5.6 Basic Computer Structure: Basic Maintenance Training Manual Module 5 Digital Technology / EIS

Sep04 / THTT
Copyright by SR Technics Switzerland

Corresponding with EASA Part-66
For training purposes only 5.6 - 1 Cat: B1.1
Basic Maintenance
Training Manual
Module 5 Digital Technology / EIS
5.6 Basic Computer Structure
Sep04 / THTT
Basic Maintenance
Training Manual
Introduction to Computers
From digital Logic, the implementation of that logic with Integrated circuits, inter-
facing Integrated Circuit logic families and the use of memory to store information
used by the ICs, we come to the computer, which is the synthesis of all these tech-
nologies. It combines a microprocessor IC, memory chips and user interface into
the modern digital computer. A computer combines both components; hardware
and a collection of programs; software, to tell it what to do. This sub module will
focus on the basic structure of the computer: its internal physical components, the
chips and how they work and interact and its external I/O devices for communica-
tion with a user.
Computer Organization
The architecture of a computer is the arrangement of its internal subsystems: the
microprocessor(s), memory, I/O and interfacing. Each subsystem may be concen-
trated on a single IC or spread between many chips. The microprocessor, also
known as the central processing unit (CPU) and usually a single chip, consists of
three parts: a control unit, an arithmetic logic unit (ALU) and temporary storage
registers. A bus a set of wires to carry address, data and control information
interconnects all of the subsystems. Most modern computers are some variation
on the basic architecture shown in Figure 1. The microprocessor, memory chips
and other circuitry are all part of the systems hardware, the physical components
of a system. The computer case, the nuts and bolts and physical parts are other
parts of the hardware. A computer also includes software, a collection of programs
or sequence of instructions to perform a specified task. Some microprocessors in-
ternally are complete circuitry. The design of general purpose computers is so
complex, however, that it is nearly impossible to design an original architecture
without any bugs. Thus many designers use microprocessors that include micro
code or micro instructions: instructions in the control unit of a microprocessor. This
hybrid between hardware and software is called firmware. Firmware also includes
software stored in ROM or EPROM rather than being stored on magnetic disk or
tape.
Computer designers make decisions on hardware, software and firmware based
on cost versus performance. Thus, todays computer market includes a wide range
of systems, from high-performance super-computers, which cost millions of dol-
lars, to the personal microcomputer, with costs in the thousands new and in the
hundreds for older used models.
Figure 1: Basic Computer Structure
The Central Processing Unit
The central processing unit is usually a single microprocessor chip, although its
subsystems can be on more than one chip. The CPU at least includes a control
unit, timing circuitry, an arithmetic logic unit (ALU) and also usually contains reg-
isters for temporary storage.
Control Unit
The control unit directs the operation of the computer, managing the interaction
between subunits. It takes instructions from the memory and executes them, per-
forming tasks such as accessing data in memory, calling on the ALU or performing
Sep04 / THTT
Basic Maintenance
Training Manual
Input/Output. Control is one of the most difficult parts to design; thus it is the most
likely source of bugs in designing an original architecture.
Microprocessors consist of both hard wired control and micro programmed control.
In both cases, the designer determines a sequence of states through which the
computer cycles, each with inputs to examine and outputs to activate other CPU
subsystems (including activating itself, indicating which state to do next). For ex-
ample, the sequence usually starts with Fetch the next instruction from memory,
with control outputs to activate memory for a read, a program counter to send the
address to be fetched and an instruction register to receive the memory contents.
Hard wired control is completely via circuitry, usually with a programmed logic ar-
ray. Micro programmed control uses a microprocessor with a modifiable control
memory, containing micro code or micro instructions. An advantage of micro pro-
grammed control is flexibility: the code can be changed without changing the hard-
ware, making it easier to correct design errors. Figure 2 and Figure 3 show
examples of hard wired control and micro programmed control of a microproces-
sor.
Figure 2: Hard wired Control
Figure 3: Microprogrammed Control
Timing
Usually, an oscillator controlled by a quartz crystal generates the microcomputers
clock signal. The output of this clock goes to the microprocessor and to other ICs.
The clock synchronizes the microcomputer subunits. For example, each of the mi-
cro instructions is designed to take only one clock cycle to execute, so any com-
ponents triggered by a micro-instructions control outputs should finish their
actions by the end of the clock cycle. The exception to this is memory, which may
take multiple clock cycles to finish, so the control unit repeats in its same state until
memory says its done.
Since the clock rate effectively controls the rate at which instructions are executed,
the clock frequency is one way to measure the speed of a computer.
Clock frequency, however, cannot be the only criteria considered because the ac-
tions performed during a clock cycle vary for different designs.
Sep04 / THTT
Basic Maintenance
Training Manual
Arithmetic Logic Unit
The arithmetic logic unit (ALU) performs logical operations such as AND, OR and
SHIFT and arithmetic operations such as addition, subtraction, multiplication and
division. The ALU depends on the control unit to tell it which operation to perform
and also to trigger other devices (memory, registers and I/O) to supply its input
data and to send out its results to the appropriate place.
The ALU often only performs simple operations. Complex operations, such as mul-
tiplication, division and operations involving decimal numbers, are performed by
dedicated hardware, called floating-point processors, or floating point units. These
may be included on the original mother board or may be optional upgrades. Now-
adays in modern microprocessors, usually a floating point unit is integrated on the
same microprocessor IC.
Registers
Microprocessor chips have some internal memory locations that are used by the
control unit and ALU. Because they are inside the microprocessor IC, these regis-
ters can be accessed more quickly than other memory locations. Special purpose
registers or dedicated registers are purely internal, have predefined uses and can-
not be directly accessed by programs. General purpose registers hold data and
addresses in use by programs and can be directly accessed, although usually only
by assembly level programs.
The dedicated registers include the instruction register, program counter, effective
address register and status register. The first step to execute an instruction is to
fetch it from memory and put it in the instruction register (IR). The program counter
(PC) is then incremented to contain the address of the next instruction to be
fetched. An instruction may change the program counter as a result of a condition-
al branch (if-then), loop, subroutine call or other non linear execution. If data from
