What is a Microcontroller?
|
 Basic microcontroller architecture |
The
microcontroller's building blocks explained
To illustrate the functions and interconnectivity of the building blocks of the microcontroller, we shall construct the microcontroller block by block:
Memory unit
Memory is
part of the microcontroller whose function is to store data.
The easiest way to explain it is to describe it as one big closet
with lots of drawers. If we suppose that we marked the drawers in
such a way that they can not be confused, any of their contents will
then be easily accessible. It is enough to know the designation of
the drawer and so its contents will be known to us for sure.
Memory components are exactly like that. For a certain input we get the contents of a certain addressed memory location and that's all. Two new concepts are brought to us: addressing and memory location. Memory consists of all memory locations, and addressing is nothing but selecting one of them. This means that we need to select the desired memory location on one hand, and on the other hand we need to wait for the contents of that location. Beside reading from a memory location, memory must also provide for writing onto it. This is done by supplying an additional line called control line. We will designate this line as R/W (read/write). Control line is used in the following way: if r/w=1, reading is done, and if opposite is true then writing is done on the memory location. Memory is the first element, and we need a few operation of our microcontroller.
Central Processing Unit
The block
that will have a built in capability to multiply, divide, subtract,
and move its contents from one memory location onto another is
called "central processing unit" (CPU). Its memory locations are
called registers.
Registers are therefore memory locations whose role is to help with
performing various mathematical operations or any other operations
with data wherever data can be found. Look at the current situation.
We have two independent entities (memory and CPU) which are
interconnected, and thus any exchange of data is hindered, as well
as its functionality. If, for example, we wish to add the contents
of two memory locations and return the result again back to memory,
we would need a connection between memory and CPU. Simply stated, we
must have some "way" through data goes from one block to another.
Bus
That
"way" is called "bus". Physically, it represents a group of 8, 16,
or more wires
There are two types of buses: address and data bus. The first one
consists of as many lines as the amount of memory we wish to
address, and the other one is as wide as data, in our case 8 bits or
the connection line. First one serves to transmit address from CPU
memory, and the second to connect all blocks inside the
microcontroller.
As far as functionality, the situation has improved, but a new
problem has also appeared: we have a unit that's capable of working
by itself, but which does not have any contact with the outside
world, or with us! In order to remove this deficiency, let's add a
block which contains several memory locations whose one end is
connected to the data bus, and the other has connection with the
output lines on the microcontroller which can be seen as pins on the
electronic component.
Input-output unit
Those
locations we've just added are called "ports". There are several
types of ports : input, output or bidirectional ports. When working
with ports, first of all it is necessary to choose which port we
need to work with, and then to send data to, or take it from the
port.
When working with it the port acts like a memory location. Something
is simply being written into or read from it, and it could be
noticed on the pins of the microcontroller.
Serial communication
Beside
stated above we've added to the already existing unit the
possibility of communication with an outside world. However, this
way of communicating has its drawbacks. One of the basic drawbacks
is the number of lines which need to be used in order to transfer
data. What if it is being transferred to a distance of several
kilometres? The number of lines times number of kilometres doesn't
promise the economy of the project. It leaves us having to reduce
the number of lines in such a way that we don't lessen its
functionality. Suppose we are working with three lines only, and
that one line is used for sending data, other for receiving, and the
third one is used as a reference line for both the input and the
output side. In order for this to work, we need to set the rules of
exchange of data. These rules are called protocol. Protocol is
therefore defined in advance so there wouldn't be any
misunderstanding between the sides that are communicating with each
other. For example, if one man is speaking in French, and the other
in English, it is highly unlikely that they will quickly and
effectively understand each other. Let's suppose we have the
following protocol. The logical unit "1" is set up on the
transmitting line until transfer begins. Once the transfer starts,
we lower the transmission line to logical "0" for a period of time
(which we will designate as T), so the receiving side will know that
it is receiving data, and so it will activate its mechanism for
reception. Let's go back now to the transmission side and start
putting logic zeros and ones onto the transmitter line in the order
from a bit of the lowest value to a bit of the highest value. Let
each bit stay on line for a time period which is equal to T, and in
the end, or after the 8th bit, let us bring the logical unit "1"
back on the line which will mark the end of the transmission of one
data. The protocol we've just described is called in professional
literature NRZ (Non-Return to Zero).
As we have separate lines for receiving and sending, it is possible
to receive and send data (info.) at the same time. So called
full-duplex mode block which enables this way of communication is
called a serial communication block. Unlike the parallel
transmission, data moves here bit by bit, or in a series of bits
what defines the term serial communication comes from. After the
reception of data we need to read it from the receiving location and
store it in memory as opposed to sending where the process is
reversed. Data goes from memory through the bus to the sending
location, and then to the receiving unit according to the protocol.
