Interfacing to Microprocessors
Chapter 12
• What constitutes a “controller” will vary from
application to application.
• It may be no more than an amplifier or a switch.
• It may be a complex system that may include
computers and other types of processors such as
data acquisition and signal processors.
• Most of the time, it is a microprocessors.
• We shall therefore focus the discussion here on
• Focus on microprocessors as general purpose,
flexible and reconfigurable controllers and the
ways sensors and actuator relate to these.
• Microprocessors are often called microcontrollers
• What is a microprocessor? What is the different
between a microprocessor and a computer or a
microcomputer and how a distinguishing set of
features is arrived at are all difficult and subjective
issues. What is a microprocessor to one is a full
fledged computer to another
The microprocessor
• A microprocessor is a stand alone, self contained
single chip microcomputer.
• It must have as a minimum:
– a central processing unit (CPU)
– nonvolatile and program memory
– input and output capabilities.
• A structure that has these can be programmed in
some convenient programming language
• can interact with the outside world through the
input/output ports.
The microprocessor
• Other important requirements:
• must be relatively simple
• reasonably small
• necessarily limited in most of its features – memory,
processing power and speed, addressing range and, of
course in number of I/O devices it can interact with.
• The designer must have access to all features of the
microprocessor – bus, memory, registers, all I/O ports,
• In short, Microprocessors are components with
flexible features that the engineer can configure
and program to perform task or a series of tasks.
The microprocessor
• Two limits on the tasks microprocessors can
• The limitations of the microprocessor itself
• The imagination (or capabilities) of the
The 8 bit microprocessor
• We will narrow down to 8 bit microprocessors
– these are the most common in sensor/actuator systems
– they are simple and representative of all microprocessor
• 16 and 32 bit microprocessors exist
• There are a number of architectures being used.
• We will emphasize the Harvard architecture
because of its simplicity, flexibility and popularity.
The architecture
• There are about two dozen manufacturers of
• All based on a few architectures.
• We shall only briefly describe here one
architecture – the Harvard architecture
• used in many microprocessors
• Simple and efficient
• The choice in smaller microprocessor
• Example: Microchip and Atmel microprocessors
The architecture
• Main features:
• Separate busses for program memory and operand
• Pipelined architecture
• Allows fetching data while another operation
• Each cycle consists of fetching the (n+1)th
instruction while executing the nth
• Integer arithmetic
• Limited instruction set
The architecture
• Bus widths vary depending on manufacturer and
on the microprocessor size.
• Example: Figure 12.1, bus architecture for a
PIC18F452 from Microchip.
• The instruction is 16bit
• Program address is 15bit wide.
• Data is 8bits and
• Operand address is 12 bits.
• These vary from device to device.
Bus architecture
The architecture
• Example, the smallest microprocessors
available (PIC10FXX) are 6 pin devices
• Summarized in Table 12.1.
• The architecture for this device is shown in
Figure 12.2.
• Here the program address bus is only 9 bits
while the instruction buss is 12 bits.
PIC10FXX microprocessors
PIC10FXX microprocessors
The architecture
• Example: one of the largest, is the
• Has an address bus 21 bits wide.
• The processor and its variants are shown in
Table 12.2
• Its architecture in Figure 12.3.
The architecture
• Architecture supports:
• Direct addressing for the first 8 bits of address
• Indirect addressing (variable pointer addressing)
for all memory space.
• Includes a CPU with associated status bits and a
set of special functions registers.
• I/O ports, other peripherals (such as comparators,
A/D converters, PWM modules, etc.)
• Timers, status indications and much more,
The architecture
• All modules available to the user.
• User writable registers are also provided.
• Microprocessors have been designed to respond to
specific needs: common to find modifications that
respond to these needs
• Example: various processors from the same family
may have a different instruction sets
– PIC10FXX has 33 instructions
– PIC18FXX20 has 77 instructions
– ATmega128 (from Atmel) has 133 instructions.
The architecture
• Memory varies from 256 bytes to over 256
• Number of peripherals, ports, etc vary from
as few as 4 to over 100
• Physical size: from 6 pin to 100 pins
• Various chip configurations (DIP, surface
mount, dies etc.)
• 8 bit microprocessors have word length of 8 bits.