memory is needed by an instruction, the address of the data is calculated and
fetched with the effective address register (EAR).
The status register (SR) keeps track of various conditions in the computer. For ex-
ample, it tells the control unit when the keyboard has been typed on so the control
unit knows to get input. It also notices if something goes wrong during an instruc-
tion execution, for example an attempted divide by 0, and tells the control unit to
halt the program or fix the error. Certain bits in the status register are known as the
condition codes, flags set by each instruction. These flags tell information about
the result of the latest instruction such as if the result was negative or positive
or zero and if an arithmetic overflow or a carry error occurred. The flags can then
be used by a conditional branch to decide if that branch should be taken or not.
Some architectures also use a stack pointer (SP) and/or an accumulator. In a
stack system, a memory location is designated as the bottom of the stack. Data
to be stored is always added to the next memory location, the top of the stack;
and data to be accessed must always be taken off the top. (This technique is
called last in, first out, or LIFO.)
Highway Structure
Figure 4 shows the highway structure, there are three buses, the address bus, the
data bus and the control bus. As we have already seen the memory consists of a
number of locations, each individually identified by an address.
The address bus is therefore used to specify the memory location or input/output
port involved in the transfer. It is a one way bus and may have anything from 4 to
64 lines depending on the number of memory addresses there are, 8 lines give
2
8
=256 addresses. The data bus is a bi-directional bus and is used to carry the
data being transferred to and from the memory or input and output transfer.
The control bus comprises input and output lines which synchronise the micro-
processors operation with that of the external circuitry i.e. read/write controls, tim-
ing signals, input/output selection. This is also a bi-directional bus.
Figure 4: Highway Structure
Sep04 / THTT
Basic Maintenance
Training Manual
Memory
Computers and other digital circuits rely on stored information, either data to be
acted upon or instructions to direct circuit actions. This information is stored in
memory devices, in binary form. This section first discusses how to access an in-
dividual item in memory and then compares different memory types, which can
vary how quickly and easily an item is accessed.
Accessing a Memory Item
Memory devices consist of a large number of memory cells each capable of re-
membering one bit of binary information. The information in memory is stored in
digital form with collections of bits, called words, representing numbers and sym-
bols. The most common symbol set is the American National Standard Code for
Information Interchange (ASCII). Words in memory, just like the letters in this sen-
tence, are stored one after the other. They are accessed by their location or ad-
dress. The number of bits in each word, equal to the number of memory cells per
memory location, is constant within a memory device but can vary for different de-
vices. Common memory devices have word sizes of 8, 16 and 32 bits.
Addresses and Chip Size
An address is the identifier, or name, given to a particular location in memory.
Since this address is expressed as a binary number, the number of unique ad-
dresses available in a particular memory chip is determined by the number of bits
to express the address. For example, a memory chip with 8 bit addresses has 2
8
= 256 memory locations. These locations are accessed as the addresses
00000000 through 11111111, 0 through 255 decimal or 00 through FF hex. (For
ease of notation, programmers and circuit designers use hexadecimal, base 16,
notation to avoid long strings of 1s and 0s.) The memory chip size can be ex-
pressed as M N, where M is the number of unique addresses, or memory loca-
tions and N is the word size, or number of bits per memory location. Memory chips
come in a variety of sizes and can be arranged, together with control circuitry and
decoders, to meet a designers needs.
Basic Structure
Memory chips, no matter how large or small, have several things in common. Each
chip has address, data and control lines, as shown in Figure 5. A memory chip
must have enough address lines to uniquely address each of its words and as
many data lines as there are bits per word. For example, the 256 1 memory has
8 address lines and 1 data line.
The control lines for a memory chip can vary. The figure shows a simple example:
two control lines, a W / R and CS. In this case, data lines transfer both inputs (when
writing) and outputs (when reading) so the W / R control line is needed to put the
memory chip in read mode or write mode. The chip select, CS, control line tells the
chip whether it is in use. When the chip is selected, it is on, acting upon the ad-
dress, data and W / R information presented to it. When the chip is not selected,
the data line enters a high-impedance state so that it does not affect, and is not
affected by devices or circuits attached to it.
Figure 5: Example of a 256 1 memory chip.
Reading and Writing
To write (store data in) or read (retrieve data from) a memory device, it is neces-
sary to gain access to specific memory cells. A small 256 1 memory chip is used
as an example. Later, this example will be expanded to a larger computer memory
system.
If we want to write a 1 to the 11th word of the 256 1 memory (such as memory
location 10 decimal or 00001010 binary), we must execute the following steps:
Sep04 / THTT
Basic Maintenance
Training Manual
1. Place the correct address (00001010) on the address lines.
2. Place the data to be written (1) on the data line.
3. Set the W / R control line to write (low, 0).
4. Set the CS control line to select (high, 1). (Many memory devices use an ac-
tive low chip select, .).
This writes the data on the data line (1) to the address on the address lines
(00001010).
The steps to read the contents of the 11th word are similar except that the W / R
control line is set to read (high, 1).
Timing
Accurate timing requirements must also be incorporated into the above steps.
While writing, the address and data information must be present for a minimum
setup time before and hold time after, the CS and signals have been activated.
This is to avoid spurious signals spraying all over the memory array. While read-
ing, address line changes are not harmful, but the output data is only valid a min-
imum access time after the last address input is stable. Manufacturers data sheets
and application notes provide the timing specifications for the particular IC you are
using.
Larger Words
The one-bit-wide memory described above provides a good introduction, but usu-
ally we want a wider memory. One way to get wider memory is to use several 1-
bit-wide memory chips. The address and control lines go to each chip, and data
from each chip is used as a single bit in the large word. It is easy to see that when
reading from address 0A (hex), the data lines D0 through D3 contain the data from
address 0A of chips U0 through U3.
An address placed on the shared address lines (called an address bus) now spec-
ifies an entire word of data. Notice that one line of the address bus connects to the
CS pin of each memory chip. This line is labelled ME, or memory enable, and sets
all four ICs active at the same time, in order to read or write data when the R/W is
