EE 407

Microwave Engineering

Lecture 15 to 16
Designing/Simulating/Testing
Microwave filters
Dr. Sheikh Sharif Iqbal
References: Text books
• Filters are two-port networks used to control the frequency response in an RF
or microwave system by allowing transmission at frequencies within the pass-
band of the filter, and attenuation within the stop-band of the filter.
• Common filter responses include low-pass, high-pass, band-pass, band-stop.
• Filters are indispensable components in wireless systems, used in receivers for
rejecting signals outside the operating band, attenuating undesired mixer
products, and for setting the IF bandwidth of the receiver. In transmitters,
filters are used to control the spurious responses of up-converting mixers, to
select the desired sidebands, and to limit the bandwidth of the radiated signal.
• Insertion loss method is a flexible & accurate way to design & analyze filters
having a specific type of frequency response. It is based on network synthesis
techniques & used in the design process of low-pass filter prototype with the
assumption that it is normalized in terms of impedance and cutoff frequency.
• Impedance and frequency scaling and transformations are then used to convert
the normalized design to the one having the desired frequency response, cutoff
frequency, and impedance level.
• Finally, Transmission line stubs (with dimensions computed using Richard's
transformation, impedance/admittance inverters, and the Kuroda identities)
OR Stepped-Impedance techniques are used to facilitate filter implementation
in terms of practical microwave components.
• Three types of practical filter responses are considered here:
• Binomial or Butterworth or Maximally flat response: can cater the need for
minimum insertion loss by providing flattest possible pass-band response.
• Equal-ripple or Chebyshev response: can satisfy a requirement for the
sharpest cutoff region.
• Linear Phase response: is important in multiplexing filters used in comm.
systems to avoid distortion at the cost of filters sharp-cutoff characteristics.
• Insertion loss or Power Loss Ratio (PLR): by definition PLR is given by;

(eq. 1) where,
Γ(ω) is the ref. coeff.
looking into the filter
Thus, the insertion loss (IL) is :
PLR of Maximally flat Butterworth low-pass filter: By definition, power
- N is the order of the filter,
-loss ratio is; (eq. 2)
- ωC is the cutoff frequency

• Note at ω>>ωC ; IL increase 20N dB per decade (*10) increase of frequency

• Maximally-flat: as first (2N -I) derivatives of above equation are zero at ω=0.

Maximally flat Low-pass filter Prototype: For a normalized two-element

(N=2) low-pass filter (of fig.), the input impedance & reflection coefficient are;

Fig.1
Thus from eq.1, the PLR of this circuit is written as;
Since,

We have,

(eq. 3)

• If LP filter of eq.2 has two elements (N=2) and is normalized (ZS=1Ω &
ωC=1): (eq. 2a) (as at ω=ωC, IL=10log(1+k2) Î 3dB, Thus k=1)

• From eq.(2a): PLR=1 at ω=0 ; substituting this in eq.3 gives Î R=1.

Also in eq. 2a, coefficient of ω2 is 0; equating this to coeff in eq.3 gives,
Thus, L=C. Similarly, the coefficient of
ω4 in eq. 2a is 1; equating this to eq.3 ¼(L2C2)=1. Thus, L=C=√2
• For filters with large ‘N’, this method to obtain ‘L’ & ‘C’ is impractical
• Thus, normalized LPF’s below are designed using given element values
(see Table 1)

Fig 2a:Circuit
begins with
shunt element

Fig 2b:Circuit
begins with
series element

generator resistance for the network of Fig 2a.

where: g0 = generator conductance for the network of Fig 2b.
inductance for series inductors
gk =
k=1 to N capacitor for shunt capacitors
load resistance if gN is a shunt capacitor
gN +1 =
load conductance if gN is a series inductor
• In previous fig’s: maximally flat LPF’s can have ‘N’ reactive elements
from ‘g1 to gN’. Elements values for N=1 to 10 are tabulated below.
• In last figure, ‘g0=1=generator impedance’ and ‘gN+1=load impedance’.

