Car electronics

Infra-red light switch

Solid-state ignition g

MSX extensions^HrjH
Public-address amplifie

Cndia

1 2-03

?2»J.JrlJi

5 \ '• !! ': i :i i

information, write or call :

0100 0100 ■
0101 1001
0100 1110
0100 0001
0100 1100
0100 1111
0100 0111

Dynalog has grown...

Starting as a manufacturer of 8 bit microprocessor trainer kits in 1 982. Dynalog has now grown to be
the only manufacturer in the world to offer sophisticated microprocessor training and development
systems based on almost all the popular and latest microprocessors.

The Dynalog range covers microprocessor training and development systems based on 8085, Z80.
8086. 8088. 6802. 6502. and the latest 16/32 bit CPU 68000. Many other advanced products like
Single Board Computers based on Z80. Universal EPROM programmers. IBM PC/XT compatible
Motherboard and Add-on cards. STD Bus compatible cards. VT 52 compatible CRT Terminals.
Graphica 1 1 85 Colour Graphic Terminals and Data Loggers are also available from Dynalog

Dynalog Micro-Systems

14, Hanuman Terrace, Tara Temple Lane,
Lamington Road, Bombay 400 007

Tel: 36242 1 , 353029 Telex: 0 1 1 -75614 SEVK IN Gram: ELMADEV1CE
Branches and representatives at: Pune. Bangalore. New Delhi. Hyderabad and Chandigarh

DEFLECTION COMPONENTS
Ef TUNER

• EHT TRANSFORMERS
(ALL MAKES)

• DEFLECTION YOKE

• LINE-DRIVER TRANSFORMER

• LINEARITY COIL

• TUNER

• V.I.F. BOARD

• MAIN TRANSFORMER FOR TV

FOR DETAILS CONTACT

Oswol electronic Co

LEADER ELECTRONICS

ALPHA NUMERIC PRINTERS

(DATA LOGGER)

A range of Erasable Programmable Read
Only Memories (EPROMs) like 2708, 2716,
require Ultra Violet Light for erasing thei
Now Professional Electronic Products
offer a compact UV Light Source UVLS-8
which can erase Eight EPROMs at a time
A 0-60 minute timer is provided on the i
to automatically shut-off the lamp.

PROFESSIONAL ELECTRONIC PRODUCTS

MREC Forges Ahead with
PR El i of West Germany

Factory : Developed Plots 8 &. I I Industrial Estate, Guindy Madras 600 032
Ph. 4322 1 4. 4322 1 5, 432768

Regd. Office : T.N.K. House 48 Anna Salai Madras 600 002 Ph. 840699
Telegram : VOLUME'

| |

M.R.ELECTRONIC COMPONENTS LTD.

MREC have finalised a
technical agreement with
the world famous %sk of
West Germany to assemble
and manufacture Control
Units for colour TVs in India

1 2-1 1

We are the leading supplier of
electronic components to the
TV-industry. Available ex-stock
in West Germany the following
as well as many other
components:

Qty. Component Price
(Cl Fair)

in$

150K TEA 120S 0,26

200 K TBA 810 AS 0,26

200 KTDA 1044 0,80

150K TV 20 0,48

200 K BU 205 0,58

100K BU 208 0,62

1000K IN 4148pe>ioo 0,88

RTV ELECTRONIC
GmbH & Co.

P.O.B. 2217

D-5632 Wermelskirchen
Federal Republic of Germany
Phone: (021 93) 37-0
FAX: (02193)3062
Telex: 8513393 rtvd

Bangalore:

VETRIL Electronics (P) Ltd.

SB-111, Industrial Estate
Peenja,

Bangalore-560 058
Phone: 384415
Cable: VETRILCO
Tx: 0845-8328

f 6 1 * (Clpsea)

200 K lOO+IOOmfd 0,36
350 V

150K 100 + IOOmfd 0,86
. 450V

150K 200 + 400 mfd 0,86
200V

B&W PictureTube
100K 20 inch 20,40

50 K 24 inch 24,80

Our selling line also includes:
Flyback transformers sawfilter
and UHF, VHF tuners, we are
also exclusive distributor for the
STOCKO-CONNECTOR
SYSTEMS, EUROPEAN CTV
Chassis and operating units
with and without remote control.

Voltage-
controlled
attenuator

Although Aphex's Type
1537A voltage-controlled
attenuator has been on
the market tor some time,
it appears that the device
is not generally known. A
pity, because it otters very
low distortion, high stab-
ility, low noise, and a wide
dynamic range. Its
attenuation can be con-
trolled precisely from 0 dB
down to —100 dB.

Fig. 1. Pin configuration
and internal circuit layout
of the 1537A.

Fig. 2. A simple practical
circuit is obtained by
combining the 1537 A with
two additional opamps.

Fig. 3. This rather more
elaborate circuit than that
in Fig. 2 has an additional
input buffer.

Moreover,
the 1537A is de-
signed for a high
slew rate resulting
in low transient inter-
modulation and a wide
bandwidth, so that it can
be used over the fre-
quency range DC to
50 MHz.

The very good perform-
ance characteristics of the
1537A make it suitable for
use in a wide range of
audio applications, such
as voltage-controlled
filters, synthesizers, com-
pressors, and others, as
well as in, for instance,
tone burst generators,
robotics, servo-controlled
machines, and many
other general electronic
applications.

In general, voltage-
controlled attenuators,
which normally use log-
antilog multiplier circuits,
exhibit high non-linear dis-

tortion and high noise. The
1537A uses new pro-
prietary techniques that
result in much lower levels
ot these characteristics.
Moreover, since It is a true
class A device, its
crossover distortion is
much lower than that ot
most other voltage-
controlled attenuators.

The 1537A is housed in a
< 14-pin dual-in-line
package. The pin con-
figuration and internal cir-
cuit are shown in Fig. 1.
Current sources Q> and
Qa, driven by the Input
signal, control amplifiers
Qi and Qe. The gain of
these amplifiers is con-
trolled via their base
voltage (since that voltage
controls the transconduc-
tance of the transistors).

The output is buffered by
transistors Q? and Q».

Some practical circuits
using the 1537A are shown
in Figures 2. 3. and 4. The
circuit in Fig. 2 is the
simplest, requiring only
two additional opamps. It
is suitable for source im-
pedances up to about
150 Q. Where the source
impedance is higher, but
does not exceed 2 kS, the
circuit of Fig. 3 should be
used. This is comparable
to that of Fig. 2, but has an
additional input buffer.

For optimum utilization of

the 1537A, the circuit
shown in Fig. 4 is
suggested. This uses even
more additional opamps,
and these should be of
the low-noise type, such as
the TL027, LF353, or NE5534.
Some typical applications
of the 1537A are shown in
Fig. 5. The voltage-
controlled resistance in
Fig. 5a has a value.
R'=R/(1-o), where a is the at-
tenuation of the 1537A.

The high-pass filter of
Fig. 5b has a cut-off fre-
quency

fc=(1-oV2nRC

where f is in hertz, and a is
the attenuation of the
device.

The circuit in Fig. 5c is a
band-stop tiller, whose
centre frequency of the
attenuation band F is
fc=1/2nv/R.RaC.C2 (Hz)
Finally, Fig. 5d shows a
suggested automatic gain
control (AGC) circuit.

Literature:

1537A voltage-controlled
attenuator

Aphex" Systems Limited
7801 Melrose Avenue
Los Angeles
California 90046
Telephone: (213) 655-1411
Ttelex: 910-321-5762

1 2-1 7

Although transistor ignition has been
available for many years now, there
are still millions of cars that have not
been provided, with the advantages
of a complete solid-state ignition
system. The solid-state ignition
described in this article will give long
and reliable service and will also
extend the useful life of your spark
plugs appreciably.

SOLID-

STATE

IGNITION

by

Hans Steeman

A full solid-state ignition system has
many advantages over conventional
systems. It will, for instance:

• enable the engine to be started
readily, whether this is cold, wet,

or hot — provided your battery is in
good condition, of course;

• ensure that even a cold or damp
engine continues to run once it

has been started;

• ensure that the spark energy is
constant and independent of the

engine speed;

• considerably reduce carbon
deposit on the spark-plug elec-
trodes, thus allowing longer intervals
between cleaning and replacement
of the plugs.

\A comparision

Fig. 1 shows a conventional coil ig-
nition system as used in most petrol
engines. The contact breaker points
are controlled by the distributor
cam: when they open, the current
flowing through the primary winding
of the coil is interrupted, which
causes a high potential to be in-
duced across the secondary wind-
ing. This voltage is high enough (10
to IS kV) to ignite the compressed
charge of air and petrol vapour in the

engine cylinder via the spark-plug.
The distributor ensures that the high
tension is applied to only one
cylinder at a time. The distributor is
driven by the engine at half engine

In the solid-state ignition system the
function of the contact breaker
points is transferred to a transistor
switch: the points merely serve to
trigger the transistor. Because of the
consequent appreciable reduction

in current flowing across the points,
these become virtually free of wear.
The timing diagrams in Fig. 2 show
the differences between the ignition
pulses generated in the two systems.
When the points in the coil ignition
system are closed, no current flows
through the primary of the coil. Note
the overshoot and ringing occurring
at the secondary immediately after
the points have closed. These
phenomena are caused by stray

Fig. I. Conven-
tional coil-ignition
system for petrol
engines.

Fig. 2. Timing
diagrams of coil
ignition (left) and
solid-state ig-
nition (right).

85128-2

Fig. 3. The c.
the solid stall

capacitance and inductance which,
since the coil is not replaced, also
occur in the solid-state system.
When the points are open, ie„ at the
moment the high tension is induced,
the two systems behave in an unlike
manner. A voltage similar to the
peak secondary voltage just before
the spark-plug fires — 4 in both
diagrams — exists across the primary
winding and dies out only after ig-
nition has taken place.

Once the required level of high ten-
sion is reached, the air gap in the
sparking plug becomes conductive,
and a spark jumps across the gap.
Therefore, during the actual ignition,
the full secondary voltage exists
across the plug electrodes. At the
primary side in the solid-state
system, there is no longer the
characteristic overshoot and ringing
of the coil system. The energy in the
secondary winding declines until it
is no longer sufficient to sustain the
spark, and this then dies out.

Circuit description

The voltage pulses provided by the
contact breaker are reshaped by
Schmitt trigger Ni — see Fig. 3 —
and then applied to monostable
MMVi. This stage has been ar-
ranged such that it provides pulses
of about 1.8 ms at its Q output (pin 6).

This corresponds exactly to the re-
quired spark duration.

In a four-stroke, four-cylinder petrol
engine, each cylinder is ignited at
every second engine revolution.
Since the four cylinders must each
fire at regular intervals, two sparks
are required for each engine revol-
ution. This means that at an engine
speed of 6000 rev/min the interval
between two sparks is S ms. As the
spark duration is 1.8 ms, the coil has
3.2 ms to restore its energy. This

period is, of course, longer at lower
engine speeds.

Monostable MMVi is triggered at its
TR input (pin 4) every time it pro-
vides a pulse at its 0 output. This
pulse is applied via NAND Schmitt
triggers N3 and N2 to parallel-
connected inverters N« to Ns that
drive the power output stages.

For safety, the output stages T1-T2
and T3T4, have been duplicated and
then connected in parallel. Diodes
Ds and D6 and zeners D7 to Dio pro-

2-20,

tect the power stages from negative
pulses and over-voltages.
Monostable MMVz is triggered by
the output pulses of MMVi and
generates for each of these a pulse
of about 0.5 s at its 0 output (pin 10).
The duration of this pulse is deter-
mined by the time constant R6-C3.
The pulse ensures that gate N2 re-
mains open to accept control pulses.
When the engine stops, the contact
breaker no longer provides control
pulses, and the gate closes after

0.5 s. This ensures that the ignition
coil cannot burn out when the
engine is not running. In that con-
dition, parallel-connected resistors
Ri and Rz allow a current of about
250 mA to prevent the contact
breaker points from corroding.

Construction

When the printed-circuit board —
see Fig. 4 — which is available

through our Readers’ services, is
used, no construction problems are
envisaged.

Collector resistors Riz and R13 get
quite hot and must, therefore, be
glued onto the inside of the lid of the
metal case.

The remainder of the construction
should broadly follow the lines
suggested in Fig. 5. If you cannot get
a cast enclosure, fit the power tran-
sistors on suitable heat sinks. Do not
skimp on the heat conducting paste!
Before the ignition is fitted into the
vehicle, it should be checked with
an ohmmeter to make absolutely cer-
tain that it is free of short-circuits.

As shown in Fig. 5, one of the sides
of the enclosure should be provided
with four insulated car-type male ter-
minals onto which the interconnec-
ting cables are push-fitted by means
of mating receptacles. These male
and female connectors are available
from most motorists' shops. It is ad-
visable to fit the receptacles with in-
sulating sleeves.

The case should be fitted under the
bonnet in a position where it is
reasonably well protected from
water ingress.

It should normally not be necessary
to alter the ignition timing. This
timing can be checked roughly with
the aid of diode D3, which should
light when the points are closed.
However, if in any doubt, the timing

Fig. 4. The
printed-circuit
board for the
solid-state ig-
nition system.

Resistors:

Ri;R 2 =100Q;5W

R3-68OQ

Rs-56 k

R7;Ra=220 Q
R9 1 0;5 W
Rio;Rn =8.2 S
Ri2;Ri3~47 Q;5 W

Capacitors:

Ci = 10 n
C2=33n
C3M70n
C«=220|i;16V
Cs 220 p;25 V
Cs = 4.7 n;630 V
C7= 100 n

Semiconductors:
Di;D?;D4;Dn 1N4001
03- LED; red
Ds;Df, 1N5406
07 to Dio zener diode
200 V;1 W
Tt;T3 = BD437
T2;T4 - BUX80
ICi =4093
IC2 - 4069
IC3M538

Miscellaneous:

Insulating plates -
preferably Teflon - for
mounting the power

terminals and mating
receptacles for fining
with M3 size self-

four insulating bushes
for use with M3 size

metal - preferably die-

PCB 85128

Fig. 5. Artist's im-
pression of how
the solid-state ig-
nition system
may be con-
structed and
housed.

85128-5

should be checked properly with a
stroboscope with the engine turning
over at constant speed.

In some cars, a resistor is connected
in series with the ignition coil. This
resistor, which is shorted out when
the engine is started, must not be
removed. Apart from the connection
to terminal 1 of the coil, i.e., that to
the contact breaker, all wiring in the
car remains as before. If the car is fit-
ted with a revolution counter, this
should remain connected to terminal
1 of the ignition cbil. M

FINALLY A WARNING: WHEN THE
ENGINE IS RUNNING DO NOT UNDER
ANY CIRCUMSTANCES TOUCH THE
TERMINALS OR COMPONENTS OF THE
IGNITION SYSTEM. BECAUSE THE HIGH
TENSION PRESENT AT VARIOUS PLACES

INFRA-RED
LIGHT SWITCH

Anyone who has ever tried to switch the tight on when
entering a dark room with both hands full will appreciate
this tight switch, it operates automatically upon being
triggered by your body heat.

