.6 + 1.9
+ 8 which is equal to
13.5 DB so this is a maximum gain that
we get but what we want is a 11 DB gain
so either we need to use a reduce the GS
or we need to reduce a GL so that is why
what we have choose is a to plot The
Circle at 2 and 3 DB so here the you
already have a 8 DB gain of the G not so
if you want 11 DB you can get the 2 DB
from the GS Max then you should make GL
Max g 2 DB from the GS and GL you should
make a zero or otherwise you can also
get
the uh this is a 3db means here
3db and then this is a zero similarly
you can make a 2 DB and 1 so as per the
plot and how the gamma s and gamma L you
can choose any value so now we will try
to plot the circle so in order to plot
the circle what you need is you need to
get the uh CSN RS CL and
RL so here your
gain GS is equal to 3 DB you already
know GS
Max so this is a small GS now you what
are the parameters that you required for
calculating the Cs and rs you need a GS
which you already have derived uh you
need a uh you need parameter that also
already been given so based on this you
can find a GS and mean Cs and rs so that
is a this is a CS you can calculate 706
and rs which is A.1 66 so when you try
to draw on the SMI chart this is a
center and then you have a CS so you
need to make one line which is the angle
of 120 and your CS is 706 so that will
come something around this
point so at this point you will get a
maximum GS Max and when you uh when you
try to plot this will come here at GS is
equal to 3 DB and there is a circle so
this point this angle will remain same
this point will always lie on this
circle so Center will lie on always on
this line and you can plot the GS is
equal to 3db Circle similarly you can
also plot a GS is equal to 2 DB Circle
where you already know GS Max your are
small GS normalized gain is 691 you will
have a CS which is a 627 angle of
120 and rs is given by
294 at the load side
so when you want to plot at the low side
that is a GL Max you can find a uh
normaliz GL by GL that is a desired GL
by GL Max so that is a normalized ZL and
based on the equation that we have seen
for the center of the circle and radius
of the circle you can find the center
and radius so the here angle is a point
uh 70 so there is another line which is
drawn for uh gamma l in order to find
the gamma L and at this point where it
is s22 conjugate you will get a maximum
GL Max and when you want GL is equal to
1 DB that will have
a a circle which is which is a center
somewhere here and you will have a
radius of
RL and you can also plot GL is equal to
0 DB so you can see that it is passing
from the uh when when you want a zero DB
it is passing from the the circle is
passing from the center of the Smith
chart and you can choose any value uh
means you can choose a value of a gamma
s or gamma L which is touching to this
circle so uh here the gamma s value is
chosen near to 50 ohm so what P
represent the center of SM chart will
represent the P ohm and always try to
use the point which is near to 50 ohm so
what is the advantage of that when it is
near to 50 ohm the passive component
that you are adding to match are in
lower values another Advantage is you
will get the uh stable means your
stability generally stab stability
curves will lie uh on on the periphery
means uh on the top of the SMI chart so
in in terms of stability also it is
better to choose the point which is near
to the uh 50 ohm so there is a less
chance of going your amplifier will go
to the unstable region so that is why
this is a gamma s that is chosen and
this is a gamma L is chosen so that
should lie on the circle of the Smith
chart so what is the gamma s so here
gamma s and gamma l so the gamma s this
gamma s will provide a gain of 2
DB and this g l will provide a gain of 1
DB so this this is a 1
DB and this is
a 2 DB so total is three so uh and if
you add a g D not then there is a 8 DB
so total gain you will get is a 11 DB
that is your desired gain now the next
part is to do the
matching so how can how can you do the
matching we have already seen in while
discussing the JY SM chart and impedance
matching that you can use a JY me chart
to do the matching and that can be
done by uh inductor and capacitor so you
can use a LC matching even you can use a
q method or t or Pi any network you can
design so for example here for 4 GHz
frequency there is a this is a JL means
here both the place it is considered JL
but here this is this is done for gamma
s and this is done for gamma l
so your gamma s is lying at this point
so this is a gamma s value and then
there is a gamma l so this uh this is at
the input
side and this is at the
output so what you need is uh here your
input means this comma JL which you want
to match that is a 50 ohm mean that is a
gamma s that will correspond to
correspond to the value of gamma s and
this is a at the S side it is a 50 ohm
and here it is related to gamma L and it
is a 50 oh
impedance so this
is uh means at the sour side if you want
to match so this is a point is gamma s
now how can you match you you need to
reach to this 50 ohm uh Circle 50 ohm
point and how can you reach to this 50
ohm point if you come down on this smid
chart now when you are coming down on
the smid chart you are using a uh Jed
that is a red lines J SMI chart and when
you are coming down that means you are
adding a capacitor in series so this is
a capacitor that you are going to add
that will be around 8 95.6 pad for
frequency of 4
GHz and when you go up so when we are
going up on the SMI chart but we are
using a blue line that is a admittance
so when you are adding a
admittance using a blue line you you
when you are traveling on the admittance
me chart you are going to add the uh
component in a parallel and that is a
inductor because we are moving up on the
S chart
so always remember when we are moving up
on the SMI chart you are either going to
