up-to-date electronics for lab and leisure

61

may 1980

U.K.55p.

U.S.A I Can. $1.75

.

contents

selektor

elektor p P's

An in depth look into the various microcomputer systems published
Elektor.

Junior Computer

At last a small, inexpensive microcomputer, for anyone interested in
learning more about computers.

coming soon

EDITORIAL STAFF

frequency lock system

One up on PLL; this system ensures better and

stable tuning than

TECHNICAL EDITORIAL STAFF

PWM amplifier

In spite of some teething problems. Pulse Width Modulation (PWM) is con-
sidered by many to be the future of audio. This month a 3 watt power
amplifier is featured.

Due to lack of space, the
MPG article has been postphoned
until the June issue.

BASIC cassette interface

Here is a cassette interface with on-board software. This will enable
SC/MP and BASIC computer users to store programs on cassette.

LCDisplays

Liquid Crystal Displays (LCD's) are an economic alternative to the well
known LED's, and LCD's consume much less power. Elektor takes a closer
look at LCD's this month and gives ideas for experimentation.

flexible intercom

There always seems to be a need for better communication, and the home
is no exception. The intercom featured here offers flexibility in location
and operation.

missing link

market

advertising index

Hot water from the freezer

The price of energy has risen consider-
ably during the past few years and it
certainly does not look as if this is ever
likely to drop. It is time to take steps to
conserve energy in a field which, until
recently, attracted attention on an
academic level only.

Currant consumption in tha freezer
Nowadays, about 47% of homes in
Britain own a freezer. Recently there
has been a drop in annual sales. This is
because freezers consume a great deal of
current. How does this compare with
other household appliances? Figure 1
shows the general energy requirements
in the home. 84% is used up by primary
energy sources such as natural gas or
coal and by secondary energy sources
such as electricity, coke or oil fuel. 10%
is required to heat water. The country's
450 million electrical household
appliances including TV, radio sets and
lighting only consume 6% between

So although household appliances use
up a relatively small percentage of the
total energy consumption, industry is
justified in its efforts to cut power
requirements. In 1978 25% of the total
went on domestic electricity. Figure 2
shows the three appliances which
consume most electricity in the kitchen.
17% alone is used up by cooling and
freezing equipment. Although their
compressors require no more than 100
to 150W, the huge amount of energy
consumed is explained by the fact that
they are almost continuously switched
on. The total current consumption may
be divided equally between the fridge
and the freezer. The consumption of the
freezer alone (6,5 GWh a year) corre-
sponds to that of agricultural machinery
or - to name another example — to
about 70% of the railway and traffic
requirements. It would, therefore, make
a considerable difference if a freezer's
energy could be 'frozen'. Even if only
part of it could be transferred in the
form of domestic heat, it would be a
step in the right direction from an
economic point of view.

Figure 3 shows the rising energy cun/e
of fridges and freezers. The latter are
usually more economical, because they
are better insulated.

Balance of energy and beet transfer

A freezer with a 300 litre capacity
- about the size recommended for a
family of four — uses up an average of
2.3 kWh a day. It is transformed into
heat across the compressor and cooling
circuit. The balance of energy is shown

5

Figure 5. A freezer including e he

6

QN = electrical energy Intake
QO = cooling measured
QC = theoretical condensation heat
QK “ loss of heat in compressor
Q\/ - useful condenser heat

Figure 6. A freezer including a heat transfer system.

in figure 4. The electricity consumption
Q|\j is used as a 100% rating. By means
of the evaporator Qo is drawn from the i
inside. The amount normally corre-
sponds to the heat which penetrates the
inside of the freezer through the door
seals etc. The theoretical condensation
heat Qc represents the entire heat
quantity as 183%. 90% of it is heat
emission Qy in the condenser, 93% is
heat emission Q« produced by the
compressor.

The balance of energy shows that it
would be quite simple to transfer part
of the heat emitted to boil water. For
this purpose the condenser is replaced
by a condenser spiral in a hot water
tank, so that heat Qy may be directly
used to heat water. An additional
heating system ensures that water is
heated to the required temperature or
to that needed during a short period or
consumption peak.

The cross-section in figure 5 shows how
such a system is built up. The freezer is
connected to the boiler by means of
tubes along which the refrigeration
substance circulates. In this way, 90%
of the electrical energy extracted from
the freezer may be used to heat water
until a temperature of about 60°C. With (
the aid of a simple technique it is I
possible to increase the level of heat by
100% or more. This because the system
operates as a water pump.

In this process the compressor plays an 1
important part. If it were insulated from
the outside air, an oil cooler could be ■
installed to reclaim 80% of the heat. ,
The performance flow diagram of such a
system is shown in figure 6. In the
diagram measuring values are recorded 1
which compressors normally achieve. The
condensation temperature Tc = 50°C
and the evaporation temperature '
To = 30°C correspond to the common
freezer values. The operation time of
the compressor is assumed to be 100%,
or continuous duty.

QN represents the electrical energy
increase with a 100% rating. The cooling
obtained Qc, 92%, is slightly higher in
comparison with the value shown in
figure 4. This is because the thermic
relationships have changed in the
cooling circuit. Of the condensation I
heat Qc, which theoretically should be
192%, about 20% is lost due to trans-
mission and radiation in the compressor.
For water heating purposes Qy = 172%
is available. The heat produced is there-
fore 1.72 times greater than the electri-
cal energy intake.

The excessive heat is drawn from the
outside air into the room the freezer is
in. As it extracts heat from its surround-
ings, it provides an ideal means to cool
cellars. An added advantage is that no
extra energy need be used. To avoid
thermic feedback between the freezer
and the boiler, these should not, of
course, be placed in the same room.

The condenser spiral, insulated com-
pressor with oil cooler including various

1980 - 5-03

control and safety installations will cost
about £ 1 25. If we suppose that approxi-
mately £ 40 a year may be saved once
the system is in operation, it will have
paid for itself within 2'h years. Com-
pared with other boiler systems, this
is extremely short. Even when water is
oil heated equivalent energy costs may
be saved.

Figure 7 shows the entire system as it
is available today. Twenty such systems
have been tested and are still functioning
well after more than eighteen months.
Figure 8 gives an idea of the machinery
involved inside a freezer. Clearly visible
are the two insulated tubes along which
the cooling medium transports the
condenser and oil cooler heat. In the
return tubes the fluid is at room tem-
perature and does not therefore require
insulating. The four tubes are the only
connections between the freezer and the
boiler. A boiler is shown with a 290 litre
capacity, as installed in a modern
home. (Recommended for a family of
four.)

Technical layout

Whereas the freezer needs to be cooled
continuously, how much hot water is
consumed depends on the time of day
and even on the day of the week
(figure 9). In the upper half of the
diagram the amount of hot water
consumed is shown with relation to the
time of day. Also recorded are the
hourly averages of consumption in a
family of four on an ordinary weekday.
There is a considerable difference in
quantity consumed at the weekend. In
a simplified manner, the total water
consumption is shown in the lower half
of the drawing. This is with relation to
the days of the week, where the con-
sumption and boiler efficiency ratio is
expressed in kilowatt hours, in other
words, as energy required by the boiler.
This graph also shows the energy which
a 300 litre freezer can produce for water
boiling purposes. With a power rating of
1.72, barely 4 kWh or about 46% of the
total energy required may be used to
boil water. The boiler capacity.

290 litres, is shown to be in proportion
to the cooling of the 300 litre freezer.
The additional heating apparatus is
exclusively used to cover the energy
consumption peaks.

Otto Koehn,

at the 15th AEG-Telefunken
technical press colloquium

(534 SI

9

Figure 9. Specific hot water requirements
for a family of four. Temperature obtained;
60°C.

5-04 - elektor i

elektor

microprocessors

No less than three microprocessor
systems have been published to
date by Elektor. To the beginner
this may seem rather confusing.

It is hoped that the following
description will serve as a guide
to anyone wishing to construct
an Elektor system.

The systems were published in the
following chronological order: the
SC/MP, the games computer and the
Junior Computer (JC). Although the
purpose of this article is to provide a
general survey rather than a detailed
discussion of the manifold possibilities
of microprocessors, practical examples
will be given by way of illustration.
(This does not imply, however, that no
other uses may be found for the sys-

First of all, let us deal with the SC/MP
(pronounced 'scamp') system. Its princi-
pal feature is its modular construction.
The microprocessor of the same name is
manufactured by National (type num-
ber INS8060). This involves a number
of printed circuit boards of the Euro-
card format (approximately lOx 16cm)
which are interconnected by means of
a bus system. The bus is nothing more
than a set of conductors connecting
all the 1 points, the 2 points, etc. Its
modular construction allows it to be
a highly flexible unit. Its smallest
version is made up of only two cards.
The system may be extended by adding
more cards to the bus printed circuit
board. These will not only provide

more memory capacity (additional
RAM and/or ROM), but a printer,
for instance, may also be installed.

The Junior Computer is constructed on
a single printed circuit board (excluding
the supply). An attempt has been made
to build the cheapest and smallest
unit possible, without eliminating any
of its 'real' microprocessor character-
istics. By means of a connector on the
board, the Junior Computer can be
coupled to the SC/MP system. The
result is a SC/MP system using an
additional processor.

The odd man out of the threesome
is the games computer. It was designed
to generate colour TV pictures directly
on the screen. The pictures are pro-
grammed to move and change in form
and colour. Thus, it is in fact a luxury
TV game device. Additional games
may be introduced (such as space
war, football and Master Mind). The
hardware (the computer itself) has
been specially adapted: It consists
of two individual keyboards of 12
keys and of a 4 key section to be used
by both players. In addition, there is an
input for two joysticks ('steering
levers') and a loudspeaker has been

1

5-05

built in for special sound effects.
Programmes may be easily changed
with the aid of a cassette recorder
which tapes them, so that they may be
played whenever required.

The games computer was not designed
for expansion since its prime purpose
was for programming. Any possible
future additions will only affect its
memory capacity.

Both the Junior Computer and the
SC/MP were designed for more general
use. Not only do they carry out specific
tasks, but games may also be played
(without a TV). Both machines are
capable of developing programmes
(already included in the standard
monitor programme) and of operating
in high-level computer languages. The
SC/MP, for instance, can use tiny
BASIC.

Every command to be carried out by
the SC/MP is then issued by a terminal.
This is a separate unit which has a
keyboard and VDU and/or printer.The
keyboard consists of figures 0 ... 9
as well as the alphabet and specific
control characters. In order to operate
this system effectively, therefore, a
terminal is essential because it enables
the computer to communicate in a high
level language. Since 'normal' words
are used, the alphanumeric keyboard
is necessary.

The Elekterminal was described in
Elektor's November and December
1978 issues. Instead of the Elekterminal
a hexadecimal keyboard and display
may be used. This is a separate module
described in the earlier series on the
SC/MP. The system is then fully oper-

ational at a machine language level,
and the Elekterminal can always be
added at a later date. Without exten-
sions, the JC is also programmed on
machine language.

The microprocessor

The first aspect to consider is which
microprocessor should be selected. The
microprocessor is at the heart of any
microcomputer system, and to a great
extent it determines its capabilities
and the speed at which tasks are carried
out. At first sight the best choice would
appear to be a microprocessor with
great potential and high speed. On the
other hand, it is very difficult to pro-
gramme hundreds of instructions. Ex-
perience has shown, that, ideally
speaking, the programmer should know
them all by heart. As far as speed is
concerned, it is of course an advantage
for the processor to be fast, without
needing a higher speed (and therefore
higher priced) memory. In practice,
however, programs tend to be held up
only when high level languages are used
or when complicated mathematical
calculations are made.

Another aspect which merits attention
is how many programmes are available.
Generally speaking, a processor may
take over other programmes, after
minor modification, provided these
were written for the same type of
processor. In this respect, the 6502 is a
good choice.

Each system hitherto discussed relies on
a different microprocessor: the games
computer on the 2650 from Signetics,

the SC/MP on the INS 8060 from
National and the Junior Computer on
the 6502 from Rockwell. Of these, the
SC/MP (8060) operates in the simplest
and slowest manner. The 6502, on the
other hand ■ is the most complex and
fastest. Between the two extremes lies
the 2650's performance. Since the
SC/MP's relatively slow operation is
sometimes considered a handicap, a
processor card has been manufactured
for it which is obtainable together with
a faster Z-80 processor (not by Elektor).
By way of conclusion, a brief descrip-
tion of the construction of each system
will be given and especially with regard
to their combination possibilities. For
further technical details, reference
should be made to articles on the sub-
ject published in Elektor.

The TV Games Computer

Number one on our list is the games
computer. It consists of a central
printed circuit board, including a
keyboard (see figure 1), a power supply
and, usually, a UHF modulator so that
a normal colour TV with an aerial input
may be used. In addition, it is advisable
to make use of a cassette recorder to
store the programmes. To facilitate
programming, the monitor is equipped
with extensive debugging capabilities
including two breakpoints. Two joy
sticks may also be added to control the
games.

The Junior Computer

For simplicity, the Junior Computer

ir Computer. The modules to the right m

is built on a single printed circuit board
(see figure 2). Of course, it also needs a
supply capable of producing +12, +5
and —5 V. The keyboard is constructed
directly onto the board. The six seven-
segment displays are mounted onto a
small auxiliary board which is soldered
in a slanting position onto the central
board. Later on, a cassette interface
and one or two cassette recorders may
also be added. Furthermore, a connec-
tion may be made to the SC/MP bus by
means of the connector which is on
the board. This can come in handy, for
instance, when more memory is re-
quired than is included in the circuit
(1 K Eprom with the monitor pro-
gramme and 1 K RAM).

The Junior Computer system operates
in a hexadecimal code: in other words,
using 0 ... 9 and then A, B, C, D, E, F.
The monitor program features a hex
assembler. When jump instructions are
encountered, the hex assembler provides
the correct byte for the corresponding
location. The monitor then passes on
the addresses to the computer.

Before more complicated tasks (using
high level languages, assemblers, etc.)
can be fulfilled, a terminal must be
connected. This can be done with the
aid of the cassette interface board. For
the necessary memory power, however,
more EPROM will have to be intro-
duced. This involves using the 8 K
EPROM + 8 K RAM card belonging to
the SC/MP.

For taping programmes, the cassette
interface board will have to be added.
Additional EPROM and/or RAM will
have to be included, whenever one
wishes to operate the available editor

assembler, disassembler or when pro-
gramming in high level languages (such
as BASIC).