• Integer data from 0 to 255 may be represented
• To address memory, usually a longer word is
• Most microprocessor have a 12 bit (4k) 14 (16k)
or 16 bit (64k) memory address but longer address
words are also used.
• Most microprocessor operate at clock
speeds between 1 and 40 MHz.
• Since often the clock is internally divided,
the instruction cycle is slower than that
• Typical values are up to about 10 MHz
cycle clock or 0.1 s per instruction
Instruction set
• Microprocessors have a small instruction set –
• sometimes no more than 2-3 dozen simple
• Varies from a minimum of about 30 to a maximum
of about 150 instructions.
• These are selected to cover the common
requirements of programming a device
• Allows one to perform almost any task that can be
physically performed within the basic limitations
of the device.
Instruction set
• Instructions include:
– logical instructions (AND, OR, XOR, etc.)
– move and branching instructions (allow one to move
data from and to registers and conditional and
unconditional branching)
– bit instructions (operations on single bits in an operand)
– arithmetic instructions such as add and subtract,
– subroutine calls
– other instructions that have to do with the performance
of the microprocessor such as reset, sleep and others.
• Some are bit oriented, some are byte (register)
oriented, some are literal and control operations
Input and output
• Input and output is defined by the availability of
pins on the package.
• Usually limited to less than about 100 pins (6, 8,
14, 18, 20, 28, 32, 40, 44, 64 and 100 pins are
• Two pins are used to power to the device
• For example, an 18 pin device can have no more
than 14 I/O pins.
• Of these, some may be used for other purposes
such as oscillators or communication
Input and output
• All microprocessor will have a number of pins
available as I/O.
• Example, a 6 pin microprocessor may have as
many as 4 I/O, a 64 pin processor can have in
excess of 48 I/O pins.
• I/O pins are grouped into ports, each addressable
as an 8 bit word (each group has up to 8 I/O pins).
• Different ports may have different properties and
may be able to perform different functions.
Input and output
• I/O pins are tri-state enabling an I/O pin to serve
as input, output or to be disconnected.
• Most I/O are digital but some may be configured
as analog as well.
• I/O pins can supply or sink considerable current –
usually in the range of 20-25 mA.
• This is not sufficient to drive many actuators but it
can drive low power devices directly or indirectly
through switches and amplifiers.
Clock and timers
• Microprocessor must have a timing mechanism
that defines the instruction cycle.
• This is done by an oscillator
• Oscillators may be internal or external.
• Usually and RC oscillator is used for internal
• A crystal is the most common way of setting the
frequency externally (this requires either dedicated
pins or the use of two I/O pins).
Clock and timers
• The oscillator frequency is usually divided
internally to define the basic cycle time.
• Microprocessors have internal timers
under the control of the user
used for various functions requiring counting/timing
At least one counter is available
larger microprocessors can have 4 or more timers
some are 8 bit timers and some 16 bit timers.
a watchdog timer is available for the purpose of
resetting the processor should it be “stuck” in an
inoperative mode.
Clock and timers
Used for
Execution of commands
Control over the functions of the microprocessor,
Status indication
• Modern microprocessors, contain three
types of memory:
• program memory, in which the program is
• data memory (RAM),
• EEPROM memory
• Note: EEPROM not available on some very
small microprocessors.
• Program memory is usually the largest
• From less than 256 bytes to over 256kBytes.
• In most cases, flash memory which means that is
rewritable at will and is nonvolatile (program is
retained until rewritten or erased).
• Data memory (RAM) is usually quite small and
may be a small fraction of the program memory
• Does not retain data upon removal of power.
• EEPROM is nonvolatile rewritable memory used
mostly to write data during execution
• Most microprocessor operate from 1.8V to 6V.
• Some have a more limited range (2.7-5.5V).
• Based on CMOS technology: This means that:
– power consumption is very modest.
– power consumption is frequency dependent.
• The higher the frequency the higher the power
Power is also dependent on
What the processor does
Which modules are functioning at any given time.
The user has considerable control over power
consumption through:
– Choice of frequency
– Mode of operation
– Special functions such as interrupt wakeup and sleep.
Other functionalities
• Microprocessor must have certain modules (CPU,
memory and I/O)
• They can have many more modules
• Add functionality and flexibility
• Many microprocessors include
comparators (for digitization purposes),
A/D converters,
Capture and Compare (CCP) modules,
PWM generators
Communication interfaces.