Timer unit
Since
we have the serial communication explained, we can receive, send and
process data.
However, in order to utilize it in industry we need a few
additionally blocks. One of those is the timer block which is
significant to us because it can give us information about time,
duration, protocol etc. The basic unit of the timer is a free-run
counter which is in fact a register whose numeric value increments
by one in even intervals, so that by taking its value during periods
T1 and T2 and on the basis of their difference we can determine how
much time has elapsed. This is a very important part of the
microcontroller whose understanding requires most of our time.
Watchdog
One more
thing is requiring our attention is a flawless functioning of the
microcontroller
during its run-time. Suppose that as a result of some interference
(which often does occur in industry) our microcontroller stops
executing the program, or worse, it starts working incorrectly.
Of course, when this happens with a computer, we simply reset it and
it will keep working. However, there is no reset button we can push
on the microcontroller and thus solve our problem. To overcome this
obstacle, we need to introduce one more block called watchdog. This
block is in fact another free-run counter where our program needs to
write a zero in every time it executes correctly. In case that
program gets "stuck", zero will not be written in, and counter alone
will reset the microcontroller upon achieving its maximum value.
This will result in executing the program again, and correctly this
time around. That is an important element of every program to be
reliable without man's supervision.
Analog to Digital Converter (ADC)
As the
peripheral signals usually are substantially different from the ones
that microcontroller can understand (zero and one), they have to be
converted into a pattern which can be comprehended by a
microcontroller. This task is performed by a block for analog to
digital conversion or by an ADC. This block is responsible for
converting an information about some analog value to a binary number
and for follow it through to a CPU block so that CPU block can
further process it.
Finally, the microcontroller is now completed, and all we need to do
now is to assemble it into an electronic component where it will
access inner blocks through the outside pins. The picture below
shows what a microcontroller looks like inside.
Thin lines which lead from the center towards the sides of the
microcontroller represent wires connecting inner blocks with the
pins on the housing of the microcontroller so called bonding lines.
Chart on the following page represents the center section of a
microcontroller.
For a real application, a microcontroller alone is not enough.
Beside a microcontroller, we need a program that would be executed,
and a few more elements which make up a interface logic towards the
elements of regulation .
Program
Program
writing is a special field of work with microcontrollers and is
called "programming". Try to write a small program in a language
that we will make up ourselves first and then would be understood by
anyone.
START
REGISTER1=MEMORY LOCATION_A
REGISTER2=MEMORY LOCATION_B
PORTA=REGISTER1 + REGISTER2
END
The program adds the contents of two memory locations, and views
their sum on port A. The first line of the program stands for moving
the contents of memory location "A" into one of the registers of
central processing unit. As we need the other data as well, we will
also move it into the other register of the central processing unit.
The next instruction instructs the central processing unit to add
the contents of those two registers and send a result to port A, so
that sum of that addition would be visible to the outside world. For
a more complex problem, program that works on its solution will be
bigger.
Programming can be done in several languages such as Assembler, C
and Basic which are most commonly used languages. Assembler belongs
to lower level languages that are programmed slowly, but take up the
least amount of space in memory and gives the best results where the
speed of program execution is concerned. As it is the most commonly
used language in programming microcontrollers it will be discussed
in a later chapter. Programs in C language are easier to be written,
easier to be understood, but are slower in executing from assembler
programs. Basic is the easiest one to learn, and its instructions
are nearest a man's way of reasoning, but like C programming
language it is also slower than assembler. In any case, before you
make up your mind about one of these languages you need to consider
carefully the demands for execution speed, for the size of memory
and for the amount of time available for its assembly.
After the program is written, we would install the microcontroller
into a device and run it. In order to do this we need to add a few
more external components necessary for its work. First we must give
life to a microcontroller by connecting it to a power supply (power
needed for operation of all electronic instruments) and oscillator
whose role is similar to the role that heart plays in a human body.
Based on its clocks microcontroller executes instructions of a
program. As it receives supply microcontroller will perform a small
check up on itself, look up the beginning of the program and start
executing it. How the device will work depends on many parameters,
the most important of which is the skilfulness of the developer of
hardware, and on programmer's expertise in getting the maximum out
of the device with his program.
Microcontrollers vs. Microprocessors:
Microcontroller differs from a microprocessor in many ways. First and the most important is its functionality. In order for a microprocessor to be used, other components such as memory, or components for receiving and sending data must be added to it. In short that means that microprocessor is the very heart of the computer. On the other hand, microcontroller is designed to be all of that in one. No other external components are needed for its application because all necessary peripherals are already built into it.