The idea behind this infra-red con-
trolled light switch is quite simple.
The body heat — which occupies the
electromagnetic spectrum between
light and radio waves, ie. 0.74 to
300 nm - is picked up by a sensitive
Fresnel lens. This lens, which has at
its focus a double differential
pyroelectric sensor, IRi, is largely
unaffected by other electrical radi-
ation. The area served by the light

switch is divided into a number of
zones as illustrated in Fig. 1. When
someone moves from one zone into
another, there is a change of tem-
perature, which is collected by the
lens as a variation in electromagnetic
energy. The sensor reacts to this
change by generating a small elec-
tric signal. That signal is processed
and used to operate the light switch.
Instead of a Fresnel lens, it is also

possible to use a simple home-made by

lens. More about this under Con- p eter Jheunissen

struction.

The circuit is suitable for operating
lamps rated at up to ISO W. Depend-
ing on the setting of the presets, the
light(s) will stay on for periods vary-
ing horn 5 seconds to 7 minutes. If
the light is required to stay on, the
normal switch should be used, or an
additional on-off switch provided in
parallel with Do. The latter switch
should be rated at 240 V. The last
solution is, in any case, to be adopted
if the infra-red switch replaces the
original light switch in the wall-
mounted box.

An additional facility is the automatic
brightness control which prevents
sensor signals being processed as
long as there is sufficient daylight in
the room.

Circuit description

The signal provided by the sensor is
so small that it must Erst be
ampliEed. This is done in A. and A2,
which together provide a gain of
47 dB. The input amplifier is pre-
ceded by an RC filter, which re-
moves any low frequency compo-
nents. Moreover, Ai and Ki function
as an active band-pass filter with a
bandwidth of 1.5 MHz to 15 MHz.
Opamp Ai is arranged as a compara-
tor. The reference voltage is applied
to the non-inverting input, pin 10, via
divider Rs-Ri. As soon as the poten-
tial at the inverting input becomes
smaller than the reference voltage,
for instance, when the sensor has
been triggered, the opamp toggles,
its output — pin 8 — goes high, and
capacitor C6 charges quickly. Com-
parator A* then toggles, which
causes T2 to switch on. Capacitor C6
discharges slowly through Pi and
Rio, and when the potential at the in-

Fig. 1. This
illustrates the
division of a
space into zones
by a Fresnel or
a home-made
lens.

Fig. 2. Typical
field-of-view
diagram in the
. x-x plane of
dual-element
sensor Type
RPY97.

>2-23

5

verting input of A« rises above the
level of the reference voltage, the
comparator toggles again, and Ti is
switched off.

The length of time the light remains
switched on depends on the time-
constant Pi-Rio-C«.

When T2 is switched off, one ter-
minal of Cto is open-circuited, and
nothing will happen, because the
triac gets insufficient gate current.
As soon as T2 is on, however, Cm is
connected to earth and begins to
charge. After a short while, the
voltage across it is sufficient to
switch on diac D«. Capacitor Cm
then discharges via the diac and the
gate of Trii, which causes the triac to
conduct. The potential across Trii is,
therefore, shaped as shown in Fig. 6,
and is reminiscent of that produced
by a dimmer switch. The
resemblance does not end there,
because, just as in the case of dim-
mer control, the lamp never gets the
full mains voltage. This is in any case
necessary to ensure that the circuit
gets sufficient power.

If the room in which the switch is in-
stalled is open to daylight, it would,
of course, be nonsense to have the
light switched on every time some-
one moved across the field of the
sensor during the day. This is,
therefore, prevented by the auto-
matic brightness switch. This con-
sists of phototransistor T> and
potentiometer P2 with which the sen-
sitivity of the circuit is set. When a
certain amount of light falls onto Tv it
conducts, which results in a larger
potential at the inverting input of A>
than at the non-inverting input. The
comparator output then remains low,
which prevents T2 from conducting,
so that the triac cannot fire. Note that
a higher resistance of P2 cor-
responds to a higher sensitivity of
the circuit and that, therefore, even a
small amount of incident light may
prevent the triac from switching on.
If the brightness control is not re-
quired, merely omit Ti and replace
Pe by a wire link.

The remainder of the circuit forms a
simple power supply that is stabil-
ized by zener diodes D3 and Ds.

Choke Li and capacitor Cr prevent
switching pulses generated by the
triac from reaching the mains supply.

Construction

The circuit is best constructed on
the PCB shown in Fig. 4, since this
has been designed for fitting into a
round electrical junction box. It may
be necessary to file the edges of the
board to make it fit snugly into the
box. Be careful, however, not to take
too much off, particularly around Cr
and C>.

The sensor may be fitted well away
from the circuit, in which case a
simple cable is needed to connect
the two. Alternatively, it may be fitted
onto the PCB, in which case the en-
tire unit should, of coutse, be in-
stalled in a suitable position as
suggested in Fig. 1.

The sensor or the entire unit,
whichever construction is chosen, is
best placed at a height of about 2 m
(6 ft 6 in) at a downward angle of
about 14° from the vertical in such a
position that the door opening is
fully covered. It should not be
placed in direct sunlight, nor above
heating appliances. Note that the
unit is not really suitable for use in
the open.

As stated earlier, the sensor may be
fitted behind a proper Fresnel lens
or a home-made one as shown in
Fig. 5. A Fresnel lens is composed of
a number of smaller lenses so ar-
ranged that they give a very short
focal length. Such lenses are used in
headlights, camera viewfinders, and
spotlights, to name but a few. If you
opt for a home-made lens, this is best
constructed from some flexible card-
board into which a number of
longitudinal slits are cut as shown in
Fig. 5. These apertures should then
be covered with clingfilm or similar
foil. The reason for this is that the
sensor should be open to infra-red
light, but not to draughts or similar
air currents. As the lens is bent into a
semicircle, the open sides should
also be made airtight.

Finally

When planning the installation, bear
in mind that the sensitivity of the
sensor is greatest for movements in
parallel with it, and least for
movements towards it. The sensi-
tivity is greatly enhanced when a
Fresnel lens is used. Further infor-
mation on this, as well as on the
sensor, may be found in the July 1985
issue of Elektor India.

WARNING. Remember to switch off
the mains when working on the PCB
and when wiring the switch into the
domestic light system. Remember
that touching the mains can be fatal!

M

Fig. 5. Suggested
construction of a
cardboard lens.

Fig. 6.

Waveshape of
the potential
across the triac.

• 2-25

car electronics

now and in the future

Alter we reported some
five years ago that the
motor industry was rather
slow in adopting micro-
electronics, there has
been a drastic reversal of
this negative attitude. In-
deed, most manufacturers,
ot cars as well as compo-
nents, now reckon that
without microelectronics
the future of the industry
would be pretty bleak. This
is because they cannot
see their way of meeting
the requirements of the
future by pure mechanical
changes. These require-
ments involve reliability,
efficiency, fuel consump-
tion, performance, and
comfort, among others. It
is, therefore, not surprising
that vast sums have been
invested in research and
development, particularly
in the field of micro-
processors.

In the early years it was
found that electronic

equipment just was not
capable of operating
reliably under the bonnet.
It could, for instance, not
cope with temperatures
varying from —40 °C to
+150 °C; vibrations up to
200 g; salt spray; dust;
sand; oil; and petrol. But
all this is history now, and
it is true to say that prop-
erly engineered electronic
units are at least as
reliable as mechanical
car parts — and they are
just as robust.

It has been estimated that
the motor industry in this
country will fit some £200
million worth ot electronic
controls and sensors in
cars this year.

None the less, some prob-
lems remain. Even elec-
tronic systems can fail
spontaneously. When they
do, the breakdown should
not present a safety
hazard. Preferably, the car
should be able to con-

tinue in spite of such
failures. Because of this
sort ot consideration, it is
becoming more and more
important that the more
complex electronic units
are provided with an auto-
diagnostic facility.

The most important areas
of research and develop-
ment are concerned with
electronic petrol injection
and ignition systems; elec-
tronic diesel fuel injection;
and electronic catalysers.

It is interesting to note that
these are all concerned
with energy consumption
and air pollution. There is
also much development
going on to retine the anti-
lock braking system (ABS),
a safety device so import-
ant that it surely should be
fitted to all new cars
within the next five years.
Last but not least, there is
the research into elec-
tronic gravity sensors for
the control of airbags and

safety belts. A pictorial
summary of these and
many other aspects is
given in Flg.1.

Fuel injection
system with
lambda probe

An interesting develop-
ment by Bosch, particu-
larly tor small cars, is the
Mono Jetronic injection
system. In this, the required
petrol-air mixture is deter-
mined from the position of
the throttle valve, and the
engine speed. This simple,
but economical, method
of engine control is op-
timized by an electronic
circuit. Deviations from the
correct mixture are
detected and quickly cor-
rected. The error signal is
provided by the lambda
probe. This system has
been in use tor some time

2-26

now and has proved itself
to be both accurate and
reliable. Bosch has recent-
ly introduced a heated
lambda probe, which, it is
claimed, has a very long
life, and provides en-
hanced accuracy at low
exhaust gas temperatures;
earlier switch-on of the
system otter the engine
has been started; and im-
proved reliability in diffi-
cult installations.

Note that a catalytic con-
verter can only work effi-
ciently if the air-tuel
mixture is finely tuned: it
must be close to what is
known as the stoichio-
metric ratio, i.e., 14.7 parts
air to one part fuel. It is
clear that the converter
works at an optimum in
conjunction with an elec-
tronic fuel injection system;
some 90 per cent of the
three main pollutants pres-
ent in the exhaust gases
can then be converted.

Battery tem-
perature sensor

Alternators and generators
in cars are generally pro-
vided with a temperature
compensation control,
which ensures that the
charging rate is higher
when the battery is cold
than when it is warm. This
control will, however, only
work satisfactorily if its tem-
perature is equal to that ot
the battery.

A temperature sensor is,
therefore, fitted to the bat-
tery case, and connected
to the compensation con-
trol unit via a two-way
cable. In this way, the opti-
mum charging voltage is
set by the compensation
regulator.

Trials have shown that this
type of control can im-
prove the state ot charge
in winter and in town traf-
fic by more than thirty per
cent. With the battery
maintained in this way, the
engine will, therefore, fire
much more readily —
even at sub-zero tem-
peratures.

Anti-lock
braking system

The anti-lock braking

2-28 elaktor India lebruaiy 1988

system, ABS, was originally
developed for aircraft, but
is now, in modified form,
fitted to a number ot stan-
dard production cars The
principle of its operation is
shown in Fig. 4. Sensors at-
tached to the hub of each
wheel monitor the number
of wheel revolutions. When
the brakes are applied
and the wheels tend to
lock, electronic circuits
controlled by the sensors
operate an hydraulic
brake pressure solenoid,
which reduces the brake
pressure to a level at
which the wheel shows no
tendency to lock. Since
the system works in-
dependently on each
wheel, optimum braking
— for the circumstances —
is achieved. As the wheels
do not lock, the danger of
skidding is reduced sub-
stantially. Many millions ot
hours of driving cars
equipped with this system
have shown that under all
conditions the braking
distance is shortened ap-
preciably compared with
that of cars not so
equipped. The security ot
the system has also been
taken care of. When the
engine is started, the en-
tire system is checked,
and all of its components
are constantly monitored
while the engine is run-
ning. If an error is

detected, the ABS is
disabled, and the normal
braking system takes over:
the driver is given an indi-
cation of this.

Anti-slip
| regulator

The anti-slip regulator —
ASR - prevents the driven
wheels from running on in
slippery conditions when
the engine moment is re-
duced in a controlled
manner, independent of
how much throttle the
driver gives. This makes it
necessary for the usual
mechanical linkage be-
tween the accelerator
pedal and the throttle
valve to be replaced by
an electronic link. The ASR
can operate in conjunc-
tion with the ABS in two
versions as shown in Fig. 6.

ASR with throttle and
brake control
In this version — see
Fig. 5a — the combined
electronics are controlled
by the signals provided by
the hub sensors If the
driven wheels tend to run
on, the throttle valves are
closed a little by the con-
trol unit ot the electronic
throttle. link, thereby reduc-
ing the engine moment. If,
because of different road
track conditions only one

Fig. 3. Comparison be-
tween cold and heated
lambda probes. (Courtesy
of Bosch)

wheel tends to run on, this
is braked by the ABS in
addition to the throttle re-
duction. This sort of combi-
nation is In a practical
sense, an electronically
controlled limited-slip
differential. The ABS
hydraulics may have to be
extended with an hydraulic
memory, but this is not
necessary if the ABS is fac-
tory fitted integrally with
the normal braking system.

ASR with throttle and
ignition control
In this version - see
Fig. 5b — the hydraulics ot
the ABS need not be
modified, and do not af-
fect the ASR. Only the ABS
electronics are extended
with the ASR parts. To re-
duce the reaction time
when the engine moment
is reduced, the ignition
and tuel injection settings
are altered by the elec-
tronic accelerator at the
same time as the throttle
valve setting. This system
does, however, need a
mechanical limited-slip
differential for safe perfor-
mance under treacherous
road conditions.

For the future:
CARiN

Philips Research Labora-
tories have for some time
been working on an elec-

Ironic co-pilol for cars,
which can plan the route,
guides the driver to this
destination, knows the pos-
ition of the car and can
specify it at any moment,
and can also provide a
number of details about
the environment or the
destination of the journey.
This co-pilot has been
given the name CARIN:
CAR Information and Navi-
gation.

In later phases of the pro-
ject, CARIN will be inte-
grated with dashboard
functions. Spoken warnings
can then be given if the
car needs to be filled with
petrol or oil, If the tem-
perature goes foo high, or
if there are battery prob-
lems. The system could
also be linked to traffic
warnings over the car
radio. This might be done
by means of the Radio
Data System (RDS) which is
now the subject of stan-
dardization discussions on
the European level, while
RDS test broadcasts are
already taking place in
this country as well as, for
example, in France,
Cermany, and Sweden.
The coupling of CARIN to
RDS, for example, would
make it possible to plan
alternative routes to avoid
traffic queues, road works
or icy patches, and to
modify the guidance pro-

vided accordingly. The
digital RDS signals are ac-
cessible to the on-board
computer and do not in-
terrupt or interfere with the
normal radio programme.
Traffic research in this
country has shown that
drivers could plan their
routes approximately
twenty per cent more
efficiently on average if
they did not merely guide

themselves by familiar
landmarks. Fuel costs and
driving time are included
in the calculation. With
CARIN one could always
reach one's destination In
the most etticient manner
possible.

