Well so before discussing about other challenges
lets see why at millimetre wave frequency power handling capability decreases And why
loss increases at millimetre wave frequency So there are many sources but because of the
reduced size of the wave guide or the wave guiding structure umm we will see that power
handling capability and the loss it increases So lets consider a rectangular wave guide
So this is the cross sectional of a rectangular wave guide lets say it is supporting TE ten
mode so for this TE ten mode if I plot the electric field strength from left to right
side So this is the electric field configuration
for TE ten mode so it is maximum on the mid plane and then if I go left or right side
the filstunt decreases Now I am considering one more rectangular wave guide but its thickness
is much smaller compare to this and it is also supporting similar mode TE ten mode so
we have perpendicular electric field components which varies simultaneously along its width Now the difference between one and two is
that only thickness it changes its width remain same so if I sent now one watt of power through
rectangular wave guide one and the same power through rectangular wave guide two so now
if I calculate the electric field values so obviously in the first case we have much smaller
electric field values and whatever current we have maximum current on the mid plane on
metal surface So obviously that will be much higher for
case two So since the current component is higher we are expecting more metal loss from
case two so not only that electric field value as well as the loss both are increasing so
power handling capability it will decrease for case two as well as loss will increase
for case two So this is one umm conceptual explanation why loss increases with reduced
height of with the decreasing cross sectional area
If I consider a micro strip line for example it showing a cross sectional view of a micro
strip line so this is the top metallic strip and this is the bottom metal which we called
the ground and in between we have dielectric material so in this case the strip width is
much smaller compare to lambda g so the energy electromagnetic energy mainly it lies just
below this strip and if I draw the electric field it looks like this So inside the dielectric I am using a small
area for energy propagation so again if I send one watt of power through micro strip
line what do you expect compare to rectangular wave guide in this case we will be having
higher electric field value and higher surface current density So power handling capability
will decrease further and loss will increase further so if I design this rectangular wave
guide now lets say at millimetre wave frequency so you know that the diemension along X it
depends on wave length for TE ten mode at start up frequency this width it is approximately
lambda not by two So if that is the case then if I design one
rectangular wave guide lets say at six gigahertz and FC is this is equal to six gigahertz case
one And in the second case FC is sixty gigahertz so what we expect for the second case the
broad side dimension of the wave guide it ten times smaller than the first case So it will provide high loss and not only
that power handling capability will decrease for the second case so the conclusion is then
the if I decrease the cross sectional area of any guiding structure in that case we will
have higher loss and lower power handling capability So now lets go back to the challenges Next
is antenna it can be resonating antenna or it can be travelling wave type antenna So
whatever we design at millimetre wave frequency it will give you very low efficiency Because
whatever power you are going to feed to your antenna so a part of that will be radiating
and another rest part it will be absorb in the antenna structure itself which is due
to the loss So as seen that millimetre wave frequencies
loss increases antenna efficiency will decrease or radiation efficiency will decrease so its
a problem actually at millimetre wave frequency so typical antenna efficiency it can be as
small as ten percent to twenty percent so that means whatever power we are feeding to
antenna only ten percent is being radiating to space So not only that if we go for on chip antenna
design lets say we are designing the antenna on silicon or silicon dielectric constant
is very high almost twelve so if you just place one resonating element on silicon so
instead of radiation in space almost all of the power will be absorb in silicon the effect
is due to surface wave generation we will see later what is surface wave So this type of problem will be facing at
millimetre wave frequencies so we have to then use some alternative ways how to reduce
this losses How to improve the gain of the antenna Next is the filtering function so
filter typically use low loss resonators and since we already discussed that if we really
on low loss resonator we don’t have any other option than increasing the volume So for small size application then filter
design of filter its a challenge Next RF electronics part so loss its increases system noise figure
so what is noise figure Okay so let me show you by using some diagram this is another
important quantity noise figure Lets say we have a two port network it can
be anything it can be a filter it can be a power amplifier it can be any two port network
This is port one this is port two So its receiving some signal at port Poneand then the signal
whatever we are collecting at port two So lets say the SNR at port one its given by
S umm sig signal to noise ratio at port one its given by SNR i and the signal to noise
ratio at port two its given by SNR O so now if I take the ratio SNR at the input and SNR at the output usually
we represent it in decibel scale we call it the noise figure So usually the noise figure
