mod05lec25
0 Comments


Well so we learned about different types of
traffic components like filter couplers power dividers T junction now a special case mesh
termination So for example let us say you are using a branch line coupler you are utilising
3 ports Port P1 Port P2 and Port P3 but isolated port you are not using some so should you
keep it as it is If you keep it open circuited then there will be reflection and the properties
of branch line coupler it will change so Port 4 isolated port it should be mesh terminated
or we should connect 50 ohms load to Port P4 Similarly mesh termination we use in many
other situations also now at millimetre wave frequencies realisation of mesh termination
is really challenging because of the reactive contribution of different guiding stem means
bends or discontinuities So whenever we are using let us say a 50 ohms
register you can use for example 1 SMD resistor so 1 end you have to be you have to solder
it to your transmission line what about the other end It should be grounded so the 2nd
end when you are grounding obviously you will be using some metallic wire and metallic wire
it is one sort of discontinuity it will give you some inductance So you want to realise
a 50 ohm you have a 50 ohm SMD but when you connect it to circuit it is not just 50 ohms
it is 50 ohm with some additional reactant then how to obtain this mesh termination at
millimetre wave frequency At lower microwave frequencies this inductance value is so small
usually you neglect it but at millimetre wave frequencies we cannot so let me give you one
example and how we can compensate this inductive effect So in this example it is realised in uhh BCB
in 45 micrometres thick BCB substrate you can see the top line metal we have CPW line
and basically and uhh but this black rectangle it shows rectangular metallic wire that connects
the top line metal with bottom plane metal and you can identify the central conductor
Now for this BCB or in silicon or gallium arsenide technology how we how would we design
resistor we cannot use SMD We use nickel chromium alloy usually the (fab) lab they will give
you a standard thickness for the nickel chromium and standard resistance per centimetre then
you have to choose a proper dimensions so you know that resistance it depends on length
and uhh the width so accordingly you have to select your length and width to realise
a given resistance value So the in this picture this RW and R L they
show the length and width of that nickel chromium section we call it thin film resistor and
right hand side you can see one long black line and on that we have 3 wires instead of
one wire they are using 3 wires to reduce the inductive effect Now if I simply take
the R L and R W so that its equivalent DC (impede) DC resistance is 50 ohms then if
we measure it input impedance it becomes different because of the inductive wires So for example
typical return loss is inferior to 20 dB above 50 gigahertz and this variation is due to
ground part inductance due to the wires so what is the compensation technique then So
for this shunt inductance we can use shunt capacitance to compensate or nullify by their
effect how to realise then the shunt capacitance The from open stub from transmission line
theory we know that if we have place one open stub with microstrip line it will give some
shunt capacitance the same technique is used they are also now the thing is that you have
to calculate what is the equivalent inductance given by the grounded wire then you have to
calculate what is the equivalent capacitance you need and for that what is the equivalent
length of the open stub But instead of this calculation step practically what we do we
simply use a full wave simulator full wave later will give you the S parameter and then
you choose the width of your open stub and simply tune the length just one parameter
and if you (simu) simulation step you will get the desired result So in this picture you can see they are using
2 open stubs with S w and length S L they will give you the required shunt capacitance
So practically the resistance value it is chosen little lower than the required one
for example this particular example let us say the required value is 50 ohms then they
start it with 38 ohms uhh this nickel chromium resistor and if they need let us say 70 ohms
then they start with 61 ohms and now you see the result If you have inductive contribution
S 11 will be very poor now since they have compensated it at some given frequency let
us say at 50 gigahertz then for 50 ohms load it is showing this black line S 11 is 27 dB
So in reflection loss return loss is more than 27 dB so it is quite good and for 70
ohms design even it is better than uhh 40dB at the design frequency so it is quite good But without this compensation if you start
with just 50 ohms the loss value is below 10 dB it is quite high So these are narrowband
designs now wideband designs is also possible there are several techniques to make it wideband
I will show you an example so in this particular example they use another shunt arm how they
realise it So one single 50 ohms they are realising by using 2 shunt resistors nickel
chromium thin film resistors so one is grounded by using chain of wires and the 2nd one it
is connected to one open stub which will keep some shunt capacitance So without the shunt