Basic configuration
The basic configuration of
CARIN is shown in Fig. 7.
Parts of the system are:

Fig. 4. Anti-lock braking
system.

Fig. 5a. ABS/ASR
controlling brakes and
throttle valve. (Courtesy of
Bosch)

Fig. 5b. ABS/ASR
controlling brakes, throttle
valve, and ignition.
(Courtesy of Bosch)

,2-29

• A modified compact
disc unit for the car

which could not only play
audio discs, but also read
road information stored on
compact discs, for
example a complete road
map ot the UK, of a town,

• A navigation system
which determines the

current position of the car.

• An on-board computer
which carries out all

processing.

• Sensors which provide
data about the func-
tioning ot the car to the
computer, for example the
temperature of the cool-
ing water, the amount ot
petrol available etc

• A car radio to receive
traffic information or

warnings.

• Communication and
control equipment con-
sisting ot a speech
module with which, via a
speech-synthesis chip, in-
formation could be given
to the user, a screen for
visual information — tor
example a road map —
and a keyboard with
which the driver can feed
information or instructions
to the computer. Here is a
brief review ot a number
ot these system com-
ponents

Large reliable storage
capacity

A Compact Disc (CD) was
originally intended to store
1 hour of music This
means that tor 3600
seconds analogue signals
are sampled 44.1 thou-
sand times a second (that
is the standard frequency)
in two channels (stereo)
and recorded on the disc
at 16 bits per sample. The
disc therefore has a ca-
pacity of:

3600x2x44.1x1000x16 bits
=5x10*9 bits (5 Gbit).

This makes available a
gigantic ROM (Read Only
Memory) which is very
rapidly accessible at any
point and on which, for
example, a complete road
map of the UK plus all
sorts of travel information
could be stored.

This idea of an electronic
digital road atlas is being

2-30 elector India februsrv 1986

elaborated further. The CD
system had to be modified
for its CD-ROM (unction by
extending its error correc-
tion capacity, so that less
than 1 bit error in 10” bits
might reasonably be ex-
pected for chance errors
on an undamaged disc.
That is a factor ot a million
better than computer
tape. In other words, even
a scratched or con-
taminated disc is more
reliable than a computer
tape.

The inclusion of an ad-
ditional fault-correcting
algorithm does cost some
storage capacity, because
extra information must be
stored to remedy errors.

This is something like sen-
ding an Important
message twice to make
quite certain that it
reaches its destination.

By cunning selection of
the error-correcting
algorithm only 6x10* bits
are lost instead of half the
storage capacity, so that
4.4x10* bits remain
available. However, if the
playing time of the CD-
ROM is increased to 66
minutes, as is intended, it
will hold 4.8x10’ bits which
corresponds to 6x10* bytes.

Economic coding
The digital cartographer is
confronted with the task of
economically transferring
to the CD a normal map
with a scale of e.g. 1:15 000
and containing about 30
colours.

An accepted method of
scanning a map point by
point makes use ot a grid
of horizontal and vertical

lines which are spaced
say, 0.1 mm apart. The
map is thus divided into
tiny squares of
0.1x0.1=0.01 mm J . each ot
a specific colour. In this
way an area of land
12x14 km (approximately
the area of Oxford) on a
scale ot 1:15 000 would re-
quire 75 million picture
elements. These would
have to be indicated in
colour. Since 5 bits repre-
sent 2*=32 colours, 75
millionx5=375 million bits
are required. That is
almost 8 per cent of the
capacity of the CD-ROM.

In addition, this method
would not make it clear to
the computer where the
roads were located on the
map

For this reason another
solution has been sought.
The method chosen is one
ot representing the roads
— which is what it is really
all about — with the help
of angle points and node
points. A dead-straight
road without side streets
has only two of these
points, one at each end. A
bending road is approxi-
mated by straight sections
with angle points at inter-
vals. An intersection is a
node point. Thirty-two bits
are needed for each
point, ie. 16 for the x co-
ordinate and 16 tor the y
co-ordinate.

The area of land and in-
land water that can be
handled with such a
coding can be
calculated as follows. Six-
teen bits give 2”=65 000
possibilities (approx.). If
mainland Britain be div-

Fig. 6. CARIN in action.

The keyboard will in future
be replaced by a touch
screen. (Philips Press
Photograph)

Fig. 7. Constituent parts of
the CARIN system.
(Courtesy of Philips)

Fig. 8. How CARIN works
on the road. (Courtesy of
Philips)

ided into two squares of
650 000 m by 650 000 m
(roughly 400 miles by
400 miles), any position
within each square, ac-
curate to within 10 metres,
can then be represented
by 2x16 bits (two times
65 000 possibilities).

An absolutely straight
road, as stated, can be in-
dicated by only 2 points. A
ring road, such as the M25
around London, needs
about 4000 points, how-
ever, to be described
properly. The average
road in British towns can
be described by 12 points,
each of which requires
32 bits. The total is thus
12x32=384 bits on average
per road. To this must be
added an address ot
32 bits to indicate where
the additional information
associated with these
points (e.g., the street
name) is located on the
disc Since names (High
Street, Station Road) are
preferable to degrees of
latitude and longitude, the
relationship between the
street names and co-
ordinates must be re-
corded, and that takes
memory capacity.

The average street re-
quires, as seen above,
384+32=416 bits. If a town
contains, say, 3000 streets,
3000x416=1.25 million bits
are needed. Adding a
similar number of bits for
the coding of the street
names gives a total of 2.5
million bits or about 0.05
per cent of the CD-ROM
capacity. That is a great
deal more economical
than the 8 per cent ob-

tained with the first
method.

These considerations, of
course, only indicate the
order of magnitude. If
more information is to be
stored, or greater accu-
racy attained, more
storage space is required.

Defining position
The CARIN system must be
able to define the position
of the car at any moment,
for which various ditterent
technical methods are
possible. An obvious
answer In the short term is
an electronic compass.
With the help of such a
compass, the direction of
motion of the vehicle rela-
tive to the earth’s mag-
netic field can be estab-
lished.

With this information, and
knowing the distance
which the car travels from
the odometer, the on-
board computer can
determine the position ot
the car. It is also able to
correct errors caused by
such things as passing
cars or bridges made of
reinforced concrete. The
mass of iron causes dn
additional magnetic field
which is registered by the
car compass. The on-
board computer corrects
these errors by regularly
comparing the infor-
mation with the digital
road map. If the calcu-
lated position is away
from the road which the
car should be following
according to the map, an
automatic correction is
made (Fig. 8). Other
solutions are being in-

vestigated for the short
term to avoid the problem
ot disruption ot the earth's
magnetic field by iron
objects.

Satellite navigation
In the longer term, satellite
navigation will be poss-
ible, e.g., with the help of
the American NAVSTAR
Global Positioning System
(GPS) which will be com-
pleted at the end of 1988
with 18 satellites in space.
The civilian part ot this
system will make it poss-
ible to detine a position at
any moment ot the day
and at any point on the
earth with an accuracy of
about 10 metres. The
satellites are located
about 20 000 kilometres
high in 6 different orbits
which are distributed
regularly around the
earth. The orbiting period
is 12 hours. Therefore, at
any moment 4 satellites
will be able to be re-
ceived at any point on the
earth. That is sufficient to
determine longitude
latitude height, and time
(with the accuracy ot an
atomic clock).

At present, 5 test satellites
have already been
brought into orbit and it is
expected that at the end
ot 1987 there will be 12,
sufficient for latitude
longitude, and determina-
tion of the time

Contact with the driver
In the design of the CARIN
system much attention has
been paid to ergonomics,
which includes ease of
use and traffic safety.

Thus, it is preferable in traf-
fic that the computer
gives its advice and infor-
mation by means of the
spoken word. The speech
synthesis chip provides the
! means of doing this,
j Another precaution is that
J the screen can only be
) consulted when the car is
standing still. It will then
be possible to look at the
map, for example, or ask
j tor tourist information.

It has already been indi-
cated that the system can
accept normal names tor
destinations, such as
The Regency, Sea Road,
Brighton, so that it is not
necessary to struggle to
specify the destination in
degrees, minutes, and
seconds. At the same time,
the system tries to discover
what the user wants with
the help of simple
questions.

For example, the following
situation might occur: the
user wants to travel from
Gloucester to the Manor
Hotel in Cheltenham. He
gets into his car and in-
serts the CD-ROM which in-
cludes Gloucester and
Cheltenham into his CD
player. After starting up
the system the following
appears on the screen:
WELCOME TO CARIN
SELECT THE DESIRED
FUNCTION:

1. ROUTE INDICATION

2. TOURIST INFORMATION

3. OTHER FUNCTIONS

The driver then keys in "1".
The system replies:

PLEASE INDICATE YOUR
STARTING POINT: (The driver
types in the street name
and the closest intersec-

tion point).

PLEASE INDICATE YOUR
DESTINATION:

TOWN? (The driver types in
"Cheltenham").

STREET OR DESTINATION?
(The driver types in "Manor
Hotel”)

The computer then selects
the best route and stores it
in its memory. If the driver
desires he can now
remove the CD-ROM from
the player and put in a
music disc

Alter the car has started,
CARIN will guide the
driver to his destination
using the speech module.
In the future the keyboard
will be replaced by a
touch screen. It will then
be possible to indicate
where one wants to go on
a map, or tind and indi-
cate a street name
alphabetically, after which
the on-board computer
does the rest.

CARIN is no dream ot the
future but a technical
reality to which Philips is
giving detailed form. M

After last month's article on the construction of the
\ graphics card, this fifth part in the series deals with the
| software, necessary for efficient operation of GDP and
associated circuits on the board.

HIGH-RESOLUTION
COLOUR GRAPHICS
CARD - 5

In the text mode, then, a screen is
normally filled from the top left-hand
comer (home position) onwards, just
as with standard video or printer ter-
minals. All standard cursor controls
such as CR (carriage return), LF (line
feed), and BS (back space) are
available and take full account of the
current character size. Thus, when a
double character size has been set,
the LF, for instance, will also have
'double size'.

Screen organization in the graphics
mode starts at the bottom left-hand
comer (origin, X=Y=0). Contrary to
the text mode, terms like pen
destination, pixel, and segment are
used rather than rows and columns.
The graphics mode has two possible
types of access: instantaneous and
provisional. Instantaneous access
implies leaving the text mode for
good, whereas provisional access
allows interpretation and execution
of all possible graphics commands
until a closing CR sign is en-
countered which will put the system
back into the text mode.

As stated before, the dominant mode
is text. This will be obvious, con-
sidering that the text mode is auto-
matically selected after a graphics
program is interrupted with the
BREAK key. When BASIC generates
the message BREAK IN LINE x, the
video interpreter must be signalled
to switch to text mode. As will be-

come evident from the discussion of
the graphics mode commands
further on in this article, the letter A
in the work BREAK has been chosen
for this purpose.

Host I/O
distribution

Although text and graphics mode
use quite different commands and
screen layouts, it is certainly poss-
ible to have both interact to create,
for instance, an inverse overlay text
anywhere on a nice background
design.

As the graphics card is also suitable
for use as an alphanumeric terminal,
the video interpreter has an input
routine for reception of characters
entered via the host keyboard. This
leads to the following functional
distinctions to be made for the video
interpreter:

• visual representation of ASCII
codes on screen with 32 lines and

80 columns:

• reception of keyboard-generated
ASCII codes and relevant move-
ment of a blinking cursor (blink rate
is programmable);

• dot by dot control of the
graphics screen, as defined by

specific instructions.

The first two tasks typically belong to
the text mode, the third to the
graphics mode. The tasks may be
mixed freely, as they function quite
independently of one another.
Basically speaking, the interpreter
will look like a super printing routine
called CHROUT (character output). It
receives all ASCII codes directed to
the graphics card and fully
autonomously puts alphanumeric
characters in the relevant screen
locations (text mode), or executes
the given graphics command
(Fig.23).

A character read routine, CHRINP
(character input), has been provided
for in the video interpreter, in case
the graphics card is used as a ter-
minal with the host computer. This
would only require modification of
input and output vectors in the host
computer to make it communicate
with the graphics card through the
video interpreter. To implement this
configuration, it would be ideal to
program an I/O distributor routine
containing all relevant system I/O
addresses in their order of priority —
Fig. 24.

The interpreter itself would require a
patch to suit the host computer
BREAK routine, but this is not strictly
necessary and will be reverted to
later.

A further important interpreter
routine involves the initialization pro-
cedure for graphics card software
and hardware. It will hardly be sur-
prising to learn that this routine is
called INITAL (initialize all).

In its present state, the interpreter
occupies less than 4 Kbytes of host
memory, which makes it possible to
put the program in a Type 2732
EPROM for this purpose. However,
the interpreter requires an additional
40 or so bytes for use as scratch and
parameter storage area. Therefore, it
would seem wiser to locate the inter-
preter in RAM area, in order to have
some free space available right
above it for scratch use. Example: the
4K interpreter RAM area C000.

. .CFFF would leave, say, CF80. .

. CFFF free for scratch purposes. If,
however, the interpreter is run from
EPROM at these addresses, some
other RAM area would have to be
reserved for scratch use, for instance
DF80. . .DFFF. Fig. 25 shows a poss-
ible memory map of the host com-
puter system.

Fig. 24. The
graphics card
and its video
interpreter may
also function as
an autonomous
graphics and/or
alphanumeric
terminal. In that
case, the inter-
preter will have
a threefold func-
tion. Two
addresses in the
host computer
software will
require

modification: one
in the output
distributor
(printer routine),
the other in the
input distributor
(reception of
keyboard
characters)

Text mode
commands

At this stage, it is useful to have
Elektor infocard 118 to hand to study
the text mode commands as listed.

:2-33

Fig. 25. Data
transport in a
system using an
alphanumeric
terminal (in this
case the VDU.
card plus associ-
ated software),
along with the
graphics card

preter. The
graphics
commands, as.
printed (1 and 1)
by a BASIC
program, are

distributor by the
BASIC inter-
preter (2 and 2),
and then simul-
taneously to the
VDU software (3)
and graphics
software (3)
Depending on
the configuration
of the input
distributor,
characters as
supplied by the
keyboard (6) are
directed to
either the VDU
or the graphics
software
(CHRINPin fig.
24)

The video interpreter text com-
mands always consist of a single
character (no parameters required)
and have ASCII values lower than
20hex (32 decimal). They are usually
referred to in hexadecimal notation,
so 0D is carriage return, 08 is cursor
left (backspace, BS). The corre-
sponding BASIC notation, however,
uses the decimal values, so CHR$(13)
and CHR$(8) respectively, preceded
by the PRINT statement.

Commands CHR$(8). . .CHR$(13) and
CHR$(26). . .CHR$(29) are standard
video control characters and require
no further comment, except that they
take full account of the current
character size.