activated additionally.
If all four memory chips were put in a single package, they would make a 256 4
IC. This IC would look like the chip in Figure 5, except that it would have 4 data
lines.
Figure 6: A 256 4 memory built with four 256 1 memory chips.
More Address Space
For even larger memory systems, the same principles as shown in Figure 7 can
be applied. Figure 7 shows a 1024 8 (1 kilobyte) memory built from four 256 8
memory chips. A kilobyte is usually abbreviated as K. Notice this is not quite the
same as the metric prefix kilo, because it represents 1024, rather than 1000.
Ten address lines are needed to address 1024 locations (2
10
= 1024). Eight of the
10 address lines, A0 to A7, are used as a normal address bus for chips 0 through
3. The remaining 2 address lines, A8 and A9, are run through a 2-to-4 line decoder
to choose between the 4 memory chips. When employed in this manner, the 2-to-
4 line decoder is called an address decoder.
To assert the CS input for one of the memory chips, ME must be 1 and the correct
output of the 2-to-4 line decoder must also be 1. When an address is placed on A0
through A9, a single memory chip is selected by ME, A8 and A9. The other 8 ad-
dress lines address a single word from that chip. The three chips that are not se-
lected enter a high-impedance state and do not affect the data lines. This example
CS
Sep04 / THTT
Basic Maintenance
Training Manual
shows that, using the proper memory chips and address decoding, any size mem-
ory with any word length can be built.
Figure 7: A 1024 8 memory built with four 256 8 memory chips
Alternate Memory Structures
Figure 8 shows how the same chip can be accessed in different ways by using two
decoders, a row decoder and a column decoder. The same 256 256 memory ar-
ray can be treated as a 64 K 1 array, a 256 256 array or other possibilities. In
fact, most larger memory chips are made as square arrays: 32 32 (1024 bytes
or 1 K), 64 64 (4096 bytes or 4 K), 256 256 (65536 bytes or 64 K), 1024 1024
(1 M), 2048 2048 (4 M), 4096 4096 (16 M) and so on. (Here the M represents
a mega-byte, which is 1048576 bytes.)
The square array makes the chips more cost effective to manufacture (easier qual-
ity control and less waste) and easier to incorporate into a printed-circuit-board cir-
cuit layout. Notice that each M N is a power of 2. So while we refer to the chips
by shorter names like 1 Mbyte, the actual number of memory cells is larger than
1000000. The product, M N, only refers to the number of memory cells in the
chip; and designers are free to choose the word size appropriate to their needs. In
fact, they may access one location as an 8-bit word and another as a 16-bit word.
For example, a computer with a Motorola MC68000 microprocessor automatically
accesses a character, such as A, as 8 bits (a byte), an integer as 16 bits (a word)
and a real value as 32 bits (a longword). (Apple Macintosh computers use the
MC68000 microprocessor.) To further complicate things, a different manufacturer
may call 16 bits a halfword and 32 bits a word. In using memory, the controller
chips and circuitry to access the memory can be just as important as the memory
itself.
Figure 8: Row and column decoders
Sep04 / THTT
Basic Maintenance
Training Manual
Memory Types
The concepts described above are applied to several types of random-access,
semiconductor memory. Semiconductor memories are categorized by the ease
and speed with which they can be accessed and their ability to remember in the
absence of power.
Volatile Memory
SAM versus RAM
Sequential-access memory (SAM) must be accessed by stepping past each mem-
ory location until the desired location is reached. Magnetic tapes implement SAM;
to reach information in the middle of the tape, the tape head must pass over all of
the information on the beginning of the tape. Two special types of SAM are the
queue and the push-down stack. In a queue, also called a first-in, first-out (FIFO)
memory, locations must be read in the order that they were written. The queue is
a first-come, first-served device, like a line at a ticket window. The push-down
stack is also called last-in, first-out (LIFO) memory. In LIFO memory, the location
written most recently is the next location read. LIFO can be visualized as a stack,
always adding to and removing from the top of the stack.
Random-access memory (RAM) allows any memory cell to be accessed at any in-
stant, with no time wasted stepping past the beginning parts of the data. Ran-
dom-access memory is like a bookcase; any book can be pulled out at any time.
It is usually faster to access a desired word in RAM than in SAM. Also, all words
in RAM have the same access time, while each word in a SAM has a different ac-
cess time based on its position. Generally, the semiconductor memory devices in-
ternal to computers are random-access memories. Magnetic devices, such as
tapes and disks, have at least some sequential access characteristics. We will
leave tapes and disks for a later section and concentrate here on random-access,
solid-state memories.
Random Access Memory
Most RAM chips are volatile, meaning that stored information is lost if power is re-
moved. RAM is either static or dynamic. Dynamic RAM (DRAM) stores a bit of in-
formation as the presence or absence of charge. This charge, since it is stored in
a capacitor, slowly leaks away. It must be refreshed periodically. Memory refresh
typically occurs every few milliseconds and is usually performed by a dynamic
RAM controller chip. Static RAM (SRAM) stores a bit of information in a flip-flop.
Since the bit will retain its value until either power is removed or another bit replac-
es it, refresh is not necessary.
Both types of RAM have their advantages and disadvantages. The advantage of
DRAM is increased density and ease of manufacture, making them significantly
less expensive. SRAMs, however, have much faster access times. Most general
purpose computers use DRAMs, since large memory size and low cost are the
major objectives. Where the amount of memory required doesnt justify the use of
DRAM, and the faster access time is important, SRAMs are common, for exam-
ple, in embedded systems (telephones, toasters) and for cache memories. Both
types of RAM are available in MOS families; SRAMs are also available in bipolar.
Generally, MOS RAMs have lower power consumption than bipolar RAMs, while
access speeds vary widely. Cost, power consumption and access time, provided
in manufacturers data sheets, are factors to consider in selecting the best RAM
for a given application.
Figure 9: (A) DRAM Cell vs. (B) an SRAM Cell
(A) (B)
bit line
address
pass
gate
storage
capacitor
row-select line
bit line bit line
Sep04 / THTT
Basic Maintenance
Training Manual
Non-Volatile Memory
Read-Only Memory (ROM)
Read-Only Memory is nonvolatile; its contents are not lost when power is removed
from the memory. Despite its name, all ROMs can be written or programmed at
least once. The earliest ROM designs were written by clipping a diode between
the memory bit and power supply wherever a 0 was desired. Modern MOS ROMs
use a transistor instead of a diode. Mask ROMs are programmed by having ones
and zeros etched into their semiconductors at manufacturing time, according to a
pattern of connections and non-connections provided in a mask. Since the pro-
gramming of a mask ROM must be done by the manufacturer, adding expense
and time delays, this type of ROM is primarily used only in high volume applica-
tions.
PROM
For low volume applications, the programmable ROM (PROM) is the most effec-
