SMART SENSORS

(21ECO746)
LMS
Key:
21ECO7
46
 SYLLABUS OF SMART SENSORS (21ECO746)
• UNIT-1: Basics of Smart Sensors and Micromachining:
Introduction, Mechanical-Electronic transitions in sensing,
Nature of sensors, Overview of Smart Sensing and control
systems, Introduction to Micromachining, Bulk
Micromachining, wafer bonding, Surface Micromachining.
• UNIT-2: MCUs and DSPs for Sensor: Introduction, MCU
control, MCUs for Sensor Interface, DSP control,
Software, Tools and Support, Sensor Integration.
• UNIT-3: Sensor Communication and MEMS: Communications
for Smart Sensors - Sources and standards, Automotive
protocols-CAN Protocol, Industrial usage of CAN, Office and
Building Automation, Home Automation, Wireless Zone
Sensing, RF-ID, Microoptics, Microprobes, Micromirrors, Field
Emission Displays.
 SYLLABUS OF SMART SENSORS (21ECO746)
• UNIT-4: Packaging, Testing and Reliability of Smart Sensors:
Introduction, Semiconductor packaging applied to Sensors,
Hybrid packaging, Packaging for Monolithic Sensors,
Reliability Implications, Testing Smart Sensors.
• UNIT-5: Implications of Smart Sensor Standards, Recent
Trends and next phase of Sensing Systems: Introduction,
Sensor plug-and-play, Communicating Sensor data via
existing wiring, Automated/remote Sensing and web,
process control over the internet, Alternative standards,
HVAC Sensor Chip, Alternative views of Smart Sensing,
Smart Loop.
 UNIT II

MCUs and DSPs for Sensor

Introduction
Several semiconductor technologies are available to
improve the accuracy and the quality of the
measurements and to add diagnostics and other
intelligence to any type of sensor. Foremost among
those technologies are MCUs, DSPs, application-
specific ICs (ASICs), and field-programmable gate
arrays (FPGAs).

The technologies also have the potential to allow for a

fully integrated (monolithic) smart sensor.
MCU Control
Single-chip MCUs
combine
microprocessor unit
(MPU) computing
capability, various
forms of memory, a
clock oscillator, and
I/O capability on a
single chip, as shown
in Figure