Table 1: Element (‘L’ &‘C’) Values for Normalized (g0=1 & ωC=1)
Maximally Flat (buterworth) Low-Pass Filter Prototype :
N g1 g2 g3 g4 g5 g6 g7 g8 g9 g10
1 2.0000 1.000
2 1.4142 1.414 1.000
3 1.0000 2.000 1.000 1.000
4 0.7654 1.848 1.848 0.765 1.000
5 0.6180 1.618 2.000 1.618 0.618 1.000
6 0.5176 1.414 1.932 1.932 1.414 0.518 1.000
7 0.4450 1.24 7 1.802 2.000 1.802 1.247 0.445 1.000
8 0.3902 1.111 1.663 1.962 1.962 1.663 1.111 0.390 1.000
9 0.3473 ].000 1.532 1.879 2.000 1.880 1.532 1.000 0.347 1.000
10 0.3129 0.908 1.414 1.782 1.975 1.975 1.782 1.414 0.908 0.313
But if the filter circuit is not given and needed to be designed:
• Order ‘N’ of a filter is dictated by given insertion-loss (attenuation) at some
frequency and using figure below, values of ‘N’ can be determined (for N<10)

Figure 3
Example 1: A maximally-flat low-pass filter is to be designed with a
cut-off frequency of 2 GHz & a minimum attenuation of 15dB at 3GHz
(a) Find the number (‘N’) of required filter elements.
(b) Find the un-scaled values of the filter reactive elements.
Sol.: Given, f=3 GHz and fC=2 GHz. Thus, |ω/ωC|-1=0.5. From fig 3
we get, N ≥ 5 to achieve minimum attenuation of 15dB (see red line)
(b) From Table 1, the un-scaled values of the filter reactive elements are;
g1 = 0.6180; g2 = 1.618; g3 = 2.000; g4 = 1.618; g5 = 0.618;
This filter circuit with five reactive elements are similar to Fig 2(a) or (b)
Scaling Designed Low pass Filter Prototypes:
• Scaling convert normalized filters to filters with prescribed impedance & ωC
• Impedance scaling: is obtained by multiplying proto-type impedances by required
source resistance (R0). Thus, RS/ = R0, RL/ = RL·R0, L/ = L·R0, C/ = C/R0,
where RS, RL , L, C are normalized values and RS/ , RL/ , L/ , C/ are scaled values.
• Frequency scaling: convert cut-off frequency of prototype from unity to required
value of ωC. Thus, scaled element values become; Lk/ = Lk/ωC and Ck/ = Ck/ωC
• For both Impedance & Frequency scaling: Lk/ = (R0·Lk)/ωC & Ck/ = Ck/ (R0· ωC)
( k=1,2,3… ) …………. Eq. 4
Example 2: If the maximally-flat LP filter has an impedance of 50, then
(1) find the scaled value of the reactive elements. (2) Draw the circuit.
Sol.: For the type of filter shown in Fig 2a; eq.4 & example1 sol. gives,
C1/ =0.984 pF, L2/ =6.438 nH, C3/ =3.183 pF, L4/ =6.438 nH, C5/ =0.984 pF