arm the shunt additional arm of dimensions is given 250 micrometer by 60 micrometer which
will give you equivalent resistance of 158 ohms this is the DC resistance So without the shunt arm uhh it shows 71 plus
J30 ohms and 100 plus J53 ohms at 100 gigahertz and 160 gigahertz respectively And now with
this compensation so this is designed by using a full wave simulator the required length
width of nickel chromium thin film resistor and the open stub of this given 59 ohms then
it becomes meshed over Y band width You can see you here the simulated results over 0
to 160 gigahertz so in this simulation also we are considering a fabrication tolerance
of plus minus 10% dimensions now you see the result So left hand side it is showing S 11
so actual S 11 value for this design given by the solid line always below minus 20dB
so 25dB from 0 to almost 150 gigahertz Now considering plus minus 10 percent fabrication
tolerance even then this dash line and dotted line even then if you can say it is below
20dB over 0 to 160 gigahertz and now right hand side it is showing real and imaginary
part of input impedance as seen by this microstrip line So imaginary part it is approximately 0 you
can see this line you can follow this line it is within plus minus 5j and the real part
it is very close to 50 it varies between 50 to 60 so overall S 11 is below 20dB Now we
have discussed about different passive components and in some of the designs we are using different
types of wave guiding structures so not only that when we go for any millimetre wave systems
we in the same system different components can be (ma) it can use different types of
wave guiding structures Some components can use microstrip line components can use rectangular
waveguide some proponents can be designing in NRD now all of this we have to integrate
in one single system so obviously we need some adapters One microstrip line component
if I want to connect it with one rectangular waveguide system we need one microstrip to
rectangular waveguide transformer uhh for transition we call Now whenever we design transition we mainly
face two problems 1st one mode matching and the 2nd one is impedance matching For example
microstrip to rectangular waveguide in microstrip line it supports Quasi TEM we know the electric
field configuration but inside rectangular waveguide we use TE 10 mode it is transverse
selective mode which has different mode property So we have to somehow transform this Quasi
TEM mode into TE mode and not only that it should be matched from both microstrip feed
point and also rectangular feed point so let me give you some popular example of these
different types of transition involving different types of guiding structures at millimetre
wave frequencies The 1st example is coaxial to microstrip line
and co axial to CPW transition frequently be use for measurement purpose So in coaxial
cable if you recall that always we use TEM mode we never use TE or TM mode we want only
mono mode propagation So for TEM mode electric field it is from the central conductor to
outer conduct so they are perpendicular to magnetic field lines Now we want to connect
it to a microstrip line so coaxial line it supports TEM mode microstrip line it is also
quasi TEM mode and it also has electric field component which is particular to ground plane
so simply we can connect one coaxial line to microstrip line directly Coaxial line let
us say its impedance is given 50 ohms microstrip line impedance is also 50 ohms than directly
it can be connected Only thing is that the central conductor it
should connect the strip of this microstrip and the outer conductor it should be soldered
to ground and we use only one hub of this so that it can excite the microstrip line
because if I use both hubs of coaxial cable then they will nullify this the effect of
microstrip line field configuration so this is the structure showing the conducting parts
only Similarly we can also excite CPW line so for CPW line you see electric field it
starts from signal line and terminates into ground so we can again use this TEM mode of
coaxial line to excite Quasi TEM mode of CPW line let me draw the field diagram So for CPW line let us say this is the signal
line and this is the ground plane Now we have electric feed lines from signal line to ground
now if I compare these feed lines with that of a coaxial cable this is the central conductor
of coaxial cable and let us say this is the outer conductor of coaxial cable and inside
we have TEM field configuration so you see then simply we can use this part
of coaxial cable we can connect this part of coaxial cable directly to CPW line we will
be soldering the central conductor to the signal line and this outer conductor this
part to the ground plane and then this component of electric field it will excite this component
and this right hand side component it will excite this component So we do not have any problem due to field
matching and in both of these 2 different guiding systems we have Quasi TEM mode and
if the characteristic impedance of both CPW line and coaxial cable is 50 ohms in that
case we can directly connect it without using any impedance transformer Next coaxial cable
to waveguide how we can excite it Now inside coaxial cable we have TEM mode and inside
