Fundamentals of Mixed Signals

and Sensors

Module 2

Prof. Adelino T. Racusa

Balancing Analog and Digital Techniques in Signal Processing

Introduction

Today's engineers face the challenge of selecting the

appropriate mix of analog and digital techniques for effective
signal processing.

Since all sensors—like microphones, thermocouples, and

strain gauges—produce analog signals, it's not feasible to
process these signals using purely digital methods.
Signal Conditioning

To bridge the gap between analog sensors and digital processing, signal conditioning
circuitry is essential. These circuits perform critical functions, including:

• Gain (Multiplication): Amplifies weak signals.

• Isolation: Prevents interference using instrumentation and isolation amplifiers.
• Noise Detection: Uses high common-mode instrumentation amplifiers to manage
noisy environments.
• Dynamic Range Compression: Implements log amplifiers and programmable gain
amplifiers (PGAs) to handle varying signal levels.
• Filtering: Applies both passive and active filters to remove unwanted frequencies.
Signal Processing Approaches

Two primary methods are used to process signals:

• Purely Analog Approach: Involves traditional signal processing

techniques.
• Digital Signal Processing (DSP) Approach: Offers flexibility and
precision but requires careful placement of the analog-to-digital
converter (ADC) in the signal path.
ADC Placement

When using DSP, deciding where to position the ADC is crucial. Placing
the ADC closer to the sensor shifts more conditioning responsibilities to
the ADC, increasing its complexity. This complexity can manifest as:

• Higher sampling rates

• Wider dynamic ranges
• Improved resolution
• Enhanced input noise rejection
• On-chip filtering and programmable gain amplifiers
Technological Advancements
Recent advancements in data converter technology have led to greater integration of
signal conditioning functions within ADCs and DACs. For example:

• 24-bit ADCs: Modern ADCs, like the AD773x series, can digitize low-level signals
(e.g., 10mV) directly, eliminating the need for additional conditioning.
• Complete Coders-Decoders (Codecs): At voiceband and audio frequencies, devices
like the AD1819B and AD73322 minimize external component requirements
through integrated circuitry.
• Analog Front Ends for Video: For high-speed applications, such as CCD image
processing, devices like the AD9814, AD9816, and AD984x series provide necessary
analog processing capabilities.
1. Medical Devices: ECG Monitors

Analog Sensors: Electrocardiogram (ECG) monitors use electrodes to

detect the electrical signals produced by the heart, which are analog
signals.
Signal Conditioning: The raw ECG signals are often weak and noisy.
Signal conditioning circuits amplify the signals and filter out noise using
instrumentation amplifiers and active filters.
Digital Processing: After conditioning, the signals are converted into
digital form using ADCs for further analysis, enabling real-time
monitoring and diagnostic capabilities.
Smartphones: Microphone Systems

• Analog Sensors: Smartphones utilize microphones to

capture sound, which produces analog audio signals.
• Signal Conditioning: The audio signals undergo
amplification and filtering to enhance clarity and reduce
background noise using analog signal processors.
• Digital Processing: The conditioned signals are converted
to digital format for processing, enabling features like
voice recognition, noise cancellation, and audio recording.
Practical Example of DSP

Consider the implementation of a lowpass filter with a cutoff frequency of 1 kHz, comparing an
analog filter to a digital filter in a sampled data system.

Key Points:

System Requirements:
An ADC/DAC is needed with adequate sampling frequency, resolution, and dynamic range.
The DSP must perform calculations within the sampling interval (1/fs).
Analog Filter:
• Realized as a 6-pole Chebyshev Type 1 filter with a ripple in the passband and none in
the stopband.
• Typically constructed using three 2-pole stages, requiring op-amps, resistors, and
capacitors.
• Maintaining performance (e.g., 0.5 dB ripple) requires precise component selection and
matching.
Digital Filter:
• Implemented as a 129-tap FIR filter with only 0.002 dB passband ripple and linear
phase.
• Offers much sharper roll-off than the analog filter, making it impractical to realize
with analog techniques.
• Advantages include immunity to component matching and drift, as clock
frequencies are crystal-controlled.

Processing Requirements:
• The digital filter requires 129 multiply-accumulate (MAC) operations per output
sample, which must be completed within the sampling interval.
• For a 10 kSPS sampling frequency, 100 µs is available for processing, necessitating
an instruction rate of over 1.3 million instructions per second (MIPS).
• DSPs, such as the ADSP-218x series, can achieve instruction rates up to 75 MIPS,
making them well-suited for this application.
Digitization Process

• Sampling: The process of converting an analog signal to a digital signal by

measuring it at discrete time intervals (e.g., measuring temperature every second).
• Quantization: The process of restricting the signal values to a discrete set of levels.
This means the continuous range of values is rounded to the nearest allowable
value.