tive choice since the data can be written after manufacture. A PROM is manufac-
tured with all its diodes or transistors connected. A PROM programmer device
then burns away undesired connections (see Figure 10). This type of PROM can
be written only once.
EPROM, EEPROM
Two types of PROMs that can be erased and reprogrammed are EPROMs and
EEPROMs. The transistors in UV erasable PROMs (EPROMs) have a floating
gate surrounded by an insulating material. When programming with a bit value, a
high voltage creates a negative charge on the floating gate. Exposure to ultraviolet
light erases the negative charge. Similarly, electrically erasable PROMs (EEP-
ROMs) erase their floating-gate values by applying a voltage of the opposite po-
larity.
ROMs are practical only for storing data or programs that do not change frequent-
ly and must survive when power is removed from the memory. The programs that
start up a computer when it is first switched on or the memory that holds the call
sign in a repeater are prime candidates for ROM.
Flash
Flash also known as FEPROM is one of the newest types of nonvolatile memory.
In this device data is erased and reprogrammed in blocks, unlike the byte by byte
altering capability of EEPROM. FLASH memory has the density of EPROMs and
the electrical erase capability of standard EEPROMs. FLASH is used in digital
cameras to store pictures, PC's store their BIOS on FLASH memory chips, and
MP3 audio recorders use FLASH to store music.
Figure 10: Programmable ROM
Nonvolatile RAM
For some situations, the ideal memory would be as nonvolatile as ROM but as
easy to write to as RAM. The primary example is data that must not be allowed to
perish despite a power failure. Low-power RAMs can be used in such applications
Sep04 / THTT
Basic Maintenance
Training Manual
if they are supplied with NiCd or lithium cells for backup power. A more elegant
and durable solution is nonvolatile RAM (NVRAM), which includes both RAM and
ROM. The standard volatile RAM, called shadow RAM, is backed up by nonvolatile
EEPROM. When the RECALL control is asserted, such as when power is first ap-
plied, the contents of the ROM are copied into the RAM. During normal operation,
the system reads and writes to the RAM. When the STORE control is triggered,
such as by a power failure or before turning off the system, the entire contents of
the RAM are copied into the ROM for nonvolatile storage. In the event of primary
power failure, to successfully save the RAM data, some power must be maintained
until the memory store is complete (+5 V for 20 ms).
Cache versus Main Memory
Memory is in high demand for many applications. To balance the trade-off of
speed versus cost, most computers use a larger, slower, but cheaper main mem-
ory in conjunction with a smaller, faster, but more expensive cache memory. As
you run a computer program, it accesses memory frequently. When it needs an
item, a piece of data or the next part of the program to execute, it first looks in the
cache. If the item is not found in the cache, it is copied to the cache from the main
memory. As you run a computer program, it often repeats certain parts of the pro-
gram and repeatedly uses pieces of data. Since this information has been copied
to the high-speed cache, your computer game or other application can run faster.
Information used less often or not being used at all (programs not currently being
run) can stay in the slower main memory.
A cache is a place to store treasure; the treasure, the information you are using
frequently, can be accessed quickly because it is in the high-speed cache. The use
of cache versus main memory is managed by a computers CPU so it is transpar-
ent to the user. The improvement in program execution time is similar to accessing
a floppy disk versus the computers internal memory.
Sep04 / THTT
Basic Maintenance
Training Manual
Peripherals
Peripherals are any devices outside the CPU. They provide additional capabilities.
One of the most common examples being communication with a user via input de-
vices and output devices. Input devices provide the computer both data to work on
and programs to tell it what to do. Output devices present the results of computer
operations to the user or another system and may even control an external sys-
tem.
Both input and output combine to provide user friendly interaction. This section dis-
cusses the most common user interfaces.
Most of these devices have adapted to certain standards and use readily available
connection cables. Thus, they can be easily incorporated into a system, and a
knowledge of the internal actions is not necessary. A knowledge of how external
memory devices work is more useful and will be discussed in more detail.
Input Devices
The keyboard is probably the most familiar input device. A keyboard simply makes
and breaks electrical contacts. The open or closed contacts are usually sensed by
a microprocessor built into the circuit board under the keys. This microprocessor
decodes the key closures and sends the appropriate ASCII code to the main com-
puter unit. Keyboards will generate the entire 128 character ASCII set and often,
with CONTROL and ALT (Alternate) keys, the 256 character extended ASCII set.
The mouse is becoming increasingly popular for use with graphical user interfac-
es. The mouse casing holds a ball and circuitry to act as a multidirectional detec-
tion device. By moving the mouse, the ball rolls, controlling the relative position of
a cursor on the screen. Buttons on the mouse make and break connections (click-
ing) to select and activate items (icons) on the screen. The trackball is a variation
of the mouse.
Other input devices include modems and magnetic disks and tapes. Magnetic
disks and tapes, discussed at length later, provide additional external memory.
Newer input devices include voice activated devices, touch screens and scanners.
Output Devices
The most familiar output device is the computer screen, or monitor. For smaller
character displays, LED arrays can be used. The next most common output device
is the printer, to produce paper hard copy. Modems and magnetic disks and tapes
are output devices as well as input devices. Newer output devices include speech
synthesizers. The output devices (except the sound device) share a common dis-
play technique: images, such as characters and graphics, are formed by tiny dots,
called pixels (picture elements). On screens, these are dots of light turned on and
off. In printers, they are dots of ink imposed onto the paper. For colour displays,
pixels in red, green and blue (RGB) are spaced closely together and appear as nu-
merous colours to the human eye.
Video monitors are usually specialized cathode-ray tube (CRT) displays. In newer
notebook computers, the monitors are being fabricated with monochrome or col-
our liquid-crystal displays (LCDs), but colour LCD displays are still quite expen-
sive. A standard TV set can even be used as a computer monitor. Two techniques
are used to turn on the screen pixels. Raster scanning covers the screen by writing
one row of pixels at a time, from left to right and top to bottom. Then, a vertical re-
trace brings the beam back to the top of the screen to begin again. Raster scan-
ning signals every pixel on or off for each screen pass. An alternative, vector
mode, only signals the pixels where something on the screen has changed.
Light-emitting diodes (LEDs) are handy when a full screen display is not neces-
sary. A single diode easily indicates on/off states such as the power light on many
devices. A popular single character display is the seven segment version, which is