rectangular waveguide we have transverse electric mode In this case we have to use a mode converter
and this is done by using the flinching field of this coaxial line so inner conductor you
can see this is the cut way view uhh this is the side view so wave inside rectangular
waveguide is propagating from left to right and left hand side we are using a shorting
plane which lambda G by 4 away from the central conductor midpoint of central conductor On central conductor on top point it is soldered
on the broadside of of top broadside of this rectangular waveguide and then the flinching
whale inside the coaxial cable when it terminates it opens into rectangular waveguide we have
flinching field inside the rectangular waveguide and that flinching field it excites the TE
10 mode inside rectangular waveguide And because of this lambda G by 4 short termination we
do not have any wave propagation on left hand side you can say somewhat like uhh if any
wave propagates in left hand side then it will be reflected from short plane and they
will be it will be out of phase and in the front side we do not have any reflecting surface
so simply it will propagate So top view how it looks you can see only the central conductor
soldered on top plane and this dotted line it shows the slots in the bottom plane so
power is coupled to the slot in to rectangular waveguide We can also design coaxial cables to SIW transition
if you recall SIW is Substrate Integrated Waveguide so for SIW inside we have dielectric
and for coaxial again we have dielectric and otherwise all the principles are very similar
Usually it is a narrowband design physically typical bandwidth 5 to 10 percent we can improve
the bandwidth by using some matching pin so simply we use some metallic wire inside the
rectangular waveguide for matching purpose to improve the input reflection coefficient
and in that way we can cover a single band of 40 percent bandwidth I am showing you another example microstrip
line to rectangular waveguide excitation at millimetre wave frequencies So here what is
done the microstrip line geometry it is fabricated on a printed circuit board substrate so this
is showing the microstrip line geometry Just below the microstrip line in the ground plane
of the microstrip line we have one aperture we call it ground plane aperture so power
from microstrip line is coupled to waveguide through this ground plane approach so this
shaded part orange colour part it is showing the ground plane metal and this dark orange
colour part it is showing the top plane metallisation of printed circuit board So you can see the termination of this microstrip
line it is terminated into one open stub radials stub and before that just before ground plane
aperture we have 2 matching stubs opens open circuited stubs it is used for controlling
the input impedance Now how it is placed inside rectangular waveguide you can see this is
showing right hand side picture it is showing the side view so this ground plane aperture
it opens inside the rectangular waveguide and above just above the strip we have a back
cavity and the cavity hide it is lambda G by 4 but it is not very sensitive to this
lambda G by 4 just like like the previous one it can be little less or more than lambda
G by 4 Then right hand side you can see the microstrip line so from microstrip line power
is now coupled into rectangular waveguide And here it shows measured performance of
a back to back transition that means from microstrip line to rectangular waveguide again
from rectangular waveguide to microstrip line and this is the measured S parameter The 1st
example over 40 to 50 gigahertz band and you see the S 21 is quite good for these 2 transitions
including a section of rectangular waveguide so its mid band loss is just 2 to 3 dB for
2 transitions And the next measured example it is on quad substrate of the PCB it is from
90 to 100 gigahertz and over this band loss is below 4dB for these 2 transitions so similar
transitions are also possible using finite ground CPW line Now excitation of substrate integrated waveguide
from microstrip line from CPW line and from coaxial cables there are different techniques
reported in literature for millimetre wave applications so for all of these 9 designs
so the very basic thing two points one point we have to convert the mode and the 2nd point
is impedance matching For an example for this 1st design you can identify left side we have
one substrate integrated waveguide structure which supports TE 10 mode and right side we
have 50 ohms microstrip line So rightmost it starts with 50 ohms and they are designed
in one single substrate so that means single board single PCB is used for the same design
and now 50 ohms microstrip line is terminated into TE 10 for field configuration field matching
or more matching For microstrip line we have the perpendicular
field components inside rectangular waveguide again we have perpendicular field components
so we do not have any problem due to freely matching however we have problem due to impedance
matching so that is why we lead one impedance transformers so how it is done here you see
you can see a linear tapering used here usually the length of this linear tapering is lambda
G by 4 at the centre midpoint mid band of frequency and by using a full wave stimulator