Advantages of Digital Signals

• Storage and Processing: Digital signals are easier to store and process using digital
computers.
• Robustness to Noise: Digital signals tend to be less affected by noise compared to
analog signals.
Information Loss

Digitizing a signal through sampling and quantization can lead to

information loss. Understanding how much information is lost and
optimizing these processes is crucial.
The theory for sampling is well-defined, while quantization is more
complex, and both depend on specific system and user requirements.
Common Signal Operations
In signal processing, several fundamental operations are performed on signals, in addition
to basic addition, multiplication, differentiation, and integration.
Sample Problem:
In a computer engineering project, you are designing a
circuit that generates a sinusoidal signal for a signal
processing application.
The specifications for the signal are as follows:

•The peak voltage (amplitude) of the signal is 5 volts.

•The frequency of the signal is set to 60 hertz, which is
commonly used in digital signal processing.
•There is a phase shift of π/4​ radians introduced by the
signal modulation technique used in your circuit.
In a computer engineering project, you are designing a circuit that generates a
sinusoidal signal for a signal processing application.
The specifications for the signal are as follows:

•The peak voltage (amplitude) of the signal is 5 volts.

•The frequency of the signal is set to 60 hertz, which is commonly used in digital
signal processing.
•There is a phase shift of π/4​ radians introduced by the signal modulation
technique used in your circuit.

•Calculate the angular frequency of the sinusoidal signal.

•Write the mathematical equation that models the voltage output of the circuit as
a function of time.
•Evaluate the voltage output at two specific time points: at t=0 seconds and
t=0.01 seconds.
Solution:

Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.01 seconds.

1. Angular frequency
W = 2 x 3.14159 x 60
= 376.99 radians / second
Solution:

Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.01 seconds.

2. Mathematical equation

1. x(t) = 5 volts x sin( 376.99 t + π/4)

= 5 x sin( 376.99 t + π/4)
Solution:
Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.01 seconds.

3. Using the formula

when t = 0
x(0) = 5 x sin( 376.99 t + π/4) π/4 = 0.7854
= 5 x sin(376.99 x 0 + π/4)
=
= 5 x sin(π/4)
= 3.53
Solution:
Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.01 seconds.

3. Using the formula

when t = 0.01 second
x(0.01) = 5 x sin( 376.99 t + π/4) π/4 =
= 5 x sin(376.99 x 0.01 + π/4)
= 5 x sin(3.7699 + π/4)
= 5sin(3.7699 + 0.7854)
= 5sin(4.5553)
Given : Solution:

Solve for:
Solution:

Given : Solve :
Amplitude = 12 volts. angular frequency
frequency = 50 hertz mathematical equation
phase shift = π/6​ radians voltage output at t=0 seconds & t=0.05 seconds.

1. Angular frequency
W = 2 x 3.14159 x 50
= 314.159 radians / second
Solution:

Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.05 seconds.

2. Mathematical equation

x(t) = 12 volts x sin( 314.15 t + π/6)

= 12 x sin(314.159 t + π/6)
Solution:
Given : Solve :
Amplitude = 12 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/6 radians voltage output at t=0 seconds & t=0.05 seconds.

3. Using the formula

when t = 0
x(0) = 12 volts x sin( 314.159 t + π/6)
= 12 x sin( 314.159 t + π/6)
= 12 x sin(π/6)
= 12 x ½ = 6 v
Solution:
Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.05 seconds.

3. Using the formula

when t = 0.01 second
x(0) = 12 volts x sin( 314.159 t + π/6)
= 12 x sin( 314.159 x 0.05 + π/6)
= 12 x sin(15.708 + π/6)
= 12 sin(15.708 + 0.5236)
= 12sin(16.2316)
Solution:
Given : Solve :
Amplitude = 5 volts. angular frequency
frequency = 60 hertz mathematical equation
phase shift = π/4​ radians voltage output at t=0 seconds & t=0.01 seconds.

3. Using the formula

when t = 0.01 second
x(0.01) = 5 x sin( 376.99 t + π/4) π/4 =
= 5 x sin(376.99 x 0.01 + π/4)
= 5 x sin(3.7699 x π/4)