good for displaying numbers.
Disk Drives
Magnetic Disc Drives
Magnetic media are essential input/output devices since they provide additional
memory. The earliest ways to store programs and data were on punched cards
and tape. Some early home computers used audio cassettes.
Disk storage is prevalent when random access is needed. In some ways, disk stor-
age is similar to that of a record player. The data is stored in circles (tracks) on a
round platter (disk or diskette) and accessed by a device (a head) moving over the
platter. Unlike the record player, the tracks are concentric rather than spiral and
the head can write as well as read.
Figure 11A shows an example of the disk recording surface. Usually the tracks are
divided into equal-sized storage units, called sectors. Also, since the disk has two
sides, most disks can store information on both sides. Therefore, locating a piece
of information in disk memory means identifying three coordinates: the side, track
and sector.
Accessing a piece of information on the disk system involves a number of wait
times until the data access is complete. First, disks may be either movable-head
or fixed-head, as shown in Figure 11B. In the movable system, a single read/write
Sep04 / THTT
Basic Maintenance
Training Manual
head is attached to a movable arm, so there is a seek time for the movable arm to
position the read/write head on the appropriate track. In a fixed system, each track
has its own read/write head, so seek time is zero since the head is immediately in
position. Second, the data must rotate into position under the read/write head. This
time is called latency. Finally, there is the normal time for the read/write to occur.
A number of types of disk technology are available. In hard disk systems, the disk
is rigid and the read/write head does not contact the disk directly. The absence of
friction between the head and disk allows finer head positioning and higher disk
speeds. Thus, hard disks hold more data and are accessed more quickly than flop-
py disks. Floppy disks enclose the magnetic-media platter in a casing, as shown
in Figure 11C, so the disk can be carried around. The floppy disk can be inserted
into a disk drive and the read/write head automatically extended; when done, the
read/write head is automatically retracted before the disk is ejected from the drive.
Variations in floppy disks include single-sided (SS) or double-sided (DS); single,
double or high density; and 3 1 /2. The density refers to the disk format used by
the disk controller. High data density allows more data to be written to the disk but
requires a higher quality diskette. Not all disks can be written as high density and
not all disk drives can read high density disks.
Dust and dirt on the disk and the imperfections in the disk surface gradually dam-
age both the disk and the head. This means that disks eventually wear out, and
the data on the disk will probably be lost.
Therefore, it is prudent to make backup copies of your disks, stored in a clean, dry,
cool place.
Figure 11: Magnetic Disk Drives
Sep04 / THTT
Basic Maintenance
Training Manual
Optical Disks
Another form of memory is the optical disk, which uses optical rather than electrical
means for reading and writing. It developed from videodisc technology during the
early 1980s. Optical disks have a greater memory capacity than most magnetic
disks; the largest ones can store 1.5 gigabytes of information, which is equal to
about 700,000 pages of printed material. Optical disks come in sizes ranging from
3.5 to 12 inches (30 cm). They are widely used as auxiliary memory when large
memory capacity is required.
CD-ROM
CD-ROM is an abbreviation of Compact Disc Read-Only Memory, a type of com-
puter memory in the form of a compact disc that is read by optical means. A CD-
ROM drive uses a low-power laser beam to read digitized (binary) data that has
been encoded in the form of tiny pits on an optical disk. The drive then feeds the
data to a computer for processing.
The standard compact disc was introduced in 1982 for digital audio reproduction.
But, because any type of information can be represented digitally, the standard CD
was adapted by the computer industry, beginning in the mid-1980s, as a low-cost
storage-and-distribution medium for large computer programs, graphics, and da-
tabases. With a storage capacity of 680 megabytes, the CD-ROM found rapid
commercial acceptance as an alternative to so-called floppy disks (with a maxi-
mum capacity of 1.4 megabytes).
Unlike conventional magnetic storage technologies (e.g., tapes, floppy disks, and
hard disks), CDs and CD-ROMs are not recordable--hence the tag "read only."
This limitation spurred the development of various recordable magnetic-optical hy-
brid storage devices; but they generally failed to penetrate beyond the publishing
world, where large multimedia files are regularly exchanged, because of incompat-
ibility with standard CD and CD-ROM players. In the early 1990s a new type of CD
became available: CD-Recordable, or CD-R. These discs differ from regular CDs
in having a light-sensitive organic dye layer which can be "burned" to produce a
chemical "dark" spot, analogous to an ordinary CD's pits, that can be read by ex-
isting CD and CD-ROM players. Such CDs are also known as WORM discs, for
"Write Once Read Many." A rewritable version based on excitable crystals and
known as CD-RW was introduced in the mid-1990s. Because both CD-R and CD-
RW recorders originally required a computer to operate, they had limited accept-
ance outside of use as computer software and data backup devices.
To handle the proliferation of ever-larger multimedia files (audio, graphic, and vid-
eo) in computer games, educational software, and electronic encyclopaedias--as
well as high-definition movies for television entertainment systems--an expanded
storage medium, digital versatile disc (DVD), was introduced in 1995.
DVD
In 1995 Philips and Sony introduced a new type of disc, known as a digital versatile
disc (DVD), which was able to store up to 4.7 gigabytes of data, such as high-def-
inition digital video files. A DVD has the same dimensions as a standard CD but
cannot be read by a standard CD player, although a DVD player can read standard
CDs. DVD players use a higher-power red laser (0.65 micrometer) that enables
smaller pits (0.4 micrometer) and separation tracks (0.74 micrometer) to be used.
Figure 12: CD-ROM vs. DVD
The DVD player uses a laser that is higher-powered and has a correspondingly fin-
er focus point than that of the CD player. This enables it to resolve shorter pits and
narrower separation tracks and thereby accounts for the DVD's greater storage
capacity.
MOD Magneto-Optical Disks
In magneto-optical disks, which can be erased and rewritten, information is written
into or read from the disk by means of the magnetic properties of spots on its sur-
face. In reading, spots with different directions of magnetization give different po-
larization in the reflected light of a low-power laser beam. In the writing process,
which erases all previous information, every spot on the disk is heated by a strong
Sep04 / THTT
Basic Maintenance
Training Manual
laser beam and is then cooled under a magnetic field. Thus every spot is magnet-
ized in one direction; in other words, every spot stores 0. Then, reversing the di-
rection of the magnetic field, only desired spots are magnetized in the opposite
direction by a strong laser beam, storing 1.
Magnetic Tape
Tape is one of the more inexpensive options for auxiliary memory. Tape access
