UNIT-1

RFM Measurements
Applications and Importance of
Microwave Measurements
 Specific uses of microwave instruments that are critical in modern
technology development are,
 Because of the severe demand on spectrum for commercial wireless
systems, it has become essential to reduce the guard bands between
channels to the bare minimum. This minimum is decided by the
phase noise in the carrier and by the selectivity of band-pass filters.
Techniques exist for accurate measurement of both, as we shall see.
 Electromagnetic interference (EMI)/electromagnetic
compatibility (EMC) measurements are critical today as
mandated by multiple internationally accepted standards.
Handheld spectrum and network analyzers have come to the market
in the past few years for such measurements in the field.
Overview of State-of-the-Art Microwave
Measurements
 State-of-the-art capabilities of common modern
microwave instruments include upper frequency limit,
frequency resolution, noise floor, amplitude accuracy, etc.
A few instruments that show impressive performance, and
yet are quite affordable, are shown here. Figure shows a
40 GHz spectrum analyzer, with up to 1 Hz resolution
bandwidth and displayed
 noise level below –130 dBm
above 1 MHz.
Vector network analyzer (VNA) with material
measurement setup
A digital storage oscilloscope up to 25 GHz.
S-Parameters and Related Black-Box
Representation
 It is well known that linear time-invariant circuits (radio
frequency (RF)/ microwave or not) can be completely
described by any of the multiport representations,
such as S-parameters, Z-parameters, etc.
 At frequencies greater than ∼100 MHz, the preferred approach
to blackbox representation is the use of S-parameters. There
are two main reasons for this.
 First, the S-parameters are the ones that are easy to measure
directly.
 Second, these are the parameters that give useful information
about the properties of components that are commonly used at
the higher frequencies
 For example, consider a low-pass filter with cutoff
frequency 1 GHz. Typical S-parameters of such a filter at
two frequencies (0.9 and 1.3 GHz) are

consider the Z-parameters of the same filter at the same

frequencies.
These are

It is not easy to relate these numbers to any aspect of filter

performance without further computations
S-Parameters
 Let us consider an N-port circuit. The definition
start by assuming that voltages and currents can be defined
at the ports, Let the voltages and currents at the ports be
denoted: V1, V2, … , VN and I1, I2, … , IN. Using the
usual frequency domain or phasor notation, these are
complex numbers that depend on frequency.

S-parameters of an N-port circuit.