The SC/MP system

Finally, it is time to consider the
SC/MP system. Because of its modular
construction, several configurations are
possible. The minimum (BASIC version)
system is based on two cards (see
figure 3). The first is the processor card
which includes an address and data bus
buffer and the possibility to connect a
terminal (RS 232 interface). For the
second the 8 K EPROM + 8 K RAM
may be chosen. If so, the monitor
programme must be part of the EPROM
and as much RAM as required (from 1 K
to 8 K) may be added. When a keyboard
is used with the 2 card system it is
able to run BASIC programs. The
processor card has an 1C (ROM) socket
specifically for this.

For programme storage purposes the
cassette interface may be added. With
the aid of the matrix printer card a
printed listing of machine language
programmes may be obtained. The
supply voltages required depend on
which EPROMs are used. The 8 K
EPROM + 8 K RAM card employs 2716
EPROMs and requires 5 V. On the
cassette interface card there is room for
EPROMs of the 5204 type and these
require +5 and —12 V. Instead of the
2716, the 2708 may also be used. As
a result, two more supply voltages (+12
and -5 V) will have to be added to the
8 K circuit and the memory capacity
will be halved (to 4 K). The existing
supply already produces the +5 V and

-12 V. On the SC/MP bus lines, all k
voltages mentioned are available. e

Apart from the modules mentioned T
above, there are still a few cards avail- 1
able which are based on a somewhat
smaller system which communicates
with the outside world by means of e
a keyboard, and eight seven-segment
displays. This model resembles the
Junior Computer in its elementary
form. With the aid of these cards, a
unit may be assembled as follows,
(figure 4).

The SC/MP processor card plus the
extension card constitute the actual
computer. The data lines are not
buffered, thereby restricting the size
of the system. In order to build up a
more complex system, a data bus buffer
must be added (printed circuit board
number 9972). The operator has 1V& K
of EPROM available (the monitor
programme) and 1 K of RAM. The
RAM may be expanded with the aid of
the 4 K RAM card. Data is written and
read by means of 26 keys and eight
seven-segment displays, installed in the
hex-l/O printed circuit board. An
auxiliary board will be needed for the
connection of a cassette recorder
(number 9905). A terminal, however,
may be fitted with another circuit
board (79101, interface for micro-
processor). The layout is shown in
figure 5.

Which one is for you?

These then are the Elektor computers
to date. Three complete systems design-
ed for different purposes. The SC/MP
for constructors with a desire for a

I large system which can be expanded
and/or modified as and when required.
I The TV Games Computer for those
■ Who want to see the results of their
t programming instantly in a visual form.
! The Junior Computer for the beginner,
? easy to construct and economical
t with great potential.

! » For the beginner . . .

' the obvious choice will be the
i Junior Computer with its excellent
teaching facilities.

* For the expert . . .

: the SC/MP will probably be the most

I desirable with its many system
t variation possibilities. Why not add

the Junior Computer to it (have a
two computer family)?

• For the programmer . . .

The TV Games is 100% FUN. De-
signed specifically for programming,
it succeeds in its purpose very well.

• For the constructor . . .

If building is your wont the SC/MP
will be fine. You can fill two kitchen
tables with it.

• For the experimenter ....

We are all one of these at heart really.
And if you fit all of the above
categories there can only be one
answer . . .

It is worth noting here that two new

computer books will be available from
Elektor, one for the SC/MP and one
for the Junior Computer.

What does the future hold?

There are two things you can be sure of
with Elektor, we always have something
up our electronic sleeve. How about a
PASCAL compiler for the Junior
Computer or a complete computer
cassette system? There are even rumours
of a new VDU system using an un-
converted TV set. Maybe even a new
computer . . . who can tell? Just watch
this space! M

5-08 - 1

jr may 1980

junior computi

There are many readers who would like
to know more about home computers
but who may not be technically minded
or who consider them too complicated
to understand. These two reasons,
coupled with cost, tend to prevent
many people from 'taking the plunge'.
To help overcome these problems we
have designed the Junior Computer (JC) .
Do not be misled by the term 'Junior' -
this computer provides the first step to
understanding large and powerful
systems. Although small in size, the
Junior Computer can be used with high
level languages (PASCAL for instance).
This is possible because it uses a
simplified method of operation and has
the advantage of various expansion
possibilities.

Junior computer

The cost and complexity of home computers is a serious deterrent to
the newcomer to computer operating and programming. We know of
many readers who would like to 'build their own' but who lack the
necassary technical knowledge. The Junior Computer has been designed
(for just this reason) as an attempt to 'open the door' to those readers
who need a push in the right direction.

It should be emphasized that, although simple to construct, the Junior
Computer is not a 'toy' but a fully workable computer system with the
capability for future expansion. It has been designed for use by
amateurs or experts, and software to be published will include a
PASCAL compiler - the computer language of the future.

The purpose of this article is to give a general description of the
operation and construction of the Junior Computer. It has been
decided to publish a more detailed description in book form.

The arrival of 'The Junior Computer' Books 1 and 2 on the market will
be announced shortly. This, however, is a preview intended to give the
reader an idea of what the computer entails.

The heart of the JC occupies no more
than a single printed circuit board which
should dispel any fears produced by
large and complicated systems. The
intention of this article is to encourage
readers to take the initial steps towards
constructing and operating their own
personal computer. Extensive and
precise details will not be dealt with
here but will be published in depth in
book form — the Junior Computer
Books 1 and 2. We can however whet
the appetite and set the ball rolling.
Specific data concerning the computer
are given in Table 1 , this is intended for
readers who are already familiar with
computers.

Block diagram

The fundamental features of the Junior
Computer are shown in the simplified
block diagram of figure 1 . The heart of
any computer system is the CPU, or
central processing unit. In this particular
case it is a 6502 microprocessor, a
40 pin chip that you can hold in the
palm of your hand - but shouldn'tl Its
purpose is to control communications
between the various units inside the
computer in accordance with the
instructions of the program. A clock
generator (oscillator) serves as a 'pace-
maker' for the processor.

A certain amount of memory is required
by the microprocessor to store programs
and data. In the JC it consists of two
sections. The first one for storing
permanent data and the monitor
program. The monitor program contains
a number of routines which perform
such chores as program loading.

elektc

lay 1980 - 5-09

debugging and general housekeeping.
The second section of memory is used
for storing temporary data and program
instructions.

The block marked I/O (input/output)
maintains contact between the computer
and the outside world including the
keyboard and display. In the circuit the
I/O appears as the PIA, or peripheral
interface adapter. It takes care of the
data transfer in two directions and can
(temporarily) store data. The operator
communicates with the computer via
the keyboard and display.

Computers are not as 'intelligent' as
some television programmes would have
us believe. In fact, they merely carry
out (programmed) instructions in a
certain (programmable) order. There are
three sets of parallel interconnections
(called buses — not the Midland Red
type!) which carry the various data and
control signals. First of all there is the
data bus to consider. It is made up of a
number of lines along which data travels
from block to block. The processor
must also be able to indicate the
memory location where data is to be
stored or removed. This is performed by
the second bus, the address bus. Last,
but by no means least, is the control bus

1

Figure 1. Block diagram of the Junior Computer.

which ensures that the CPU is able to
control the internal status, for instance
the nature and direction of data transfer
and the progress of successive program
sections.

This then very briefly covers the various
blocks, their functions and their
interconnections. We can now move on
to look at the circuit in greater detail.

Circuit diagram

The circuit diagram of the entire Junior
Computer (except for the power supply)
is shown in figure 2. Now that the block
diagram has been examined, each
section should be easily recognisable.
The 6502 microprocessor is IC1. Below
it is the clock generator formed by N1,
R1, D1, Cl and the 1 MHz crystal. The
system uses a two-phase clock, shown in
the circuit diagram as signals 01 and 02.
The memory is constituted by IC2, IC4,
IC5 and part of IC3. The monitor
program is stored in IC2, a 1024 byte
EPROM (Erasable Programmable Read-
Only-Memory). This is the basic
program in the computer (not to be
confused with BASIC — a high level
computer language). The RAMs
(Random Access Memory) IC4 and IC5
serve as user memory and together have
a capacity of 1024 bytes.

In the PIA, IC3, there are another
128 bytes of RAM. The PIA constitutes
a data buffer which controls all the data
transfer passing in either direction
between the computer and ports A and
Photo 1. The completed Junior Computer looks like this. The keyboard and display can be B - The port lines are fed OUt to a 31 pole

clearly seen, the microprocessor and other components being on the other side of the printed connector . IC3 also contains a pro-
circuit board. grammable interval timer.

The displays (Dpi . . . Dp6) and keys

elektor may 1980 - 5-1

Figure 3. The power supply which produces the three voltage levels required by the Ji

I (SI . . . S23) are at the bottom of the
circuit diagram. Of these keys, sixteen
are for the purpose of entering data and
addresses in hexadecimal form and the
remaining seven have various control
functions. Data to the displays and from
the keyboard is transferred across seven
lines from port A. The information on
the displays is controlled by the software
in the monitor program, which also
ensures that key function signals are
recognized. IC7 multiplexes the displays
and periodically checks the state of the
rows of keys to see which one, if any, is
being depressed. With the aid of switch
S24 the display may be switched off.

The display may be used in two different
I ways. Usually, the four left hand displays
, will indicate an address and the two
right hand ones will indicate the data in
the address location concerned. As a
second possibility, the two left hand
displays can show the (hexadecimal)

; code of an instruction while the others
display the address of the data corre-
J sponding to this instruction. This makes
program entry much easier.

The address decoder, IC6, provides chip
select signals for each of the various
I memory blocks. These appear as K7, K6
and K0 for the EPROM, PIA and the
RAMs respectively. The other five
| selection signals are available externally
for memory expansion. The RAMs also
require a R/W (read/write) signal. This is
made available via gate N6 and is
generated by a combination of the R/W
signal in the 6502 and the 02 clock
pulse (02 = data bus enable). Another
I control signal is the reset signal RES,
which places the microprocessor and the
PIA in the correct initial condition for
the monitor program. A reset is generated
when key RST (SI) is pressed and half
of a 556 timer (IC8) is used to suppress
any contact bounce this key might
| produce.

j There are two ways in which a program
being run can be interrupted by means
of the NMI (non-maskable interrupt).
The first one is provided by the STOP
key S2 (which uses the other half of IC8
! for contact bounce suppression) and the
second is provided by the STEP switch
S24 when this is in the 'ON' position.
When the output of N5 then changes
from high to low, the IRQ (interrupt
request) connection causes the program
being run to be interrupted, for instance
j by programming the interval timer in
IC3. Also present on the control bus are
the two clock signals 01 and 02 which
control the PIA and the RAM R/W
signals. These determine the direction of
data transfer. Finally, lines RDY, SO
and EX provide possibilities for future
expansion.

All the address, data and control signals
are fed to a 64 pole expansion connector
which, as its name suggests, is meant for
the purpose of expanding the system
further at a later stage. Figure 3 shows
the power supply for the Junior
Computer. This produces three voltages:
+5 V for all the ICs and the displays.

and +12 V and -5 V for the EPROM
(IC2). Capacitors C5 . . . C14 ensure the
necessary decoupling.

A few remarks

Before work is begun on the construc-
tion of the Junior Computer, two more
aspects have yet to be considered. The
entire system is built up on three
printed circuit boards of which one is
double sided with plated through holes.
It is advisable to check all the through
connections with an ohmmeter to make
sure that both sides of the circuit are
well connected. This will avoid problems
later, for after soldering it is very
difficult to trace any breaks.

Normally, of course, the 2709 EPROM
will not have been programmed when it
is bought. The monitor program (or
'hex dump') is given, so that the reader
who has a PROM programmer at his
disposal may program the 1C himself.
Alternatively, pre-programmed 2708s
can be purchased from the retailers
listed at the end of this article.

How to build the Junior Computer

Construction of the Junior Computer is
not difficult by any standards. If it is
assembled carefully (paying particular
attention to solder connections) and the
instructions are followed to the letter,
very little can go wrong. The three
sections of the JC are each constructed
on a separate printed circuit board: the
main board (including keyboard) the

display board and the power supply.

The smallest of the printed circuit
boards is the display board (figure 6) .
This is connected to the main board by
means of thirteen wire links. The seven-
segment displays can be soldered
directly onto the printed circuit board.
The main board is double sided and is
shown in figures 4 and 5. With the aid
of the component overlay it is possible
to see on which side to mount the
various components. First resistors
R1 . . . R20 and diode D1 are mounted,
then capacitors Cl . . . C13, followed by
all the 1C sockets. It is advisable to use
1C sockets especially for IC1 . . . IC3. Be
sure to use a top quality type with gold
contacts.

The other side of the board can now be
assembled. Switches SI . . . S23
(Digitast) and LED D2 (remember the
LEDs polarity) can now be mounted.
Two holes remain free next to the
keyboard for switches S24 and S25.
These switches are connected to the
main printed circuit board using short
lengths of insulated wire. A single wire
link is placed on the main board to
connect the 'D' input of IC6 to the zero
volt rail. The other connection indicated
between D and EX is meant for future
expansion. The 31 pole connector is
mounted on the keyboard side,
followed by the 64 pole connector
which is positioned on the other side of
the board.

The display printed circuit board can
now be connected to the main board.
The distance between the two boards

if mimWiT ^ rtirijj j^

should be about 5 mm. All that remains
to complete the computer board is to
solder the 1 MHz crystal in place, and
finally, fit IC1 . . . IC3 (the expensive
ones) into their sockets. The main board
is now complete.

The power supply has been left until
last. The simple construction should not
give anyone any headaches. All
components are mounted according to
figure 7, not forgetting the mica insu-
lating plate (with a smear of heat-sink
compound) under IC2. Connections
between the power supply and the
computer can be made using a four wire
cable to the 64 pole connector as
follows:

+12 V to pin 17c

+5 V to pins la, 1c

—5 V to pin 18a
0 V to pins 4a, 4c

It would be wise to make absolutely
sure that these connections are correct.
An error here can be very costly.

This completes the construction of the
Junior Computer and now we approach
the moment of truth.

Switch on

Just before you do that, one more
check -over would not be a waste of time.
Are all the chips the right way round?
Are there any cut offs of wire lying on
the boards? A final thorough inspection
could save you money. Now switch
on . . . and of course nothing happens,
the display remains unlit. There is no
reason for alarm yet, everything is
exactly as it should be. Now press the
RST key and random hexadecimal
characters appear on the display. This is
quite in order and as good a proof as
any at this time that your JC is func-
tioning correctly. It can now be fitted
into the case of your choice.