Thus,

Fig 4

• Same technique is used to obtain elements for the type of filter in fig 2b
PLR of Equal-ripple Chebyshev low-pass filter: By definition, power-
-loss ratio is; (eq. 5) - TN(x) is Chebyshev
polynomial of order ‘N’
• For normalized filters, ωC=1; where, T1(x) = x,
Thus for 2 element filter in fig 1; T2(ω)=2ω2 -1 T2(x) = 2x2-1,
T3(x) = 4x3-3x,
Eq.5 Î (eq. 6) T4(x) = 8x4 -8x2+1,
• Again for equal-ripple filter, using eq.6 & eq.3 lets find ripple level (k2)
If (eq. 3) is,
for even ‘N’& at ω=0, eq.3 & 6 gives ripple level, k2=(1-R)2/(4R) (eq. 7)
• Using eq’s 3, 6 & 7, we can find the two elements of the filter in Fig.1.
• For large ‘N’, given element tables are used to design LPF prototypes:
N g1 g2 g3 g4 g5 g6 g7 g8 g9 g10
1 0.699 1.000
Table 2:
2 1.403 0.707 1.984
For equal-ripple LFP
3 1.596 1.097 1.596 1.000
with 0.5 dB Ripple &
4 1.670 1.193 2.366 0.842 1.984 g0=1 & ωC=1
5 1.706 1.230 2.541 1.230 1.706 1.000
6 1.725 1.248 2.606 1.314 2.476 0.870 1.984
7 1.737 1.258 2.638 1.344 2.638 1.258 1.737 1.000
8 1.745 1.265 2.656 1.359 2.696 1.339 2.509 0.880 1.984
9 1.751 1.269 2.668 1.367 2. 724 1.367 2.668 1.269 1.750 1.000
10 1.754 1.272 2.675 1.373 2. 739 1.381 2.723 1.349 2.524 0.884
Table 3: Element (‘L’ &‘C’) Values for Normalized (g0=1 & ωC=1)
equal-ripple Low-Pass Filter Prototype (with 3.0 dB ripple)
N g1 g2 g3 g4 g5 g6 g7 g8 g9 g10
1 1.995 1.000
2 3.101 0.534 5.810
3 3.349 0.712 3.349 1.000
4 3.439 0.748 4.347 0.592 5.810
5 3.482 0.762 4.538 0.762 3.482 1.000
6 3.504 o. 769 4.606 0.793 4.464 0.603 5.810
7 3.518 0.772 4.639 0.84 4.639 0.772 3.518 1.000
8 3.528 0.77 5 4.658 0.809 4.699 0.802 4.499 0.607 5.809
9 3.534 0.776 4.669 0.812 4. 727 0.812 4.669 0.776 3.534 1.000
10 3.538 0.777 4.677 0.814 4.743 0.816 4.726 0.805 4.514 0.609

• Again for equal-ripple filter; the order ‘N’, dictated by given insertion-loss
(attenuation) at some frequency, is found using next two figures (for N<10)
For designing equal-ripple LP Filter Prototype (with 0.5 dB ripple):

Figure 5
For designing equal-ripple LP Filter Prototype (with 3.0 dB ripple):

Figure 6
PLR of Linear Phase low-pass filter prototype: The phase response is:

and

• Using same technique, elements values for linear phase LPF is obtained
• For large ‘N’, given element tables are used to design LPF prototypes:

N g1 g2 g3 g4 g5 g6 g7 g8 g9 g10
1 2.000 1.000
Table 4:
2 1.577 0.422 1.000
3 1.255 0.552 0.192 1.000
For linear phase LFP
4 1.059 0.511 0.318 0.110 1.000 with g0=1 & ωC=1
5 0.930 0.457 0.331 0.209 0.071 1.000
6 0.837 0.411 0.315 0.236 0.148 0.050 1.000
7 O. 767 0.374 0.294 0.237 0.177 0.110 0.037 1.000
8 0.712 0.344 0.273 0.229 0.186 0.138 0.085 0.028 1.000
9 0.667 0.320 0.254 0.218 0.185 0.150 0.111 0.068 0.023 1.000
10 0.630 0.300 0.238 0.206 0.180 0.153 0.124 0.091 0.055 0.018
Comparison of the three responses in terms of attenuation :
Comparison of the three responses in terms of phase:
MW/RF filter implementation using Stepped-Impedance LPF’s:
(1) (2) (3) (4)

Among above steps:

- (1) is usually given
- (2) & (3) can be
implemented using
the method described
- (4) implemented via
two methods:
• trans-line stubs
• stepped impedance
Stepped-impedance Filters:
• Relatively easy (but inefficient) way to implement MW/RF LPF/HPF’s
• Mainly used in filters which do not require sharp cutoffs (mixers)
• Using [ABCD] or [T] matrix (see examples later), a transmission line with
characteristic impedance of Z0 and length l can be expressed as,

where,

• Again for equivalent circuit of the TL, [ABCD] or [T] matrix is given by

• equating above expressions:

and and
• For a TL with (l<λ/4 or θ<π/2) and (Z0>>1); XΑ+’ and BΑsmall’
and the transmission line approximates a series inductive impedance
Thus, X = ωL = Z0 sin(θ) = Z0 sin(βl) ≅ Z0 βl and B ≅ 0 ;
• Using scaling equation (for impedance & frequency) to un-scale (LN) gives:
ωL=ω{(LNR0)/ωC} ≅ Z0 βl , where Z0=Z0h or impedance of thin TL section
So at ω=ωC, βl = (LNR0)/Z0h , where R0 is equivalent to filter impedance