 + and – signs for voltages and arrows are used for
currents
 The interpretation of these symbols is
1. The voltage is (potential on the terminal marked +) –
(potential on the terminal marked –). The choice of + and
– is in principle arbitrary and at the user’s discretion, but
in practical circuits, as we shall see soon, a convenient
choice is obvious.
2. The current direction is as shown by the arrows. Current
is taken to be flowing into the circuit on the terminal
marked +. Notice that having selected the terminals
marked + for the voltages, there is no further choice for
the current.
 Now the mathematical definition of S-parameters is
simple. Define the variables (complex) a1, … , aN and b1,
… , bN by,
Spectra of Commonly Encountered
Signals
 The mathematical theory behind the spectra normally
encountered is,
 1.Distribution of power at different frequencies in a power
signal
 2. Power spectral density for deterministic and random
signals
 3. Commonly encountered power spectral densities in the
case of deterministic signals
 4. Commonly encountered power spectral densities for
random signals
Microwave Filters and Directional
Couplers
 Microwave filters are essential components in most
measurement setups.
 Low-pass and band-pass filters are the ones commonly
used, while high pass and band-stop filters are rare in
measurement setups.
 Low-Pass Filter
 consider a low-pass filter from Mini-Circuits, part
number BLP-450+.
A low-pass filter with BNC co-axial
ports.
 The important features are:
 The 3 dB cutoff frequency is 440 MHz, and below 360
MHz the insertion loss (defined as –20 log |S21|) is below
0.5 dB, and above 565 MHz, the insertion loss is >30 dB.
While there is no uniform convention as to how pass-band
and stop-band are defined, there is no doubt that 0–360
MHz belongs to the pass-band, and frequencies
above 565 MHz belong to the stop-band.
A band-pass filter with SMA co-axial ports.
Here the pass-band is small, and the group delay
is not of much concern. The major issue with
this filter is the insertion loss going down from
4 to 5 GHz, and this is likely to give a second
pass-band around 6 GHz. This can be corrected
by again cascading with a low-pass filter with
cutoff around 5 GHz. Finally, note that the
insertion loss here is -3 dB, which is very high
compared to the insertion loss of the low-pass
filter. filter
This is a common feature of
resonant circuits, and the narrower the
fractional bandwidth (higher Q), the
more is the loss due to parasitic
resistance in the inductors and
capacitors used to realize the filter.
Directional Couplers-four-port
component (Passive and ideally lossless)
 S-parameters have the following properties:
 1. The diagonal elements are 0.
 2. Each column has an additional 0 (other than the
diagonal).
 3. The magnitude of a nonzero element in any column is
either C (the coupling) or T (the through-port transmission
or main line loss).
 The number C (the coupling) is the main specification of
the coupler.
 Usually it is specified in dB: coupling = –20 log10 |C|.
 Ideally, |C|2 + |T|2 = 1, but in practice there are losses,
and |C|2 + |T|2 < 1.
 Two other parameters are also used in describing couplers:
 1. Isolation:
 The off-diagonal element in any row that is supposed to
be 0 will not be 0 in practice. If its value is some I, then –
20 log10 |I|
is called isolation. Ideally, this should be ∞, but its
minimum value in practice (specially over the band of
operation) should be significantly higher (in dB) than the
coupling.
 2. Directivity: This is isolation coupling. In practice this
should be at least 10 dB, and most vendors offer products
with better than 20 dB directivity.
 A directional coupler with co-axial ports.
1. The input should only be
at port 1.
2. Ports 2 and 3 should see
close to 50 Ω impedances.
For measurement purposes,
couplers are mostly used
under these conditions.
With the above numbering,
we get:
S21 = T, S31 = C and, by
symmetry, S32 = S41 = I
A waveguide directional coupler
Microwave Attenuators
 The attenuator is perhaps the simplest microwave
component. It is a two port component and is simply
described by the S-parameters: S11 = S22 = 0;
S21 = S12 = A. A is a real positive number less than 1,
and (–20 log10 A) is the attenuation factor. For a 6 dB
attenuator, A = 0.5. For a 20 dB attenuator A = 0.1.
 An attenuator with fixed A can be easily realized with a
three-resistor T or pi network.
A microwave SPDT switch using two
FETs and an inverter.
Conceptual schematic of a digitally variable
attenuator
Today, attenuators are mostly realized using FETs. The Figure shows a basic
architecture of a 6-bit stepped attenuator which can give up to 31.5 dB
attenuation in steps of 0.5 dB. It uses six resistive attenuators with the given
attenuation factors and 12 (2 for each resistive attenuator)SPDT switches.
Input is at RF1 and output at RF2.
Microwave Connectors and Adaptors
 The most commonly used microwave co-axial connector
is the SMA. This is usable until 18 GHz. The diameter
of the Teflon dielectric is around 4.5 mm. To understand
the terms inner conductor and outer conductor, the
electric field is present between the inner and outer
conductors in a co-axial cable.
(a) SMA male connector. (b) SMA female
connector.(Usable until 18GHz)
The diameter of the Teflon dielectric is around 4.5 mm.
A 3.5 mm co-axial adaptor.
The SMA connector is compatible with a
slightly better connector called the 3.5 mm
series.Here the dielectric is air and the
diameter of the outer co-axial conductor is
3.5 mm. This is shown in Figure. This type
is usable until 26.5 GHz. The SMA and 3.5
mm types are roughly compatible with the
2.9 mm or K connector. However, repeated
connections between these types may
damage the K connector, and this is not
recommended. A picture is shown in Figure.
Here the outer conductor diameter is 2.9
mm. This works spurious mode free until
40 GHz.
Traditional Measurement Techniques
 The Power Meter
 Measuring the strength of a microwave signal is the starting
point for microwave measurements-strength of a signal is