Something wrong, after all? I

Unfortunately, (due to Murphy's Law I

no doubt) there is a possibility that T9 ' "

pressing the RST key will depress the 1

operator rather than cause anything to

appear on the displays. This will of

course occur with an unprogrammed

2708 (IC2). A survey of the most

common errors and how to deal with

them are given below. Figure 4. Component

First verify that the supply voltages at board (EPS 80089-1 )

the 64 pole connector are as follows:

5-16 - elektor

nay 1980

pressed. If this is not the case, the error
will be in:

- the timer IC8

- the pull-up resistor R2

- the RST key SI.

With the supply switched off, the
resistance between pin 12 of IC6 and

0 V (connector pin 4a) can be measured.
If there is no 'short' between these two
points the wire link will have been
placed in the wrong position on the
main board.

The last check to carry out involves the
clock generator and for this an oscillo-
scope will be required. The CPU
produces two clock signals which are
fed to the expansion connector: 01 on
pin 30a and 02 on pin 27a. With the aid
of the 'scope it can be seen whether a

1 MHz square wave is present at both
points (minimum RMS value 3 V). In
the event of the oscillator not operating
or showing a defect, this will probably
be due to capacitor Cl, diode 01 or IC9.
Of course, other faults are possible, but
the above checks should clear most
problems.

For readers who have facilities for pro-
gramming their own EPROM (IC2) the
monitor dump is given here (figure 8).
There are 64 rows of 16 bytes each,
a total of 1024 bytes. The first column
gives the hexadecimal address for the
byte in col 0.

Your Junior Computer is now rearing to
go and it is possible to begin your
programming lessons. Each section of
the Junior Computer Book is clearly
illustrated with examples that can be
put into practice on your very own
computer. As mentioned earlier, there
are plans afoot for the publication of a
number of programs and a PASCAL
compiler for the JC. Look out for
further details. M

1C00
1C10
1C20
1C30
1C40
1C50
1C60
1C70
1C80
1C90
1CA0
1CB0
1CC0
1CD0
1CE0
1CF0
1D00
ID 1 0
1D20
1D30
1D40
1D50
1D60
1D70
1D80
1D90
1DA0
1DB0
1DC0
1DD0
1DE0
1DF0
1E00
1E10
1E20
1E30
1E40
IE 50
1E60
1E70
1E80
1E90
1EA0
1EB0
1EC0
1ED0
1EE0
1EF0
1F00
1F10
1F20
1F30
1F40
1F50
1F60
1F70
1F80
1F90
1FA0
1FB0
1FC0
1FD0
1FE0
1-FF0

85 F3 68
F4 86 F5
83 1A A9
F2 D8 78
F0 F6 20
FA 48 A5
A9 03 85
C9 12 D0
0B A5 EF
85 El A 4
4C 7A 1C

85 FA 4C

86 E8 84
2A 20 6F
D3 IE A0
D9 20 5C
20 IE 10
10 BB 20
A9 C9 12
20 83 IE
F9 A9 03
00 B1 E6
FB 20 8E
5C ID C9
10 10 04
8D 81 1A
FA 20 CC
1A A0 03
D0 F5 A0
4A 4A 4A
B9 0F IF
1A A0 06
D0 07 E0
FA 8A 29
20 6F ID
F0 12 20
10 04 85
00 B5 F9
A0 01 C9
C9 20 F0

84 F6 60
00 91 EA
EB C5 E9
Bl EA A4
F0 10 38
AE IE 60
F6 85 E8
E8 A5 E9
E7 69 00
79 24 30
02 02 01
7A 1A 6C
0A 88 Bl
60 38 A5

85 EE 20
C8 Bl E6
EC 88 84
30 D3 20
C9 20 F0
03 85 F6
91 E6 D0
91 E6 4C
ID 10 F2
FB 85 F9

85 FI
BA 86

04 85
20 88
F9 ID
FI 48
FF D0

09 E6
85 FA
FF D0
A2 04
7A 1C
E9 A9
ID 10
00 Bl
IE 20
C9 20
5C IE
D0 07
20 EA
85 F6
95 F9
ID F0

10 10

05 FE
A2 08
ID 88

A2 I
06 8C
20 DF
8D 80
8C 82
27 D0
0F 4A
10 21
6F ID
F9 A2
91 E6
00 F0
0C 29
A5 E6
E6 EA
D0 E6
F6 91
A5 EA
A5 E2
A5 E9
E9 00
85 E7
19 12
02 02
7E 1A
EC AA
E4 E9
D3 IE
A 4 EE
EE 20
D3 IE
12 29
4C 33
E6 C8
AA IF
85 FB
C6 F9

68 85 EF
F2 A2 01
Fl A9 03
ID D0 FB
C9 13 D0
A6 F5 A 4
14 C9 11
FA D0 02
A5 F0 85
0D Bl FA
06 FA 26
20 D3 IE
77 A0 00
F7 85 FB
E6 C5 FB
F8 IE 30
47 IE F0
20 F8 IE
20 F8 IE
IE 4C CA
20 8E ID
C8 CA 10
FB 20 8E
11 0A 0A
A2 FF 60
A 4 F6 A5
F0 05 A5
A9 FF 8E

82 1A 09
ID 68 29
1A 8E 82
1A E8 E8
F5 A9 15
AA 98 10
85 FB 20
10 0F 85
FF 60 20
CA C8 C4
1A C9 40
IF C9 19
85 EA A5
D0 02 E6
60 A5 E8
EA A5 EA
E9 01 85
85 E6 A5
69 00 85
85 E9 60
38 A5 E6
02 78 00
02 01 01
Bl E6 A0
88 Bl EC
FF 85 EC
20 5C IE
91 EC 88

83 IE 20
20 5C IE
IF C9 10
1C C8 20
20 35 IF
D8 A9 00
20 6F ID
4C DE IF

85 FA 68

86 FF 4C
85 FF 85
20 88 ID
13 A6 F2
F4 A5 F3
D0 06 A9
E6 FB 4C
FB 4C 7A
0A 0A 0A
FB CA D0
A4 E3 A6
91 E6 20
20 6F ID
D0 07 C8
E9 10 3E
Cl C9 13
A5 FD 85
30 A0 10
1C A9 EE
D0 FB 4C
F8 20 5C
ID F0 F6
0A 0A 85
A0 00 Bl
FB 20 CC
F9 20 CC
82 1A E8
80 49 FF
0F 20 DF
1A A0 7F
60 A2 21
60 A0 FF
03 18 69
60 IE 84
FA C6 F7
A6 IE 20
F6 D0 F6
F0 16 C9
F0 06 29
E7 85 EB
EB A5 EA
85 EA A5
C5 E6 D0
EA A5 EB
E3 85 E7
E9 60 38
18 A5 E6
E5 E8 A5
10 08 03
02 01 01
FF C4 EE
A0 01 60
A5 E5 E9
A0 00 Bl
A5 E7 91
EA IE 4C
A0 00 Bl
F0 1A 20
35 IF F0
F0 E0 38
85 FB 85
10 EB 85
FF FF 2F

BCD

85 F0 85
33 1C A9
F6 A2 FF
F0 FB 20
9A A5 FB
40 C9 10
00 85 FF
33 1C C9
1C C9 15
0A 05 El
F9 A5 FA
E2 E8 D0
4D ID C9
10 F0 85
Bl E6 C5
C9 10 D0
D0 14 20
F6 20 47
0D C9 11
85 FB 85
CA 1C A2
IE 20 8E
20 F9 ID
FE 20 5C
FA 85 F9
ID 88 F0
ID A9 00
E8 2D 80
60 48 84
ID A 4 FC
88 10 FD
A0 01 20
0A B0 03
07 CA D0
F7 84 FD
F0 07 20
DC IE A2
60 A0 00
60 F0 12
0F AA BC
A4 F6 Bl
C5 E8 D0
E9 85 EB
06 A5 EB
E9 00 85
60 18 A5
A5 E8 E5
65 F6 85
E7 E5 E9
46 21 06
03 03 03
F0 0D D1
88 88 88
00 85 ED
E6 C9 FF
EC 88 A5
65 IF 20
E6 C9 4C
F8 IE 30
EE 91 E6
E5 E6 38
FA 85 F9
FA 18 A5
IF ID 1C

FB 84
IE 8D
9A 86
88 ID
48 A5
D0 06
F0 0A
14 D0
10 EA
91 FA
05 El

01 C8
14 D0
FA 20
FA F0
0A 20
20 IE
IE F0
D0 09
FA 85

02 A0
ID D0
60 20
ID C9
A9 7F
0D A5
8D 81
1A 88
FC 4A
60 A8
8C 80
B5 ID
C8 10
FA 60
C6 F7
6F ID

02 A0
Bl E6
A0 03
IF IF
EA A0
EC A5
A0 00
C5 E7
EB 4C
E8 65
F6 85
E6 A5
60 40
0E 02

03 6C
EC D0
D0 E9
A9 FF
D0 ID
E6 91
F8 IE
F0 16
E6 A9
8A C8
E9 02
20 6F
FA E5
32 IF

coming

soon

Pest Pester

Perpetually plagued by mos-
quitos? This summer, protect your
person with the 'Pest Pester'.

•X-

Disco Lights

Brighten up your parties to the
beat of the disco greats. Not
Travolta, but the next best thing
— Disco Lights.

•X-

8K RAM + 8K EPROM
SC/MP and Junior Computer
owners who are short on memory
may extend their systems with an
8K RAM + 8K EPROM board.

Figure 8. The JC's hex

•X*

Summer Circuits 80
As usual the July/August issue
will have more than 100 circuits
to keep you busy over the sum-
mer.

•X-

5-1 8 — elektor

one up on PLL

Even though the superhetrodyne re-
ceiver is an invention dating from the
earliest days of radio, the principle
behind it nevertheless remains valid.
Almost every radio operates on an
oscillator which is mixed with the
aerial signal to produce an IF signal.
Early tuners had terrible drifting prob-
lems. In modern units however, the
oscillator gets a little electronic help.
This is necessary because it is by no
means easy to continue oscillating for
long periods without the penalty of
some drift.

This help may come in the form of a
circuit which subjects the differential
output of the mixer to an electronic
check. It must continually ask itself:
is the IF frequency right? If a deviation

frequency dividers are engaged in the
operation.

One important drawback of the fre-
quency synthesiser is the oscillator's
inability to generate any given fre-
quency (within its range); it can only
produce those which are equal to a
fixed frequency, multiplied by a whole
number. Thus, a frequency synthesiser
cannot be tuned continuously — its out-
put jumps step by step. That is the price
that has to be paid for stability! A price
we can well afford, as long as we make
sure the steps are small. Furthermore,
the signals to be received are not just
anywhere on the frequency band: for
broadcast transmitters, the band is also
divided into fixed steps.

The constant comparison made between

frequency lock system

One of the basic requirements of
any radio receiver is a stable
front-end (or tuning heart), with
the ability to tune a station in,
and hold it without drifting.
Automatic Frequency Control
(AFC) was one of the first
methods found to solve this
problem, but it ran into trouble
when trying to tune in a weak
station, adjacent to a strong one.
Then PLL came along. The PLL
(Phase Locked Loop), while
solving the AFC pitfall, was
tricky to design and complicated.
There is now a system that
promises to be the future standard
of tuners.

seems imminent, the oscillator is
instructed (by means of a control volt-
age) to alter its frequency so that a
correct IF is regained. This is basically
how AFC works. With the aid of AFC,
a highly stable oscillator frequency may
be achieved; that is to say, provided that
the receiver has an input signal.

However, it is equally possible to regu-
late the phase rather than the frequency
of the oscillator signal. This is what a
PLL system does. The oscillator is tuned
for a certain transmission frequency.
A control loop ensures that the oscil-
lator signal remains in a strict phase
ratio to the HF signal. The control sig-
nal for the oscillator, which is obtained
from the phase ratio of the oscillator
and HF signal, therefore varies according
to the modulation. In other words,
the control signal also produces the
desired LF signal.

Just like other types of feedback (that's
what it isl) a phase locked loop often
suffers from instability. Feedback ampli-
fiers may oscillate under unfavourable
circumstances; similarly, a PLL control
loop may suffer from PLL 'jitter'. This
occurs when the oscillator frequency
fluctuates so rapidly that, although its
average remains the exact value of the
HF signal, the result produced is unfit
for use. This is the reason why a PLL
system may turn out to be quite dis-
appointing in practice. Things are
different, fortunately, when a frequency
synthesiser is used.

A crystal (clear) comparison

A frequency synthesiser is quite a com-
plex circuit which continually compares
the oscillator signal with that of a highly
stable crystal oscillator. The synthesiser
checks whether the oscillator frequency
remains as constant as the crystal
frequency. For this purpose, all sorts of

the oscillator frequency and a stable
crystal frequency also takes place in the
frequency lock system described in this
article. The block diagram in figure 1
demonstrates the basic principle.

The input signal of the frequency lock
is generated by the oscillator in the
receiver's tuning section (HF): f osc .

The frequency lock system provides the
oscillator with an output signal in the
form of a control voltage U c . The cir-
cuit controls U c in such a way that f osc
is exactly equal to one of a series of
'scanning frequencies': frequencies
which are separated by a fixed distance.

How it works

The heart of the frequency lock circuit
is the D flipflop FF. This operates as a
harmonic mixer with two inputs: the
D input and the clock input of the flip-
flop. The input signals are symmetrical
square waves, at frequencies f OS c and
f c | respectively. A symmetrical square
wave also appears at the Q output of the
flipflop with a frequency f q . The latter
frequency is, of course, dependent on
the two input frequencies, according to
the following formulae:

fq = I fosc — c - f c l I and f q < ’/a f c |
The two vertical lines refer to the absol-
ute value. In other words, the number
not preceded by a plus or a minus sign
(otherwise f q would turn into a negative
frequency, which would be ludicrous).
The figure c in the formula is a positive
whole number. Let us suppose f osc
equals 2005 kHz and f c | equals 20 kHz.
From the condition fq<’/af c | it then
follows that c must be equal to 100 and
fq will be 5 kHz. If f osc is equal to
2010 kHz, c has two possible values,
namely c=100 and c=101 (f q will
then be 10 kHz). Therefore, a raster of
IS fcl (the clock frequency) is created.

- 5-19

frequency lock system

Variable c is also called the harmonic
number.