simply you can tune the terminating width to obtain a better S 11 and S 21 usually this
design is wideband having more than 40 percent bandwidth So similar design is possible and
right hand side instead of open circuit they start with a short circuiting valve it somewhat
looks like how we feed rectangular patch antenna it is using some infant feeding And the rightmost one instead of just simply
inset it is actually capacitively coupled we are using uhh uhh one thinker type capacitive
coupling Next CPW to SIW transition so in all these examples you can see we are not
using conventional CPW usually we avoid it at millimetre wave frequencies because a simple
CPW line it sits on substrate dielectric substrate it will excite surface wave mode rather we
prefer boxed CPW or Brownback CPW line So right hand side you can see the wires of the
ground back CPW line and left hand side we have the SIW structure Now in SIW we have
perpendicular electric field components and right hand side for the CPW we have both microstrip
line mode and CPW mode For CPW mode you remember that it is inside the slot and it will be
then parallel to top plane and for the microstrip line mode it is perpendicular so in the same
direction as of uhh TE 10 mode So again we have good field matching the only
thing is that we have to obtain good impedance matching and that is being done here by using
short circuited slot termination of different shape In the next row this is example of coaxial
to SIW transition this 1st example actually we discussed for an angular waveguide you
can see the coaxial central conductor shown by this black dot which is lambda G by 4 away
from the shorting valve and 2 matching pins they are used to improve the input impedance
matching And instead of a shorting valve we can also use one open circuited valve In this
case you can say that the central conductor is directly soldered to one open circuited
end so we do not need any extension but again we are using similar matching wires here and
right hand side again a shorting valve but with different geometry for impedance matching Next is NRD NRD it can support both longitudinal
section magnetic and longitudinal section Electric In longitudinal section magnetic
mode in the propagation direction mainly we have the magnetic field components So for
LSM 11 if I concentrate on electric fields electric fields they are maximum on the central
plane and they are parallel to the ground plane If you recall we are using a dielectric
slab in between 2 ground planes so now if I want to excite this LSM 11 mode by using
a rectangular waveguide so right hand side it shows a TE 10 field configuration of rectangular
waveguide We have electric field perpendicular to broadside so simply then we can place a
rectangular waveguide on top of this so that electric they are in same direction so then
it shows the top view actually I am chewing a cut way view So uhh you you see LSM 11 mode it will have
parallel electric field components parallel to ground inside this and then you have to
turn this rectangular waveguide to have the electric field of the TE 10 mode in the same
direction we do not have any problem due to field matching Next problem is impedance matching
because left hand side we have transverse electric mode right hand side we have longitudinal
section magnetic mode So how it is done rectangle waveguide it is air filled so you see its
dimensions we are using a tapering slowly (())(29:01) and right hand side we have the
dielectric slab it is now extended into rectangular waveguide and again we are using a tapering
And by controlling the tapering length we can tune the input impedance now hard to excite
LSE 11 mode For longitudinal section electric if I look
at electric field configuration you see it is coming from left hand ground plane than
going into propagation direction From right hand side is coming then from on mid place
it is in propagation direction so we cannot simply place a rectangular waveguide on top
of it because we have both left hand side and right hand side components and they will
cancel each other away So we have to use a different scheme rather than electric field
we can excite this LSE 11 mode from an any aperture from the ground plane and using uhh
let us say magnetic field coupling Here this is the geometry so instead of TE 10 we are
again turning into have the TE 01 mode and left hand side it is terminated with a short
plane similar to go coaxial to rectangular waveguide it is lambda G by 4 away from this
uhh midpoint And you can see this rectangular waveguide
section and right hand side we have the NRD section now we have coupling through magnetic
field from uhh the rectangular waveguide to LSE mode of this NRD guide Again we can change
we can tune the tapering length for proper impedance matching So we have discussed about
the different types of passive components mesh termination and also transitions frequently
used at millimetre wave frequencies Next we will start some active devices and then its
applications typically electronic switch because electronic switches very popular for controlling
the performance of wireless systems and then some examples of millimetre wave systems so
for today let us stop here thank you

Leave a Reply

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