CHRS(17) and CHRS(18) switch the
graphics card to the text and

CHR$(20) resets character size to
minimal. Starting from an 8 x 5 matrix,
characters may either be enlarged
vertically or horizontally with
graphics command S Text command
code CHR«20), or 14he«, will effect
immediate return to minimal charac-
ter size, irrespective of current size.

CHRS(4) forces all ASCII codes in
between this command and the CR
code to be executed as being
graphic instructions. The CR will
restore operation in the text mode.
Code CHR$(4) is particularly useful
for mixing drawings or background
colours with text overlays, or for
changing character size within a

Suppose, for instance, that the text
mode is set and that a program menu
is to be put onto the screen. The
characters are to be double sized
and in red. Underneath the menu
must appear the prompt 'enter your
choice, please' in blue, normal sized
characters. Programming character
size can only be effected in the
graphics mode, so that the current
text mode will have to be temporarily
left to insert relevant graphics com-
mands, preceded by CHR$(4):

10 REM colour red (=6)

20 REM double sized characters
30 REM normal height
40 PRINT CHR$(4)"C6,S2,1"

50 REM final CR will
60 REM restore text mode
70 PRINT "MENU. . ."

90 REM end of menu
100 REM colour blue (=3)

110 REM normal sized characters
120 PRINT CHR$(4)”C3,Sl,r
130 REM final CR will
140 REM restore text mode

150 PRINT "enter your choice,
please"

Alternatively, lines 120 and 150 could
be written as:

120 PRINT CHR$(4)"C3"

150 PRINT CHR$(20)" enter your
choice, please"

Note that use of CHR$(17) and
CHR$(18) for this purpose is far less
elegant than CHR$(4), which instruc-
tion results in a more compact
program.

CHRS(1) transfers video buffer con-
tents to GDP for execution at high
speed.

CMRS(2) opens or closes the video
buffer area (toggle function).

These two commands require some
further explanation. In a mono-
chrome system, one screen equals
16 Kbytes of GDP memory; 48 Kbytes
are required with 8 colours (RGB);
64 Kbytes with 16 colours (RGBI).
Remember that these memory areas
are located on the graphics board,
not in the host computer. Transfer of
graphics card memory to host com-

puter memory would be convenient
if a copy of the current screen were
to be stored on disc or tape. How-
ever, reading a copy of the screen
contents into host computer memory
would require it to have a 64 Kbyte
RAM area available for this purpose,
which would seem a bit large, even
for a simple 16 colour image. Obvi-
ously, it would be far more efficient
to copy the sequence of GDP in-
structions generating such an image,
rather than the image itself as it is
present in GDP memory. Even more
space could be saved if all insignifi-
cant ASCII codes, such as LFs and
spaces, are not written into this com-
mand buffer. With this method of
copying, more than some 16 Kbytes
of host memory will hardly ever be
needed for storage of even the most
complex image in the form of a rel-
evant graphics command sequence.
Once the sequence has been stored
in host memory, it is available for
inspection, modification, or for sav-
ing it onto disc, to be reloaded later
and executed using the PRINT
CHR$(1) instruction. Note also that
from an aesthetical point of view, it is
more pleasing to see a graphic
image develop according to its

architectural logic rather than to see
it completed by a scanning line from
top to bottom of the screen. More-
over, the GDP has an execution
speed which fully justifies the com-
mand sequence approach.

Fig. 26 shows that it is useful to
reserve a dedicated RAM area in the
host computer for video buffer pur-
poses. Note that the buffered and
disc-stored instruction sequence
may also be loaded into another
6502-based computer equipped with
the present graphics card.

CHR$(2). By now it is time to explain
how the GDP instruction sequences
must be stored in the video buffer.
Theoretically, this buffer may have
infinite length, but a 4... 5 Kbyte
area would appear sufficient for
most purposes.

Fig. 26a shows that a buffer that has
not been opened, and then closed at
the end with a CHR$(2) instruction, is
as if non-existent and can not be read
by the video interpreter. This effec-
tively protects the area from being
written into inadvertently. Once
opened with a CHR$(2) command,
the buffer will receive all and only
the video interpreter commands
meant to control the GDP, and
exactly in the order in which they are
issued, be it in the text mode or in
the graphics mode. This process of
copying will stop once a further
CHR$(2) command is encountered,
which will effectively lock the buffer
area — Fig. 26b. The buffer contents
are now available for use. as outlined
above.

CHRS(l). This command enables
execution of the GDP instructions
contained in the video buffer by the
video interpreter — Fig. 26c. After
receiving a PRINT CHR$(2) com-
mand, the interpreter searches for
the buffer start address, reads the
sequence of instructions and
executes them one by one with
dedicated subroutines, until the end-
marker is encountered. Execution
runs at a considerable speed, as all
intermediate (BASIC) x,y coordinates
calculation time is no longer
required. -Thus, a first test-run of a
BASIC graphics program may show a
relatively slowly developing graphic
image. However, with the buffer
opened before this run, it would
afterwards contain only the
necessary video interpreter instruc-
tions for very fast execution by a
PRINT CHR$(1) command.

Graphics mode
commands

Like commands for the text mode,

Fig 26. In Fig.

26a the video
interpreter
controls the
graphics card,
but CDP
instructions are
not saved. Fig.
26b makes clear
that every GDP
instruction is
copied into the
host's video

especially
opened for this
purpose with a
CHR$f2)
instruction. Fig.
26c sche-
matically shows
how a ‘packet of
GDP instructions '
is taken out of
the video buffer
for transfer to
the graphics
card. This
process runs
under video
interpreter con-
trol and is
started off with a
PRINT CHRSfl)
command in
BASIC.

>2-35

those for use with graphics consist of
a single byte (ASCII code A. . .Z) to
specify the relevant command. How-
ever. certain instructions require
one, two, or even three parameters to
follow it. lb create a ring or a circle,
for instance, it is necessary to specify
the radius, thickness, and sectors),
all neatly separated by commas.

In the text mode, reception of an
ASCII value 42ho» will result in a B
printed on the screen. This same
value, however, received in the
graphics mode will change the
background colour into the colour as
specified by the parameter following
the B command.

Syntax rules for the graphics com-
mands are as follows: spaces in
between or before commands and
parameters are ignored by the inter-
preter, except for the P command.
Parameters must be separated by
commas; but there should be no
comma between a command and its
first parameter. In case a command
follows a parameter, a comma should
be inserted. The command string is
always closed with a CR.

Certain (recursive) commands allow
a series of parameters to follow them
without the need for a repeat of the
relevant command code; refer, for
example, to commands D J, or X.

If an expected parameter is missing,
the interpreter assigns default value
0 to it. Thus, B CR or B equals B0 CR
or B0

Whenever there is a parameter too
many, or any other syntax error
exists, the wrong instruction is
entirely ignored and consequently
not executed.

Note that the interpreter only
recognizes integers, so 15.6
becomes 156.

Certain instructions must be fol-
lowed by CR, in absence of which
they will not function properly. This
is notably so in the case of the P
command.

After these general remarks, the par-
ticulars of every graphics command
will be examined. Refer also to
infocard 118.

A : return to text mode. Whenever a
graphics program is interrupted, the
BASIC interpreter will print BREAK
IN LINE. . . on the screen, of which
message letter A is used as the com-
mand to return to text mode.

Syntax: A (CR)

A , (next text mode
command)

B : set background colour. The only
parameter expected after this com-
mand is the code number of the
background colour to be set. After
its initialization, the interpreter
defaults to a black background.

Syntax: B n CR)

B n, next graphics mode

command)

Note that the B command must never
be issued in the RMW mode, as then
the background is not erased. Yet, is-
suing it may result in quite pleasing
effects.

C : set or combine pen colour. The
only parameter expected is the code
number for the desired pen colour,
its initial colour being white. This
command may be used without
restrictions in the RMW mode.

Note that the parameter for com-
mands B and C may be positive (in
which case the former colour is
ignored), or negative (in which case
the new colour is combined with the
current colour by an AND logic
operation).

D : draw from current position to xy
destination. At least two positive
parameters are expected to define
the absolute destination for the pen
to draw a line to. starting at its current
position. X.Y coordinates are given
from absolute origin (X=Y=0).
Syntax: D xy CR

D xy. next command
Recursive syntax:

D xy.xy. . ,x,y 'CR:

D x,y,x,y. . ,x,y, next command)

G : draw geometrical shape. The let-
ter G stands for geometry and pro-
vides access to two fixed
geometrical shapes as selected by
the first parameter. The sign of this
parameter defines whether or not
the shape is open or solid. The xy
parameters define the size of the
shape in terms of x and y units on
their respective axes, starting from
the current pen position. When x
and /or y are negative, the shape will
be drawn behind and/or below the
current pen position.

Syntax: G + n,xy (CR)

G + njcy next command)

H : home pen to current origin (see
I), irrespective of its current position,
and without drawing. No parameters
are expected.

Syntax: H (CR)

H. next command)

I : set current location as new absol-
ute origin. No parameters are
expected.

Syntax: I (CR )

I, (next command)

J : this command is the relative
counterpart of the D command. At
least two parameters are expected to
define the relative xy destination of
the drawing pen. The relative coor-
dinates are defined with respect to

current pen position, and either one
of them may be negative to indicate
that the pen destination is behind
and/or below its current position.
Syntax: see command D

L ; set line type n. One parameter is
expected to specify the type of line
the GDP is to use for drawing. Note
that the type of line not only affects
outlines, but also plain surfaces like
circles, squares, or rectangles and
triangles, which may produce highly
interesting effects.

Syntax: see command C.

M : move to absolute location xy
without drawing. Coordinates x and
y must be positive.

Syntax: M xy (CR)

M xy (next command:
often: M„I = M 0,0,1

to home pen to absolute
screen origin

N : plot dot xy. Two parameters are
expected to define absolute pen
destination xy and one for colour c
of the plotted point.

Syntax: N c,xy (CR)

N c,xy, (next command)
Recursive syntax: see command D

O : draw a circle, ring, or section of
it with current absolute origin as
centre. Depending on the type of
circle or ring, the pen need not
always return to the centre at the end
of the drawing.

Syntax: 0 s,r,t (CR)

0 s,r,t, (next command)
where s is the code for the
drawn part or sector of the
circle, r is the radius, and t
the thickness of the ring.
When r equals t, a disc is obtained
(quickest drawing); when r > t, a
ring is drawn with thickness t;
appearances are deceptive in this
case, as this shape requires less lines
to be drawn, and yet is slower to

P : print alphanumeric characters
without leaving the graphics mode.
This command functions like its text
mode counterpart CHR$(4). All
codes following command P are
printed as alphanumeric characters,
starting from the current pen
position.

Syntax: P characters (CR)

Example:

10 E = 12

20 PRINT "M 128,15,1"

30 PRINT "P example number ”;E
The text "example number 12’ is
printed from location X= 128, Y=15
onwards without leaving the
graphics mode. Note that P is the
only graphics command in which
spaces are significant.

0 : set the print direction. One
parameter is expected to select
horizontal or vertical printing direc-
tion for characters following the
above command P.

Syntax: see command B.

R : move to relative xy position from
current pen position, without draw-
ing. This command is the relative
counterpart of command M. Thus, x
and y may be negative
Syntax: see command M.

S : set character size. Two expected
parameters x and y define the
alphanumeric character size
horizontally (x) and vertically (yj. The
parameters are independently
specified to enable printing very
wide, yet short characters, or,
alternatively, very tall, yet narrow
ones on the same screen. Parameter
values must be in the range of 0. .

. . IS, and Sl',1 resets normal size.
Syntax: see command M.

T : set character type. One
parameter is expected to select nor-
mal or italicized characters, irrespec-
tive of current printing direction.
Syntax: see command B.

U : select pen/eraser activity
up/down. Two parameters are
expected with this command: the
first selects either pen or eraser, the
second determines whether the pen
or eraser is active ('on the paper’) or
inactive (‘off the paper'). This com-
mand allows drawings to be
generated by means of recursive
algorithms which alternately lift the
pen and put it down.

Syntax: U p,u (CR)

U p,u, (next command)

Note that the interpreter is always in
the ‘pen down' mode after its
initialization. Thus, it is not necessary
to issue command U,l,l at the start of
every graphics program if it is
remembered to run the INITAL
routine first.

V : get pixel status. Two parameters
are expected to specify the xy coor-
dinates of a pixel whose status is to
be read from the screen memory.
Syntax: see command N.

The four colour bits of a specified
pixel become the lower nibble (bits
0. ..3) of a buffer called PIXBUF.
Access of PIXBUF will be reverted to
later.

W : set read-modify-write (RMW)
mode. One parameter is expected to
select/deselect RMW operation, as
explained in the first part of this
series of articles.

Syntax: see command B

X : draw coordinate axis. Three
parameters are expected to specify
direction a, increments s. and mark-
ing intervals i.

Syntax: X as.i (CR)

X as.i, (next command
Recursive syntax:

X an,i,. . . .as.i CR)

X a_s.i, . . . asj, next
command

Z : select page. One parameter is
expected to select a screen page.
Depending on the set vertical resol-
ution, two or four pages of screen
memory are available. The Z com-
mand enables the user to smoothly
thumb through the screen memory
pages to look for certain data or
drawings.

Syntax: Z p (CR)

Z p. (next command
where p has a value between
0 (page 1) and 3 (page 4).

Parameters and
variables

The above summary of available text
and graphics commands shows their
conciseness and suitability to be
effectively used in any BASIC
program, as they can be printed as a
so-called string of characters. This
implies that all parameters may be
thought of as variables and that their
signs (plus or minus) remain signifi-
cant. In extreme cases, one variable
could be assigned to a whole
sequence of graphics commands to
create recursive algorithms and,
consequently, compact programs
like the ones shown below.

The first program draws one of the
figures as shown on page « in the
January 1986 issue of Elektor

India . . The second program does
the tubular figure as shown on the
front cover of the November 1985

Graphics card
and Junior
Computer

As an expansion of the DOS-
equipped Junior, the graphics card
and its video interpreter may be

used as a dedicated graphics ter-
minal, together with the VDU card
for alphanumeric I/O (Fig. 23). How-
ever, it is also possible to use the
graphics card in the text mode, thus
obviating the VDU card and software
altogether.

It is possible to use the CHRINP
routine, but this is not obligatory
(note that Junior’s DOS supports
several output devices, but only one
input device, ie. usually the
keyboard). In the source listing of
the video interpreter, the BREAK-test
subroutine has been adapted to suit
the Junior. What remains to be done
is modification of two addresses in
the Junior output distributor the
address of the CHROUT routine
(video interpreter) is placed in
between the addresses of VDU
driver and Centronics printer
routine.