Basic MCU block

diagram
MCUs for Sensor
Interface
In addition to the basic features of an MCU, a number of
custom modules are integrated on the same chip to
increase the utilization of the process, reduce printed
circuit board space, and increase the functionality for a
specific application.
These include analog input capabilities, A/D conversion
techniques, Processing bandwidth, Electronically
programmable trim, Onboard memory, Power
conservation, and Improved electromagnetic
compatibility (EMC) and Control of radio frequency
interference (RFI).
For purposes of discussion here, the MC68HC05 family of
8-bit MCUs is used to explain the various MCU features.
The main features/components are:
1. Peripherals
2. Memory
3. Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
The main features/components are:
1.Peripherals
2. Memory
3. Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
Existing MCU technology contains a variety of peripherals, or
hardware options, that enhance the capability of the MCU.
Peripherals enable the MCU to obtain information from sensors and
control output devices. Some of the most common peripherals are
general-purpose I/O ports, timers, and serial ports.
Timers usually measure time relative to the internal clock on the
chip or an externally provided clock signal. An on-chip oscillator
that operates up to 4.0 MHz at 5V or 1 MHz at 3V controls the clock
on the chip.
Two basic serial ports are the serial communications interface (SCI)
and the serial peripheral interface (SPI).
The SCI is a simple two-pin interface that operates asynchronously.
Data is transmitted from one pin and received on the other. Start
and stop bits synchronize communications between two devices.
The SCI port is a universal asynchronous receiver
transmitter (UART) that can be used with an RS-232 level
translator to communicate with personal or other types of
computers over fairly long distances.
The SPI port requires a third pin to provide the synchronizing
signal between the control chip and an external peripheral.
This type of communication is usually on the same board.
Standard SPI peripherals are available from many
manufacturers and include ADCs, display drivers, EEPROM,
and shift registers.
The main features/components are:
1. Peripherals
2.Memory
3. Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
Various types of memory can be integrated on a chip,
including RAM, read only memory (ROM), EPROM, EEPROM,
and flash memory. Semiconductor memory is based on a
single transistor or cell that is on or off to generate a bit that is
either a 1 or a 0. Memory is classified as either volatile or
nonvolatile. Volatile memory is not stored when the power is
disconnected to the MCU. Nonvolatile (NV) memory is stored
when power is disconnected.
RAM can be read or written (changed) by the CPU and is
volatile. The write/read (W/R) endurance is a key parameter
for RAM.
ROM, which can be read but not changed, is nonvolatile
memory and is included in the design (masked layout) of the
chip.
The main features/components are:
1. Peripherals
2. Memory
3.Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
I/O is a special type of memory that senses or changes based
on external digital elements and not the CPU [3]. I/O ports
connect the external elements to the CPU and provide control
capability for the system. I/O can be either parallel,
transferring 8 bits at a time to the MCU, or serial, transferring
data 1 bit at a time.
General purpose I/O connections (pins) can be used as either
an input or an output. A number of pins are typically grouped
together and called a port.
The program determines the function of each pin. Program
instructions evaluate the logic state of each input and drive
outputs to logic 1 or 0 to implement the control strategy.
The main features/components are:
1. Peripherals
2. Memory
3. Input/Output
4.Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
An ADC is frequently integrated with the MCU. For example,
the typical ADC in the HC05 family of MCUs consists of an 8-
bit successive approximation converter and an input-
channel multiplexer.
The successive approximation register (SAR) is the most
popular method of performing A/D conversions because of
its fast conversion speed and ease of use with multiplexed
input signals.
The main features/components are:
1. Peripherals
2. Memory
3. Input/Output
4. Onboard A/D Conversion
5.Power-Saving Capability
6. Local Voltage or Current Regulation
7. Modular MCU Design
The advantage of MCU hardware and software is the variety
of power saving approaches. Varying the processing speed or
stopping processing altogether can have a significant effect
on overall power consumption. In addition, the ability to
operate at lower voltages also reduces power consumption.
The 68HC05 MCU family has two additional modes of
operation to reduce power consumption, called the WAIT and
STOP modes.
In the WAIT mode, the processor bus is halted but the on-
chip oscillator and internal timers are left in operation. The
device returns to normal operation following any interrupt or
an external reset.
In the STOP mode, both the processor and the on-chip
oscillator are halted. The device returns to normal operation
only after an external interrupt or reset.
The main features/components are:
1. Peripherals
2. Memory
3. Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6.Local Voltage or Current Regulation
7. Modular MCU Design
Onboard voltage or current regulation is important to sensors
that are not ratiometric, because the variation in supply
voltage over a -40°C to +125°C operating range can be
greater than ±5%.
At higher levels of integration, such analog voltage and/or
current regulation can both reduce component count and
improve accuracy. Furthermore, those regulation schemes can
be dynamically altered to improve functionality or reduce
power consumption.

