Scribd
Upload
93 views9 pages

DSP Case Study

The document discusses the design of highpass digital filters using different methods. It compares the step-invariant design method to the bilinear z-transform method. The step-invariant method designs a digital filter to match the step response of the analog filter, while the bilinear z-transform method directly maps the analog transfer function to the digital domain. The document presents an example of designing a simple highpass filter and shows that the step-invariant method can closely approximate the bilinear z-transform method while avoiding non-linear distortion. It also demonstrates how to convert a step-invariant filter to satisfy a desired frequency response specification.

Uploaded by

Suhail Jain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
93 views9 pages

DSP Case Study

The document discusses the design of highpass digital filters using different methods. It compares the step-invariant design method to the bilinear z-transform method. The step-invariant method designs a digital filter to match the step response of the analog filter, while the bilinear z-transform method directly maps the analog transfer function to the digital domain. The document presents an example of designing a simple highpass filter and shows that the step-invariant method can closely approximate the bilinear z-transform method while avoiding non-linear distortion. It also demonstrates how to convert a step-invariant filter to satisfy a desired frequency response specification.

Uploaded by

Suhail Jain
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 9

DIGITAL FILTERS

A Case study
By TARUN 2K17/EE/220
Suhail Jain 2K17/EE/212
Digital Filters
● A digital filter is a system that performs mathematical operations on
a sampled, discrete-time signal to reduce or enhance certain aspects of that signal.
● A digital filter system usually consists of an analog-to-digital converter (ADC) to
sample the input signal, followed by a microprocessor and some peripheral
components such as memory to store data and filter coefficients etc. Program
Instructions (software) running on the microprocessor implement the digital filter by
performing the necessary mathematical operations on the numbers received from
the ADC.
● In some high performance applications, an FPGA(Field-Programmable Gate
Array) or ASIC(Application-Scientific Integrated Circuit) is used instead of a general
purpose microprocessor, or a specialized digital signal processor (DSP) with specific
paralleled architecture for expediting operations such as filtering.
About the case study

● The purpose of this case study is to present a development of the step-invariant


approach to highpass digital filter design .
● I will try to show that the transfer function, G(z)sif , obtained via the step invariant
design method, may be made to approximate closely to the transfer function, G(z)
obtained via the bilinear z-transform method.
● Also I will try to show that it is possible to convert a time domain (step-invariant)
filter, G(z)si , to one that satisfies a frequency domain specification ,G(z)sif . This can
be achieved by observing certain conditions and by employing a suitable gain term.
The validity of the method is demonstrated using a practical example of a simple
highpass filter and a digital phase-advance network.
The approach

● As we know that the impulse-invariant design method for bandlimited filters, is based on
the application of standard z-transforms, whereby an analogue filter transfer function,
G(s), is transformed to an equivalent digital filter transfer function ,G(z), that is

● For the analogue filter, the impulse response, g(t), is defined as L-1[G(s)] Similarly, for a
digital filter, the impulse response, gk, is defined as Z-1 [G(z)]. To be impulse-invariant
gk= g(t) for t = 0, T, 27, ..., where T is the sampling period. Furthermore, the frequency
response of the digital filter ,G(exp(jωT)), will approximate to the frequency response of
the analogue filter,G(jω), if aliasing errors have been minimised by band-limiting [G(s)]
and by correct choice of sampling period T.
The approach(continued…)

● A common approach to the design of non-bandlimited digital filters (highpass or


bandstop) is to use the well-known bilinear z-transform, that is, by directly
substituting 2 / T[(z -1) / (z + 1)] for s in [G(s)] ,a corresponding transfer function G(z),
is obtained.
● However, non-linear distortion (warping) may be introduced into the filter
representation because of the non-linear relationship between the analogue filter
frequency scale and the digital filter frequency scale. This problem is resolved by
using prewarping techniques.
● An alternative approach to the design of non-bandlimited filters is to use the step-
invariant method described in this case study
HIGHPASS FILTER DESIGN

● A simple analogue highpass filter has a transfer function : G(s) = s /(s + α), where ω = α rad/ s is
the cut-off frequency of the filter. This is non-bandlimited and consequently the impulse-
invariant design method is excluded, however, employing the suitable bilinear z-transform we
obtain

● Referring to this equation it is seen that G(z)bl has a gain term equal to 1/(1+ αT/2), and the
corresponding z-plane representation has a zero at z = 1 and a pole at z =1- αT/2/(1+ αT/2)
that is, the pole is at z = (J - αT + α2T2 /2 – α3T3 /4 + …)
HIGHPASS FILTER DESIGN(continued…)

● Now consider where Y(s) is the Laplace transform of the filter


response and X(s) is the Laplace transform of the filter input signal. For the step-invariant
design method the step input signal is assumed to have an amplitude of A for t = > 0, i.e. X(s)=
A/s therefore Y(s) = (A/s)*(s/s+α) = A/(s+α).
● Taking the inverse Laplace transform of equation above we obtain the corresponding step
response of the analogue filter, thus [Y(s)] = L-1 [Y(s)] = Aexp(-αt) and Transforming equation
mentioned above into the z-domain (standard z-transform) yields Y(Z) = A*z/(z - exp(- αt)).
● The standard z-transform of X(s)= AI s is X(z) = A*z I ( z - 1), therefore the corresponding
transfer function of the digital filter is
HIGHPASS FILTER DESIGN(continued…)

● Taking the inverse z-transform of equation mentioned in previous slide we obtain the
corresponding step response of the digital filter, thus yk = Z-1 [Y(z)]= Aexp(- αkT).

● Comparing equations for y(t) and yk we see that yk is equal to y(t) at the sampling instants, and
therefore the digital filter defined by equation for G(z) in previous slide is step-invariant.
● Now returning to the bilinear z-transform of the analogue filter (equation for G(z) bi) we allow
for prewarping by making the substitution
Thank You

You might also like