it is represented in decibel scale ten log ten SNR at the input divided by SNR at the
output in DB So for a lossy system for any given passive network lets say filter it will
increase the noise figure why Because whatever signal is coming at the input port due to
the loss the signal component will be attenuated further And at the same time this component will add
some noise so at the output total noise contribution is increasing but total signal strength is
decreasing so as a result what we expect that noise figure will increase so at millimetre
wave frequency this term noise figure is very popular by these we can understand umm how
noisy the component is It may be any active component like power
amplifier low noise amplifier mixer or it can be also passive element like filter coupler
So for passive element usually the noise figure this is equal to the insertion loss insertion
loss of that component for passive element so insertion loss is nothing but a Stwoone
in DB So lets go back so then we see that at millimetre
wave frequency loss increases so it increases the noise figure of the component So one way
to decrease the no overall noise figure of any transmitter or receiver is that a first
element of the chain of the component of the lets say for receiver if I start from antenna
Just after antenna we have low noise amplifier then mixer component so the first element
should provide minimum noise figure and it should be placed at close to antenna as possible So that we can decrease the overall noise
figure of the system and finally the system must be cost effective because if we use group
three group five semi conductor materials like gallium arsenide to fabricate millimetre
wave system it is not cost effective So thats why till now it is popular only with the defence
sector who can afford this expensive system So if we really want any consumer application
we have to find out low cost solution and that is possible if we fabricate the same
system using silicon germanium technology of CMOS by CMOS technology So already some
of the university group and some research lab they did it at sixty gigahertz using silicon
technology So next is packaging another important issue
So whatever components we are designing we have to put it under some packaging to protect
it from severe weather condition and we have different packaging systems one is called
multi chip module So where basically different chips or different functions are integrated
together in one single chip and from and then it is put under one single package So from outside you cant understand that inside
there are many chips So this MCM or multi chip modules it can be fabricated using low
temperature co-fired ceramic technology or more popularly known as LTCC using microwave
chip on flex so basically its a one type of flexible laminates also we can use conventional
printed circuit board technology or laminates to integrate this multiple chips and next
then how to umm connect the chip with the other component So for that we have to use some sort of bonding
so wire bonding or it can be flip chip mount we will see it later At millimetre wave frequency
usually wire bonding is lossy flip chip mount is less lossy but it has some additional problem
of power leakage and coupling and generation of surface wave modes So this is showing a typical scenario one
on the chip antenna you can see the copper part fabricated on silicon but the problem
is that if we directly fabricate one antenna on silicon most probably it is not going to
radiate anything So almost all of this power umm it will be absorbed by the silicon So then how we can make it radiate so one
alternative solution is that use one AR cavity just below this antenna structure so this
right hand side picture actually showing that AR cavity and this antenna then it is placed
on the AR cavity so now it will radiate into space so this is one solution how we can improve
the gain of the antenna so now the second picture this is showing a silicon based low
cost gigahertz transceiver So this antenna now it is being integrated
in the single chip where we have other components like low noise amplifier and switching and
everything all are millimetre wave components so when we integrate antenna inside the same
chip we have to take care of many factor for example this antenna is resonating structure
it has someelectric field in silicon and air so it can interact with other metallic part
of this chip itself So then antenna it should be placed as far
as possible from the chip not only that if there is any metal part on this chip itself
then we have to use some sort of insulator in between antenna and the chip part or simply
we have to increase the separation between antenna and this chip part Now if I look at this chip then umm the passive
components typically the inductors it occupies very large area and usually they are fabricated
on the top metal layer of the chip so the coupling between the inductors and the antenna
it is expected to be maximum so sometimes thats why to avoid that coupling between the
inductor and the antenna we use some sort of guard ring around the antenna structure
which will minimize the coupling between antenna and the inductors So you can also see this is since umm silicon
umm substrate so they are using that AR cavity so this is the radiator and just below it
we have this four hundred micrometre thick AR cavity so it just below the AR cavity we
have antenna reflector cavity this is just metallisation on the cavity wall so which
behaves as a reflector and further increases the gain Now lets say we are going for a practical
design We want to design some millimetre wave components at using some printed circuit board
technology so for that we have to choose a printed circuit board first What is printed
