Lecture 52 - Oscillator Design - II

IIT Roorkee July 201
61 followers
Mar 4, 2024
Concepts Covered: Three Port Oscillators, Voltage Controlled Oscillators, Discrete Tuning, CE BJT based Oscillator, Colpitts OScillator, Hartley Oscillator, CG FET based Oscillator, Microwave Oscillator.
[Music]
[Applause]
[Music]
uh welcome once again to Arab
transceiver design course uh we are
discussing the oscillator design and in
last class we have stopped at the uh
cross coupled oscillator so in order to
design cross coupled oscillator your GM
RP should be greater than or equal to 1
now there is a three point oscillators
that are also used uh
that by tipping out the inductance at
one of the point which is a C1 and
C2 and uh we can obtained by grounding
each of the transistor terminal so this
is a common source configuration and
this is a common gate configuration so
any of these can be used to generate the
oscillation here if C1 equal to C2
transistor must provide a sufficient
trans conductance so that it will
satisfy a GM RP which is greater than or
equal to 4 uh this kind of oscillator
has
drawback because you required a very
high Q inductor and it is only
generating the single-ended output for
certain uh low power application or in
uh some of the application where uh you
have a off chip inductor that there you
can use this kind of configuration but
for onchip cross coupled oscillators are
widely used because it is providing a
differential output and and you can have
a
oscillation uh with uh less Q inductor
as
well so now we are moving towards the
voltage controlled oscillator because we
need to tune our oscillator frequency at
different range that is a requirement
for transceiver design and we have seen
while discussing the basics of the
communication so what what is the input
for this kind of voltage control
oscillator V control is a input then
there is a vco your output frequency is
Omega out so this output frequency if we
try to draw the graph output frequency
versus V control we are changing the at
voltage V1 we are getting Omega
frequency Omega 1 and as we are
increasing the voltage V2 it is going
going to provide the frequency Omega 2
so we can write Omega out which is a
function of K vco this is a k vco which
is a slope of the characteristics or the
gain of the
vco or we can also mention a
sensitivity and which can be uh the unit
will be radian by Herz into voltage
because the frequency angular frequency
omega as a unit of radian per Herz uh so
k v into V Control Plus Omega not so
this is how we can uh Define the uh vco
in terms of the
equation now if we want to add the uh
voltage control means if we want to
change the oscillation how can we change
we we need to change uh one of the
parameter so we know the Omega is equal
to 1 by under root L and C so which
parameter we can change easily we can
change uh inductor is fix because we
cannot change the inductor but we can
change the capacitor and this capacitor
will help us to uh change the
oscillation
frequency so for that uh vco we can use
the Moss
veror to change the uh pre uh capacitor
so this is already uh C1 uh based on the
C1 it is already oscillating at one
frequency but this C variable is added
that does not that that is provided by
using this Moss variable capacitor so
how can we get that drain and Source
will be connected to the V control and
gate will be connected to this so here U
as as I told it is difficult to vary the
inductance electron elect electronically
we only vary the capacitance by means of
VOR so this C VOR if we try to plot uh 0
to uh we if we try to increase the
voltage it will uh the capacity will
reduce to the C mean so most veror are
more commonly used than PN Junction uh
for a low voltage design so this is how
we can have
a vco voltage control oscillator and
this Moss uh capacitor has also it is
not a linear capacitor that we can see
it will not linearly change but because
of the C1 which is uh connected it try
it mostly try to add with the C1 and we
can have a oscillation which is a
linear so we if we try to get a CW which
is a V at voltage V1 at voltage V2 C
where 2 and at vdd by 2 this is a
voltage so for LC vco with white un
range vco with a uh continuous
tuning here the common mode level is
simply given by a gate Source voltage of
a diode connected transistor carrying a
current of ID by 2 so vgs 1 and 2 which
is equal to given by this
equation so he here what we can see is
uh we can also put the current Source at
the top
and uh try to uh we will try to make
sure that equal current will be
distributed on the both the
branch V using a capacitor coupling to a
VOR so here what what has been done this
uh there is a capacitor coupling C S1
and
CS2 and this
voltage of the
veror will be controlled means this VB