If the input signals of the harmonic
mixer, f osc and f c |, are kept constant,
the formula gives the value of the out-
put frequency fq. However, there is
another way of getting the same result:
by keeping the output frequency
together with one of the input fre-
quencies at a constant level. Let us
suppose that fq is equal to 250 Hz and
f c | is equal to 1000 Hz. Which values
may f osc assume without invalidating
the formula? Let us start by substituting
a 1 for c. Then:

250 = | f 0 sc -1 • 10001.

That is correct where f osc = 1250Hz
and (remember we're talking about an
absolute value) where fpsc = 750 Hz.
Now let us replace c with a 2. The
formula then reads:

250= | f 0 sc- 2- 1000 |,
and that is correct where f 0 sc = 2250 Hz
and f 0 sc = 1700 Hz. Any whole number
('integer') may replace c, with the result
that f 0 sc may assume the following
values: 750 Hz, 1250 Hz, 1750 Hz,
2250 Hz, 2750 Hz, 3250 Hz . . . etc.
Exactly the series of raster frequencies
required!

The question is now: how can we main-
tain a constant single input (f c |) and
output frequency (fq)? The input fre-
quency should not give any problems,
for it may be derived from a stable
crystal oscillator. In the block diagram
f c l is shown to originate by dividing the

Figure 1. A block diagram of the frequency regulator. An important part is played by the
D flipflop FF, which operates here as a harmonic mixer. The frequency regulator continually
compares the oscillator frequency with a highly stable crystal frequency.

frequency of a crystal oscillator by n.
Keeping the output frequency fq level is
no easy task, for it is impossible to
influence it directly. The only other
frequency which can be directly affected
is f 0 sc- That is vvhy an automatic con-
trol system is used. F 0 sc is controlled
in such a way that fq remains constant.
For this purpose fq is continuously
compared to a stable reference fre-
quency f r . Both signals are fed to simple
pulse generators, one of which produces
a positive pulse at every period, the
other a negative pulse. The output
signals of the two pulse generators are

added together and summed by an
op-amp 1C. The output signal of the
mixer/buffer produces the control volt-
age U c which regulates the frequency
fosc-

If fq is equal to f r , the average output
voltage of the counter (IC2) will be nil
and so will the control voltage U c - At
the input of the mixer there will be as
many positive as negative pulses. If fq is
too high for one reason or another,
more negative than positive pulses will
reach the input of the mixer. After
some time the output of the mixer will
also become negative. This will cause U c

to rise and, therefore, f osc to drop until
fq is again equal to f r .

Quite a complicated business, but the
result is worth it; with crystalline
precision f osc can be made to equal one
of the frequencies in the raster fre-
quency series.

In practice

Now that the block diagram has been
dealt with in detail, few words need be
said on the practical layout given in
figure 2. In any case, the block diagram
can easily be recognized in it. One
highly advantageous aspect comes to
light immediately: in spite of the cir-
cuit's complicated operation, it is very
reasonable in price. Such a remarkable
piece of electronic ingenuity requires
only a few IC's and one or two com-
ponents. It is much simpler than the
PLL system.

IC1 contains a crystal oscillator and a
fourteen bit binary counter. For the
crystal, a 4.43 MHz type has been
chosen, like the one used in colour TVs.
It is inexpensive and easily obtainable.
It can be replaced by a crystal of a
different frequency without any diffi-
culty, provided that it is between 1 and
6 MHz. Only the difference in individual
distance between raster frequencies will
show that a crystal of another fre-
quency is being used.

The signal at pin 3 if IC1 has a fre-
quency of approximately 270 Hz.
This is the input signal of a second
counter, IC2. There the frequency is

divided once more by four (output Q2
on pin 11), so that a signal of some
70 Hz is generated. This is indicated
as signal f c | in the block diagram. Thus,
the raster frequencies come to be
separated by approximately 70 Hz.
A signal is also derived from the Q3
output of IC2. Its frequency is divided
by two by means of FF1 so that a
frequency regulation of approximately
17 Hz is achieved.

FF2 is the harmonic mixer. The two
pulse generators are each simply con-
structed from a diode and a resistor
(D1/R2 and D2/R3 respectively). To
avoid having to use a split supply
they both operate at one half of the
supply voltage. Op-amp IC4 is wired as
a mixer and operates as the one given
in the diagram together with the
inverting buffer amplifier. Its output
voltage is the control voltage U c - The
'reset' push button S has been provided
to interrupt the entire control process.
This is important when tuning to
another frequency.

It is absolutely essential for IC4 to be
the type indicated, because it has a
high impedance FET input.

Indicator

The control voltage U c is not only used
to regulate the frequency of the oscil-
lator in the HF section, it is also desir-
able to make its value apparent in one
way or another. The ring circuit of the
frequency regulator is only operative
when U c does not deviate too much

from its zero level and only then is the
receiver optimally tuned. For this
reason U c is made visible with the aid of
the indicator circuit of figure 3. This is
called a window comparator. When U c
is 'in range', the green 'lock' LED (D6)
lights. The receiver is then tuned. If one
of the two red 'out of range' LEDs
lights, it is advisable to alter the re-
ceiver's tuning to prevent the frequency
regulator from becoming inactive (when
D4 lights it moves to a lower frequency,
when D7 lights, it moves to a higher
one).

Connecting it to the HF oscillator

There are many different receivers and
as many different high frequency oscil-
lators. There are, therefore, various
ways to connect the frequency regu-
lator.

The classic model operates with the aid
of a resonance circuit consisting of a
coil and a variable capacitor in parallel.
The oscillator signal is across the reson-
ance circuit. The oscillator frequency
may be regulated by the control voltage
if a varicap diode is wired parallel to the
tuning capacitor (see figure 4).

The oscillator signal across the reson-
ance circuit is derived by means of an
amplifier stage with a dual gate MOS-
FET. Due to the high input impedance
of the MOSFET, the resonance circuit
is lightly loaded. The amplifier stage has
two outputs: one for the f 0 sc signal
sent to the frequency regulator and one
with the same signal to feed the input of

greater simple sound effects

1980-5-21

greater simple
sound effects

a frequency counter, which may be
included. The latter output is of no
importance to the frequency regulator,
but is included, as an option.

D9 is the varicap diode. It is connected
to the tuning capacitor in series with
C9. Otherwise, the slightest variation in
U c would lead to an enormous change
in the oscillator frequency and result in
instability.

If the tuning loops of the receiver are
already provided with varicaps, another
one will, of course, not be necessary.
The control voltage U c may then be
added to the tuning voltage at the
varicap of the oscillator circuit. D9. C9
and R22 (figure 4) can then be omitted.

> If, as more often than not is the case,
the receiver is provided with a 'counter'
output (at which the oscillator fre-
quency is available), figure 4 may be left
out altogether. The input of the fre-
quency regulator is then connected to
the 'counter' output.

Figure 4's circuit must be attached as
closely as possible to the oscillator. It is,
as it were, part of the oscillator. The
input of the amplifier stage is highly
sensitive and prone to interference.

Experimenting

A circuit like the frequency lock is ideal
for hobbyists who enjoy experimenting.
The layout offering the best results
differs from case to case. It depends on
the drift features of the receiver oscil-
lator and in some cases it may be worth-
while to change the original layout a
little.

For instance, the clock frequencies of
FF1 and FF2 may be chosen at a higher
or lower level by deriving them from
IC2 outputs other than the Q2 and Q3
now used. A prerequisite is, however,
that the two IC2 outputs used be next
to each other (the clock frequency of
FF1, therefore, is half that of FF2).
Choosing a higher or lower level clock
frequency will affect the distance be-
tween the raster frequencies (they will
be either closer to or further away from
each other) and also the speed at which
the frequency regulator operates.

If the clock frequencies of the flipflops
change, capacitors C3, C4 and C5 will
also have to change. At twice the fre-
quencies they will be half their value.
In stead of enlarging C5 (for which an
electrolytic capacitor may not be used),
R4 and R5 may have a higher value.

The frequency regulator may also use
an external reference frequency. This
could, for instance, be derived from the
time base of a counter. Such a fre-
quency (for instance 1 MHz) may be
fed by means of a capacitor of 39 pF
to pin 1 1 of I Cl. In that case, R1, Cl,
C2 and the crystal are no longer re-
quired. If a low frequency reference
AC is available (of say, 100 or 250 Hz),
IC1 may even be omitted entirely. The
reference frequency can then be directly
connected to pin 1 of IC2. Any external
reference frequency must, of course,
be (crystal) stabilised. H

H. Thienel

In the simple sound effects generator
published in Elektor May 1979, only
eight of the twelve outputs of IC2 are
used. This modification, using the last
four outputs, adds a further dimension.
Only five resistors and one transistor are
required as the circuit diagram shows.
Transistor T2 has a similar function to

T1 in the original circuit. The base drive
forT2 is provided by outputs Q9 . . . Q12
of IC2 via resistors RIO . . . R13. This
addition to the circuit causes the
fundamental frequency to fluctuate
continuously. The result is an even
better(?) noise. Try it and see. H

W T1»

6-22 - elektor

1980

PWM amplifier

The subject of digital audio has been
dealt with in Elektor before and cer-
tainly will be again. The benefits are
impressive, so much so that virtually
all major manufacturers of audio
equipment are investigating its possi-
bilities. Recording companies too are
aware of the potential of a digital
system (digitally recorded records are
already available commercially).

Until quite recently, the performance
of PWM amplifiers was disappointing
due to the poor quality semiconductors
used. With the introduction of modern
high speed switching transistors, PWM
is now coming of age.

fed to an op-amp I Cl . This is used as
a comparator and is followed by a I
number of schmitt triggers in parallel.
This has two purposes. Firstly the
waveform needs to be 'square' and
secondly sufficient base drive current
is needed for the output stage which
uses two ordinary but fairly fast tran-
sistors (BD 137/138).

The entire amplifier oscillates and pro-
duces a square wave. This is because one
of the inputs of the comparator (IC1) is I
connected to the output by means of
an RC network. Both inputs of IC1 are
biased to one half of the supply voltage
using voltage divider R3/R4. Whenever

FWM amplifier

In spite of some initial teething
troubles. Pulse Width Modulation
(PWM) is considered by many to
be the next step in audio circuit
design. The PWM amplifier de-
scribed in the Elektor September
1979 issue has been used as
a model for the following article.
Although it has only a modest
3 watt output, it is a practical
and efficient amplifier.

E. Postma

The PWM amplifier

In Elektor's December 1978 and
September 1979 issues, a fair amount
was said about PWM amplifiers. How-
ever, it might be a good idea to recap
the principles briefly. A PWM amplifier
contains a symmetrical square wave
generator. The duty cycle of this
square wave is then modulated by the
audio signal. The output transistors
do not operate linearly but function
as switches, that is, they are either
full on or off. Under quiescent con-
ditions the duty cycle of the output
waveform is 50% which means that
each of the output transistors is fully
saturated (conducting) for an equal
amount of time. The average output
voltage is therefore zero. It therefore
follows that if one of the output
switches is closed for a longer period
than the other, the average output volt-
age will then be either negative or
positive depending on the polarity of
the input signal.

It can be seen then that it is the average
output voltage that is proportional to
the input signal. Since the output
transistors function exclusively as
switches, very little power loss occurs
in the output stage.

The September 1979 issue discussed a
variation of the above principle. This
was a self oscillating PWM amplifier in
which the square wave generator, the
pulse width modulator and the output
stage formed a single unit. This pro-
duced an efficient amplifier with only
a very small number of components.
A modified version of that circuit with
a printed circuit board is described here.

The circuit diagram

The circuit of the complete amplifier
is shown in figure 1 . It can be seen that
a PWM amplifier need not be very
complicated at all. The input signal is

the output of IC1 is low and the emit-
ters of T1/T2 are high, capacitor C3 is
charged by way of R7 and the voltage
rises at the non-inverting input. If it
rises above the level of the inverting
input, I Cl's output changes low to high
and the emitters of T1/T2 change from |
high to low. As a result, C3 is now
discharged through R7, the voltage at
the plus input drops below that of the j
minus input and the output of IC1
switches back to a low state. The result
is a squarewave output; the frequency
of which is determined by R7 and C3.
The values given result in an oscillation 1
at 700 kHz.

Provided Murphy doesn't get in the
way, we should have an oscillator.
Now we have to pulse width modulate
it. The level at the inverting input of I
IC1, which is used as a reference, does
not remain constant but is determined
by the audio signal. The point at which
the output of the comparator changes,
is also determined by the amplitude. As
a result the width of the square waves
is constantly changed (modulated) by
the audio signal.

At the output of the amplifier, filtering
is required: it is not supposed to act
as a 700 kHz transmitter! An LC/RC
network is used, consisting of L1/C6 ,
and C7/R6.

With a load of 8 ohms and a supply
voltage of 12 volts, the amplifier pro-
duced 1 .6 watts. At 4 ohms, 3 watts
were measured. Cooling the output
transistors was not necessary. The
harmonic distortion proved to be
surprisingly low for such a simple
design. Less than 0.32% total harmonic
distortion from 20 Hz-20 kHz was
measured.

Figure 2 shows the printed circuit board
and parts layout for the amplifier. Its
construction requires little time and
money, so it offers an excellent oppor-
tunity for anyone wanting to become
better acquainted with PWM. K

J °J

& »

interface and software on one Eurocard

No cassette interface is included
in the BASIC microcomputer
described in Elektor, May 1979;
furthermore, NIBL doesn't
include su itable cassette routines.
The obvious solution is to
combine these two missing links
on a single p.c. board: a handful
of ICs for the interface hardware,
and Vx K worth of software in
EPROM.

With these extensions, programs
for the BASIC microcomputer
can be stored on and retrieved
from tape.

The combination of cassette interface
plus the necessary software on a single
p.c. board offers several interesting
possibilities:

• Users of a (normal) SC/MP system
can use this p.c. board to keep

certain special programs close at hand.

• Users of the BASIC microcomputer
can use this board (without the

components for the interface) for
permanent storing of BASIC programs
— control routines, say.

• Even without the NIBL ROM, the
BASIC computer board makes a

good CPU card, with complete in- and
output buffering. The EPROM section
can therefore be used to store a 2 K-
byte monitor routine (we are working
on this!), located on page 0. The BASIC
card already contains a TTY interface,
so that a TTY or VDU (the Elekter-

minal, for instance) can be used for
developing programs in machine
language.

• With this card added to the existing
SC/MP system, it becomes relatively

easy to use other CPU cards instead of
the original SC/MP card. In this way,
the system can be converted to any
other microprocessor - the Z80, for
instance.