circuit board It is nothing but a dielectric slab and its comes with a top and bottom coppermetal So if you take the cross section the dielectric
layer it is sandwiched between two metal layer and what is the thickness of dielectric we
have some standard thickness available and just if you companies actually make this substrate
we call it substrate and some standard dielectric constant also available so one of the popular
company name is Rogers corporation form USA So if you go through Rogers website you can
see that its coming with some standard thickness as well as with some standard dielectric constant
so for example if I consider a typical substrate its name is RT duroid five eight eight zero
its dielectric constant is two point two and if you go through Rogers website you will
find out some typical thickness like 5mil one mil equal to one thousand of umm one inch So ten mil twenty mil thirty two mil some
standard thickness are only available you don’t have any value in between so you have
to choose first your substrate So how to choose your substrate first So we have to keep may
factors many parameters in mind when we are going to choose your substrate so for example
if I increase dielectric constant of the substrate what will happen The components size will
decrease Why Already we discussed that lambda g this is
the call to approximately lambda not by root epsilon R so if I design a resonator lambda
g by two length at least say in a substrate with dielectric constant ten point two and
we have one more substrate of dielectric constant two point two so obviously it will be of small
size in ten point two dielectric constant substrate So then obviously we have some advantage
we can umm miniaturise component But at millimetre wave frequency already the
components size is small and not only that if we go for miniaturisation we face one more
problem What is that Loss increases as well as power handling capability decreases Different
substrate provides different loss value that loss we want we we quantify by using a parameter
we call it lost tangent of the substrate so we will see later So loss tangent that is another important
factor for any given substrate higher at the loss tangencies higher the dielectric losses
so when you are choosing your substrate for your application you have to keep all this
points in mind and now lets visit through Rogers umm page and find out how the specifications
are given So for example I am considering a typical
RT duroid five eight seven zero substrate and RT duroid five eight eight zero substrate
so this page I am showing from Rogers corporations website this is the data sheet so you can
see dielectric constant epsilon R some suggested values are given they are calling it the design
value For five eight seven zero its two point three three for five eight eight zero it is
two point two And not only that if I go to right the suggested
frequency range is eight gigahertz to forty gigahertz so they are specifying this dielectric
constant value for design from eight to forty gigahertz if I increase the frequency further
so that means dielectric constant may vary it may not be fixed at two point three three
or two point two So similarly the dissipation factor or loss tangent another name is tan
delta so you see for five eight seven zero they are suggesting two values point zero
zero zero five this is at one megahertz and below point zero zero one two this is at ten
gigahertz for five eight eight zero these values are point zero zero zero four and point
zero zero zero nine So you see as the frequency increases what
they are showing the measured loss tangent values they are increasing but we will see
theoretically loss tangent should take place but practically if I look at the measured
value loss tangent is increasing with frequencies There are many other factors like what is
the thermal co efficient of epsilon R so epsilon R its changing with frequency not only that
epsilon R its also a function of temperature consider any space based application In space temperature can be as low as minus
sixty minus seventy degree centigrade or if any component is under constant sunlight illumination
so the temperature can be as high as hundred degree or even more than that we have a wide
range of temperature this component should support is variation So dielectric constant whatever we are using
to design that components then it should not vary with temperature in fact there are only
few umm space certified components umm for the space certified components the dielectric
constant or whatever change in dielectric constant its really small negligibly small So depending on application you have to choose
your substrate very carefully for example if you go for any space based application
you have to only choose space certified umm substrate So with this some there are additional
parameters like what is the young modulus of your substrate What is the vapour absorption
co efficient of your substrate because if the dielectric it absorbs vapour so water
is very lossy and if it absorbs vapour dielectric substrate it will becomes very lossy So there are many other factors so we have
to keep in mind So in the next part we will see what are the different material properties
at millimetre wave frequencies and not only that even at terahertz frequencies So sometime
we use the same material for microwave frequency designs circuit design And as well as for the millimetre wave frequency
design but then since already we know that dielectric constant its a function of frequency
then how are these material behaves at millimetre wave or sub millimetre wave frequency Okay
so we will take a short break then we start again

Leave a Reply

Your email address will not be published. Required fields are marked *