Attention! The addresses to be
placed into the distributor are always
one byte lower than the actual rou-
tine addresses. Thus, if the video
interpreter was assembled to run
from B003 onwards, the byte to be
placed in location 2313hcx or 23I5he«
is 02, not 03:

2313 : 02
2315 : B0

l "10 .02" or 10 ,03"

\ "10 ,04" <

'TO , 05"

2315 : 02

2316 : B0

The character read routine, CHRINP,
is placed in locations 2301 ho* and
2302h» as B005hcx (= B006-1). The
address of routine INITAL is B000 hc ,.
This routine is accessible with disc
command DISK!"GO B000", and
ought to be run every time before a
graphics program is started to clear
all preceding variables. Attentive
readers will have found out by now
that the graphics interpreter is
assembled into a 4Kbytes block of
6502 object code starting at B000hcx.
This memory area lies within the
RAM work space of the BASIC inter-
preter, so a modification of V3.3 disc
track 0 is called for. The procedure
to effect this has been described in
Elektor India issue of
October 1983, page 10-45. Refer-
ring to table 2 on that page, change
the byte in location A218h« trom BF
into AF. Write modified track 0 back
onto disc. The modification with the
I/O distributor is also done on track
0, the corresponding addresses are
either (A)313 and (A)314 or (A)315 and
(A)316.

The video buffer containing the GDP
command sequence extends from
7000h.ex right up to AFFFhex (16 K).
This implies that 12 Kbytes are left for
the BASIC program, starting from
3A79hex. This will be large enough
for most purposes. K

its modest dimensions and battery operation make this
amplifier eminently suitable for use out of doors: at
| garden parties sporting events, boating galas to name
but a few it might be classed as a good-quality
PA (public address) amplifier in view of
its better frequency response than
| usually found in run of the
mitt PA units

MOBILE AF
POWER AMPLIFIER

by

Arno Sevriens

Table I.
Correlation of
input voltage,
output power,
and supply
voltage.

In contrast to hi-fi power amplifiers,
which are designed for low distor-
tion, high slew rate, low noise, and
good damping factor, public ad-
dress amplifiers must meet different
requirements. Among these are high
output power, good reliability, and
robustness.

Generally, low supply voltages result
in relatively low power outputs. This
is, however, not the case in the pres-
ent amplifier, as shown in Table 1.
The figures in this table become
even more impressive if the modest
dimensions of the amplifier are con-
sidered. Note that the supply
voltages in the table are either asym-

metrical or symmetrical. The differ-
ence between the two will be
reverted to later.

Circuit description

Designing a power amplifier used to
be a complicated and complex job. .
Nowadays, however, the complexity
is contained entirely in proprietary
ICs to which only a few external
components need to be added to ob-
tain a first-class amplifier. There used
to be a lot of scepticism about these
'black boxes', but over the years they
have more than proved themselves,
so that any doubt as to their oper-
ational qualities is quite misplaced.
The only drawback with them is that
there is so little to explain about the
final circuit.

Figure 1 shows that the proposed
design is of the double push-pull
type, which is about the only way
that a large output power can be ob-
tained from a low supply voltage.
This arrangement ensures an
available power output double that

of a single push-pull amplifier. If
then the output impedance is
halved, the output power is doubled
again, so that four times as much
power becomes available. Note that
the power developed by an output
amplifier is determined by

P=u pp j /8Rl [W] (1)

where P is het output power, u PP is
the maximum peak-to-peak voltage
excursion across the load, and Rl is
the load (i.e., loudspeaker) im-
pedance.

A noteworthy aspect of the circuit is
that, although the loudspeaker is
direct-coupled to the amplifiers, an
asymmetrical supply may be used.
This is made possible by the virtual
earth created by R« and R«. It is, of
course, true that this is high im-
pedance, but this is of no conse-
quence here because the output
current does not return via earth.
However, since it is permissible —
and often preferable, as will be seen
later — for a symmetrical supply to
be used, the circuit makes provision

2-38 .1*10. in

>2-39

Fig. 3. The
printed-circuit
board for the
power amplifier
can conveniently
be combined
with a 200mm
long heat sink
for the power
stages.

Resistors:

R>;R>;R«;Ri>=2B2
Ri;Ri;Ru = 100k
R«;R# = 2k2

Capacitors:

Cr = 1n5

C». . .Cs;Cr = 220n
Ce = 10jr;16V

Ti;Ti = BD250

T«T. - BD249
Di...D4 = IN4001
ICi;ICi = TDA2030(A)

Miscellaneous:

100 * 200 <25 mm, e.fl.
Fischer SK42 type

Fig. 4.

Completed as-
sembly of PCB
and heat sink.

2-40 «lek!or india f

on the printed-circuit board shown
in Fig. 3. This has been designed to
mate with a 200 mm long heat sink
for the ICs and power transistors,
which obviates any long connecting
wires between these components
and the board.

The driver ICs and power transistors
should be mounted onto the heat
sink with the aid of insulating
washers (provided with these com-
ponents) and ample silicone grease.
It is recommended to tap three M3
threaded holes in the flange of the
heat sink to facilitate mounting the
PCR The completed assembly is
shown in Fig. 4.

The amplifier may be fitted in a
suitable case of its own, but in many
cases it would be more sensible to
combine it with an appropriate pre-
amplifier. Whatever housing is used,
however, it is important that there is
adequate cooling of the heat sink.
Screened cable should be used for
connecting the pre-amplifier to the
power amplifier. Power lines and
loudspeaker connections should
have a cross-sectional area of not
less than 2.5 mm*.

Use of a pre-
amplifier

As stated earlier, the power amplifier
may be powered by an asymmetrical
12 to 36 V supply or a symmetrical
+6 to ±18 V one. The supply may
consist of a 12 V car battery, or two or
more of such batteries in series, or it
may be a simple mains supply. In the
latter case, a symmetrical version is
obtained by the use of a centre-
tapped mains transformer.

If the supply is used to power the
pre-amplifier also, care should be
taken to ensure that the signal earths
in the two amplifiers are at the same
potential. If, for instance, the pre-
amplifier is intended to be powered
by an asymmetrical supply and is
then connected to the power ampli-
fier supply, the e line in the power
amplifier will be shorted to earth as
shown in Fig. 2a. The only solution to
this problem is to use a symmetrical
supply and power the pre-amplifier
from one half of this as shown in
Fig. 2b. If the pre-amplifier is in-
tended to be powered symmetri-
cally, there are, of course, no
problems — see Fig. 2c. H

SCS application

The banks are closed and
you want to check how
much money is in your ac-
count. You put your plastic
card into an automated
teller machine and the
bad news is printed out.
Does anybody else know?
Your employer's personnel
department has just put
all its records on com-
puter, including details ot
your employment history,
your health, your edu-
cation, and even family in-
formation. Once It was
kept in a locked cabinet
and only one person had
the key. Now terminals all
over the company, even in
other branches, are con-
nected to the same com-
puter. Who can read your
personal files?

Your company has in-
stalled an up-to-the minute
office automation system.
Senior staff dial it up every
evening and use their own
home computers to
analyse sales data and
work on new product
plans. How do they know
that rival companies are
not dialling the same
computer and stealing In-
formation?

Espionage or
accident

In each case it is imposs-
ible to give an accurate
answer. Because of its very
nature, no one knows how
much computer "es-
pionage" goes on. What is
certain is that, as
companies put more and
more of their vital data —
and information about
their employees and
customers — onto com-
puter, the risks are grow-
ing. Information thieves
can be just as clever as
the people running the
computer system.

But it is not only deliberate

Protecting data
from prying eyes

theft of information that
computer operators have
to guard against. If many
different departments use
the same system — or if
there are links between
systems — there is the risk
of accidentally straying on
private information, rather
like hearing something
spicy on a crossed tele-
phone line.

It is relatively easy to pro-
tect information in old
fashioned filing cabinets.

A lock on the cabinet and
another on the door may
be enough. In any case, it
is obvious to anyone when
someone is rifling through

a drawer. But who knows
who is tapped into a
computer?

Protection by
taw

Protection of computer
data is no longer optional
in Britain, one country
where electronic data is
now very widely used. It
has become a legal
obligation, certainly as far
as personal information is
concerned. A Data Protec-
tion Authority, established
by the British Government,

Dr John Yardley. managing director of JPY Associates,
which produces the data protection software package
Data-Lock.

by Alan Burkitt

is drawing up rules that
will apply to everyone us-
ing a computer to store or
process such information
— in theory, even owners
of home computers.
Fortunately, however, there
are means of protecting
information. The personal
identity number needed
to key into a bank
machine is helpful, but to
be easily memorized such
numbers usually have only
four digits, so that a deter-
mined spy could perhaps
discover it by trial and er-
ror. In any case, that does
not protect the information
zipping along the tele-
phone wires from the
bank's computer, which
can be tapped by anyone
with simple equipment.
More sophisticated means
are needed and here the
computer industry has
taken lessons from govern-
ments that want to protect
secrets by using codes.
Confidential information is
converted into a secret
code by an encryption
unit using a system known
as the Data Encryption
Standard or DES. It can
then be stored in a com-
puter or transmitted down
a telephone line. Either
the system is pre-set with
an individual code or key,
or it is typed in by the
operator at the time.
Encrypted data bear no
resemblance to the
original and, because of
the vast number of poss-
ible keys, only one of the
world's leading Intelli-
gence agencies would
have the ability to
decipher it.

Special

software

But the system is expens-
ive, or it has been until
now. A DES encoder /de-

coder costs about £3000
in the United Kingdom
and a separate unit has to
be bought for every bank
branch and every
company computer.
However, it is possible to
use the computer itself as
the encryption system. All
that is needed, instead ot
a "black box”, is special
software that programs the
computer to encode or
decode information as it
goes into or comes out of
the memory.

That means companies
can ensure the confiden-
tiality of their information
at only a fraction of the
cost and, as the DES code
is used, the security is
every bit as good. Pion-
eering the software ap-
proach is a three year old
company, JPY Associates
of New Malden, which has
invented a system called
Data-Lock.

John Yardley, JPY's manag-
ing director, said: "So far.
our customers have
wanted to protect either
personal data or commer-
cially sensitive information.
We have had a lot of en-
quiries from people
developing office auto-
mation systems, where
they want Data-Lock writ-
ten into the software."

Among special devices to
allow handicapped
children to play and learn
with computers are a work
station with voice syn-
thesizer for the blind, a
helmet projecting an elec-
tronic beam for those with
difficulties in controlling
their limbs, and a con-
troller operated by
movements of the eye for
the severely disabled.
Simpler aids include extra-
large switches or levers for
children wo have difficult-

2-42 elektor India feDfuary 1986

Speeding up
the process

Compared with the tra-
ditional hardware solution,
this might work out at
under £500 for the
cheepest version — a big
saving. JPY has started off
by producing the software
to run on PDP-11 and VAX
computers made by the
Digital Equipment
Company.

A potential drawback of
software encryption is that
it can be slow: after all,
some complex mathemati-
cal processes are applied
to the data. "We have put
everything into our soft-
ware to make it go fast,"
said Dr Yardley.

"Data-Lock processes three
or four blocks of 512
characters each per sec-
ond." In other words this
whole article could be en-
crypted in about four
seconds — not too long to
wait if the information is
confidential.

"The encrypted infor-
mation can be computer
programs, databases, or
anything. It really does not
matter, and we think we
can get the system to
operate even faster."

Even the engineer main-

taining a computer can-
not obtain the data
without the key, which is a
sequence of eight
characters, numbers or let-
ters. "It could be a name
or a telephone number,
although ideally it should
be random,” Dr Yardley
said.

Limited keys

In one approach, the
operator enters the key at
the terminal so that data
can be coded or decod-
ed. But the key, itself en-
coded, can be stored in
the system and the
operator keys in another
sequence to operate it.
With this method, users
can label their keys with
familiar names, without
compromising security.

This system is called Key
Manager.

What happens if the
operator forgets the se-
quence, or even dies? "In
co-operation with the
owner of the computer we
can get the key back if it
is on a key manager
system,” Dr Yardley said.
"Otherwise it is lost for all
time."

Obviously this can create

Educational software
for the handicapped

ies with standard switches;
and various computer pro-
grams tailored specifically
for those with different
learning disabilities.

All this hardware and soft-
ware has been produced
by the four Special Edu-
cation Microelectronics
Research Centres
(SEMERCs) covering
England and Wales. The
centres are a task force
that collaborates closely
with the regional and na-
tional offices ot Britain's

highly successful Micro-
electronics Education Pro-
gramme (MEP), established
in 1981 to foster computer
awareness among
teachers and pupils.

Power and
flexibility

MEP’s national co-
ordinator for special edu-
cation is Mary Hope, who
is based at the Council for

problems, and JPY is think-
ing of a way round them.
One possibility is to limit
the operator to a certain
number of possible keys —
so that they can be tried
in turn.

"It may take a week to
come up with the right
one," said Dr. Yardley. That
of course may not be
secure enough for really
vital commercial infor-
mation, but few intending
spies have the resources
to devote that length of
computer time to ordinary
material.

And it will mean we can
be confident no
unauthorized eyes are
looking at our bank ac-
counts or other personal
details stored on
computer.

by Jack Cross

Educational Technology in
London. In 1982 she wrote:
"We are just starting to get
some idea of how we can
help children with learn-
ing difficulties; to enable
the severely physically
handicapped to com-
municate and express
themselves despite their
handicaps; to help deaf
children overcome their
speech and language
problems; and to minimize
some aspects of the han-
dicap of blindness. The

power and flexibility of the
new technology is enor-
mous and must be one of
the most excisting areas ot
development in edu-
cation."

Well balanced
team

The best known SEMERC is
probably the one at
Newcastle, which covers
northern England and is
conviently close to the
MEP Directorate’s offices.
Colin Richards, the centre's
manager, calls micro-
electronics "a caring
technology which opens
doors that, without it,
would remain firmly shut".
He and his team en-
thusiastically demonstrate
the capabilities of some of
the many devices held by
the SEMERC emphasizing
that each has to be
evaluated as a teaching
tool.

It Is not difficult for the
layman to perceive why
so many of these are de-
signed to operate without
the familiar typewriter
keyboard. It is not suitable
for the blind or near-blind,
or for young people who
suffer from a severe motor
impairment. In any case,
the aim is to assist in
general teaching/learning
activities, not to impart
touch typing skills. So most
of the hardware can be
controlled by such actions
as the flick of an eyelid,

sucking and blowing, us-
ing a large simple switch
— or whatever method
suits each handicapped
child.

For the very young, men-
tally or multiple-
handicapped children,
there are simple and
cheap electronic inter-
faces that allow them to
use a large switch to
operate an ordinary toy —
from a drum playing fluffy
rabbit to a laser gun. This
system, developed by the
Manchester SEMERC can
be operated by hand,
foot, breath pressure, voice
or grasp. It can be used to
encourage attentiveness
or as a reward for perform-
ing in response to varied
commands trom the com-
puter, such as: "Hit the
blue switch three times
and the rabbit will beat
his drum.”