• Similarly from, ωL ≅ Z0βl ; L ≅ (Z0hβl)/ω ≅ (Z0hl)/(λgf) (shown before)

• Typically,the impedance of the inductive microstrip line (Z0h≈Zh) is set

to be as high as physically possible (thinnest practically fabricated line)
• Similarly, using the TL equivalent model below, ‘βl’ & ‘C’ values for
capacitive microstrip line (Z0l=Zl ≈ as high impedance as possible) is obtained

jX jX

As where,
before:
• For TL with (l<λ/4 or θ<π/2) and (Z0<<1); B Α+’ and X Αsmall’
and the transmission line approximates a shunt capacitive impedance
Thus, B = ωC = Y0 sin(θ) = Y0 sin(βl) ≅ Y0 βl and X ≅ 0 ;
• Using scaling equation (for impedance & frequency) to un-scale gives:
ωC=ω{CN/(ωCR0)} ≅ (βl)/Z0 , where Z0=Z0l or impedance of thick TL section
So at ω=ωC, βl = (CNZ0h)/R0 , where R0 is equivalent to filter impedance

• Similarly from, ωC≅(βl)/Z0; C ≅ (βl)/(ωZ0l)≅ l/(λgf Z0l) (shown before)

• Typically, the impedance of the capacitive microstrip line (Z0l≈Zl) is set

to be as low as possible (thickest fabricated line allowed the ckt. size)
• Typically, the ratio Zh/Zl should be as high as possible (practically fabricated)
Problem: Design a stepped-impedance low-pass filter with maximally
flat response and a cutoff frequency of 2.5 GHz. It is necessary to have
at least 25 dB attenuation at 4.0 GHz. The filter impedance is 50 ohm;
the highest practical line impedances is 150ohm & the lowest is 10 ohm.
(Sol: βl1=5.9° , βl2=27° , βl3=22.1° , βl4=36.9° , βl5=16.2° , βl6=9.9°)
Problem: Find the [ABCD] matrix for a lossless transmission line of
length (l) and characteristic impedance (Z0) as shown in Figure below.
Sol. For a lossless TL, the output
voltage or current has the same
magnitude as the input but lags
behind in phase by e-jβl. Thus, the
following can be written:
Since ZL=Z0,
v2=-Z0i2

Then,

Thus, the [ABCD] matrix can be written as:

Microwave Filter

Using

Transmission line stubs

(Summery of a student project)
 How to implement filter by this
way?

Give this circuit as an example for 0.5 equally

ripple chepysheve LPF:
 The filter is designed by the
following steps:
Step1.
Lumped element low pass prototype (from
tables, typically)
Lets take 0.5dB equally ripple of degree 3
 Step2.
Convert series inductors to series stubs and

shunt capacitors to shunt stubs by Richard’s

rule
Richard’s rule
 Step3.
Using Kuroda's Identities:
the series inductive stub in the diagram here
can be replaced by a shunt capacitive stub
on the other end of the λ/8 line, with
different values of characteristic impedance
Kuroda's Identities

k = 1+Z1/Z2
 Step4.
Add lum/8 lines of Zo = 1 at input and
output
 Step5.
Transform design to 50Ω and fc to obtain
physical dimensions
 Step6.
Rules of width
Final Design
 Filter Transformation
(Summery of a Student Project)

• First step: find the normalized values from the table

– i.e. g1,g2,g3,…,gn
ω
• Second step: Impedance and frequency ω ←scaling
ωc
• Frequency ω Scaling for low pass jB
filter: ω
jX k = j ( ) Lk = jωL' = j ( )Ck = jωCk '
ω
k
ωc
c
Ck
Lk where Ck ' =
Where Lk ' = ωc
ωC
 Filter Transformation
Impedance Scaling: Multiply each inductor and resistor by Ro
and divide each capacitor by Ro