time is slow, since the data must be accessed sequentially, so tape is primarily
used for backup copies of a systems memory.
Tape is available in cassette form (common sizes are comparable to the cassettes
for a portable tape player and VCR tapes) and on reels (diameter is approximately
one foot).
Digital audio tape (DAT) is replacing other forms of tape backup system in newer
computer systems.
A single 4-mm-wide DAT cartridge, which fits in the palm of your hand, can hold
over 2 gigabytes (GB) of data (1 GB = 1024 MB).
Aircraft Application
In an old type of Digital Flight Data Recorder, airplane operating parameters with
relation to time are recorded for later use during performance evaluation or during
accident investigation. This recorder has a write- and a erase head and stores the
last actual flight parameters.
The old Quick Access Recorder (QAR), or also known as Performance Mainte-
nance Recorder (PMR) uses a 'Tape Low' indicators on it. Recording of parame-
ters begins on track one of the tape. At the end of the track (starting with one), the
recorder senses the end-of-tape, changes direction and records on the next track
in the opposite direction and so on, until the sensing end-of-tape is detected. Be-
cause this recorder type has only a write head, the cassette has to be changed
before reaching the end of tape.
Sep04 / THTT
Basic Maintenance
Training Manual
Computer Self Monitoring Methods
Today, computers are making decisions that are affecting our lives everyday. We
are finding that computers are being used to help us cook meals, heat our homes,
drive our cars and fly our airplanes. When the computer makes a decision, it must
be the RIGHT decision every time. This becomes very important when human life
is at stake. Even as great as they are, computers or something in their system can
become defective, so we must have some way of detecting this when it does oc-
cur. one of the common methods used is called self-monitoring, self-checking or
just simply selfcheck. These methods vary considerably from one designer to an-
other.
Self-monitoring is used to help ensure that the computer or its system is working
properly and is capable of performing the function for which it was designed. There
are provisions for some kind of action upon the detection of a failure.
Categories of Self-Monitoring
A self check can be one simple check or a very complex set of checks, containing
many different types of checks'. This usually depends on many factors, some of
which can be cost, application and degree of safety needed.
Generally speaking most self-check's are software generated in one form or an-
other. The Central Processing Unit will follow a program with the routines neces-
sary to complete all of its self-check's.
There are many categories of self-check's but they will usually be one of the fol-
lowing categories or a combination of one or more.
These categories are:
CPU circuit (including microcomputers),
Unit or module circuit or system circuit self-check's.
Figure 13: Categories of Self-Monitoring
CPU
Module / Unit
System
Airplane
Sep04 / THTT
Basic Maintenance
Training Manual
CPU Circuits
This method is one of the more important self- check's, due mainly to the fact that
most failures can be detected almost immediately. The faster the problem is de-
tected, the faster the failure action can be taken. The other methods can have
some form of delay built-in that can delay response time. Although all self-check's
involve the CPU in one way or another, this method is usually distinguished by it
being incorporated into one or more of the CPU's STATUS, CONTROL, AD-
DRESS or DATA BUS lines.
The following are the three most commonly used forms of CPU self-check's. They
are conversion, timing and data validity.
Figure 14: CPU Circuits
Conversion
This method of self-checking can be used to verify the ROM. There are different
ways this can be accomplished, depending on the designer.
Some of the names used for checking ROM (usually EPROMS) are checksum,
sumcheck or most often just ROM check. This check is used to see if the program-
ming has been changed or altered in some way. We can do this by the checksum
method of calculation. An example of this is a data stream from the EPROM which
is calculated in the checksum register and this checksum is added to the program.
The result is then compared to a stored value to get a pass or fail condition.
Figure 15: Conversion
Sep04 / THTT
Basic Maintenance
Training Manual
Data Validity
To check the areas of the computer, like the RAM, the registers in the CPU, etc.,
we can use a method of self checking called data validity. Data in the form of ones
and zeros is put into memory and register locations to see if they will "toggle"
(change from high to low, or from low to high). An example of a RAM test is to write
all ones into all memory location then read them to make sure they are all ones.
Next we write the complement in all locations, and read if they are all zeroes. The
registers can be checked by loading the registers with a "walking" pattern of one/
zero and then by reading back the pattern.
There are many other types of checks that can be done with this method of testing.
The examples shown are but a few.
Figure 16: Data Validity
Timing
Timing in a electronic circuit is usually a very important consideration but it be-
comes even more important in a computer circuit. Computers must process infor-
mation at a high rate of speed and tasks must be completed at the proper time.
Timing includes the internal timing of the CPU and timing of its support circuitry,
like ROM, RAM or anything that provides information to the CPU.
The methods of checking computer timing are many, but one of the most common
is comparing the timing to a separate timing circuit. This circuit has many names
and designs but the most used is the "watchdog timer".
The basic design of a watchdog timer usually uses a running clock that must be
stopped or reset before a certain number of clock pulses have occurred. The CPU
will have an output to stop or reset the clock. If the clock circuit "times out", the tim-
er circuit will reset or interrupt the CPU to indicate a problem has occurred.
Figure 17: Timing
Sep04 / THTT
Basic Maintenance
Training Manual
Module Self-Check
Just as the CPU circuits needed a type of self-check, so does one step above the
CPU at the module level. The CPU will still be utilized to execute the instruction of
the self-checks. Instead of checking the components directly associated with the
CPU, such as the ROM and RAM, the circuits of the module will be checked.
There are many methods of module self-check. Some of the more common ones
used are feedback, wrap- around, and comparison. Each will be discussed individ-
ually.
Feedback
The most widely used form of self-checking is to use feedback. Feedback is the
process of using a portion of the output fed back to the source to allow monitoring
of that output. One example would be a portion of an analog signal fed back
through a A/D converter and into the CPU for comparison. The program would al-
ready contain the ratio of the output to the feedback and the monitoring would take
place any time there was an output. Figure 18 illustrates the use of feedback self-
checking.
Figure 18: Feedback Self-Check
Wrap-Around
Another form of feedback is called wrap-around. This method usually employs the
full output being brought back to the CPU for comparison rather than just a portion
of it as in the feedback method. Quite often when the data bus is used as the out-
