1. Interfacing to Microprocessors
Chapter 12

2. introduction
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 microprocessors.

3. introduction
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

4. 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.

5. 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.

6. The microprocessor
Two limits on the tasks microprocessors can perform:
The limitations of the microprocessor itself
The imagination (or capabilities) of the designer.

7. 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.

8. The architecture
There are about two dozen manufacturers of microprocessors
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

9. The architecture Main features:
Separate busses for program memory and operand memory.
Pipelined architecture
Allows fetching data while another operation executes.
Each cycle consists of fetching the (n+1)th instruction while executing the nth
Integer arithmetic
Limited instruction set

10. 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.

11. Bus architecture

12. 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.

13. PIC10FXX microprocessors

14. PIC10FXX microprocessors

15. The architecture Example: one of the largest, is the PIC18FXX20
Has an address bus 21 bits wide.
The processor and its variants are shown in Table 12.2
Its architecture in Figure 12.3.

18. The architecture Architecture supports:
Direct addressing for the first 8 bits of address space
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,

19. 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.

20. The architecture Memory varies from 256 bytes to over 256 kbytes
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.)

21. Addressing 8 bit microprocessors have word length of 8 bits.
Integer data from 0 to 255 may be represented directly.
To address memory, usually a longer word is needed.
Most microprocessor have a 12 bit (4k) 14 (16k) or 16 bit (64k) memory address but longer address words are also used.

22. Speed
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

23. Instruction set Microprocessors have a small instruction set –
sometimes no more than 2-3 dozen simple instructions.
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.

24. 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

25. 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 common).
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

26. 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.

27. 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 mA.
This is not sufficient to drive many actuators but it can drive low power devices directly or indirectly through switches and amplifiers.

28. 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 oscillation
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).

29. 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.

30. Clock and timers Registers Used for Execution of commands
Control over the functions of the microprocessor,
Addressing
Flagging
Status indication

31. Memory Modern microprocessors, contain three types of memory:
program memory, in which the program is loaded,
data memory (RAM),
EEPROM memory
Note: EEPROM not available on some very small microprocessors.

32. Memory 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

33. Power Most microprocessor operate from 1.8V to 6V.
Some have a more limited range ( V).
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 consumed

34. 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.

35. 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.

36. Other functionalities
One or two comparators are provided on many microprocessors.
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 processors.
Serial interfaces such as UART, SPI, two wire interface (I2C), synchronous serial and USB ports are available

37. 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)

38. Programs and programmability
A microprocessor is only useful if it can be programmed.
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.

39. Programs and programmability
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.

40. Programs and programmability
Instruction sets for microprocessors are small and based on the assembly language nomenclature.
Microprocessors have been designed for integer operations.
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.

41. Programs and programmability
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.

42. 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.

43. Interfacing Issues Three basic modes:
1. Continuous dedicated monitoring of the sensor by the microprocessor
2. Polling the sensor
3. Interrupt mode

44. 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

45. Poling mode Sensor operates as if the microprocessor did not exist.
Its output is monitored by the microprocessor
The microprocessor reads the sensor’s output at a given rate or intervals - poling
Output is then used to act

46. Interrupt mode Microprocessor is in sleep mode
Outputs of the sensor are not being processed
Upon a given event, microprocessor wakes up through one of its interrupt options
The sensor activates the interrupt

47. Notes: 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

48. General Interfacing Requirements
Microprocessor input interfacing requirements
Microprocessor output requirements
Errors introduced by microprocessors

49. Input interfacing requirements
Signal level
Impedance and matching
Response, frequency
Signal conditioning
Signal scaling
Isolation
Loading

50. Output interfacing requirements
Signal levels
Power levels
Isolation

51. Input signal levels Basic level: zero to Vdd No dual polarity signals
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

52. Impedance P are high input impedance devices
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

53. Response and frequency
Most sensors are slow devices
Can be interfaced directly
No concern for response and frequency range
Some sensors are part of oscillators
Frequencies may be quite high
Need to worry about proper sampling by the microprocessor

54. 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

55. 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

56. 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 polarity.
AC signals may sometimes be coupled through capacitors to eliminate dc levels

57. Offset Example Thermistor: 500 at 20ºC
Varies from 100 to 900 for temp. between 0 and 100ºC

58. Offset At 500ºC At 0ºC At 100ºC V varies between 3.428V and 5.684V
5.684V is above the 5V operating voltage of the microprocessor

59. Offset 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 2.842V
Increase the resistor from 1000W to, say, 1500 W. This will reduce the output and will vary from 2.526V to 4.5V

60. Offset - other solutions
For ac signals
Rectification
Only appropriate if signal is unipolar
Bi-polar signals produce negative signals
Cannot be used with microprocessors

61. 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

62. Scaling By amplification By attenuation Amplifiers are preferrable
Operational amplifiers
By attenuation
Resistance dividers
Transformers (for ac)
Amplifiers are preferrable
Dividers introduce errors
Transformers are noisy and big

63. Isolation
Two basic methods
Transformers
Optical isolation

64. Loading Microprocessors load the sensor
Not an issue with low impedance sensors
Must be buffered for high impedance sensors
Solution: voltage followers with FET input stages
An error due to loading should be taken into account

65. Output Interface Most microprocessors: 1.8 to 6V
20 to 25 mA per output pin
Can power small loads directly (LEDs, small relays)
Protection diodes on all outputs

66. 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

67. Output pins
MOSFETS:
Driven

68. Output pins connection of loads
Sourcing current
Sinking current
The two are somewhat different:

69. Errors and resolution Errors introduced by the microprocessor:
Due to resolution of A/D, D/A
Sampling errors
These come in addition to any errors in the sensor/actuator

70. Resolution Digital systems have an inherent resolution:
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

71. Resolution of modules A/D - n bits resolution, meaning:
a 10 bit A/D, digitizing a 5V input has a resolution of:
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

72. Resolution of modules PWM (Pulse Width Modulator)
Given a clock frequency fosc, the PWM resolution is:

73. CPU errors Most microprocessors are 8 bit microprocessors
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 high)

74. 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 times
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