• The main purpose for developing this
new p.c. board was to add a cassette

interface to the BASIC microcomputer.
However, there is also room for a
hexadecimal monitor program. This
means that programs in both BASIC
and machine language can be developed
on the same computer. After all, NIBL
allows for running machine-language
routines as part of a program in BASIC
(by means of the LINK command).

cassette interface for BASIC microcomputer

The interface

The hardware for the interface consists 1
of an FSK modulator (FSK = Fre-
quency Shift Keying) and an FSK
demodulator, as shown in figure 1.

When a logic 'V is applied to the input
of the FSK modulator, a 2400 Hz
sinewave appears at the output. A
logic 0 at the input is coded as a 1200
Hz signal. These 2400 Hz and 1200 Hz
tones are recorded on tape.

When the tape is played back, the
demodulator must obviously convert
the 2400 Hz and 1200 Hz signals back
| to logic Is and 0s. This digital signal
• is applied to the serial input (Sj n ) of
I the BASIC microcomputer, the soft-
’ ware takes care of the conversion from
serial to parallel mode and stores the
data in the correct memory locations. Figure 1.

elektor may 1980 - 5-25

SOUT

\SLTL

SC/MP

_n_n

sin

lie principle of the cassette interface described.

2 ;

LIST

3 PR"WHAT IS THE FIRST ADDRESS";
3 INPUT A: REM INPUT:

3 PR"WHAT IS THE LAST ADDRESS".
3 INPUT B: REM INPUT:

:: REM INPUT: # XX

hill e-language programs by

Table 2. This program can be used for
adjusting the modulator.

Table 3. A program that will prove useful
when calibrating the demodulator.

FSK modulator

The FSK modulator is virtually a text-
book recipe: Take one 1C . . .'. In this
case, a function generator — the
XR2206 (see figure 2b). Two supply
voltages are required (+5 V and —12 V),
both of which are available on the
existing supply board. Since the input
signal is at TTL logic level, transistor
T2 is included as a level converter.

The output level is set by means of
P4; it should be adjusted to suit the
input sensitivity of the recorder used.
P2 and P3 set the two output fre-
quencies (1200 Hz and 2400 Hz,
respectively). The easy way to do this
is to use a frequency counter. However,
it is also possible to use the computer
itself as a calibration aid, as described
in the next section.

Calibrating the modulator

Since all timing in the existing SC/MP
system is derived from a crystal oscil-
lator, it is possible to obtain extremely
accurate reference frequencies by means
of a fairly simple program. Table 1
lists a program in BASIC that can be
used in order to load the program
given in Table 2 by means of a TTY
or a VDU. This second program is used
to generate the actual reference fre-
quencies. Once this program has been
loaded, it can be started by means of
a LINK command. At Flag 0 (pin 14c
on the connector) a squarewave will
now appear, with a frequency of either
1200 Hz or 2400 Hz. The actual fre-
quency depends on the numbers stored
at locations 'AA' and 'BB' (see Table 2).
Above this table, the numbers for both
locations are given, both for the BASIC
microcomputer (4 MHz clock frequency)
and for the Elektor SC/MP system
(2 MHz clock).

The complete procedure is as follows.
When using the BASIC microcomputer,
the first step is to load Table 1. As soon

cassette interface for BASIC microcomputer

as it is started ('RUN'), the computer
will ask for the first address; this must
be entered in hexadecimal: #0C 00, for
instance. Then the last address is
entered in the same way. The computer
will now proceed to request data for
the addresses - note, however, that it
will specify each address as a decimal
number: 0C00 = 3072. The display will
therefore initially read '3072 = ?'; the
opcode for the first address can now be
entered (# C4). Once the complete
program has been entered in this way,
it can be started: 'LINK # 0C00'. The
desired reference frequency will now
appear at Flag 0.

The reference signal and that from the
FSK modulator are both applied to the
simple test circuit given in figure 3. The
output can be connected to a high-
impedance headphone, or to a tape
recorder with some kind of level
indicator. The correct numbers are
loaded into the program for a 1 200 Hz
reference tone, and a logic 0 is applied
to the input of the modulator. P4 in
the modulator is set to maximum. If
headphones are used, three frequencies
will now be heard: the 1200 Hz
reference, the output from the modu-
lator, and the difference (beat) fre-
quency. P2 is now adjusted until the
beat frequency becomes zero. When
using the level indicator on a tape
recorder, P2 and the potentiometer in
figure 3 are both manipulated in turn,
in such a way that the signal level
becomes as low as possible — the refer-
ence frequency and the modulator out-
put will then be almost identical.

The program can now be modified to
produce the 2400 Hz reference fre-
quency, and a logic '1' is applied to the
input of the modulator. Using the same
procedure as that described above, P4
can now be adjusted until the two
output frequencies are (virtually)
identical.

FSK demodulator

The circuit of the FSK demodulator is
given in figure 2a. The input signal is
first passed to a trigger circuit (inverters
N1 and N2). This converts the sinewave
output from the tape into a symmetrical
squarewave. Two differentiating net-
works (N4 and N5) produce short spikes
at both the positive- and negative-going
edges of the squarewave, since N3 is
included as an inverter in the feed to
N4. These spikes are passed to a low-
pass filter R7 . . . R9 and C4 . . . C6; the
voltage across C6 is therefore pro-
portional to the frequency of the input
signal. A comparator circuit, consisting
of T1 and N6, converts this 'smoothed'
voltage to the corresponding TTL
logic levels: 0 or 1 for 1 200 Hz or
2400 Hz, respectively.

The only adjustment point in the
demodulator is PI in the comparator
circuit. For this calibration, a tone
generator could be used. Alternatively,
a simple program will allow the com-
puter itself to do the work.

cassette interface for BASIC m icrocomputer
Aligning the demodulator

If a 'symmetrical' input signal is applied
to the demodulator, the output should
also be symmetrical. For the input
signal, 'symmetrical' means that 1200
Hz and 2400 Hz signals are applied
alternatively during exactly equal
periods of time - in other words, two
periods of the 1200 Hz signal must be
followed by four periods of 2400 Hz,
then two at 1200 Hz, and so on. The
output will then be a symmetrical
squarewave, switching between logic 1
and logic 0.

The program given in Table 3 will

elektor may 1980 - 5-27

produce the desired 'symmetrical' input
signal for the demodulator. As before,
it can be loaded into the BASIC micro-
computer with the aid of the program
given in Table 1 — in the Elektor SC/MP
system it can be loaded directly, of
course. The signal is again present at
Flag 0 (pin 14c of the connector).
Having connected this signal to the
input of the demodulator, PI is ad-
justed until a symmetrical output
signal is obtained. No complicated
measuring equipment is required here:
the average value of a symmetrical
squarewave that is switching between

EPROM :

5-28

1980

CASSETTE SOUTINES OPERATING PBCCEOJRE
OPTKMAL SPEED SELECT:

OV and the supply voltage is equal
to half the supply voltage. A DC volt-
meter is therefore connected to the
output, and PI is adjusted so that
the meter reads 2.5 V. This will occur
over a small part of the range of the
preset potentiometer, and the 'ideal'
setting is in the centre of this range.

For those who are interested, the test
signal consists of two periods of the
1200 Hz signal, followed by four
periods at 2400 Hz, then two at 1 200
Hz, and so on. This corresponds to a
transmission rate of 600 Baud.

The EPROM section
Up to four EPROMs, type MM5204Q,
can be mounted on this p.c. board. The
complete circuit, including address
decoding, is given in figure 4.

The address decoder (IC4) is arranged so
that this complete 2 K memory section
(4 x Vi K EPROM) can be located on
any 'page' from 0 to F. One complete
page corresponds to 4 K of memory, so
the 2 K contained on this board only
fill the lower half of the page - from
address x000 to address x7FF. The
remaining lines from the address
decoder (corresponding to the remain-

ing half page) can be brought out to the
connector by means of wire links. Note
that this should not be done if the card
is to be used in the original Elektor
SC/MP system: those bus lines are
already in use!

IC3 is the 'page' decoder; in conjunction
with N8 it determines on which page
this memory section is to be located.
IC4 takes it from there, subdividing
the page into eight equal sections of
% K. Both IC3 and IC4 are used as
three-to-eight decoders; the selection
between the lower eight and the upper
eight pages is done by means of wire
links, as shown.

The data output lines from the EPROM
section are buffered by means of IC10.
The NRDS signal from the SC/MP
ensures that the data only appears on
the bus when it is actually required.

The p.c. board

All circuits given in figures 2 and 4 are
mounted on a single printed circuit
board. This board, with its component
layout, is given in figure 5.

Once it has been built and adjusted, as
described above, it can be plugged
straight into the existing bus of either
the BASIC microcomputer or the

5-30 - alt

may 1980

interfa

ace for BASIC microcomputer

Elektor SC/MP system.

Software

The cassette routines given in Table 5
can be used, in principle, on any SC/MP
system. To 'dump' a program, the
'begin' and 'end' addresses must first
be specified. This is not necessary when
'loading' a program, since these addresses
are already specified on the tape.

If the program is run as it stands, the
transmission rate will be 600 Baud.
Alternatively, the data at address 1FF5
can be modified for different Baud
rates: IE, 50, and FE give transmission
rates of 600, 300 and 110 Baud, re-
spectively - when used in conjunction
with the BASIC microcomputer, that is
(4 MHz clock). The Elektor SC/MP
system uses a 2 MHz clock, so that the
same data values will give 300, 150 and
55 Baud, respectively. Obviously, other
values can be used to obtain different
transmission rates, as required.

The first and last addresses, and possibly
the data value for the desired trans-
mission rate, must first be stored at the
corresponding memory locations. This
means that the system must include a
simple monitor program, at least. The
'Kitbug' monitor from the SC/MP
introkit, for instance, or the NIBL
BASIC interpreter. Although both of
these programs also contain their own
in- and output routines, it seemed
advisable to include these routines in
the cassette software given in Table 5.
This avoids any problems that might
occur when incorporating these routines
in the program - especially when using
it with the NIBL interpreter, since
different versions of this program exist.
The in- and output routines are located
at different positions in the memory!
It is quite possible, of course, to use
existing in- and output routines - only
a few modifications are required in the
program given in Table 5 to specify the
new addresses (specifically in the
sections under $2 and $4).

The new 'load' and 'dump' routines are
based on the original Elbug versions.
This means that tapes can be recorded
on one system and played back on the
other without any problems.

The 'instructions for use' of the cassette
routines are given in Table 4; this
actually shows a Load and a Dump
procedure. The first step in the Load
procedure is to jump to the cassette
routines by means of a LINK instruc-
tion (assuming that the BASIC micro-
computer is used). Then an L is keyed
in, to select the Load routine, and the
program will be transferred from tape
to memory. The example given also
shows what happens when an error
occurs.

The same procedure can also be used
when the tape was recorded at a differ-
ent transmission rate; however, the
'speed byte' at address 1FF5 must first
be modified in this case. For 300 Baud,
say, the data value # 50 must be
stored here. Then the LINK instruction

interface for BASIC micr

elektor may 1980 — 5-31

CHECKSUM
BLOCK COUNTER
BYTOUT-BYTE
BIT COUNTER
DUMMY VARIABLE
SPEED BYTE

TEMPORARY SPEED BYTE
NOT USED
NOT USED
NOT USED
END ADDRESS HIGH
END ADDRESS LOW
BEGIN ADDRESS HIGH
BEGIN ADDRESS LOW
SAVED P3 (L)

SAVED P3 IH)

SUBROUTINE ADDRESSES
PUT C: 81 CE Send one c

GECO: 81 8F Retrieve or

send 'echo

BYTOUT: 81 2B Transmit o

LDBYTE: 8155 Retrieve oi

ratchpad' on Page 1. The i

is given, followed by the start address
8006, this address was listed in Table 4.
Finally, Table 4 gives an example of a
Dump procedure. First, the begin and
end addresses of the program to be
'dumped' are specified. Then the
program is started, by means of a
LINK command, either at address
8000 (for autostart) or at 8006 (after
specifying the desired transmission rate
in address 1FF5). The tape is started,
and a D is keyed in — the Dump routine
will run and the program is stored on

When using NIBL, the end address of
the program to be dumped is found by
giving the command 'PR TOP'. The
computer will respond by printing a
decimal number; this must be converted
to the equivalent hexadecimal value
before it can be specified for the Dump
routine. The begin (and end) address
depends on the Page used: #111E for
Page 1 , # 2000 for Page 2, # 3000 for
Page 3, and so on. Page 1 is the only one
with a 'peculiar' begin address; the
reason is that NIBL uses the first 11D
memory locations for storing data.

Scratchpad

To avoid the need for an additional
RAM card, locations 1FF0 to 1FFF on
Page 1 are used as scratchpad for the
cassette routines. The data stored at the
various addresses are listed in Table 6.
Note that, since this section of memory
is used as scratchpad, the end addresses
on Page 1 should never be higher than
1FEF (4080 in decimal).

Table 6 also lists the various subroutine
addresses, with a brief indication of
what they do.

More on the EPROMs

At present, '/> K EPROMs are not as
'readily available' as could be wished.
If the MM5204Q proves difficult to
obtain, a 2 K EPROM can be used
instead. The details of this modification
are given further on - note that the
same idea may prove useful in many
other applications as well!

However, assuming that the MM5204Q
is to be used, there is still one minor
problem. These EPROMs are not quite
fast enough for use in a 4 MHz system
such as the BASIC microprocessor! For

this application, a minor modification
of the CPU card will also be required.
Before going into detail, it seems
advisable to summarize the various
possibilities for the EPROM section:

• four Vi K EPROMs, type MM5204Q,
When used in combination with the

Elektor SC/MP system (2 MHz clock),
these can be mounted without any
further modifications. For use with
the BASIC microcomputer, however,
the 'slow memory access' described
below must be added to the CPU card.

• one 2 K EPROM, type 2716. This
alternative is discussed further on;

it is equally suitable, without further
modifications, for both 2 MHz and
4 MHz systems.

• one 4 K RAM card, used as ROM.
This is the way the software was

originally tested, and it works on all
systems.

Slow memory access

As explained above, this modification
to the CPU card is only required if the
MM 5204 is used as EPROM in con-
junction with a 4 MHz system such as
the BASIC microprocessor. In all other
cases, it is unnecessary!

The problem is that these EPROMs
are not quite fast enough. For this
reason, the microprocessor's 'read' cycle
must be lengthened slightly. This can
be achieved by using the NHOLD input:
when this control pin is connected to
supply common, the SC/MP is 'frozen'
so that it maintains the current status.
A read cycle (or a write cycle, for
that matter) can be extended for as long
as required in this way. For the present
application, the read cycle must be
extended by 250 . . . 500 ns. Since
Page 0 is fully used for NIBL, there is
no need to extend the read cycle there.
The write cycle can remain unaltered
for all pages of memory.