put from the CPU, the output data is latched into registers. The CPU then checks
the condition of these registers and compares their contents with the output just
sent out. The program dictates what happens if the comparisons don't compare.
In another wrap-around method the data bus is used to control an output through
a buffer. When the computer changes an output, the signal from the output device
is wrapped back to the data bus where the computer can monitor the state of that
particular output. In the case of a digital output being converted to an analog one,
the analog signal is wrapped back to the input section. There it is converted back
to digital from analog and compared with the digital output from the CPU.
Figure 19 illustrates the use of the wrap-around self-check method.
Figure 19: Wrap around Self-Check
Sep04 / THTT
Basic Maintenance
Training Manual
Comparison
Certain circuits within a module that performs the same function are compared as
a form of self-monitoring. The outputs from both circuits are sent to a comparison
device. If one of the outputs doesn't correspond to the other, a fault condition is
detected and displayed. Figure 20 shows an example of circuit comparison as
used for self-checking.
Figure 20: Comparison Self-Check
System Self-Check
Self monitoring or checking at a systems level usually involves more complex and
lengthy check methods. Usually a mechanical action, in combination with an elec-
trical stimulus, is used in self-checking the workings of a system. In an airplane, a
systems self-check is usually more expensive to incorporate because of the addi-
tional hardware needed to provide the feedback to the computer. Some of the
hardware involved might be transducers, discrete switches and relays, plus all the
associated wiring and connectors. Because of this additional hardware, and
weight, it may often be more economical to have the self-test performed at the
module level.
The methods for self-checking a system are similar to the methods used in self-
checking a module. Feedback loops are usually the type which allows the compu-
ter to check the position of a driven device within the system. Wrap-around may
also be used with a system to convert the digital data sent out with a wrapped-
around input. The comparison method of self check may be used to compare the
drive signals to the flaps of the airplane, making sure both were equal.
One example of a system self-monitor is a Digital Analog Adapter which converts
ARINC 429 information to discrete and analog signals. Heading is brought in and
converted from digital to analog information for a synchro transmitter. After the dig-
ital to analog conversion, the wrap-around enables the converted analog signal to
be looped back to the input section of the computer. There, it is converted back to
digital and compared to the digital output from the computer. If the two are not the
same, the fault flag is set.
Figure 21: System Self-Check
Sep04 / THTT
Basic Maintenance
Training Manual
Power-Up
Every circuit, regardless of its complexity, size or function must start from a power
down state and go to a power-up state to perform its task. What better time to
check the operation of the circuitry than when it is initially turned on?
A typical system which employs a power-up self- check may be a yaw damper sys-
tem. Once the power is applied, the CPU performs tests on the associated circuit-
ry, such as ROM, RAM and timing circuitry. Input conditions may be set to certain
states so the CPU can check digital and analog signals, and/or discretes.
Devices connected to the outputs such as actuators may also be driven to certain
positions to ensure they will move when directed in the execution of the main pro-
gram. Other discrete outputs, such as lights and indicators may also be reset.
Power supplies and their reset signals are checked. Once the complete power-up
self check is completed and all passes, the main program is entered. If there are
failures, the power-up self check may restart the system and record the failure in
the maintenance memory.
Figure 22: Power Up Test
Continuous Self-Check
Monitoring during the execution of the main program of a system can be termed
"continuous". The urgency of the system usually dictates whether this type of self
checking should be used or not. If conditions aboard the airplane exist in which the
failure of that system would be a safety factor, continuous monitoring is used to
constantly check the system performance. Failure of this continuously monitored
system usually results in a redundant system coming on-line and the first system
is set to a power down condition.
Continuous self checks perform partly the same self tests as at power-up. Usually
the RAM and ROM are checked, as well as the power supply monitor and input/
output circuitry. Continuous self checks are usually much shorter than the power-
up checks because of the large amount of processing which must take place be-
tween the self-checks.
Figure 23: Continuous Self-Check
Sep04 / THTT
Basic Maintenance
Training Manual
Types of Computer Self-Monitoring
Computers used in the airplane industry today may contain one, two, or more CPU
circuits. Each may be processing the same, similar or completely different infor-
mation all within the same computer.
We can have a computer that contains one CPU monitoring itself but if there are
two or more, they can check themselves or each other. Within a computer we can
have one or more of these types of computer self-checking.
The three types are single, interacting and monitor circuit.
Single
When a computer contains only one CPU we are limited to the CPU self-checks
discussed elsewhere in this section. This will usually include ROM check, RAM
check, timing check and internal CPU check. Other checks may be included but
will vary with the design requirements.
Figure 24: Single Computer Self-Monitoring
Interactive
The computer that contains more than one CPU may be processing different infor-
mation and for all practical purposes can be considered single CPU's contained in
one housing. The ones that are processing the same or very similar information
can use interactive checks, in addition to the normal CPU self-check's.
When two or more CPU's check each other, this is called interactive checking.
These can take the form of "Handshaking", "Watchdog timing", "information ex-
change" to name a few. The complexity of these checks depends on the design
requirements.
Figure 25: Interactive Computer Self-Monitoring
CPU
Internal
Timer RAM ROM
CPU#1
CPU#2
Sep04 / THTT
Basic Maintenance
Training Manual
Monitor
A computer can have a separate CPU installed for the sole purpose of checking
that the CPU is performing its designed function. The monitor CPU can be con-
nected to one or more processing CPU's even if they are processing different in-
formation. The operation of the CPU and MONITOR CPU is similar to the
interactive self-checking.
The monitor checking offers another way of self-checking in that it can process the
same information as the main CPU and then compare its processed information to
the main CPU's processed information for accuracy. This type of checking is usu-
ally expensive and used most often in critical systems.
Figure 26: Monitor
MAIN CPU
MONITOR CPU
OUT IN
COMPARE
Self-Check
Interactive
Sep04 / THTT
Basic Maintenance
Training Manual
Computer Technology as applied in Aircraft Systems
Reference Computer
Although computers can be classified by hardware as analog, digital, or hybrid,
they are more often classified by their tasks or application. A computer which may