power available from a source.
 power available from a source (with a finite source impedance)
is defined as the maximum power that can be dissipated in an
appropriate load connected to the source. The maximum power
transfer theorem tells us that “appropriate” here means
conjugate matched. For such a matched load, half of the power
drawn from the voltage source is dissipated in the load, while
an equal power is dissipated in the series impedance.
 Usually measuring instruments are designed to have an
input impedance of a standard value (50 Ω is most
common), and it is assumed that the source impedance is
also of this same value, but these are never very accurate.
Moreover, the instruments are normally calibrated to
display the power absorbed in them, so the available
power is the same as the displayed value only if there is a
true impedance match.
 The component used for this measurement is a power detector,
which essentially converts a microwave signal to a more easily
measured quantity— usually a constant voltage. There are three
types of such detectors that are,
 the thermocouple-based, thermistor-based, and
 diode-based types.
The thermocouple and thermistor sensors cover a range from
approximately 1 μW to almost 1 W, while the diode sensors
cover roughly from 0.1 nW to around 100 mW. Higher ranges
are available using attenuators, while amplifiers (and
heterodyning) can sometimes be used for measuring even
lower values.
 Traditionally the sensor has been assembled as one unit, while
the rest of the circuits, the display, and peripheral components
(such as power supply) have been assembled as another unit.
Thermocouple-Based Power Detector
 If two different conductors form a junction that is heated while the
other ends of the two conductors are kept cold, a potential difference
develops between the two cold ends. The potential difference is
roughly proportional to the temperature difference between the hot
junction and the cold ends. For practical use, at least one of the
conductors should not have a very good conductivity (i.e., a copper
+ aluminum thermocouple is not preferred).
 This is because the whole effect depends on the generation of
additional free carriers (electrons) due to heat, and in metals with a
very large free carrier concentration to start with, heat does not alter
the number substantially.
 Additionally, if the heat is generated by dissipating a microwave
signal in a resistance located at the junction, the rise in temperature
(and hence the potential difference) is proportional to the power of
the signal. This is the basis for the thermocouple-based power
detector.
 For making a practical device,
 1.The microwave signal to be dissipated as heat should
see a matched load (conventionally 50 Ω) throughout the
band of operation, which should be very wide.
 2. The generated dc potential should be isolated from the
microwave signal.
 3. Circuits should be implemented to measure the dc
potential accurately, even when it is extremely small
The microwave signal reaches
the junction through the gold
contacts, and since the contact
area is very narrow (n+
silicon and TaN) compared to
the rest of the conductors
(gold and n+ silicon, gold
and TaN), most of the heat is
dissipated across n+ silicon
and the TaN junction
Different layers in the complete thermocouple structure.
Building blocks of a thermocouple-based power meter
Circuit using the thermocouples
A simple chopper using FETs.
 The chopper effectively connects the input to the output when
the trigger is high and gives 0 output when the trigger is low.
 It is quite possible that there may be nonidealities in the circuit,
and what we get at the output when we expect 0 is not quite 0,
the input voltage at the output, again, it may not be quite the
input voltage.
 But the difference between the two levels will be proportional
to the input signal. The output of the chopper can be amplified
by ac-coupled amplifiers (the amplifiers before the detector are
of this type), and since there is no question of dc drift now,
these amplifiers can have ample gain, in order to boost the
weak dc signal to a desired level. The synchronous detector
reverses the operation of the chopper. This gives back a dc that
is equal to the amplitude of the input ac (square wave). The dc
signal can be further amplified if required before the display.
The display will of course be calibrated suitably to display the
power absorbed by the sensor.
Transmission Measurement
 Measurement of |S21| of a two-port component
 One basic measurement is of gain, which roughly relates to the
ratio of output to input. measurement of |S21| of a two-port
component that is assumed to be linear and time invariant. The
first point that should be clear is the impedance value with
which the S-parameters were defined. If Rs = RL = R0, then P1
= |b2|2/(2RL) and P0 = |a1|2/(2RL) = |Vs|2/(8RL), which
gives

If the device under test (DUT) is a series capacitor of value 1 pF,

then with R0 = 50 Ω, the true value of |S21| is 0.53, while the value measured
by using eqn.if Rs = 20 Ω and RL = 100 Ω is 0.6, and for Rs = 40 Ω and
RL = 60 Ω it is 0.532. So, the simple approach illustrated in Figure 3.18 is
indeed very effective. To further enhance the accuracy, when Rs and RL are
significantly different from R0, attenuators are used.
 If a load is connected to port 2 of this attenuator, and we
try to estimate the reflection coefficient looking into port
1, the incident wave will be attenuated by A before it
reaches the load, and the reflected wave from the load will
be again attenuated by A before it emerges from port 1 as
the reflected wave.
 So, the reflection coefficient of the load is reduced by a
factor A2. Another way of putting this is that the
impedance looking into port 1 is close to R0, whatever the
load. This is exactly the behavior we want. But leads to
the loss of power in the attenuator; if the DUT has a very
low |S21|, then adding attenuators may reduce the value
of P1 below what can be measured with the available
power meter. This is the main limitation of this
measurement