This read cycle extension can be achiev-
ed as shown in figure 6. The NRDS
signal from the bus is passed through
N1 to N2, where it is gated by the signal
on connector pin 30c. This is one of
the address lines: it ensures that the
NRDS signal is only passed when one
of the pages 1 ... 15 is selected — not
for page 0, in other words. The output
from N2 triggers a monostable multi-
vibrator (N3/N4) that delivers an output
pulse of approximately 0.5 /is. The
result is that the NRDS pulse is
lengthened by this amount. The effect
on the software timing is so small that
the in- and output routines will still
function as intended. No modifications
are required in the software.

The four components needed can be
mounted on a very small printed circuit
board. The board and component
layout are given in figure 7. As can be
seen in the photo (figure 8), this board
is mounted near the connector on the
CPU board by means of short wire links.
The actual connections are shown in
figure 9: they are all either wire links
on the existing board or pins on the
connector. Note that the wire link
marked 'X' must be removed.

An alternative for the MM5204

4 x 5204 = 2716. This may seem pecu-
liar arithmetic, but it is actually a good
alternative solution if the MM5204
should prove difficult to obtain. The
idea is that four 'k K EPROMs can be
replaced by one 2 K version. The

interface for BASIC m

cassette routines given in Table 5 can be
located in the first Vi K, and the re-
maining IVi K can be used for other
software — a monitor program, for
instance (currently under develop-
ment!).

As shown in figure 10, the chip select
connections to the Vi K EPROMs are
decoded by means of eight logic gates.
Four of these (N5 . . . N8) re-encode
the four chip select signals into the
two-bit data required to address IC3.
Gates N1 . . . N4 ensure that the
EPROM is put out of action when no
chip select signals are present.
Admittedly, these gates are not strictly
necessary: all the control signals re-
quired are already present 'somewhere'
on the cassette interface p.c. board.
However, that little word 'somewhere'
is the reason for investing in the two
additional ICs. It is now possible to

mount the complete circuit on a small
p.c. board that simply plugs into the
1C sockets on the existing board. No
messing about!

In principle, you would assume that
any 2 K EPROM labelled . . 2716
could be used, but unfortunately this
is not quite true. The limitation is that
the 1C must work off a single +5 V
supply. The Texas Instruments version,
for example, needs three supply volt-
ages, for this reason, it cannot be used
in this circuit.

The printed circuit board for the
'EPROM alternative' is given in figure
1 1 . After mounting the ICs and resistors,
short and stiff connecting wires must be
soldered into place to form the connec-
tions to the existing 1C sockets on the
cassette interface p.c. board. This is
where the two rows of holes on each
side of IC3 come in. The idea is that

each piece of wire is looped through
two holes, so that a fairly rigid mechan-
ical construction is achieved (see figure
12). The two rows of wire 'pins' are
then inserted into two 1C sockets, as
shown in the photo (figure 13). The
2 K EPROM is located above the 1C
socket for IC6; the connecting wires
are inserted into one row each of the
sockets for IC5 and IC7. The four
remaining connections to the extension
board are connected to four wire links
that carry the chip select signals; this
is clearly shown in the photo.

In conclusion

As anyone who is actively interested
in the 'hardware' side of micropro-
cessors will know, what is written today
may well be out of date tomorrow.
This is certainly true when you attempt
to design 'general purpose' additions,
such as the cassette interface described
here. Against all odds, we have tried to
make this module suitable for all
SC/MP systems . . . Furthermore, we
have attempted to solve the availability
problem of EPROMs by offering several
alternatives.

If you can build and install any of the
variations described here, we will have
succeeded in our aim. Even if you only
succeed after working out some other
alternative according to the principles
laid out, we feel that this article is not
wasted. After all, the idea is to provide
you with a cassette interface! And,
provided you can obtain any of the
EPROMs listed, this should certainly
be the case if you are the proud owner
of either of the SC/MP systems de-
scribed in Elektor. M

rrrr r 1

5-34 - elektc

>'SP'ays

Liquid crystal displays are an economic alternative to the well
known LED. They combine high readability with versatility. As far
as amateurs are concerned, however, judgement is still being suspended.
This is because, until recently, LCD's have always been difficult to
obtain, expensive, and involved complex operation. Now, at last, they
are in a new phase of development and the prices have come down
dramatically.

IXlHsplitys

A little current leads to a lot of contrast.

The world's most well known optical illusion.

During the past two years LCDisplays
have been catching up with their LED
counterparts. In fact, it almost looks as
if LED's will soon be considered old-
fashioned. This is hardly surprising,
when both types of display are com-
pared. LCD's consume approximately
1000 times less current than LED's.
Contrast under bright light improves
rather than deteriorates. Furthermore,
LCD's are extraordinarily versatile.
They can be transparent and allow for
great flexibility in size and form.

Before the above advantages could be
capitalized on, a few initial problems
had to be overcome. This was done
successfully with the result that high
quality LCD's are being mass produced.
They now have a satisfactory lifespan
and temperature range.

One beneficial effect of the rising
quality of the product is that it is
becoming more and more in demand in
industry and therefore more readily
available on the retail market.

Inside LCD's

A detailed knowledge of the technologi-
cal background of LCD's is not strictly
necessary in order to be able to use
them. Readers who take an interest in
this particular aspect are referred to the
bibliography which is given at the end
of this article.

An LCD Display basically consists of
two very thin glass plates between
which there is a liquid crystal layer
some 10pm thick. This layer consists
of a crystalline molecular structure.
What is essential is that the molecular
structure changes under the influence
of an electrical field. Depending on the
direction in which the molecules are
organized, the liquid crystal layer
becomes either transparent or reflective.
The inside surface of the two glass
plates is coated with a transparent,
conductive layer and this forms the
electrodes. A voltage applied to them
creates an electrical field which causes
the molecules in the liquid crystal
layer to change direction. The plane
affected (or segment of a digital display)
then alters in transparency.

Figure 1 shows the basic construction
of an LCD. The Si02 layers given in
the figure should be mentioned. These
insulate the electrodes from the effects
of the liquid crystal and the two
polarizers (polarisation filter discs). The
alignment of the crystalline structure
is such that transparency will not
change until a voltage is applied. The
organisation of the crystal molecules
in the electrical field is shown in figure
1. When an (alternating) current is
applied between the two electrodes,
the crystal molecules will be arranged
horizontally. As can be seen, the lower
half has no drive current and so the
liquid crystals are in a vertical configur-
ation.

In an unenergized state in a reflective
LCD, a vertical and a horizontal polar-

elektor

1980 - 5-35

izer are laminated onto the liquid crystal
cell at right angles (or 90° rotated) to
each other (see figure 2a). Vertically
polarized light entering the front of the
cell (A) follows the rotation of the
crystal alignment as it passes through
the cell again rotating 90 degrees, the
polarized light passes through the
horizontal polarizer to the reflector (E).
The light is then returned through the
cell again rotating 90 degrees, and passes
out of the LCD through the vertical
polarizer.

In an energized state, however, (see
figure 2b) across one or more of the
character segments the crystal molecules
in the segment align themselves with
the electrical field. Rotation does not
therefore occur in the energized seg-
ment. The vertically polarized light
from the energized segments cannot
pass through the horizontal polarizer,
but is rather absorbed by it. The seg-
ments therefore appear as dark images
against a light background. The opposite
happens with parallel polarisation fil-
ters, the powered segments are trans-
parent and appear as bright images on a
dark background.

Things are different when a semi-
transparent mirror is used as a reflector
(figure 3b). It results in 'transflective'
displays which can be illuminated
from in front as well as from behind.
When current consumption is of minor
concern, in mains power equipment
for example, the light source behind
the display can be constantly on. If
the surrounding brightness is greater
than the light intensity effected by
the built-in lighting, the display operates
in a reflective manner. If the external
brightness is less, 'transillumination'
or transmission occurs.

There are also displays which operate
exclusively on a built-in light source,
that is to say, producing transmission
without a reflector (see figure 3c).
These are called transmissive displays.
Recent developments seem to favour
reflective and transflective displays,
whereas the transmissive type tends
to be pushed into the background. The
former types nearly always display
dark characters on a bright background,
whereas the transmissive type features
transparent characters on an opaque
(dark) background.

Characteristics

The prime feature of the LED is bright-
ness, while that of the LCD is contrast:
the main criterion for readability.
Contrast involves a certain light/dark
ratio of segment brightness during the
'on' and 'off' state when the external
light is constant and it is seen from
the same angle. The ratio is between
1:10 and 1 : 20. A good example of
this effect is the text of this magazine
where black and white contrast is sharp.
The operational ratios also have an
influence on the contrast, especially on
the viewing angle and on the triggering

Figure 1 . Basic construction of an LCD. The layer of liquid crystal is hermetically enclosed
between two glass plates. The glass plates contain transparent, conductive electrodes. As shown
in the layout, the direction of the molecules changes under the influence of an electrical field.
In combination with the externally adhesive polarisation filters, 'capsizing' the molecules
between the triggered electrodes causes a change in the transparency of the corresponding

2

Figure 2. According to the position of the polarisation filters the following takes place:

Figure 2a. In an unenergized reflective LCD, the segments are transparent when the filters are
parallel to eachother. Polarized light is rotated 90' by the liquid crystal material.

Figure 2b. The triggered segments become opaque (dark! when the filters are at right angles to
eachother. Rotation does not occur in an energized reflective LCD.

5-36 - elektor

1980

LCDispla

Figure 3. Depending on the construction of the display, there are the following types of LCD:
Figure 3a. Reflective operated. At the rear a reflector has been incorporated.

Figure 3b. Transf lective display. A semi-transparent reflector enables it to be illuminated from
behind as well.

Figure 3c. Transmissive display.

(static or multiplex). The viewing
angle is shown in figure 4. LCDisplavs
achieve a viewing angle of up to 160°,
where the light/dark ratio is 1 : 3.

The contrast is also dependent on the
operating voltage. For maximum con-
trast a certain field intensity between
the segment electrodes and the back
plate electrode is required, which
relies on a certain voltage. Figure 5
shows the typical voltage curve. When
the voltage rises the liquid crystal
molecules are realigned gradually. The
contrast at a certain voltage depends
on the percentage of molecules in the
field which have already changed
direction. When contrast is at a maxi-
mum, this will be around 100%. If the
voltage is further increased, the contrast
will remain constant rather than in-
crease. This may be a disadvantage if
multiplex applications are sought. Con-
trary to multiplexing LCD's, the shorter
'on' period (analogous to the increase
of segment current with an LED) is
not compensated.

The level of operating voltage required
may be freely chosen. On the one hand
it is determined by the basic material
used and on the other by the density
of the liquid crystal layer. The thinner
the layer, the higher the field intensity
(at the same voltage level) and the
lower the operating voltage required.
Nowadays, LCD's are being designed
with operating voltages in the region
of 1.5 V to 20 V. The contrast curve
shown in figure 5 is temperature de-
pendent. At higher temperatures, con-
trast is achieved at a lower voltage.
The curve then becomes more pro-
nounced. If the temperature is low, the
opposite happens: the curve than
flattens out. Again, this may cause
problems if a multiplex operated system
is used.

The switch-over times of an LCD rely
on the voltage and temperature. Figure
6a shows the time lapse of contrast
when the LCD is switched 'on' and
'off' respectively. It features a relatively
long switch-on delay (*d in the figure)
of 100 ms, before any change in con-

trast takes place. If the contrast is to
reach 90% of the maximum value,
another 70 ms (*r) will be required.
When it is switched off, the contrast
starts to fade immediately but takes
about 230 ms (*f) for it to be complete.
Depending on the type of material
used, the turn 'on' time with a rising
operating voltage, becomes markedly
shorter, whereas the turn 'off' time
lengthens only slightly.

Temperature is also an important factor.
Generally speaking, when the LCD's are
in a warm environment, they switch
more rapidly (see figure 6b).

Lifespan and temperature range

Both aspects are closely connected. A
great deal is known about the lifespan
of LCD's, nevertheless, it merits a few
words here.

What is a lifespan? It all depends on the
type of display used (reflective, trans-
missive or transflective). A 50% drop in
contrast leads to various results. The
lifespan is also dependent on the num-
ber of operating hours until a failure
rate of 50% occurs.

No matter how the lifespan is defined,
it is certain that during the past few
years great progress has been achieved.
A life expectancy of more than 50,000
hours (almost six years of operation!) is
now quite normal.

In the early stages of development,
problems affected the LCD's resistance
to ultra violet light, humidity and
foreign debris. As glass plates used to be
stuck together with adhesive material,
the liquid crystal was not hermetically
sealed and therefore had a lifespan
of only one to two years. This was
solved by the introduction of special
laminating material for glass. By coating
the plates with a thin layer of quartz,
the liquid crystal remains unimpaired
and at the same time the electrodes are

insulated against it.

More stable substances are now being
sought to extend the temperature
range and improve the switching times.
The chemical stability of a few of the
most recent liquid crystals is of such a
high quality that it has once again
become feasible to use the old adhesive
technique. This is an important step on
the road towards large surface displays
with a view to the alphanumerical
LCD’s of the future.

Polarizers have not, however, undergone
a similar development. Light polar-
ization takes place in a polyvinyl
alcohol foil, which is stretched to a
maximum and then soaked in an
iodated compound. The foil is very
thin (25 /tm) and must be pasted to a
carrier foil. Polarizers tend to bleach
in high temperature and humid sur-
roundings, which may result in a loss
of contrast. A solution would be an
LCD with 'sun glass' (darker polarizer)!
By using impermeable protective foil
and improved adhesive and solidifying
processes, polarizers would then be well
protected against humidity.

Operational and storage
temperatures

As mentioned above, the performance
of LCD's slows down when the tem-
perature drops. At temperatures of
about — 10°C they even freeze up
altogether with the result that the liquid
crystal becomes a solid. At the other
temperature extreme the liquid becomes
thinner until it loses its crystal structure.
A distinction should be made between
the operating and the storage tempera-
ture ranges. If the working temperature
exceeds its range, the display will
become inoperative. It is only when the
storage temperature range is exceeded,
however, that permanent damage is
done.

Liquid crystal material currently in use
has a working temperature range with
a lower limit between —5° and -15°C
and an upper limit between 40° and
80°C. The storage temperature range
has a lower limit of —20° to — 40°C
and an upper limit of 60° to 85°C
(depending on the liquid crystal used).