be used for a source of information or data can be called a reference computer.
Reference signals from this computer may be self-contained and only provides
outputs.
An Inertial Reference System (IRS) is one example of a reference computer. This
system is a laser gyro and accelerometer based reference system used to gener-
ate such outputs as airplane attitude, heading, acceleration and angular informa-
tion. Other than for initialization purposes, the IRS needs no inputs to perform its
task.
Some of the units utilizing this information as a reference are the Flight Control
Computers, the pilot's Horizontal Situation Indicators, and the Flight Management
Computer.
Figure 27: Reference Computer
Informational, Warning and Display Computers
A computer system that collects data from various places, processes it, and for-
mats it for display and warning can be called an Electronic Instrument System (Air-
bus) or an Engine Indication and Crew Alerting System (Boeing).
The main task of such a computer system is to collect data and display it in a cen-
tral place. During the different phases of a flight, from power up through touch-
down, the flight crew is often in need of information concerning a certain airplane
system.
Information needed may include for example total air temperature, engine oil lev-
els, hydraulic pressures, and engine vibration levels. On the ground, the mainte-
nance personnel often need to recall certain events that occurred during the flight,
such as out of normal parameters on an engine (overspeed), or Auxiliary Power
Unit voltage information.
The flight crew has various types of information available to them before, during
and after a flight. Parameters used to set and monitor engine thrust are displayed
on a cathode ray tube (CRT) full time and the remaining engine parameters may
be selected for display by the crew.
Maintenance information can be displayed when required by maintenance person-
nel. Airplane configuration, equipment cooling and status, electrical/ hydraulic pa-
rameters, performance data and engine exceedance are some of the types of
maintenance information available.
On Figure 28 the Airbus Electronic Instrument System is shown. It is divided in an
Electronic Flight Instrument System EFIS, and an Electronic Centralized Aircraft
Monitoring ECAM. The EFIS provides flight information and the ECAM provides
system and warning information. The Electronic Instrument System comprises
seven computers:
three identical Display Management Computers (DMCs)
two identical Flight Warning Computers (FWCs)
two identical System Data Acquisition Concentrators (SDACs).
The DMCs comprise two independent parts: one for the EFIS function and one for
the ECAM function.
Sep04 / THTT
Basic Maintenance
Training Manual
Figure 28: Example of the Airbus Electronic Instrument System
Controlling Computers
A computer with the primary task of controlling something can be called a control-
ling computer or controller. This is one of the largest categories of computers.
In industry today nearly anything that can be controlled by a computer. Computer-
ized controllers range from simple temperature controllers to entire systems for
controlling a complete factory.
Airplanes have a variety of systems, surfaces, and devices needing control during
operation, both in the air and on the ground. It is impractical to have the flight crew
manually control all of the necessary systems, so computers are used to lighten
the crew's workload by providing automatic control.
An example would be the control of the slats and the flaps on an Airbus A330:
The Slat Flap Control Computer (SFCC) provides a means to monitor the slat and
flap lever position and to control the slat and flap position on the wings.
Position of the surfaces is selected from a control lever which transmits the de-
mand to the Slat Flap Control Computers through a Command Sensor Unit.
The SFCCs, which are identical, ensure the control and monitoring of the slat and
flap system. On lever demand, the SFCCs send signals to a Power Control Units
to energize the valve blocks.
Two hydraulic motors in each Power Control Unit provide hydraulic actuation.
Each of them is powered by a different hydraulic system and has its own valve
block and pressure off brake. The valve blocks control the direction of rotation and
the speed of their related PCU output shaft.
The hydraulic motors move the transmission through a differential gearbox. Then
torque shafts and gearboxes transmit the mechanical power to the actuators which
drive the surfaces.
Wing tip brakes are provided in order to stop and lock the system when certain
types of failure are detected.
Feedback signals sent by dedicated Position Pick-off Units are used by the com-
puters to control and monitor the system. A Feedback PPU provides signals con-
cerning the PCU output shaft position, while two Asymmetry PPUs send
information about surface actual position. In addition signals sent from an Instru-
mentation PPU are used for position indicating on the ECAM.
Figure 29: Slat Flap Control System
Sep04 / THTT
Basic Maintenance
Training Manual
Interactive Computers
Some computers can perform several different tasks depending on operator in-
puts. Such a computer is called an interactive computer. Interactive computers
typically display information to the operator and then manipulate the data based
on the interaction between the operator and the computer.
The Flight Management Computer (FMC) is an example of an interactive compu-
ter. The flight crew interfaces with the computer by means of the Control Display
Unit (CDU) to input performance data, initialization data and route structure.
The computer calculates optimum cost profiles for climb, cruise and descent used
by the autopilot and auto throttle for automatic flight control. All computed values
are also automatically displayed allowing the crew to fly an optimum profile using
manual control.
Figure 30: Interactive Computers
Airplane Digital Systems - Summary
Airplanes typically have many computers to control, monitor, provide references,
and make available information. These computers can be either analog, digital or
hybrid. It can also be noted that computers are typically different combinations of
the types of computers as categorized by application.
Those are interactive, reference, information, warning, display, and controlling,
computers.
As in the example of the interactive computer, the flight management computer is
also used as a control computer by controlling the autopilot and other systems. It
can be an Informational computer by providing the flight crew with route Informa-
tion.
The flight management computer also acts as an information computer by storing
information to be used by other airplane systems. It also acts as a warning com-
puter, In that it monitors many of its inputs and stores this information for further
reference. Computer types and applications are as varied as their tasks.
Sep04 / THTT
Basic Maintenance
Training Manual

5.6 Basic Computer Structure: Basic Maintenance Training Manual Module 5 Digital Technology / EIS

Uploaded by

Copyright:

Available Formats

5.6 Basic Computer Structure: Basic Maintenance Training Manual Module 5 Digital Technology / EIS

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

5.6 Basic Computer Structure: Basic Maintenance Training Manual Module 5 Digital Technology / EIS

Uploaded by

Copyright:

Available Formats

Sep04 / THTT

Copyright by SR Technics Switzerland

You might also like

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.