R0 L
L' =
ωC
C
C' =
R0ωC
Rs ' = R0
RL ' = R0 R
 Filter Transformation

• Impedance and frequency scaling for high pass filter

• Frequency Scaling: ⎛ ωc ⎞
ω ← ⎜− ⎟
⎝ ω⎠
− ωc 1 − ωC 1
jX k = j ( ) Lk = jBk = j ( )Ck =
ω jω C k ' ω jωLk '
1 1
Where C '= where Lk '=
k
ωC Lk ωc Ck
 Filter Transformation

• After Impedance Scaling

1 Ro
Ck ' = Lk ' =
RoωC Lk ωc Ck
 Filter Transformation

• frequency scaling for band pass filter:

ω0 ⎡ ω ω0 ⎤ 1 ⎡ ω ω0 ⎤
ω← ⎢ − ⎥ Or ω← ⎢ − ⎥
ω2 − ω1 ⎣ ω0 ω ⎦ ∆ ⎣ ω0 ω ⎦

ω2 − ω1 ω0 = ω2ω 1
Where ∆=
ω0
 Filter Transformation

j ⎡ ω ω0 ⎤ 1
jX k = ⎢ − ⎥ Lk = jωLk '+
• By substitution: ∆ ⎣ ω0 ω ⎦ jωC k '
Lk ∆
where Lk ' = And Ck ' =
∆ω0 ω0 Lk

j ⎡ ω ω0 ⎤ 1
jBk = ⎢ − ⎥Ck = jωCk '+
∆ ⎣ ω0 ω ⎦ jωLk '
∆ Ck
where Lk ' = And Ck ' =
ω0Ck ω0 ∆
 Filter Transformation

• Impedance Scaling:

Multiply each Lk’ by Ro and divide each Ck’ by Ro

 Filter Transformation

• Filter Transformations:

• frequency scaling for band stop filter:

−1
⎡ ω ω0 ⎤
ω ← ∆⎢ − ⎥
⎣ ω0 ω ⎦

ω2 − ω1
Where ∆= ω0 = ω2ω 1
ω0
 Filter Transformation

It can be shown that for band stop filter:

• Each series inductor is converted to a parallel LC circuit with values:

∆Lk 1
Lk ' = and C k ' =
ω0 ∆ω0 Lk
And
• Each parallel capacitor is converted to a series LC circuit
with values:
1 ∆C k
Lk ' = and Ck ' =
∆ω 0 C k ω0
• Impedance Scaling: Mutliply each Lk’ by Ro and divide
each Ck’ by Ro
 Filter Transformation
Summary of the transformations (elements values)
ω2 − ω1
∆= ω0 = ω2ω 1
ω0

Low pass High pass Band Band stop

filter filter pass filter filter
Lk ' =
Lk R o L R ∆
R o Lk 1 Lk ' = k o
Series Lk ' = Ck ' = ω 0∆ ω0
ωc RoωC Lk ∆ 1
elements Ck ' = Ck '=
R o ω 0 Lk ∆
R o ω 0 Lk
Ro∆ Ro
Lk ' =
Lk ' =
Shunt Ck Ro ω 0C k ∆ ω 0C k
Ck ' = Lk ' =
Elements RoωC ωc Ck Ck ' =
Ck
C ' =
∆C k
R o ω0 ∆ Roω0
k
 Filter Transformation