Voltage control

LCD segments are triggered into
operation by applying an alternating
current. It must be a frequency of
over 30 Hz (to prevent the display
from flickering.) This is essential and
it makes no difference whether the
electrodes have been insulated against
the effects of the liquid crystal or not.
If they have not been insulated, the
application of a direct voltage will
result in electrolysis thereby destroying
the electrodes. If the electrodes are in
fact insulated, the ions in the liquid
crystal are shifted. This breaks down
the electrical field and the display
fades at once.

If the supply is DC (as in battery
powered equipment), an AC waveform
will have to be generated by means of
an oscillator. To prevent the display
from visibly flickering, the frequency
range is limited at its lower end. The

LCDisplays

upper end is limited by the resistance
of the electrodes and the capacitance
(the RC time constant) of the segments
in the display. In an equivalent circuit
an LCD segment represents the parallel
connection of capacitor C and of a
high-valued resistor R. The capacitance
is primarily determined by the size of
the segment surface. For instance,
its capacitance per digit depends on the
height of the digit and on the LC
material and will be between 150pF
(8 mm digit size, high quality LC) and
4 nF (maximum value for 25 mm
digit, standard LC).

The resistance is dependent, among
other things, on the segment's surface
and on the quality of the electrodes'
insulation. 'In the above examples the
corresponding values for the direct
voltage resistance would be 1400MJ2
(8 mm) and 8 Mfi (25 mm high).

If only alternating current is applied,
the resistance in the segment may be
disregarded. The current consumption
will then rely on capacitance and
frequency (figure 7). In the case of a
display with a very small surface area,
it is possible to reach a working fre-
quency of up to 1 kHz; with larger
displays, however, there is little point
in having an operating frequency of over
100 Hz. Usually, manufacturers indicate
an operating frequency of 32 Hz at
50 mV.

How does it work?

The next distinction which must be
made is the difference between static
operation (direct segment control) and
multiplex operation (switched segment
control). As its name suggests, static
operation provides each segment with
its own drive, and one common elec-
trode may be used by all the segments
(and usually is). Thus, in this respect it
is like the seven segment LEDisplays
(common cathode or common anode).
As opposed to multiplex operation,
static operation is uncritical with regard
to contrast, tolerance and temperature.
Figure 8a shows a simple control
circuit for a segment with a push-pull
transistor stage. The transistors are
part of a CMOS inverter 1C, a CD4007
or CD4009, for instance. The inverter
receives a square wave of 30-50 Hz at
its input and switches at its output
between +Ub and 0 V. The peak value
of the alternating current applied to the
segment is equal to half the operating
voltage.

Capacitors are expensive and take up
a lot of space when compared with
1C gates, so it would be an advantage
if the circuit could be built without
any discrete components as shown
in figure 8b. After inversion, the square
wave at the rear electrode is 180° out
of phase with the one at the segment
electrode. Between the two electrodes
lies an alternating current with a peak
value equal to the supply voltage U^.
This principle can be put into practice

9

*juul

Figure 9e. A complete direct drive circuit with EXOR gates as segment drives.
Figure 9b. The pulse diagram shows that an AC is present at the triggered segments.

5-40 - elektor

1980

LCDispla

in an elegant manner with the aid of
EXOR gates of the CMOS type (for
instance, CD 4030 or CD 4070). Figure
9a shows the circuit. A gate is required
for every segment. To one of the inputs
of each gate and the display common,
a constant low frequency alternating
current is applied. The other gate
input then controls the segments. If
there is a logic 1 at the control input
the square wave at the segment elec-
trode will be out of phase (with refer-
ence to the display common) and in
phase if it is a logic 0. This is clearly
shown in the diagrams in figure 9b.
Because the signals are in phase when
the segment is powered, no difference
in voltage occurs. When they are out of
phase, the AC rises with a difference
in potential of twice the amplitude of
the square wave (between the triggered
segment electrode and the common
electrode).

This must of course be taken into
account when the supply voltage of
the LCDisplays is fixed. On the data
sheet this is usually given as the
effective value of the AC waveform. The
effective value of the waveform is equal
to its peak value and this is equal to the
operating voltage Ub of the CMOS
gates. For an LCD specifying an oper-
ating voltage of 4 to 6 V the CMOS
drive circuit will be fed with 5 V.

Multiplexed operation

The threshold values of the LCD con-
trast curve may also be multiplexed,
although this will be limited to a few
steps. The reasons for this are:

— The contrast is not pronounced.

— The contrast curve is temperature
dependent.

— As opposed to LED's the contrast
may not be increased by means of
short interval overdrive.

— If the system is direct voltage (as
opposed to multiplexed) controlled,
these problems are avoided, although

LCDisplays

1980 - 5-41

the high number of connections to
the display and drive circuit is often
a drawback.

Commonly used LCD's include the
'three step multiplex'. In this type up
to three segments are attached to a
single connection. Figure 10 shows a
seven segment display for a three step
multiplex operation with a matrix
organisation for two digits. This ex-
ample does not include two matrix
points which could be added. If all
the matrix points are to be optimally
exploited, only 2 + 3 connections will
be required for 18 segments. The system
becomes operative in three chronologi-
cal stages.

First, all the segments at the back
electrode COM 1 are triggered, then
those at COM 2 and those at COM 3,
after which the cycle starts again. In
order to trigger the back electrodes
(COM = 'rows' in the matrix) and the
segments groups (columns in the matrix)
square waves are used which supply an
AC to the triggered segment. Further-
more, the control signals have to be
such, that the AC is in phase for the 'on'
segments and out of phase for the
switched 'off' segments. The row
and column signals have to differ in
amplitude. Usually, the higher voltage
is applied to the back electrodes and
the lower to the segments. Figure 11
gives a practical example of digit 4
which is lit in the seven segment display
shown in figure 10. The triggered
segments are given in the matrix as
shaded circles.

The corresponding pulse schedule shows
from top to bottom: clock, COM
signals, column signals and the differen-
tial signals UcOM — UcOL which be-
come operative in segments dp, nc, G
and C.

One multiplex step corresponds to one
clock period. The column signals are
obtained when a square wave is con-
nected through to a clock signal, and
to an equal voltage for the rest of the
time (the two subsequent clock periods)
to the COM row concerned. The pulse
at the COM transmission activates
the rows concerned. Whether the
segments on the row (matrix points) are
'on' or 'off', depends on the phase
layer of the column signal at that
moment. For an inoperative point,
the column signal is in phase and for a
triggered one it is out of phase to the
column signal. In the pulse diagram,
for example, column signal COL 1 is
out of phase to the common signal
COM 1 during the first multiplex
step (pulse on COM 1). The decimal
point (dp) is switched 'on' during the
first step. This can also be seen in the
differential signal (COM 1 -COL1).
The voltage operated at both segment
electrodes is added to the COM and
COL signal. This is not true of the
untriggered segment nc on the first
row. Here the column signal COL 2
is in phase to the COM 1 signal. The

Photo 3. Alphanumeric 48 digit LCD unit. This ready-to-incorporate module by GEET has a
display surface of 142 x 22 mm, two rows of 24 digits, where every digit is made up of a
5x7 matrix (a total of 1680 display dots). The module already contains a multiplex drive
circuit and consumes 2 mA. The printed circuit at the rear of the display module is optional
and has a character generator. ASCII input bus and display interface.

result at segment nc is an AC which
is definitely smaller than at the
triggered dp segment, because the
COM and COL signals are now sub-
tracted. The value of the AC remains
below that of the minimum operating
current of the LCD. The untriggered
segment will of course not be activated.
The column signal is generated by
means of a shift register, at each output

of which an EXOR gate has been con-
nected. The second input of all the
EXOR gates is at the clock for direct
triggering. This is how the information
('1' or '0') at the shift register's output
can determine the state of the square
wave at the output of the EXOR gate
(inverted or non-inverted). After the
gates, CMOS analogue switches follow,
which switch the voltage values when

5-42 - elektor may 1980

LCDispla

11

Spl Sp2 Sp3

Figure 11. Segment control for digit - 4' + dp on a
a clock, COM signals, column signals and the diffe

-TLr_" i r. n_-u i r

the column signal is generated. The
optimum ratio of row voltage to column
voltage is V 0 pt = n/f* where n is the
number of multiplex steps. For a three
step multiplex the ratio is \/3 = 1.73.
Figure 12b shows the required voltage
values for three step multiplexes to be
generated and the corresponding voltage
phase for the COM and COL signals.
The voltage U 0 is the starting voltage
(for a 10% contrast) of the display and
this is indicated on the data sheet.
Usually 1.05 V is enough.

Conclusion and outlook

More information can now be displayed
with multiple segment displays.
They are available with 1120 light

spots (32 alphanumeric characters in a
7x5 format). The portable, battery
driven data terminal with an LC screen
is no longer an illusion. The complex
control systems behind such displays
can be simplified by means of integrated
drive circuits. As the number of multi-
plex steps in LCD’s is, technologically
speaking, limited, the LCD has to
become ’active’ for average to large
quantities of information. That means
that at every intersection of the control
wires there is an active semiconductor
element, such as an FET.

The rear of the display consists of a
large area chip, on which the corre-
sponding transistor matrix has been
etched.

A display of this type was recently

introduced by National Panasonic
(Matsushita). It was demonstrated in a
prototype pocket TV set with a flat
LCD screen (see photo). The reflective
LCD contains 57,000 (240 x 240)
displays dots on a chip measuring
44 x 56 mm.

Figure 13 shows the basic construction
of the screen. Every matrix point on the
silicon substrate consists of a capacitor
and of an FET. 1 10,000 transistors and
capacitors on a single chip!

The sample TV consumes barely 1 .5 W
with a battery voltage of 4.6 V (2 lithium
cells). It is not likely to be mass pro-
duced until a further reduction in
dimensions and current consumption
has been achieved. At any rate, the
example shows that a flat LCD screen

can already be produced at this stage in
development. As far as multi coloured
LCD's are concerned, however, the
production of these cannot be expected
in the near future.

VALVO: 'Liquid crystal display
elements', VALVO technical infor-
mation for Industry, March 1978,
nr. 780329.

FAIRCHILD: 'LCD 78', LCD brochure
of the Fairchild Camera and Instrument
Corporation.

Martin Bechteler: 'Liquid Crystal
Displays highly reliable components'
SIEMENS components report 17(1979),
Volume 3.

Paul Smith: 'Multiplexing Liquid-
Crystal Displays' Electronics, May 25,
1978, p. 113.

D. Davies, W. Fischer, G. Force,

K. Harrison and S. Lu: 'Practical liquid
crystal display forms forty characters'.
Electronics, January 3, 1980.

Figure sources:

Figures 1, 4, 6b, 7, 9b, 1 1, 12b, photo 2 end
photo 3: Siemens.

Figures 2a, 2b Fairchild Camera and Instru-

Figure 6a: VAL VO.

Figure 10: Data Modul.

Figure 13, photo 4: National
Panasonic/Matsushita.

Photo 1: HAMLIN H

5-44 — elektor may 1980

f lexit

flexible

P. Deckers

Certain requirements must be met
if an intercom system is to be fully
flexible and efficient. It is essential
that any station can call any other
without the need for a master
station. The number of inter-
connecting wires should be as few
as possible, conversations between
any two stations should remain
private and the standby current
drain needs to be low. It would be
useful if the system could also
operate as a babyphone without
blocking the line.

The intercom described here
complies with all of these
requirements while being flexible
with regard to station location.

intercom system

a mobile communication system

Number 98 in the Elektor Summer
Circuits (1979) issue reached a respect-
able twelfth position in our recent
competition, reason enough for us to
look at the circuit in greater detail. It
achieves everything that is required of an
intercom system and consequently
remains unchanged. (A minor error did
occur in the original circuit diagram,
however, resistor R20 should have a
value of 1S28 and not 1k8).

The intercom system is designed to have
a maximum of five stations with com-
plete security between any two. Further-
more, any station can be wired as a
babyphone. The system operates on a
four-wire ring cable laid in any con-
venient manner, that is, two or more
stations can be connected in the same
length of cable 'in series', or individually
by a 'spur', or in any combination that
happens to fit the 'bricks and mortar'.
For further flexibility suitable sockets

can be placed in any desired position
and if all stations are equipped with
plugs they then become completely
mobile. The only criterion in the cable
network is that the power supply should
be connected to it — at any convenient

Block diagram

Figure 1 shows the block diagram of
one of the stations (number two)
together with the power supply. The
four wires of the ring circuit carry the
positive and negative of the 15 volt
supply, the audio signal, and the control
signal (S). Depending on the station's
number, one of four reference voltages
can be switched to the control line
while a fifth reference voltage is con-
nected directly to one input of a
window comparator. In this way the

comparator will receive the second
lowest reference voltage from station 2
and the highest, for instance, from
station 5.

As the comparator's other input is
connected to the control line, when the
voltage on the S line is the same as the
reference voltage fora particular station,
the electronic switch (ES) will close
thereby supplying power to the preamp
and power amplifier stages. When
inoperative, the 'push-to-talk' switch
IS2) will be in the 'listen' position, so
that the audio signal on the LF line
reaches the loudspeaker by way of S2
and the power amplifier. To reply, push-

button S2 is pressed. The loudspeaker
wiil then be connected to the input of
the preamp and will function as a micro-
phone. The output signal of the preamp
is fed to the LF line via S2b.

When there is a reference voltage on the
S line, a 'line busy' indicator will light at
every station.

To call a particular station, the corre-
sponding key (Sla . . . Sid) is pressed
causing the relevant reference voltage to
appear on the S line. As all five switches
are mounted in an interlocking group,
pressing one of the keys SI a... Sid
will cause the Sle key to drop out and
to feed the amplifiers with supply volt-

age. Again, S2 can be used to switch
between transmission and reception. As
the S line voltage no longer corresponds
to the reference voltage of the station
itself, the ES switch will remain open.