will be fixed and we can try to get the
uh capacitor uh change in the capacitor
by changing this voltage of VB or a v
control So to avoid the VOR
modulation due to the noise of the bias
current source that this bias current
Source
Noise We return to the tail bias
topology but so in the previous example
we have used the tail bias topology uh
to uh reduce the noise of the bias
current but what will happen employ AC
coupling between the VOR and the coal so
as to allow positive and negative
voltage across the
vors and what is the principal Tropic of
the circuit stems from the parasitics of
a coupling capacitor so this will also
add a par itics and that will try to
reduce the maximum capacitor so your
tunability will uh
reduce vco you we can use a uh using a
coupling capacitor to ere parasitic
capacitance to the substrate so here the
choice of Cs which is equal to 10 into C
Max so if we try to choose CS uh this
capacitor CS which should be higher
then reduces the capacitance range by
10% but introduce a substantial
parasitic at X and
Y or at P andq because integrated
capacitor suffers from the parasitic
capacit capacitance to the substrate so
we can see that uh metal six uh if
there's a capacitance a and this is a
capacitance B There are a various uh
parasitic capacitance to the substrate
so this
CB by CB by C AB for uh if your number
of layers 2 3 4 5 6 uh typically exceed
the
5% so that is why uh this uh does not
have a very wi tuning
range so we can use a discrete tuning
the What is the drawback it cannot have
a sufficient tuning range so how can we
improve the tuning by adding a discrete
tuning so in discrete tuning what we
will do uh we will put a more Capac one
one we will use for a fine tuning so
this is because it has a variation less
so this is used for a fine
tuning and this will be used for a
course
tuning
so this uh here you can see that for n
into CU so n number of capacitor we can
add so the first capacitor if it is
switched on and then with we control we
can change the uh oscillation frequency
again we can change the uh by tapping
the another capacitor on so in
application where substantially wider
tuning range is necessary we can use
this discrete tuning
method so uh to achieve a capacitance
range well beyond C Max by c mean that
we have seen the limitation of only
using a vector diode the lowest
frequency is obtained if all of the
capacitor are switched in and the vector
is at its maximum value so that Omega
minimum frequency is given by 1 by L1
into C1 + C Max into ncu where all will
be switched
on and maximum frequency we will get
when there is
a only C1 plus C Min so the highest
frequency occurs if the unit capacitor
are switched out and the VOR is at
minimum value so for uh tuning tuning uh
higher for tuning the course tuning we
will switch on the capacitor and then
fine tuning will be done by a v control
signal so this is how discrete tuning
can be done but there is some
disadvantage of this also it is it has a
gap of certain
frequencies that we
can have a variation in a fine tuning
range so if we expect Delta oscillation
to be greater than Delta oscillation 2
this Delta oscillation 1 uh to be
greater than Delta oscillation 2 so uh
this ncu switch into the tanks and the
vector sees a large constant capacitance
so Delta Omega oscillation 1 = to 1 by
L1 C1 into c maxus c uh mean by 2 C1 so
this is uh Delta Omega oscillator 2 if
we try to find out the difference Delta
Omega oscillator 1 by Delta Omega
oscillator 2 it will be given by this
equation so this variation in K weco
proves undesirable in the pl circuits so
we it is not uh linear and it it is not
good to use this kind of v in the
pl so what is the issue of a discrete
tuning it has a blind Zone and that
blind Zone should be removed so the
oscillator fails to recover the range of
Omega 2 and Omega 3 if we try to see
this for any combination of fine or
course control So to avoid the blind JNS
each of the two consecutive tuning
characteristics must have a overlap so
if we overlap this uh this tuning range
with this tuning range and we can get
the same frequency at these two
point so this pre uh to aoid the blind
Zone what we do is each of the two
consecutive tuning characteristics must
have some overlap this precaution
translates to a smaller unit capacitor
but a larger number of them enhance a
complex layout so your layout complexity
will increase
so this is uh all about all about the
voltage control oscillator which we can
uh which we have discussed now we will
try to see the
oscillator on the microwave
perspective and some of the widely used
oscillator which can be designed on the
uh the way of microwave circuits
so and uh how the BJT base oscillator
can be
design so if we try to uh consider any