Why?
⎛ ωc ⎞ High Pass Filter
ω ← ⎜− ⎟
⎝ ω ⎠
ω0 ⎡ ω ω0 ⎤
ω← ⎢ − ⎥
Band Pass Filter
ω2 − ω1 ⎣ ω0 ω ⎦
−1
ω2 − ω1 ⎡ ω ω0 ⎤
ω← ⎢ − ⎥ Band Stop Filter
ω0 ⎣ ω0 ω ⎦
 Simulating MW/RF filters
designed by Stepped-impedance method
using
‘High Frequency Structural
Simulator’ (HFSS-5.6)
Discussed Topics
• Introduce the Software HFSS
• Key features of HFSS
• Basic Steps for using HFSS
• Consider simulating the designed stepped
impedance low pass filter.
Introduction to HFSS
• Typically HFSS is used for modeling electromagnetic
fields in passive 3D structures
• Originally developed by HP. Currently purchased by
Ansoft from Agilent Technologies.
• Can work in UNIX or Windows Operating systems
• Can cater for Isotropic/Anisotropic Materials
(ANSOFTÎ can simulate Tensor or ferrite material )
• HFSS is a finite element (FEM) based software
• Trade off exist between mesh size of the FEM solution
and the desired accuracy in the results
Key Features of HFSS
• Uses FEM technique to solve Maxwell's Equations
and can and can include dispersion.
• Can simulate unrestricted geometries that contains
unlimited dielectrics and ports (Symmetries, Z-mul)
• Contains a model parts library of standard
microwave devices (filters, bends, circulators etc)
that can be used to model complex devices.
• Can calculate scattering parameter response for
multiple modes
• Can solve for electric-field distribution, including
far-field antenna radiation patterns.
Key Features of HFSS (cont.)
• Can calculate impedance and complex propagation
constants at each port for an unlimited number of
modes (Given Impedance/Calibration/Polarization
lines, as required)
• Can display vector and contour plots
• Quantitative versus distance graphs for field solutions
• Smith Chart plots
• Shaded and Animated field plots
• Data table
Basic Steps for using HFSS
• Modeling the Structure:
- Drawing the Geometry
- Defining the Material
- Defining and Calibrate the Ports
- Defining the Surface or Boundaries
• Analyzing the EM behavior
- Solving for S-parameters
- Analyzing the results using the post-processor
 Fabricating MW/RF filters
designed by Stepped-impedance method
using
Printed Circuit Board plotter:
‘Protomat C60’
LPKF’s Protomate C60 Machine.

¾ Rapid PCB prototyping for SMD boards

¾ High precession PCB, directly from CAD data

¾ Support auto-contact and vacuum mechanism for

multi-layer PCB’s (top and bottom layers only)

¾ Prototype Microwave PCB’s (FR3/Drude/Teflon..)

Software's supplied with Protomate C60
¾ ‘CircuitCAM’ software:
- import/edit PCB data (Garber, Excellon format)
- Draw MW PCB prototype
- Prepare the production data for plotting
(isolation channel, Rubout, Contour routing)
- Export machining data (LMD format)
¾ ‘BoardMaster’ software:
- Import LMD data from ‘CircuitCAM’
- Adjust board size and position of prototype
- Select the order of production phase
- Smart tool management (drill/mill/dispense/vacuum)
- Enable ‘auto-contact’ or ‘Vacuum’ operation
Tools used by Protomate C60
¾ Universal Cutter : engrave isolation (min 0.2mm)

¾ RF Cutter : for RF/µ-wave structures (RF motor)

¾ Spiral Drill : for drilling (0.3 mm-2.4 mm diameter)

¾ Micro Cutter : engrave isolation (min 0.1mm)

¾ Contour Router : milling contour (like guillotine)