Head

movements

The "Photonic Wand" is a
kind of plastic helmet
which projects an elec-
tronic beam to control a
moving spot on a visual
display screen. Though
designed — by Dr John
Cole of Chester — tor
people capable of mak-
ing well controlled head
movements, the wand is a
source of motivation even
for sufferers from (airly
severe cerebral palsy. It
comes complete with pro-

grams dealing with
literacy, number work,
music and painting.

The Newcastle SEMERC
group is trying to find out
what kind of software
works best with Twinkle.

The device utilizes eye
movements which, by way
of sensors affixed to the
child's forehead, give the
user access to the
computer.

The staff is particularly ex-
cited about the potential
ot the Adventure Board
produced by Britain's Edu-
cation Development
Centre at Walsall. A
magnetized playing piece
is moved along grooved
pathways cut into a solidly
built flat surface, actuating
a series of 16 switches
which produce messages
on a screen. The demon-
stration model portrays a
village, with shops, a farm,
animals and trees - the
kind of small toys found in
any home or infant school.
A typical "adventure"
might read: "I set off from
the farm gate. . . I saw a
duck and three duckl-
ings. . .1 crossed the
bridge over the stream" —
and so on. The program
can be stored, returned to,
and, with the appropriate
equipment, printed out.
Teachers can prepare
their own boards and •
write their own stories
about anything from the
route to school to a
treasure island. Walsall
Education Authority, which
holds the patent on
Adventure Board, hopes to
go into commercial pro-
duction soon.

Overlay

keyboard

Teachers particularly prize
devices that allow them to
present their own material
without having to have
any programming skills.
The overlay or Concept
keyboard, now coming
into general use, is essen-
tially a flat board, pro-
duced in various sizes,
inlaid with rows of touch
sensitive pads. On this is
placed a piece of paper

with up to 256 pre-
produced items — they
can be words, pictures,
diagrams, numbers, tex-
tured surfaces or passages
in braille. The pupil gains
access to this "prompt,
cheap, microprocessor" as
Mr Richards describes it,
by pressing the relevant
section of the overlay.

There are several devices
that use graphics to assist
speech training and make
it enjoyable. Micro Mike
has been developed by
the Manchester SEMERC
and the Northwest
England Computer Club.
Children can draw pic-
tures or control racing
vehicles on the screen by
making the correct sounds
at the right volume. Visi
Speech (Jessop Ac-
coustics) displays sounds
in shapes, which the
speech impaired child Is
invited to match. C-Speak
(Rank Stanley Cox)
operates In much the
same way, using a radio
microphone and a dis-
play on an ordinary tele-
vision set.

Viewscan has been de-
signed to help visually
handicapped people
read text. He or she moves
a miniature scanning
camera over the surface
of the page; its content is
magnified up to 64 times
and can appear as dark
on a light background, or
the reverse, In selected
degrees ot intensity.

Totally blind people can
learn to use Opticon,
which translates printed
words into tactile signals
received by the reader
through the fingertips.

Voice syn-
thesizer

Britain's Royal National In-
stitution for the Blind has
particularly welcomed a
work station developed
with MEP funding by
Dr Tom Vincent of the
Open University at Milton
Keynes. A keyboard
microcomputer turns
words into print and then,
using a voice synthesizer,

I Wendy Kerton operates a Braille keyboard microcom-
puter which turns her words into print and then into
I speech.

into speech. It received an
award from the BBC radio
"In Touch" programme as
the year’s best invention to
help the visually disabled.
MEP Touch Screens are a
welcome addition. These
fit to the front of a com-
puter screen and are a
way of bypassing the stan-
dard keyboard. Mike
Bostock, technology
manager for MEP, writes:
"For many children this is
not a serious inconve-
nience, but such distrac-
tion can become acute
for very young children,
those with manipulation
problems, short concentra-
tion spans, or acute learn-
ing difficulties,” With this
system, they simply have
to point to the required

Much has been written on
the changes brought
about by the Introduction
of robotics in vehicle pro-
duction, but little is heard
about the tremendous pro-
gress which the use of
computers and robots has
brought to the initial
stages of vehicle concep-
tion — the design stage.
Technological advances
in recent years have led
to an accelerated vehicle
renewal rate. But develop-
ment time is long. First, the
general style and
mechanical

characteristics of the ve-
hicle must be defined,
feasibility studies carried
out, production methods
determined, production
plants modified, tooling
manufactured, and so on.
The production of a new
vehicle is like a giant jig-
saw puzzle which relies on
the input of many
specialist technologies
before each piece falls
into place.

object on the screen and
the computer responds
directly.

The SEMERCs hold many
thousands of software pro-
grams. Some are commer-
cially produced, others
were developed by MEP
groups and may be
copied, without charge in
Britain, for educational
purposes. Mr Richards and
his team in Newcastle are
continually evaluating
new devices and software
in discussion with teachers
and other users' groups,
and demonstrating their
potential classroom use in
an area running from the
Scottish border southeast
to the boundary between
Lincolnshire and Cam-
bridgeshire

Since it is British Govern-
ment policy to put as
many handicapped
children as possible into
mainstream education
(only 2% go to special
schools), many of those
who attend the SEMERC
courses are non-specialist
teachers from ordinary
primary or secondary
schools.

Special Education Micro-
electronics
Research Centre.
Newcastle

Polytechnic Coach Lane
Campus

Newcastle upon Tyne.
England, NE7 7XA.

CAD in practice
at Renault

The robot as
sculptor

Once past the heavily-
guarded gates of the
Renault Technical Centre
at Rueil. outside Paris, the
visitor to the styling
department is in for a sur-
prise. Looking through a

glazed partition protecting
! a battery of screens and
computer keyboards, he
would see a life-size car
taking shape in plastic -
under the expert "hand"
of a robot. The articulated
robot "arm" automatically
obeys the commands of a
computer programme. In
just two days, the life-size

Micro-electronics Edu-
cation

Programme's Directorate.
Cheviot

House. Coach Lane Cam-
pus

Newcastle upon Tyne,
England. NE7 7XA.
Council for Educational
Technology,

3 Devonshire Street.
London, WIN 2BA.

model will be ready for
finishing and painting.

This robot sculptor is just
one of the aspects of what
is known In Renault als
CAO (Conception Assistee
par Ordlnateur) or Com-
puter Aided Design (CAD).
The use of computers at
the design stage may
seem surprising — almost
shocking. What happened
to the artistic talent, flair,
design abilities and
knowledge, which no
mere machine can pro-
vide? But it is not really a
contradiction in terms: an
electric typewriter enables
the author to put his
thoughts on paper more
quickly — it does not pro-
vide him with the plot for
his story.

From drawing
board to
model

The first step in the design

of a new vehicle is taken
by the stylist, who pro-
duces the initial rough
sketches. But a vehicle is a
three-dimensional object
and nothing can be
finalized until a model is
prepared — first at 1/5th
scale and then, if initial
results seem satisfactory,
life-size.

For accurate comparison,
several models must be
made based on different
sketches. But to make a
model, sketches must first
be translated into ac-
curate plans from which
the model will be con-
structed. Before the in-
troduction of CAO, all
these stages involved long
manual processes. First
step for Renault was the in-
troduction ot computers
able to convert drawings
into plans. Progress has
been rapid.

The battle of
pictures

Touch sensitive screens,
sculptor robots, and 3-D in-
terpretation ot drawings
are just some ot the CAO
applications at Renault.
But now a new field of
development promises
rich rewards: the synthetic
computer picture. "Star
Wars" fans will be lamiliar
with these animated im-
ages created directly on
a screen by a computer
programme — the tech-
nique also used for cer-
tain television program-

Formerly, it was not until
the lite-size model was
completed that a realistic
3-D representation ot the
vehicle was obtained. But
the touch sensitive screen
image, good as it is, is still

One computer pro-
gramme can interpret the
stylist's drawing in plan
form on to a screen and
on to a tracing console
This programme not only
interprets a drawing — it
can also handle volume
(small model) or work from
a sketch drawn on a
graphic console (touch
sensitive screen).

Once the plans have
been prepared, the life-
size model can be con-
structed by the sculptor
robot, again using the pro-
gramme which controlled
the preparation of the
original plans.

only a line drawing, and
for stylists and decision-
makers it is vital to be
able to "visualize" the pro-
ject from every angle as
early as possible, before
committing themselves to
costly model production.
The synthetic computer im-
age answers this need.

Twice as fast

Using traditional techni-
ques, it took 12 to 16 weeks
to make a model. With
CAO and the use of
plastics, the time has

been reduced by half —
to between 6 and 8 weeks.
While reducing costs
(again by half), the pro-
cess also liberates
designers from what was a
time-consuming pro-
cedure. The Renault com-
puter system used for all
these operations is known
as Unisurf. Matra Datavi-
sion, who developed a
three-dimensional logistics
system known as Euclid,
recently signed a co-
operation agreement with
Renault, as a result of
which the Renault Unisurf
and Surfapt systems will be
progressively integrated
with Euclid.

A means of
communi-
cation

All this may sound like
magic: project an image
of the vehicle on to the
screen, select the desired
angle, choose the colour
— and there you have it.
But one can go turther by
calling for a close-up of a
detail (wheel arch, wind-
screen, etc) and this
presents many ad-
vantages

The image thus obtained,
apart from its aesthetic
quality. Is rigorously ac-
curate and makes it poss-
ible to check volumes and
line quickly. It is also an
efficient means of com-
munication with other
departments concerned,
and enables them to be-
gin their work on the pro-
ject at an early stage.

The synthetic image is in
full development. Soon it
will be possible to project
the image, life-size, on to a
screen . . .to modify a
curve, a volume, or a
detail straight on to the
screen. . .to develop a
programme which will
animate the image so that
it can be seen in motion
against a chosen
background.

The field of research is
vast, but it depends on in-
tensive work in pro-
gramme development.
Renault is tackling the task

in co-operation with uni-
versities specializing in
these techniques.

But the synthetic image,
like the sculptor robot, is
only a tool — fascinating,
powerful, and enabling
designers to devote their
talents entirely to creativity
by liberating them from
long and repetitive tasks.
Just as the fountain pen
assisted writers by saving
them the time formerly
spent sharpening their
quills, CAO assists creativi-
ty but is not a substitute for

Renault UK. Limited
Western Avenue. London,
W3 ORZ Telephone:
019923481

The concept of MSX makes it possible
to design programs and add-on
units that will work with any MSX
computer. In a short series of articles
\ we will offer you designs for a
number of useful extensions intended
to enhance the computer's abilities to
control external equipment.

EXTENSIONS - 1

The coming of the MSX series of
computers has been a success in so
far that incompatibility problems
among different computer types
have largely been overcome, both in
the software and in the hardware
field. With standardization finally at a
practical level, it is now possible to
develop programs and add-on units
that will work on any MSX computer,
irrespective of its make. The first part
in this series of articles will deal with
a versatile MSX input /output bus and
a digitizer plus 8-bit I/O port to plug
into it, thus considerably enhancing
the computer's abilities to control ex-
ternal equipment.

Microsoft Extended BASIC (MSX) is
now the software base for a whole
group of computers with largely
identical hardware features. Com-
pared with the well-known Com-
modore C64, they feature the
following improvements:

• a further expanded BASIC with
improved string handling

capabilities and a larger number of
graphic commands:

• built-in printer interface to the
Centronics standard;

• cartridge slots to plug in hardware
expansions or game/utility packs;

• disc operating system (MS-DOS)
standard available.

This article will concentrate on the
modifications to the universal I/O
bus designed to operate with the
C64 computer and as described in
the June 1985 issue of Elektoi
India. These modifications are
necessary because MSX computers
generally use the Z80 micropro-
cessor. The add-on boards for this
bus published so far will also require
slight modifications; the digitizer
A/D and D'A converter is featured in
the June 1985, and the 8-bit I/O port
in the January 1986 issue of Elektor

I/O bus
modifications

For use with the Z80 processor, the
universal I/O bus will require some
minor modifications because of the
IORQ, RD and WR signals of this
CPU, which also uses another selec-
tion of I/O channels as compared to
the 65XX t ype.

The IORQ signal will have to be in-
verted if it is to replace the 65XX <t>2
clockpulse. For this purpose, IC2 is
replaced by an inverter Type
74LS240. This modification will in-
evitably invert all signals passing
through IC2, but it is quite useful
because the bus RD (read) signal is

2-46 .i»to

inverted as well, and may now be
used to enable data transfer onto the
bus without contention problems.
Because addresses Ao. . .A3 are also
inverted, slot numbers 1. . .4 will be
in the reverse order, as will the four
addresses available within each in-
dividual slot. Programs for the
digitizer and the 8-bit I/O port need
no modifications, however, for these
boards are selected by SS (Slot Sel-
ect) only, and do not use any address

decoding within their slots.

A further advantage of the Z80 IORQ
signal is a simplification of the ad-
dress decoding circuitry. As this

CPU is able to select I/ O cha nnels
0 . . . 2S5 by means of IORQ and
A0...A7, compone nts 1C 3, Ri.Rs
and Si are removed. IORQ is connec-
ted to I/O bus line A10 while CPU
A1...A7 go to I/O bus lines
Ati. . . Ais. Connect pin 18 with 19 of
the empty ICi s ocket to enable ICi
with the IORQ signal and install
jumpers b, d, and g as indicated in
Fig. 1.

The Z80 processor uses sep arate
read and write signals (RD and^WR)
instead of a combination (R/W), so
the IRQ line had to be sacrificed to
make room for the Z80 WR. This im-

plies that the digitizer and the 8-bit
I/O port will have to take their WR
signal from the former IRQ bus con-
nection, but more about this later.
As MSX computers do not use the
BUSACK signal, Tt, R13, and Ru may
be removed. Fig. 2 shows the
modified PCB.

Add-on board
modifications

The digitizer and the 8-bit I/O
boards will have to be slightly

Fig. 1. Orig-
inally designed
for use with the
Commodore
C64. this
universal I/O
bus is easily
modified to
work with any
type of MSX
computer.

.2-47

Fig. 2.

Modified PCB
layout for the
I/O bus.
Remember to
connect pin 18
to pin 19 of the
empty ICi
socket.

modified for use with the Z80 WR
signal. As for the former, Fig. 3
shows these modifications in heavy
lines . 1C? is now clocked direct by
WR. lb make this possible, pin 6 of
IC« is cut off from the IC body. A
wire is installed from IC; pin 7 to
pin 6 of the bus connector — see
Fig. 4.

The 8-bit I/O port is modified as
follows: disconnect pin 8 of ICi by
cutting it off. Fig. 6 shows how
socket pin 8 of IC> is connected to
bus pin 6; do not forget to fit the
marked wire link on the component

side. Fig. 5 shows the relevant part of
the circuit diagram for reference
purposes.