small signal model of uh transistor and
if we connect one feedback Network which
is a admittance
y1 admittance
Y2 admittance Y3
so this is connected in this form at
the where the input conductance is of
the transistor either BJT or F is given
by GI and this is GM into V1 minus V2 uh
that is this voltage is one of the
voltage is a V1 and another is a
V2 and your collector or drain is a
denoted as a V4 and your emiter or
source is denoted as a
V2 so in order to have the
oscillation if this
uh admit uh if this parameter with uh
voltage V1 V2 V3 V4 equal to zero we
will get
the uh
oscillation so how this this is y1 + Y3
+ C which is so what is here it
is this is a y
parameter so if we want to write the KCl
or
KV we'll get
y1
plus this at this voltage y1 V1 voltage
that is y1
[Music]
Y3 and GI which is Multiplied with the
V1 so y1
+ Y3 +
GI multiply with
V1
plus there is a V2 voltage if we see at
V2 it is a minus
y1 so Min -
y1
minus
GI then that is
a uh
minus into
V2 Plus at
V3 that is
V1 V2 this is V2 and then V3 voltage
that is
this
Min -
Y3 into V3 plus there is
no 0 equal to 0 so this is what we have
written uh KCl at each node which is
equal to
zero now we will write uh another minus
y1 + minus y if we consider this node V2
then V2 by - y1 + CM + C I um y1 + GM or
this GM plus gi gi
I into y1 + Y2 y1 + Y2 g i
G+ GM into
V2
and uh
minus
Y2 - c - Y2 into V3 and minus C so
similarly we can write for all an equate
equal to zero what is your aim is to
find the value of Y3 y1 and Y2 to have
the
oscillation so oscillator if we use a
common AER BJT in this case so for
common Emer BJT your V2 is equal to
Z your V2 if we have common emiter if we
connect this with the
ground and if we connect a V3 to
V4 if we connect V3 to
V4 then output admittance is G which is
equal to zero you can write the we can
decompose that admittance Matrix in this
form uh that is V1 uh so you will have a
uh V3 V4 and uh which is and V1 because
V2 is equal to zero so that is denoted
by V so if we try to make it zero you
will get a y1 which is equal to J B1 Y2
which is equal to J B2 if we replace
here y1 is equal to J V V1 Y2 is equal
jb2 and Y3 is equal to jb3 we'll get
a uh g i Plus jb1 + B2 Min - J B3 GM
minus J B3 - J B2 + B3 is equal
to0 in order to have the oscillation
this 1 by V1 + 1X B2 + 1 by B3 is equal
to 0 and X1 + X2 + X3 is equal to 0 and
if we decompose 1 by B3 + 1 plus GM by g
i 1 by B2 = to 0 and X1 which is GM by g
into X2 so by this way we can get
the uh this is a condition to have the
oscillation so we can design a callit
oscillator for which X1 which is equal
to minus Omega not C1 we replace
this uh with the X1 then X2 which is
minus
Omega uh
C2 and X3 which is a Omega into
L3 so if you write it down - 1 by Omega
1 by C1 + 1 by C2 plus Omega L3 we'll
get Omega not
frequency and uh you you should have a
C2 by C1 which is equal to GM by g i so
if both if this condition is satisfied
you will get oscillation frequency this
so this is a widely known oscillator
which is a Pulpit oscillator by using
the
BJT another oscillator that is a heartly
oscillator which is Omega into L1 so X1
which is equal to Omega into L1 X2 which
is equal to Omega into L2 and X3 which
is a - 1 by Omega into C3 so we can make
a uh X1 X2 either capacitor or inductor
so in heartly oscillator it is a
inductor so in on chip design we
generally try to avoid this but off chip
we can use this configuration your L1 by
L2 which is equal to GM by g i and your
oscillation frequency is given by this
equation now we can use the same uh
method to find for the common gate f
so for common gate F in cpit
configuration this uh Omega not which is
equal to 1 by L3 C1 + C2 by C1 C2 and
this is a GM by G and for heartly
oscillator only the C and L is replaced
this is GM by G
not so in Practical uh there is a
resistance also involved and uh that
resistance uh will
uh add the because of this resistance
the Omega not frequency will deviate and
it is given by this equation and your C1
and relation of C1 Dash where C1 Das is
this and R by GI is given by this
equation so this is how heartly and cpit
oscillator we can
design another approach in microwave
oscillator design
is uh by considering
the reflection coefficient comma so here
if you see this is a resonator on the
left
side and uh it is a impedance J L and
there is a active device which is the
impedance of J
in and noise is added plus or minus ET
it is mentioned by ET and there is a a i
n which is a uh A I N Signal which is
transferred from this to this side a l
which is transfer from active devices to
this so you will have the gamma in which
is a uh noise which is
generated and Gamma L uh this is a a
node and this is the a in node we have
already seen the signal flow graph
in the signal flow graph if we want to