add inductor in series or parallel and
here we are adding in parallel because
it is a uh y smid chart you are using a
admittance smid chart so then you need
to add around 2.5 Nan of inductor in
parallel so you can easily match by
using a simple LC matching Network and
this both value 2.5 nry and 8
95.6 P parad ptop parad are easily
realizable at the onchip by using onchip
inductor and onchip capacitor so it will
not take much area as well and you can
design your matching Network on chip
similarly at the output
side here uh there is this is a point
gamma L that we have already seen from
the previous me chart this is a gamma L
and Gamma
s so if you want to find out the
impedance matching you go on the SM
chart and when you are going down on the
SM chart and you are using a admittance
me chart you need to add a capacitor in
parallel that will be around 595 parad
and then inductor means when you're
going up in order to meet to the 50 ohm
line
when you are moving up you are adding
inductor in series and that will be
around one n en of value so this is the
way you can get uh the LC matching at
the input side and output side and when
you put all the s parameter and the
matching on the simulator you will get
the uh battery input matching output
matching and the gain of 11 DB at the
desired frequency you can also so this
is the value of gamma s and gamma L that
is33 angle of 120 GMA s uh and Gamma L
is 22 angle of 70 you can also use a
stop matching means you can use a
transmission line we already have seen
how you can use a convert inductor
capacitor in terms of transmission line
you can also use a stop matching uh to
do the impedance matching so in the if
you take a previous case there is a
capacitor and inductor
so uh here there is a
capacitor and
inductor and this is a
[Music]
input and at the
output if you see uh you have a a again
capacitor is in
parallel and inductor
in series this is
a
output so this is the way you can uh
design the amplifier of the desired gain
in the last class we have seen the
amplifier for the gain what is the what
means gain that we required for
the in the last class what we have done
is we have gone for the maximum gain now
in this class we are going for the uh
means in this class we have discussed
that if we desired a certain value of
gain that is yes or uh means that is a
11 DB less than the maximum value then
you can uh draw the gain circles and by
using the gain Circle you can find out
the value of GMA s and GMA l so this is
how the amplifier design for desired
gain and gain circles should
draw
now uh now we we will move to the low
noise amplifier
design because we we have already
discussed the amplifier design at micro
frequency so now it is important to go
for important to design a low noise
amplifier and we already have seen the
low noise
amplifier is a first component of the
receiver and that will try
to because it will improve the
sensitivity of your receiver and it will
add very less noise because your
incoming signal is already the noisy
signal and when you are adding a less
noise by using a low noise amplifier
then what is happened due to the freeze
formula
your uh to Total F noise factor F is
depend upon the F1 plus F2 -1 by G1 so
the it is important that you should have
a minimum noise figure or noise factor
for the LNA and it should have a
sufficient going sufficient gain so that
you will have
a higher gain it will SL down the noise
effect of the incoming
stages so here if we want to design a
low noise amplifier how we can consider
uh there is a one noise Spector equation
which is widely used to design a lone
wise amplifier in microwave circuit also
or in RF uh IC as well so what we will
considered here that there is a two two
Port Network and uh we already have
discussed that if there is a two Port
Network and for this two Port network if
you want to define the noise Factor it
is equal to
SNR in
by
SNR
out now SNR in you can write SI by
ni you can also write SNR out here n
o by s o because it is in
denominator now because there is a gain
involved we are considering the
amplifier design we can also write this
as a SI
I by ni into G uh into n
o by g into SI so this SI here will be
can cancel out so what you will get is
a output noise divide by
gain divide by 1 by n i so here you are
dividing output noise means total noise
which is at the output dividing it by
gain so that when you are dividing it by
gain that is called input referred means
you are you are not considering gain
whatever output noise is there you are
uh taking it out so output noise you are
uh deriving it as
a uh at the input side so when you are
bringing the output noise at the input
side you can represent it as a noise
current I or you can also represent it
as a noise voltage VN so uh that
equivalent noise current at the input
side is called input referred noise
current ion it is also VN at the uh
noise which is voltage noise at the
input side due to the two Port Network
it is a input reper noise
voltage so this is then this will become
a noisess two Port Network so this will
uh be a noisess TW Port Network and you
already have represented the noise
generated by the twort network as a
voltage source and a current source so
here the total output noise power is
proportional to the mean square of the
short circuit current at the input Port
so this is of the noise pre amplifier so
the noise power due to the source alone
so here this is a source so source is
represented in terms of the admittance
Ys and current is and then you can write
F which is i c squ because then no
longer you need need to worry about this
total short circuit current ISC Square
divide by the is square that is a your
Source current is noise current is so
we'll discuss more about this uh
equation how we we are coming to the
final equation of noise factor in the
next
[Music]
class
[Music]
a