¾ End Mill : wider insulation or fine milling

The Versatile Isolation Options
¾ Standard Isolation : the fastest production time for simple boards
¾ Pad Clearance Isolation : for easier soldering of SMD components
¾ Micro Cut Isolation : Micro cutter used only where necessary
¾ Track Clearance Isolation : wider spacing for high voltage or capacitance
¾ Rubout : complete or partial removal of residual copper
¾ Spike removal : as 1st item but removing unstable small areas
 Testing MW/RF filters
using
‘HP8510C Network Analyzer’
Vector network analysis is a method of accurately characterizing
microwave components (amplifiers, diodes, mixers, duplexers, filters,
couplers, attenuators etc.) by measuring their effect on the amplitude
and phase of the swept-frequency and the swept-power test signals
• Measuring both magnitude & phase of components are important: to
fully characterize a linear network and ensure distortion-free
transmission; to measure complex impedance in-order to design
efficient matching networks; to develop accurate models for
computer-aided-engineering circuit simulators; etc……….
• HP 8510C vector network analyzer (VNA) can measure and display
important network characteristics for frequencies of 45MHz - 30GHz
• VNA hardware consists of: (a) a sweeping signal source (usually
internal), (b) a test set to separate forward and reverse test signals,
and (c) a multi-channel, phase-coherent, highly sensitive receiver.
• Reflection measurements: are made by comparing reflected signal
to incident signal & used for following characteristics of test network
• Transmission measurements: are made by comparing transmitted
signal to incident signal & used to find following device characteristics
Minimum hardware configuration of 8510C:
• Source:The source provides the RF signal using an Agilent 8360 Series
synthesized sweeper, or an 834x Series synthesized sweeper, or an
835x-Series sweep oscillator with
an appropriate plug-in.
• Test Set: The test set separates the
source signal into an incident
signal (sent to the DUT) and a
reference signal (used to compare
transmitted and reflected signals)
• Network: is the Display/Processor
and the IF/Detector (Receiver). It
processes and display the signals.
• Printer or plotter and disc drive. Source not shown
Vector Network Analyzer Calibration:
• Why needed: To compensate for the frequency dependent phase and
amplitude characteristics of the cables and receiving system, it is
necessary to calibrate using the known responses of standard loads,
such as open-circuit, short-circuit 2.4 mm coax system.
& matched-load (50Ω) terminations
• Type of available calibrations:
(1) RESPONSE
(2) RESPONSE & ISOLATION
(3) S11 for 1-PORT
(4) S22 for 1-PORT
(5) ONE-PATH for 2-PORT
(6) FULL 2-PORT
(7) TRL 2-PORT
• Errors corrected by each type of Calibration:
• Response Calibration: provides signal path frequency response error
correction for the selected parameter. This model may be adequate for
transmission measurements of well matched, low loss devices and for
some reflection measurements where vector normalization of magnitude &
phase frequency response errors provides enough measurement accuracy.
• Response and Isolation Calibration: This error model adds correction
of the leakage term, either directivity for reflection measurements, or
crosstalk for transmission measurements. This model is appropriate for
transmission measurements of well matched, high loss devices and for
reflection measurements where mismatch is not a large factor.
• 1-Port Calibration: error model provides directivity, source match, and
reflection signal path frequency response vector error correction for
reflection measurements. This model is best applied to high accuracy
 -reflection measurements of one-port devices.
• Full 2-Port & TRL 2-Port Calibration: provides error model for full
directivity, isolation, source match, load match, and frequency response
vector error correction for transmission and reflection measurements of
two-port devices on S-Parameter test sets. This model provides best
magnitude and phase measurement accuracy for two-port devices but
requires measurement of all four S-parameters of the two-port device.
• S11 Frequency Response Calibration:
(1) Press CAL Î 2.4m coax ÎS11 Î for _ CAL 1 to start calibration process.
(2) At Port 1, connect a short, a open circuit and a load as required.
(3) When trace is correct, press SHORT, OPEN & LOAD. (S11 data is measured)
(4) Press DONE Calibration, then select a CAL SET (1 through 8).
(5) Calibration menu is displayed with _CORRECTION ON_ .
(6) Corrected S11 data is displayed
(7) Connect the test device & measure S11 (of the designed L.P. filter)
Making a measurement using a Network Analyzer:
• Two port S11 measurements (in log format): in frequency-domain:
1. Connect the DUT (LP filter) between Port 1 and Port 2.
2. Read the Return Loss of the DUT.
(a) Press CHANNEL 1 ; S11 ; LOG MAG MARKER.
(b) Use the RPG knob to position the marker at any point on the trace.
The measured return loss, magnitude and value at the marker
position is displayed above the graticule. The frequency value is
displayed in the active function area. See the next page figure for
an example of a typical display.
4. To read the measured phase angle, press PHASE
5. To position the trace automatically for viewing, press AUTO.
Vector Network Analyzer Display of S11:
Calibration Table of Vector Network Analyzer :