**(Ratiometric means that an output signal changes in direct proportion

to a change in input or supply voltage. Ratiometric sensors typically
operate on a supply voltage of 5 VDC, with a tolerance of +/- 10%.)
The main features/components are:
1. Peripherals
2. Memory
3. Input/Output
4. Onboard A/D Conversion
5. Power-Saving Capability
6. Local Voltage or Current Regulation
7.Modular MCU Design
A methodology has been developed that allows custom
MCUs to be designed to specifically address the
requirements of a particular application.
The customer-specified integrated circuit (CSIC) approach
differs from an ASIC in the performance and density that can
be achieved.
DTMF-- Dual Tone Modulation Frequency.
VHF-- Vacuum Fluorescent Display
VPW-- VPW is another name for Pulse Length or Pulse
Duration Modulation
DSP Control
 DSP vs MCUs
 DSPs have hardware arithmetic capability that allows the real-
time execution of feedback filter algorithms.
In contrast, MCUs use lookup tables to approximate filter
algorithms with inherent limitations of flexibility and accuracy.
A typical DSP can execute instructions in less than 100 ns. That
capability allows a peak execution rate of 20 MIPS, which is 10
to 20 times the performance of conventional MCUs. DSPs are
often rated in million of operations per second (MOPS).
On the other hand, generating code for a DSP, in general, is
more difficult than generating code for an MCU.
The need for real-time processing / real-time application in
several systems is causing control system engineers to evaluate
and use DSP technology.
DSP vs MCU
The features of a specific 16-bit DSP, the
DSP56L811, include the following:
40-MHz, 20-MIPS operation from 2.7V to 3.6V;
One-cycle multiply-accumulate shifter;
16-bit instruction and 16-bit data word;
Two 36-bit accumulators;
Three serial ports;
Sixteen I/Os, two external interrupts;
Power dissipation of 120 mW at 40 MHz.
DSP technology is evolving in two ways:
1) First, DSP-like performance can be approached when a
time processor unit (TPU) is included in an MCU design,
which is possible in higher performance MCUs. The TPU is a
programmable microstate machine that addresses
requirements for computation and greatly reduces the
overhead on the main processor. That allows the main
processor to calculate strategy-related items and not have
to decide when a specific activity is initiated.
2) Second, in high-end applications, such as an automotive
near-obstacle detection system or noise-cancellation
system, a dedicated 24-bit DSP is used.
24-bit DSP and its architecture:
The device has program memory, two data memories, and
the same kind of instruction set that is common throughout
the Motorola MCU family.
A DSP has been used to develop an intelligent sensor for
checking the pitch of tapped holes in motor blocks. The
sensor replaced a mainframe computer that was checking
one hole per minute. By providing DSP power in each sensor,
the inspection system increased the rate to 100 holes per
minute.
Algorithms Versus Lookup Tables:
MCUs use lookup tables to store values that are accessed
when the program is running.
Algorithms are used to correct for variations from expected
results and to implement a control strategy.
The speed of accessing the information from a table or
performing a calculation determines the response time of the
sensor input to the MCU/DSP portion of the system. That can
be the limiting factor to initiating a change to the output in
an MCU/DSP-controlled system.
Software, Tools, and
Support
Creating a new approach or an alternative to existing control
technologies requires much more than the architecture.
The software used to program the control portion and the tools
that allow the system to be developed are equally important.
Portable code (software compatibility) is essential if future end
products require migration to a higher performance MCU.
One main advantage of using existing MCUs or DSPs is that the
development tools already exist and allow the designer to
quickly and easily develop both system hardware and software.
For example, data entry, program debugging, and programming
of an MCUs OTP, EPROM, and EEPROM can be accomplished by
utilizing an evaluation module (EVM).
MC68HC05EVM development tool for 68HC05 MCU
Sensor Integration
The basic sensor element is interconnected in close
proximity to the enhanced MCU in the same package.
The fourth level in advancing integration is to migrate the
developed MOS interface onto the MCU.
The fourth level is also the optimum point at which to
consider inclusion of any output-drive capabilities required
by the specific sensing and control application onto the MCU.
At level IV, there is a definite transition from standard-
product MCUs into customization for a specific application or
market.
 The cost trade-off
To demonstrate the costs of combining a sensor with an MCU:
Consider a small 0.100″× 0.100″ MCU processed at 0.8 defect/cm2
and a 0.100″× 0.100″ sensor processed at 4.0 defect/cm2.
The example MCU will have a relative cost of 1.05. The sensor will
have a relative cost of about 1.29.
That gives a combined two-chip relative cost of 2.34.
If the sensor and the MCU are combined on a single die:
 Using a sensor process that creates 4.0 defect/cm 2 , the resulting die
would be about 0.128″× 0.128″ and have a relative cost of 2.50.
That cost increase of only 1.07 times the combined die cost may be
less significant than the added assembly and test costs of two
separate devices.
The cost tradeoff changes dramatically as the MCU die size
increases:
 If the same 0.100″× 0.100″ sensor with relative cost of 1.05 is
combined with a larger 0.300″× 0.300″ MCU with relative cost of
15.15, the relative cost of the combined two-chip set rises to 16.44.
Combining the two chips into an equivalent die size of 0.311″×
0.311″ and processing at the higher defectivity sensor process will
yield a relative cost of 75.28, or 4.58 times the two-chip solution.
Integrating a (typically) higher defectivity sensor process onto
another chip becomes increasingly more costly as the total die size
increases.
The cost penalty of adding a sensor on a chip decreases if the
defectivity of the sensor process decreases.
Typically, MCUs use larger wafer size and a higher density process
than the sensor. As improvements in processing technology occur,
the added cost can be traded off against packaging and assembly
costs for the two-chip approach.
An interim (temporary/provisional ) approach to the fully integrated
sensor on an MCU is shown. The sensor interface for the
accelerometer is initially a standalone circuit that uses the same
design rules and process as the MCU. That provides a shorter
design cycle and lower initial cost for the sensor.