have oscillation uh we need to have a
negative resistance uh oscillator and
that we already know then in the last
lecture we have seen that negative
resistance is required in order to have
a
oscillation if we want to design
oscillator this is a model of
oscillator there is a
transistor and a terminating Network
which is a lossless uh here this is a
gamma out of the transistor the way we
have designed the amplifier while
discussing the amplifier similarly we
can also design the
oscillator here this is denoted because
we are terminating the network that is
denoted by the gamma T the total is a
gamma in reflection coefficient and
Gamma L you need to design a uh input
matching Network work in this case so
that it will sustain the oscillation so
this combination of gamma in and Gamma L
uh needs to have a uh satisfy the
oscillation so what is this negative
resistance model here this is a given by
JL Omega circuit negative uh resistance
and this is negative resistance added by
the uh active
circuit so
from active circuit we'll get a negative
resistance and this is a passive
resonator which is generating the uh
which is a reflection coefficient given
by the gamma l
so if we try to see in terms of the
impedance or
admittance uh same we can represent
gamma L and Gamma in so your total
current if we see in terms of the
voltage V LT means in terms of the
voltage VT Plus V int equal to VT is
equal to 0 and in this current if it is
a series resonance it is a we can
consider it as a voltage if it is
parallel it is I t+ I in which is equal
to I to equal to
zero so this is how we can use a
negative resistance model to have the
oscillation so we already have seen this
speedback model so transfer function uh
we V out by V in given by a j Omega by 1
minus beta J Omega A J Omega same thing
we can write L by a in so if you write a
L by a in which is equal to reflection
coefficient gamma in uh Omega into
a/ by
1us gamma L Omega into gamma in Omega
because uh if we try to see there is
only one Loop uh that is g GMA in by
gamma L and if we want to reach from if
you want to find the L by a in it is it
has only one path which is a gamma in
Omega a so one gamma in Omega a by 1
minus gamma L Omega a into gamma in
Omega a so in the absence of any input
signals for the oscillation to sustain
itself the denominator should be zero
and that we already have seen in uh in
this case if we have a denominator is
equal to0 you will have this is equal to
1
so uh this 1 - beta J Omega A J Omega uh
is equal to 1 so this is equivalent to
the expression on this GMA L into gamma
in is equal to uh 1 - GMA L into gamma
in Omega is equal to0 so you should have
a gamma L into gamma
in is equal to 1
so this is a condition to have the
oscillation so what is the condition the
two condition mention above can be
satisfied only you have a J in a minus
ZL and Omega the this already have seen
while discussing the one port oscillator
and where a and infinity are the stady
state amplitude and a frequency
respectively here uh which which is
further mention as a gamma L Omega gamma
in Omega a is equal to
1 however that for a less than a not
when the amplitude is less than a stable
amplitude again to provide so that R in
Omega a should be greater than RL Omega
so this is the uh different way of
representing the Baran criteria this is
the same Baran criteria but here we are
uh trying to see the Baran criteria in
terms of the reflection coefficient
gamma so hence to start up when a is
equal to0 your R in Omega is equal to0
should be greater than RL into Omega so
for start and similarly for a greater
than a the amplitude is more than the
stable amplitude so attenuation has to
be provided so for that your R in
Omega comma a should be less than RL
Omega so if we try to uh have the
graphical
analysis this is JL Omega which is
plotted and this is a minus Z in
a so if we try to plot the J in means
negative resistance Z in and the
jedel in the same plane impedance plane
so we can see that interestingly angle
uh we can find the angle measured
clockwise from the device line Arrow
Direction so if we try to find the angle
between JL and minus Z in this is a uh
this this angle is
unstable
because it has a u the face shift which
is very high and if we have a angle
Which is less than
180 that is this which will provide a
stable operating point so the
intersecting angle measured a clockwise
from the device line Arrow Direction set
by Jed in a for increasing a to the line
impedance Locus Arrow direction must be
less than 180° for the stable point in
order to have a stable
oscillation this is how uh we
can have a stable
oscillation so we'll discuss more about
the design of this uh kind of oscillator
uh in the next class thank
[Music]
[Applause]
[Music]
you
[Music]
a