This completes the adaptations of
the digitizer and the 8-bit I/O port for
MSX applications.

Ribbon cable
connection

Depending on the make of MSX
computer, it either features a so-
called cartridge slot, a 50-way I/O
connector, or both. Connectors for

the cartridge slot are hard to obtain,
however, because they are in fact no
more than a projecting piece of the
PCB on which the expansion circuit
(EPROM, interface) is mounted. Car-
tridges usually come in a small
plastic housing, and a cheap one
may be salvaged and carefully taken
apart, so that it will still function after
the flat ribbon cable is connected to
it for signal extension towards the
I/O bus. A means for switching off
the cartridge (ie. the contained
PROM) will be needed when the
wires to the I/O bus are connected

2-48 .i.

and cartridge protection. The
necessary circuitry for this protec-
tion, however, may not be incor-
porated at all, so it is still sensible to
switch off the computer before plug-
ging in or removing any cartridges.

Testing

After completion of the wiring job
using Table 1 as a guide, an initial
test of the I/O bus may be per-
formed. The digitizer and the 8-bit
I/O must be removed for this
purpose.

The Z80 processor features 256 I/O
addresses, and every slot on the I/O
bus takes four of them. With all
switches on block S2 closed, the
computer will find the slots in the ad-
dress range 0 ... IS. lb check the
correct function of the modified ad-
dress decoding part, two bits of
every slot are hard-wired to its SS
signal (pin 7), thus creating a byte to
be read by the computer at the ap-
propiate address. At slot 1, connec-
tor pins 17 and 18 are wired to pin 7.
At slot 2, pins 15 and 16, at slot 3 pins
13 and 14 and, finally, pins 11 and 12
go to pin 7 of slot 4. Remember that
the slot numbers refer to the new,
reversed order, and that the com-
puter may only read these addresses
when the wires are present.

Once the computer addresses a slot,
its SS signal goes low, together with
the connected databits. This results
in the following decimal values to be
read from the slots: slot 1 = 63; slot 2
= 207; slot 3 = 243; and slot 4 = 252.
Listing 1 shows a program to test all
I/O locations from 0...255. It will
skip any addresses reading 255, i.e.
with all bits high. The range from
0. . .127 is reserved for the I/O bus,
as MSX itself uses 128 and onwards.
Locations 152 and 162, for instance,
contain the VDP (Video Display Pro-
cessor) and PSG (Programmable
Sound Generator) status, respect-
ively. In case the indicated decimal
values are not read correctly by the
program, some wiring error may
have been made, or the jumpers b, d,
and g are not yet in their correct pos-

If the first test-run is successful, the
modified digitizer board may be
plugged in and tested using
program Listing 2. Fig. B shows how
the digitizer is tested by connecting
the wipers of eight preset poten-
tiometers to the appropriate inputs.
The potentiometers are connected
between the +5 V supply and earth.
Program Listing 2 will start the A/D
converter eight times to sub-
sequently read the converted values.
Turning the presets will verify the
correct function of the digitizer

2-50 elektor India febfuary 1986

10 OPEN" grp : " FOROUTPUTAS# 1
20 AC0L=9:BC0L=8:CC0L=14
30 R 1=8 :R2=8:U 1=4:112=4
40 SCREEN2,2
50 COLOR 15, 4, 4
60 CLS

70 LINE<32,0) -(32, 176)

80 LI NEC 3l,0)-(31, 176)

90 LINESTEPC + 0 , + 0) -STEP< + 2 10 , ♦ 0)

100 LINESTEPC +0 , + 1) -STEP<-210,+0)

110 FORY= 155T0 15STEP— 20
120 PRESET ( 0 , Y-2)

130 PRINT#1,MID*<STR*< 155-Y) , 1) i
140 LINE! 3 1 ,Y) -< 36, Y* 1)

150 NEXTY

160 GOSUB 180

170 0=0: GOTO 160

180 FORA= 1T08

190 OUT0.A-1

200 H< A) =INP( 0) / 1 . 6

210 A*<A)=CHR*<A+48)

220 NEXT

230 F0RN=56T0232STEP24
240 Y= 160 :X=N
250 Q=Q+ 1 : H=H< Q)

260 A*=A*<Q)

270 I FNOT < H<=M ( Q > - 1 0RH>M< Q> ♦ 1 ) THEN320
280 PRESET < X-5 , Y+ 24)

290 PRINT tt 1 , RI GHT*< AS , 2) ;

300 I FP= 1ANDMC Q> =>4THENGOSUB350ELSEGOSUB380

Listing 3.
Graphic
presentation of
the input
signals to the
digitizer is
possible after
keying in this
program writ-
ten in Microsoft
Extended
BASIC.

340 RETURN

350 LINECX- 1 , Y-U 1-H) -<X-R 1 , Y-U 1) .ACOL.BF
360 LINECX , Y-U2-H) -<X- 1*R2 , Y-U 1 ) .BCOL.BF
370 GOTO400

380 LINE<X-1,Y-U1-H)-<X-R1,Y> .ACOL.BF
390 LINE<X,Y-U2-H)-<X-1+R2,Y> .BCOL.BF
400 LINECX-Rl ,0) -<X-1+R2,Y-U1-H> ,3,BF
410 PRESET <X,Y-H)

420 DRAW" c=c col i a2m+=r 1 ; , +=u 1 ;m-=r2 ; , +=u2 ; "
430 PRESET <X- 1 , Y-H-U 1-U2)

440 DRAW” c=cco! ;<<0m*=r 1 ; , +=u 1 ;m-=r2; ,+=u2j*
450 PAINTCX , Y-U 1-H) , CCOL
460 PRESET (X , Y>

470 DRAW"c=b;1+=rl;u+=ul;m+=rl; ,+=uls"

480 PAINTSTEPC -4 , - 1) ,0
490 PRESET <X,Y> ,0

500 DRAW"r+=r2iu*=u2;m-=r2; ,+=u2j "

510 PAINT STEPC *4,-1) , 0
520 RETURN

/O

•5 s'

^ *

= 74LSOO

board. Because Listing 2 uses slot 8-bit I/O board. All inputs are con- ULN 2803. By changing A in line 20
address 112, switches 12, 13, and 14, nected direct to the outputs by eight to NOT A AND 2S5, the test program

as indicated on the PCB, will have to wires. The program will read back will print two neat columns of 0. . .7.

be opened. the sent values 0. . .7 as 255. . .248

This program is also useful to test the because of the bit-inversion by the

Practical example

The input signals to the digitizer are
highly suitable for graphic presen-
tation on the screen, as they are sim-
ultaneously visualized. Listing 3
suggests a program to implement
this. The value at each input is shown
as a vertical bar on the screen. A
scale enables the values to be read
off. Lines 20 and 30 of the program
allow experiments with colours and
bar sizes.

The vertical axis is drawn in lines 70
and 80, the horizontal one in lines 90
and 100. The scale is set up in lines
110... 150, completing the basic
layout of the screen. Line 170 is the
central starting point in the program,
the remainder of which consists of
two subroutines. Lines 180 . . . 220
first read the input data of the
digitizer before lines 230. . .340 put

I B

Fig. 3

Modified cir-
cuit diagram of
the digitizer
board. One of
the necessary
alterations in-
volves the
removal of
pin 6 from IC*

Fig. 4. PCB of
the digitizer, in-
cluding the ex-
tra wire link.

Fig. 5. The
heavy lines in-
dicate the
modifications to
the 8-bit I/O
board.

Fig. 6.

Modified PCB
of the 8- bit I/O
port. Just as
with the
digitizer, an IC
pin has to be
removed; in
this case pin 8
of I Ci.

Fig. 7. It pays
to make one's
own printer
cable. Eleven
wires in a
cable between
two connectors
do the job.

them on the screen using another
subroutine from line 350 onwards
which merely converts the data into
coloured bars. To erase any
previously higher values, line 400
puts a green part on top of every bar.
As a finishing touch, lines 410. . .510
provide bevelled edges to the bars.

A printer cable

As soon as the computer starts to

grow into a so-called 'system', cables
are required to connect peripheral
equipment like a disc drive, a
modem, or a printer. Considering
the cost of a ready-made cable, it is
certainly worthwhile to make one
yourself for connecting a printer.
Fig. 7 lists the necessary connec-
tions in the form of a table; the 14-way
connector goes into the computer
and the 36-way type (known as a
Centronics connector) into the
printer.

More interesting extensions of MSX
computers will be featured in future
issues of Elektoi India. N

2-52 elekto

Chapter 2 of our Digi-Course II had described different
versions of the set-reset Flipflop. Operation of the D-
Flipflop was described in-detail. The gates M and N are
wired in such a way that the condition of the C input
decides whether or not the input D will be passed on to
output Q or blocked. The actual Flipflop part consists of

The NOR-version of the D-Flipflop is shown in figure 2.
In this circuit, the value present at input D is stored by
the Flipflop on the "0" to "I" transition of the signal on

Digi-Course II

Chapter 3

From the truth table we can see that the Q and U

circuit that we started with, in Chapter 1. It is reversed in
this case and we interchange Q and 5.

In practice, this simple D-Flipflop does not meet all the
requirement of an advanced design for functioning as a
storage element. In the application of Digital technology,
storage elements are required which take up and retain
information in a precisely defined time frame. In a D-
Type Flipflop the moment of storing the input signal level
is precisely defined, but before that moment, the output
Q takes on all the values that appear at the input.
Practically this problem is solved by using an additional
Flipflop. This additional intermediate Flipflop takes on all
the values appearing at the input, but passes on the
value to the output Flipflop ony when a command to do
so is given to it. This combination of Flipflop is called a
I Master Slave JK Flipflop.

The figure 4. given above shows one such complicated
circuit. This circuit requires all the gates so far installed
on our Digilex Board. The contents inside the dotted
rectanles are actually two independent RS Flipflops, with
status control. Two output indicators are used for each
Flipflop and the conditions of both Master and Slave
Flipflops can be seen simultaneously on these LEDs. A
logical "1 ” on the C input (L9) causes all the values at
the inputs J and K to be taken up by the first Flipflop.

The second Flipflop reacts only to the values provided by
the first Flipflop at the time when input C goes to "O"
from ”1 ”. The actual response of this circuit can be
studied by observing the conditions of the output
indicator LEDs B/C and E/F.

1 2-53

selex

The control signal at C (L9) is also called the 'Clock'
signal. As the Flipflop reacts only at the moment when
input C goes to "O" from "1", it is said to be edge-
triggered. triggered by the negative going edge of the
input clock. As the edge defines an exact point of time
during the sequence of events, this type of arrangement
is practically very useful. In the computer technology,
mostly the triggering is done on the positive or negative
going edge of control signals. Some computers are so
fast in operation, the rise time and fall time of the
positive or negative going edge also matter and affects
the operation.

A few trials on the circuit will be enough to clarify the
actual operation of the JK Flipflop. The circuit functions
properly only when the data remains stable during the
clock pulse. For proper operation data must be placed on
the inputs before the clock pulse is applied, and should
remain unchanged during the clock pulse.

The input conditions O/O on J/K do not affect the
output, the condition 1 /I on J/K reverses the output
conditions on Q/0 at every clock pulse.

For a practical application, one need not construct the
JK Flipflop using so many descrete gates. A single chip
containing two such JK Flipflops is commercially
available. The number of the TTL 1C is 74L576.

The circuit symbols for a Flipflop are shown in figure 5
and the pin diagram and internal connections are
illustrated in figure 6.

DIN EleMor/Elex

Figure 7 shows the Pin diagrams and internal
connections of some other TTL Flipflop ICs.

JK Flipflop with three inputs. PRESET end CLEAR

2-54 elektor India lebru

uraitHttm

Chhotani Building, 52-C, Proctor Road,
Grant Road (East), Bombay 400 007

selex

Wet-Finger-Test

The

Digilex-PCB

| is now available!

The Digilex-PCB is made from- best quality Glass-
Epoxy laminate and the tracks are bright tin plated,
I the track side is also soldermasked after plating,
i Block schematic layout, of components and
I terminals is printed on the component side.

Price:

I Rs. 85.00 + Maharashtra Sales Tax.
Delivery charges extra: Rs. 6.00

Send full amount by DD/MO/PO.

Available from:

precious 6

ELECTRONICS CORPORATION

,2-55

selex

Experiments
with a
Transistor

Transistor is the most
important component in
the history of electronics.

It is the most important
semiconductor device
designed so far. Even the
integrated circuits which

descrete circuits, are
nothing but chips with
thousands of transistors
integrated on them (in
many cases.the number is
very very large!)

The transistor is basically
a current amplifier, and a
few simple experiements
should illustrate the
functioning. The
components required for
these experiments will be
easily available from an
electronics shop.

The list of components
required is given below.

1 4.5 V Battery Pack
1 Transistor BC 547 or any
other small signal NPN

transistor like BC 107.
BC 550 etc.

1 Light Emitting Diode
(LED)

1 Resistance 220A
(1/8 Watt)

1 Potentiometer 100k A

In the experiments that we
are going to conduct, the
LED is used to give a
rough indication of the
level of current flowing
through the circuit,
because the brightness of
the glow depends on the
amount of current flowing
through the circuit.

First, let us see how
brightly the LED glows,
when directly connected in
series with a battery pack
and a 220Aresistor. The
figure 2 illustrates how
this is done. As the circuit
is complete, current flows
through the resistor as

Introducing the transistor
in the circuit as shown in

not much different initially.
The Base-Emitter junction
is connected in such a
way that it just behaves
like a diode connected in
forward direction
The make understanding

shows the transistor in a
simplified form as a
connection of two diodes
connected back to back
(Remember, this is not a
simple physical joint of
two diodes but it is an
integrated junction inside
the structure of the
transistor).

If the potentiometer is
connected to collector of
the transistor instead of

longer completed as the
two diodes are in reverse
direction and now the
Base-Collector junction
behaves like a diode
connected in reverse

shown in figure 6. The
collector is connected to
the junction of
potentiometer and LED.
and the center pin of the
potentiometer is connected

transistor. The behaviour
of this circuit is now
totally different. The

potentiometer setting now
prominently affects the
glow of the LED. As we
had already seen in figure
3 a current can flow
through the LED and'the
Base- Emitter junction of
the transistor. (Base
current).

This current causes a
much greater current to
flow through the Base-
Collector junction.

(Collector Current). Both
the Base current and
Collector current flow
simultaneously through

pole of the battery. The

acts as a current amplifier
and the current flowing
through the collector is
proportionally much
greater than that flowing
through the base.

We can summarise the
observations as follows:

1. A small base current
flows through the 202A
resistor, the LED and
the potentiometer. This

and leaves through the
emitter. This current
must flow through the
Base-Emitter junction.
The potentiometer
setting can be adjusted

2. The current flowing
through the Base-
Emitter junction decides
how much current
should flow through the
Base-Collector junction.
This property is called
the current
amplification, as the
Base-Collector junction
carries much greater
current than the Base-
Emitter junction.