Other functionalities
• One or two comparators are provided on many
• Depending on the microprocessors 8 or 10 bit A/D
converters are provided, usually in multiple
channels (4 to 16).
• PWM channels (up to 8) are common on some
• Serial interfaces such as UART, SPI, two wire
interface (I2C), synchronous serial and USB ports
are available
Other functionalities
• Many microprocessors provide multiple
interfaces, all under the user’s control.
• Other functions such as analog amplifiers and
even transceivers are sometimes incorporated
within the chip.
• The I/O used for these functions are either digital
I/O (for communication for example) or analog
I/O (for A/D for example)
Programs and
• A microprocessor is only useful if it can be
• Programming languages and compilers have been
designed specifically for microprocessors.
• The basic method of programming
microprocessors is through the Assembly
programming language
• Can be, and very often is done through use of
higher level languages with C leading.
Programs and
• These are specific compilers, adapted for a class
of microprocessors.
• They are based on a standard C compiled (such as
ANSI C) and modified to produce executables that
can be loaded onto the microprocessor.
• Most microprocessors can be programmed in
circuit allowing changes to be made, or the
processors to be programmed or reprogrammed
after the circuit has been built.
Programs and
• Instruction sets for microprocessors are small and
based on the assembly language nomenclature.
• Microprocessors have been designed for integer
• Programming for control, especially sequential
control is simple and logical.
• Floating point operations and, are either not
practical or difficult and tedious.
• They also tend to require considerable time and
should only be attempted if absolutely necessary.
Programs and
• There are both integer and floating point
libraries freely available.
• Floating point operations are only practical
on the larger microprocessors because they
require much memory.
Examples of microprocessors
PIC10FXXX (low level, 6 pin),
PIC16F62X (midrange, 18 pin),
PIC18FXX20 (high level, 64 or 80 pin),
Atmega128 (high level, 64 pin).
A comparison of these typical processors
will reveal most of the properties and
capabilities of microprocessors.
Interfacing Issues
• Three basic modes:
– 1. Continuous dedicated monitoring of the
sensor by the microprocessor
– 2. Polling the sensor
– 3. Interrupt mode
Continuous mode
• Microprocessor is dedicated for use with
the sensor
• Its output is monitored by the
microprocessor continuously
• The microprocessor reads the sensor’s
output at a given rate
• Output is then used to act
Poling mode
• Sensor operates as if the microprocessor
did not exist.
• Its output is monitored by the
• The microprocessor reads the sensor’s
output at a given rate or intervals - poling
• Output is then used to act
Interrupt mode
• Microprocessor is in sleep mode
• Outputs of the sensor are not being
• Upon a given event, microprocessor
wakes up through one of its interrupt
• The sensor activates the interrupt
• Interrupts can be timed
• Interrupts can be issued by sources other
than the sensor
• The microprocessor may be involved in
other functions, separate from the sensor,
such as control of an actuator
• Feedback from actuators may also be
used to perform interrupts
General Interfacing
• Microprocessor input interfacing
• Microprocessor output
• Errors introduced by
Input interfacing
• Signal level
• Impedance and matching
• Response, frequency
• Signal conditioning
• Signal scaling
• Isolation
• Loading
Output interfacing
• Signal levels
• Power levels
• Isolation
Input signal levels
• Basic level: zero to Vdd
– Must scale signals if necessary
• No dual polarity signals
– Must translate/scale as necessary
• Direct reading or A/D
• Can read voltages only
– AC or DC
– Limitations in frequency
 P are high input impedance devices
– ~ 1 - 10 M
– Input current - < 1 A.
• Ideal for direct connection of low
impedance sensors (magnetic,
thermistors, thermoelectric, etc.)
• High impedance sensors (capacitive,
pyroelectric, etc.) must be buffered
– Voltage followers
– FET amplifiers
Response and frequency
• Most sensors are slow devices
– Can be interfaced directly
– No concern for response and frequency
• Some sensors are part of oscillators
– Frequencies may be quite high
– Need to worry about proper sampling by
the microprocessor
Response and frequency
• Example: 10 mHz P, cycle time of 0.4
s. (most processor divide the clock
frequency by a factor - 4 in this case)
• Any operation such as reading an input
required n cycles, say n=5
• Effective frequency: 0.5 MHz
• Sampling cannot be done at rates
higher than 250 kHz
• Any sensor producing a signal above
this frequency will be read erroneously
Response and frequency
• Some solutions:
– Divide the sensor’s frequency
• Reduces sensitivity
• Must be done externally to the P
– F-V converter
• Introduces conversion errors
• Must be done externally
– Frequency counter at input
• Use output of the counter as input to mP.