The ratio between the
Collector Current and the
Base Current is the factor
by which the current is
amplified. This is called
the Gain of the transistor.

For a small signal
transistor like BC 547. the
current gain lies between
100 and 500; the heavy
duty transistor have a
lesser current gain... about
20 to 150.

The simplified
representation of a
transistor shown in figure
5 shows two diodes
connected back to back.
This junction can also
have exactly reverse
polarity. The first one is
called an NPN transistor
and the one with reversed
polarity is called an PNP

selex

Transistor Tester

Transistors are current
amplifiers. This property
has been stressed many
times so far. We have
even seen a very simple
way to give a rough check
to the transistor by using
the 'Wet-Finger-Test'.

For those who are
interested in going a step
ahead, and do some circuit

Figure 2 shows the
current paths inside a
transistor. The base
current flows into the base

emitter. The collector

collector and leaves
through the emitter. The
collector current is much

current and the ratio of
collector current and the

© # •

amplification factor or the
gain of the transistor. The

compared to a valve for
current through the
collector, which is opened
by the base current.

sense. The value only
controls the opening for
the flow, whereas in our

| ordinary valve the flow
also depends on the
pressure, which is not true
| in case of the transistor.

| The collector current is
| decided only by the base
| current and not by the
supply voltage between
' collector and emitter.

The current gain of the
1 transistor is expressed as
Collector Current

| same type number. The
value deviates
considerably from one
transistor to another. A
typical value can be
generalised for a particular
type number, and in
designing practical
circuits, care must be
taken to allow for this
deviation so that one
transistor can be
substituted with another
equivalent.

The circuit described here
is suitable for measuring
the gain of all the normal
NPN transistor types.
(Alternative circuit for PNP
transistors is suggested at
the end of this article.)

The basic circuit is given
in figure 3, to explain the
function of the transistor
tester. The transistor
under the test is
designated by TUT
(Transistor Under Test).

The constant current of 10
micro amperes, which is
the base current. This
must produce a current
through the collector
which is the 'B' times the
base current. The collector
current 1C flows through

calculate the value of B
accurately. To measure
this voltage using a meter
would be an expensive
proposition because the

selex

The Practical Circuit :

in figure 4, The constant
current source here is
made up of R1, R2. 01

The zener diode D1 gives a
constant voltage across
the emitter resistance R1 .
A constant voltage across

rough tf

I current is negligible

npared to the collector
| current, the emitter

irrent is almost same as
e collector current and
| hence the circuit results in
a constant colector current
for transistor T1 . The
| values have been selected

| amperes.

This constant current of
10 microamperes flows
through the base of the
test sample to be minus
pole of the battery. A
constant base current of
10 micro amperese causes
a collector current which
is given by the following

1C = BX10>t A

This current flows through

R3 and causes a voltage

= BXlO.nA * 1 KO
This voltage now appears
on one of the inputs of the
comparator. IC1 . Other
input of the comparater is
fed with the voltage given
by the voltage divider
potentiometer PI . The

thus giving a range c
4.7 V on the voltage
divider output. The
potentiometer
adjusted in

comparator are at same
voltage level.

When the inputs to the
comparator are equal, its
output is zero and the
zero voltage at one end o

through

14. Tl
r LEE

brightly.

The comparator IC1 is an
operational amplifier and
amplifies the voltage
difference between its two
inputs. The gain of this
operational amplifier is
several hundred thousand.
A small difference in the
two input voltages is
amplified to a high voltage.
Even a small difference in

microvolts will be
amplified to such a level

extinguished. The LED

aiy 1986 2-59

selex

practically glows only

voltages are equal. The
potentiometer knob can be
fitted with a dial and it can

directly.

Construction
Details :

The entire circuit of the
tester is so small that it
fits on just half of the
SELEX PCB. A list of
components and the
connection diagram is
given for simplifying the
task of construction. As
usual remember to solder
the passive components
first and then the
semicondutor components.
Jumper wires should be
insulated and properly
soldered. Single strand
jumper wires will be
preferable as they have no
risk of splicing up and
touching other component
leads or PCB tracks like
the stranded wires. The
assembled circuit can be
fitted inside a standard
case similar to the one
shown in figures 1 and 5.
LED is fitted using an LED
holder. A standard knob
with dial can be used for
the potentiometer. Three
wires with three different
colours are used for Base.
Collector and Emitter
connections to the TUT.
These wires should

crocodile clips at the other
end. The openings through

out from the case should
be marked B, C and E for
convenience in connecting
the TUT.

The operational amplifier
specified is LF 356. but
can be substituted with TL
081 or 741 in case of non
availability of other two
ICs. However, using 741
will hamper the
performance of the tester
and transistors with B
values below 120 will be
impossible to test. The
heavy duty power
transistors generally have
the B values in this range.
Finally a word about the
1C. Insert the 1C in its

j direction. The 1C pins may

inwards if the 1C is brand
new, before inserting it
into the socket.

Calibration:

Once you have your tester
working, the scale of
potentiometer PI can be
calibrated. Figures 6 and 7
shows the connections
required during calibration.
First the multimeter is
connected in place of the
TUT as shown in figure 6.
The actual current is
measured using the DC
micro-amperes range of
the multimeter, to see if it
deviates from the desired
10 micro-ampere value.
After this the multimeter
is connected with a 100
KT1 potentiometer (Linear),
as shown in figure 7. With
the test circuit different
collector currents between
0.25 mA and 4.5mA are
simulated in 0.25 mA and
0.5 mA steps. At every
step the position of PI for
balancing the voltage
across R3 is marked. The
markings will correspond
to B values between 25
and 450 if the base
current was really 10
microamperes. If the base
current measured during
the first test was different,
the simulated values of
collector current should be
suitabl/ changed.

For example, consider a
case where the base
current is 1 1
microamperes instead of
10.

The collector current
simulated for B value of
25 would be 0.275 mA
instead of 0.25 mA.

Alternative Circuit
for PIMP
Transistors.

suggested alternative
circuit for testing PNP
transistors. The
functioning of the tester is

described for NPN

Digital

Sequential

Alarm

Annunciator
Panel (Ador
Sequence-
guard Series)

Advani-Oerlikon solid
state Digital Sequential
Alarm Annunciator Panel
has been designed and
developed for the first
time in the country for
use in thermal power
plants and other heavy
engineering industries
wnere continuous
process systems are
required to be
monitored.

In the standard product,
it is not possible for the
operator to identify the
point that has given the
alarm first, and the
subsequent sequence.
Now, with the use of
Advani-Oerlikon Digital
Sequential Alarm
Annunciator Panel, the
operator can
numerically identify the
sequence of faults
indicated digitally with
the help of seven-
segments LED displays
mounted along with
every window. The first
nine faults are
annunciated in this
style.

The Panel is modular in
construction.

For further details please
quote ref. No.
P/26/L/SHG and write
Advani-Oerlikon Limited
Post Box No. 1546
Bombay 400 001.

Hand-held
Plastic
Moulded
Instrument
Case: T-28

"Comtech - T-28" is a
well designed plastic
moulded instrument
case basically suitable
for many Hand-Held ins-
trument having overall
dimension of 28mm W x
85 mm x 163.5mm L. The
front Bezel (20.8mm x
51.8mm) is suitable for
3!6 digit display. The ins-
trument case can
accommodate a pcb of
79mm x 74mm on the
four projected lugs
provided on the inside of
the front cover, by using
No. 4 self tapping screws.
The 0.5mm deep recess
on the front cover will
accept an aluminium
lable of 111 mm x 71.5
mm for various control
indications and a simitar
recess on the back cover |
(64mm x 46mm) can be
used for manufacturer's
details.

The moulded case
features a separate
battery compartment
suitable for 9V flat cell,
such that the cell can
be replaced without

opening the instrument.
The two antislip rubbers
at the back, preserve the
case from scratches. The
case offered in-grey and
black colour, is suitabe
for DMM. Temperature
Indicators & remote-
controllers etc.

For n

e information w

M/s. Component Technique
8. Orion Apartment
29 -A. Lallubhai Park Road
Andheri (West)

Bombay - 400 058

PLA Digital
Multimeter
Type : DM-
14B1

Pla introduced cheapest
indigenous DMM type
DM-14B1 in their
handheld DMMs family.

It measures AC/DC volts.
DCI and ohms with
superb overload

protection on all inputs.
Push button switches for
easy single hand
operation, plug-in LSI/
LSD/ICS. single 9V
battery tasting 1000 hrs.
These all features
available at rock bottom
prices. Very useful for day
to day work, rugged and
compact. Please enquire
for price, available ex-
stock.

For further details contact
PLA Electro Appliances
Private Limited
Thakor Estate.

Kurla Kirol Road.

Vidyavihar - (West),

Bombay - 400 086.

Desoldered

Without

Diamond

Diam ond

Desoldering

Pump

■ Desolders thoroughly.

■ Sturdy construction

■ Replaceable Teflon Nozzle

■ Largest selling in India

■ Widely accepted in (J.K., CI.S A,
West Germany and Singapore.
Export Enquiries Welcome

Manufactured by:

Industrial Electronic
& Allied Products

1423, Shukrawar Peth,

Off. Bajirao Road,

POONA 41 1 002, India.

Phone: 446241, Gram: SEFOTAKE.

Distributors:

Electronics Corporation

>r Road Gram Road (E). Bombay-400 007.

Chhotani Buili

DATA BOOKS

signotics ® gs** ™

niNnns^Dd m

sSh Texas Instruments

^ , SIEMENS

ziiog nciijntei ©SANYO

as ^ philips nm

SGS W PHILIPS [HU

"W TELEDYNE BEIIHXL0 UICIIIC

HswinxJpAcmD IPMfl

COOK BOOKS. OSBORNE BOOKS. SYBEX BOOKS.
TAB BOOKS- SAMS BOOKS- TOWER BOOKS -
ELEKTOR BOOKS- APPLE COMPUTER BOOKS AND
MICROPROCESSOR COMPUTER BOOKS etc

Almost Everything For
Sinclair ZX

SPECTRUM

Computers

o Guaranteed Repairs.

e Latest Software Rs. 50/- for two programmes
including cassette and manuaUBet our free list.

ELTEK BOOKS-N-KITS

6 RITCHIE STREET. MOUNT ROAD
MADRAS-600 002. S. INDIA
GR AM :"ELTEK"PHONE: 844405 TELEX 41-6734 TEX IN

Pioneer Electronics

2-62 el.

™ ' For further details write to:

THE MOTWANE MANUFAC-
TURING COMPANY PVT. LTD..
MOTWANE R 0 . a4 |^ a V 01
Tel.: 86297/86084 Telex: 752-247 MMPL IN
Grams: MOTWANE or Gyan Ghar, Plot 434 A,
14th Road. Khar, Bombay-400 052. Grams-
: MOTESTEM.

. ADSMMC

classified ads.

Advertisers

Index

CONDITIONS OF ACCEPTANCE OF CLASSIFIED ADVERTISEMENTS

1) Advertisements are accepted
subject to the conditions
appearing on our current rate
card and on the express
understanding that the
Advertiser warrants that the
advertisement does not
contravene any trade act
inforce in the country.

2) The Publishers reserve the
right to refuse or withdraw any
advertisement.

3) Although every care is taken,
the Publishers shall not be liable
for clerical or printer's errors or
their consequences.

4) The Advertiser's full name
and address must accompany
each advertisement submitted.
The prepaid rate for classified
advertisement is Rs. 2.00 per
word (minimum 24 words).

Semi Display panels of 3 cms
by 1 column. Rs. 150.00 per
panel. All cheques, money
orders, etc. to be made
payable to Elektor Electronics
Pvt. Ltd. Advertisements,
together with remittance, should
be sent to The Classified
Advertisement Manager For
outstation cheques
please add Rs. 2.50

AFCO I. & C. LTD

APEX

APLAB

APPOINTMENT

ARUN ELECTRONICS

ATRON ELECTRONICS

BAKUBHAI AMBALAL

COMPONENT TECHNIQUE . .

COSMIC

DEVICE ELECTRONICS

DYNALOG MICRO SYSTEM . .

ELETEK BOOKS N KITS

ELCIAR

GRAFICA DISPLAY

IEAP

KAYCEE

LEADER

LEONICS

LOGICS

MOTWANE

M R. ELECTRONICS

OMLILA'S

OSWAL ELECTRONIC CO.

PADMA

PECTRON

PIONEER

PLA

PROFESSIONAL ELECTRONIC
RAJASTHAN ELECTRONIC .

RAJKUMAR

R0CHER ELECTRONICS

SAINI ELECTRONICS

SIEMENS 2.0

SOLDRON

SUCHA ASSOCIATES

TESTICA

TEXONIC

UNLIMITED

VASAVI

VISHA

ZAC ELECTRONICS
ZODIAC

8085 Microprocessor Ir
case Rs. 2970/- Buil
PROGRAMMER power
CMOS RAM with dry-
expandable to 8K. 12K.
New Age Electronics th

Components

are normally available with the following companies:

INTEGRATED ELECTRONICS
INSTRUMENTS

8-2-174 Red Cross Road
Secunderabad 500 003 Phone: 72040

SIUKON ELECTRONICS

315. Lamington Road. Shop No. 4
Kalpana Mension, Opp. Police Station
Grant Road. Bombay - 400 007
Ph . 350644

VISHA ELECTRONICS

1 7. Kalpana Building,

349. Lamington Road

Bombay • 400 007 Phone: 362650

DYNALOG MICRO SYSTEMS

FOR THE HOBBYIST & ELECTRONIC BUF?

WE HAVE GOT IT ALL >. I..—

Also we have wide range of TRANSISTORS in 2SA, 2SB, 2SC, BC, BF & 2N series

TELEPHONE EXCHANGE

Complete kit as described in Elektor - India
Jan 1986 issue, kit for Rs. 1300.00 only.

(Kit contain the PCB and components that

Molody bell in semi assembeld module
(with speaker) for Rs. 60.00 only.

RESOURCING SUPERIOR QUALITY COMPONENTS
FROM TREASURY SUPPLIERS

Vi5HA

1 7, Kalpana Building. First Floor
349. Lamington Road
Bombay-400 007

R.N No. 39881/83

MH/BYW-228
LIC No 91

Watts News?

After 30 years of communicating our
finest efforts to you— we still have
more news for you.

Cosmic is now breaking every sound
barrier in maintaining its sophisticated
electronic image by touching perfection
in the manufacturing of its Tape decks/
recorders. Stereo Systems. Amplifiers.
Turntables. Head-Phones.

A single dominant factor has
encouraged us to keep expanding and
that is consumer satisfaction.

Stay tuned to us.

cosmic

We are Sound!