• Expensive
– Faster microprocessors
Input signal conditioning
• Offset
Primarily for dc levels
Can be offset up or down
Usually done to remove the dc level
Sometimes needed to remove negative
– AC signals may sometimes be coupled
through capacitors to eliminate dc levels
• Example
– Thermistor: 500 at 20ºC
– Varies from 100 to 900 for temp. between 0
and 100ºC
• At 500ºC
– V = (12/1500)*500 = 4 V
• At 0ºC
– V = (12/1400)*400 = 3.428 V
• At 100ºC
– V = (12/1900)*900 = 5.684 V
• V varies between 3.428V and 5.684V
– 5.684V is above the 5V operating voltage of
the microprocessor
• Some solutions
– Remove 3.428V through an inverting amplifier
– Reduce the source voltage from 12V to, say
6V. This will change the range from 1.714V to
– Increase the resistor from 1000W to, say,
1500 W. This will reduce the output and will
vary from 2.526V to 4.5V
Offset - other solutions
• For ac signals
– Rectification
• Only appropriate if signal is unipolar
– Bi-polar signals produce negative signals
• Cannot be used with microprocessors
Offset - other solutions
• Bridge connection
– Battery must be floating
– Output: 0V at 0ºC to 2.3V at 100ºC.
– Offset of arbitrary value can be added
• Done by decreasing the value of lower-left resistor
• 1V offset with 285.7 resistor
• By amplification
– Operational amplifiers
• By attenuation
– Operational amplifiers
– Resistance dividers
– Transformers (for ac)
• Amplifiers are preferrable
• Dividers introduce errors
• Transformers are noisy and big
• Two basic methods
– Transformers
– Optical isolation
• Microprocessors load the sensor
• Not an issue with low impedance sensors
• Must be buffered for high impedance
• Solution: voltage followers with FET input
• An error due to loading should be taken into
Output Interface
• Most microprocessors:
– 1.8 to 6V
– 20 to 25 mA per output pin
– Can power small loads directly (LEDs, small
– Protection diodes on all outputs
Output Interface
• Large loads:
– Must add circuitry to boost current, power
– MOSFETS are ideal for this purpose
– Inductive loads: must add protection against
large spikes
– Often necessary to isolate output
– Very often necessary to translate voltages for
Output pins
– Driven
Output pins connection of
• Sourcing current
• Sinking current
• The two are somewhat different:
Errors and resolution
• Errors introduced by the
– Due to resolution of A/D, D/A
– Sampling errors
• These come in addition to any errors
in the sensor/actuator
• Digital systems have an inherent
• LSB - least significant bit
– Any value smaller than the LSB cannot
be represented
– This constitutes an error
– LSB is inherent in any module as well as
in the CPU itself
Resolution of modules
• A/D - n bits resolution, meaning:
a 10 bit A/D, digitizing a 5V input has a resolution
5V/1024 = 4.88 mV
The A/D can resolve down to 4.88 mV
Can represent data in increments of 4.88 mV
(a 14 bit A/D resolves down to 0.3 mV)
For a 1V span on a sensor, this is approximately
0.5% error
Resolution of modules
• PWM (Pulse Width Modulator)
• Given a clock frequency fosc, the PWM resolution
PWM res . =
log (fosc /fPWM )
log (2)
CPU errors
• Most microprocessors are 8 bit
• Integer arithmetics
• Largest value represented: 256
• Roundoff errors due to this representation
• Special math subroutines have been
developed to minimize these errors
(otherwise they would be unacceptably
Sampling errors
• All inputs and outputs on a microprocessor
are sampled. That is:
– Inputs are only read at intervals
– Outputs are only updated at intervals
– Intervals depend on the frequency of the clock, operation
to be executed and on the software that executes it
– Sampling may not even be constant during operation
because of the need to perform different tasks at different
– Errors are due to changes in input/output between
sampling to which the microprocessor is oblivious
– Errors are not fixed - depend among other things on how
well the program is written

PowerPoint Presentation - Interfacing to Microprocessors