Circuit diagram

Figure 2 shows the complete circuit
diagram of one station. The five refer-
ence voltages are derived from the
supply via five zener diodes (D 1 ... D5)
which are connected in series. Resistor
R1 ensures that a current of approxi-
mately 12 mA passes through the zener

' pole on/off :

it may 1980 -5-47

diodes. The call up voltages for the
other four stations are selected by means
of switches S1a...S1d via diodes
D6 . . . D9. The remaining reference
voltage is connected to the junction of
diodes D12 and D13 which form one
input of the window comparator (ICIa
and ICIb). As it is station four which is
shown in the diagram, the reference
voltage will be 8.4 V (4 x 2.1 V). There-
fore the voltage levels on the non-
inverting inputs of ICIa and ICIb are
7 V and 9 V respectively. When the volt-
age on the S line is somewhere between
these two levels, the output of ICIa will
be low and the output of ICIb will be
high. Transistor T1 (the electronic
switch of figure 1) will start to conduct
thereby providing the preamp and
power amplifier (IC2a and IC2b respec-
tively) with power. A call up voltage
greater than 1.4 V on the S line is
detected by ICIc which will light Dll
to indicate that the audio line is busy.
Sla . . . Sid together with Sle form a
row of interlocking keys. When one
of these is pressed, any key which
was depressed earlier will spring back.
Switches Sla ... Sid are wired so that
contact is made when they are de-
pressed. Switch Sle on the other hand,
is wired so that contact is made when it
is not depressed (t'other way round!).
Sle should be depressed whenever one
has no intention of conveying a message.
This will put the unit in the receiving or
'listen' mode.

Switch S3 enables the intercom to be
used as a babyphone. In the baby's
room the intercom station is switched
over to 'babyphone'. This makes the
unit's preamp slightly more sensitive
since it bypasses resistor R24. The out-
put of the preamp is fed continuously
to the LF line via S3b. Every other
station can 'listen in' to this room
simply by pressing the corresponding
button (and at the same time conver-

sation can be held between the other
stations as normal).

Construction and setting up

The printed circuit board and com-
ponent layout for the flexible intercom
are shown in figure 3. There are four
mounting holes in the centre of the
board, apart from the ones at each of
the four corners. This enables the board
to be mounted as a single unit. An
alternative is to saw it in half and mount
the two halves one on top of the other.
Both halves are connected by a pair of
wires.

During construction the correct pos-
itioning of diodes D6 . . . D9 and the
connection of the key switches must
not be omitted. This is because there are
five reference voltages and only four
connections to the S line. The fifth is
connected directly to the junction of
D 12/D 13. This means that there is one
wire link on each of the five printed
circuit boards and in each case it is in a
different place.

The power supply can be virtually any
15V/1A type. A suitable circuit is
shown in figure 4. The power supply
can be connected to the ring cable at
any convenient point.

It should be mentioned that the 2.1 V
and 1.4 V zener diodes may prove
difficult to obtain, in which case they
may be substituted for green LEDs and
red LEDs respectively. For this purpose
the LEDs should be forward biased.
Resistor R12, the pull down resistor on
the control line, is only required in one
of the stations.

The intercom system needs very little
adjustment. The sensitivity control PI,
should be adjusted while the circuit is
switched to the babyphone mode,
whereas the sensitivity of the intercom
during normal operation is determined
by the value of R24. H

mis iiii:

Modifications to
Additions to
Improvements on
Corrections in

Circuits published in Elektor

Car Collision Alarm

Elektor no. 51/52, July/August 1979 issue

is saturated, so that the alarm goes off cor-
rectly. T3's emitter will have to be connected
to the 5 V supply instead of the 1 2 V supply,
R9 should have a rating of 2W and D1 will
have to be replaced by a wire bridge.

Top Preamp

Elektor no. 57. December 1 979 issue p. 1 2-1 5.
On the p.c. board parts layout IC5 should be
rotated 180°.

Choro synth

Elektor no 59. March 1980 issue p. 17-23.
The parts list for our Chorosynth proto type

updated one. We apologise for the error, and
include a complete and corrected parts list

One last note, the regulator 1C 17 should be
rotated 180° on the parts layout in figure 5.
Resistors:

R1 = 10 M
R2.R3, R6.R10,

R21 . . . R43.R64.R68,
R74.R75.R82.R84 = 10 k
R4,R73 = 100 k
R5 = 1 k
R72 = 5k6
R7.R77.R78 = 1 M
R8.R9.R19.R20,

R66.R70.R80 = 47 k
R11.R53. . . R62= 22k
R12 . . R15 “ 18 k
R16 = 15k
R17 = 4k7
R18- 82 k
R44 . . . R52 = 1k8
R63.R67 = 39k
R65.R69 * 6k8
R71.R97B= 330 Si
R76“ 100 o
R79 > 470 k
R81 = 2k2
R83 = 1 k
R85 = 220 O
R86A.R1 128= 5.6 0
R86B.R102B ■ 0.56 O
R87A = 6.8 n
R87B.R90B ■= 150S2
R88A.R89A. R92A = 10 O
R88B.R108.R1 12A = 2217
R89B.R1 1 0A,

R1 1 1 A,R1 13A = 27U
R90A,R91,R94A,R109B = 8.2 0
R92B.R98B = 68 O
R93A.R96A.R97A,

R99A.R1 00A = 12 O
R93B = 39 O
R94B.R106B = 1.6 0
R95 A, R98A, R 1 01 A, R 1 02A,
R105A.R109A ■ 15 0

(parts list continued overleaf)

Figure 4. A simple power supply circuit for the intercom system. Virtually any 15 V/1 A
type supply will suffice.

5-48 — elektor may 1 980

Capacitors:

Cl = 470 n
C2.C3.C4 = 3n3
C5 = 2n2

C6.C7.C8.C9.C1 8 . . C26.

C41 = 10pF/35 V tantalum
CIO . . C14.C17.C38.C39,
C40.C42 . . . C45,

C50.C51 = 100 n
C53 = 47 n

C15.C36 = 4 M 7/35 V tantalum

C16 = 2jj 2/35 V tantalum

C27.C28.C31 ,C52 = 10 n

C29.C30.C35.C37 = 47 n

C32 = 1 2 n

C33 = 22 n

C34 = 27 n

C46 = 1000 m/35 V

C47 " 330 m/35 V

C48.C49 = 330 n

Semiconductors:

D1 . . . D4 = DUS

D5.D6 • 1N4001

T1 . . . T5.T8.T10.T13 = BC 557

T6.T7.T1 2 = BC 547

T9.T1 1 - BF 256B (or BF 245B)

IC1 - LF 356

IC2.IC4 . . . IC7= 555

IC3 = 741

IC8.IC9 = 4520

IC10.IC1 1.IC12.IC18 = 4011

IC13.IC14 = TL084

IC15 = 78L15

IC16 = 78L12

IC1 7 • 79L15

Trl = 16 V/150 mA transformer
S1.S5 . . -S15 = SPDT
S2.S3.S4 = DPDT
LI . . . L4 = 5 turns of 0.2 mm
dia. enamelled copper wire on

What makes a StopSlip elastomeric pad so
useful is its incredibly high coefficient of

brought very close to vertical, and flat objects
simply placed on it - not stuck on — will stay

the material; it does not gradually decrease,
nor is it affected by repeated wetmopping.
Cobonic L td.. Knapton Mews. Seely Road,
London SW179RL.

Telephone: 1011 7676780

(1522 M)

The word's lowest noise audio
preamp 1C?

the HA12017 now in stock at AMBIT is
sufficiently superior device to warrant such i
billing.

The SIL housed 1C uses Hitachi's new silicoi
surface process, to provide an exceptional!!
low noise characteristic, that is reliably
repeatable in mass production. Under stan
dard pickup conditions, a S/N ratio of 82.6 di

just 0.185 mV. " mPU

A new generation of cases

The Bocon range of instrument cases from

West Hyde Developments have been widely

acclaimed for their impeccable tooling, and

the latest two types will certainly be no

The Bocon 'Desk' series is made in black ABS

in four sizes. These beautiful mouldings have

an ingenious stepped tongue and groove
construction, with highly polished surfaces
and flat, textured areas on the top. The one-

piece front panel is natural anodised alu-

minium, angled to provide three separate
surfaces. Inside there is provision for p.c.

boards or chassis.

The Bocon 'Commander', BOC690. is a

keyboard and display enclosure made in black

foam plastic. The housing is designed to
accept most proprietary keyboards, and has a

similar finish to the rest of the Bocon range.

However, as well as this phenomenal feature
the 1C has less than 0.002% THD over the
audio band of 20 Hz to 20 kHz, at an outpu:
level of 10V (RMSI - again with the measure
ment carried out under RIAA conditions.

The SIL package means that isolatior
between stages is kept to a maximum.

200 North Service Road.

Brentwood.

Essex. CM 1 4 4SG.

Telephone: 10277) 230909.

(1512 M

UK 16 - elektor may 1980

DC switched and tuned AM radio
unit

The new 91072 from Ambit International is
available in three stages of complexity,

(uniquely) both switched and tuned by DC
connections only.

Switching uses a 'ground -to-make' system,
enabling easy control from MPU bus lines if
required.

The standard bands are: Longwave 150-
400 kHz, Mediumwave 510-1620 kHz.
SW1 5-10 MHz and SW2 1.6-4 MHz.

Any frequency span of approximately up to
3:1 ratio can be accommodated in the region
100 kHz to 30 MHz to special order.

The unit is intended for broadcast radio
reception, and is fitted with a 6-8 kHz
bandwidth multi-element ceramic filter. A

buffered local oscillator output, together with
DC switchoff through a high impedance drive

Tuning a complete 3:1 frequency span is

unit may readily be interfaced with any Ambit
tuning synthesiser system.

The board is normally supplied in a screening
can, with edge connector terminations, or
may be supplied as a bare PC8 for incorpor-
ation into larger enclosures (as illustrated).
The antenna for LW/MW is a ferrite rod. but
the other two bands are intended for long

wire termination. If required, the two SW

bands may be substituted with wire fed

MW/LW.

Ambit International,

200 North Service Road,

Brentwood, Essex, CM 1 4 4SG.

Telephone: 102771 230909

(1514 Ml

414-digit multimeter

Gould Instruments Division has introduced a
new 4 54 -digit multimeter, the DMM12, which

accuracy of 0.05% and a built-in electronic
technique for making true root-mean-square
(r.m.s.) measurements on AC signals. Using
the latest solid-state circuitry and components
specifically selected for high stability and low-
noise performance, the DMM12 has 27

current and resistance, and is also available
with optional probes for radio-frequency and
high-voltage measurements.

The Gould DMM1 2 digital multimeter has an
ergonomically designed front panel using the
latest international symbols. Maximum reading
is 19999. and maximum resolutions on

are 10 mV, 10 nA and 100 mil, respectively.
The liquid crystal display incorporates separate

decimal point. Overrange and "battery low"
are also indicated using the display.

The true r.m.s. sensing AC/DC converter
used in the DMM12 can accept waveforms
with a crest factor (peak/r.m.s. ratio) of up to
4:1 at full scale, and a combined AC/DC
facility is available to measure AC waveforms
with a DC content. The true r.m.s. value

measures the energy content of an AC

waveform, and hence makes the DMM12
ideally suited to power-system measurements.
The DMM12 is housed in a rugged case and

meets IEC348 and VDE specifications.

Standard models are mains/line powered but
option BP12 gives true portability with
rechargeable cells.

Roebuck Road.

Essex. 1G63UE
Telephone: lot 1 500 1000

(1518 M)

Portable capacitance meter

The new Model 820 portable capacitance

economical multi-range instrument combining
digital accuracy with complete portability. Its
ten ranges cover capacitances from 0.1 pF to

1 Farad. Accuracy is 0.5% or 1% of full scale.

and resolution down to 0.1 pF. according to

In use the capacitor leads are simply inserted
into a pair of slots and the capacitance is

indicated on the clear 4-digit LED display. A

flashing display provides overrange indication.

Provision is also made for using jack plugs

when measuring in-circuit capacitances.

The Model 820 is ideal for production line or
laboratory use. It has a robust and attractive
moulded case but weighs only 675 g (1.51 lb).
It will operate with rechargeable or disposable
cells and there is provision for a charger. A tilt
stand, spare fuse and 26-page operating
manual are supplied.

Unit 3..

Westfieids,

Portsmouth Road,

Telephone: 107051 596020

(1520 M)

A new magnetic tape head

Being introduced by Monolith Electronics is a
new magnetic tape head for compact cassette
machines.

The C44RP2ES01 is a four channel cassette
head for record and playback, which also has
combined two independent half track erase
sections, thereby providing for full stereo
autoreverse record/playback and erase all in
one unit.

This head is produced to the standard "EIAJ"

mounting format, making it suitable for most

mounting holes, and measuring 12 mm from
hole centres to front face. Wiring is facilitated

Each record/playback channel has an im-
pedance of 650 ohms at 1 KHz, with a head
gap in the order of 1.5 microns giving a

playback frequency response of +10 dB over

the range 8 KHz/333 Hz.

The erase sections have an efficiency of better
than 55 dB on a 1 KHz signal.

The C44RP2ES01 was designed for auto-

reversing stereo recorders, but may also be

suitable for certain data recording purposes.

The Monolith Electronics Co. Ltd..

5/7 Church Street, Crewkerne,

Somerset TA 18 7 HR,

Telephone: 104601 74321

(1521 M)

elektor may 1980 — UK 17

Miniature chokes

The new 8RBS series of fixed inductors adds
a fourth member to TOKO's range of signal

The 8RBS spans from 1 00 pH to 15 mH in a
diminutive package, based on 5 mm pin
spacing, with Q's as high as 80.

The major applications are in transient sup-
pression in sensitive logic circuitry, and as a
general decoupling device to keep equipment
RF interference levels within the recently
introduced international standards.

Ambit International.

200 North Service Road,

Brentwood.

Essex, CM 14 4SG.

Telephone: (0277) 230909

(1508 M)

Miniature encoder switch
produces BCD signals

A miniature rotary encoder switch, having
10 positions and a four line binary coded
decimal output, has been introduced by
Impectron Limited.

Measuring only 23x 23x 19 mm overall
(excluding mounting pins and spindle) the
BCM 23 switch will enable designers to
fit a low cost encoder into tight corners. The
device will prove particularly useful where
manuel controls or mechanical assemblies
have to be interfaced to electronic logic.

The unit has five signal pins and two fixing
tags for PCB mounting, and may also be front
panel mounted using a threaded spindle bush
and nut. One signal pin is for voltage supply,

while the others represent 1. 2, 4 and 8. On
position “O' no contact is made, but as the
spindle is clicked around its remaining 9 pos-
itions the four output pins represent the
number of the position selected in BCD form.
The spindle may be continuously rotated, and
in this mode may be used as a shaft encoder
with 36° resolution.

The BCM 23 is manufactured to high stan-
dards, although it is a relatively low cost
device. Rated operational life is greater than
10’ rotations. (2 million steps) and operating
torque is as low as 500 gem.

The electrical rating of the switch is such that
it may be directly connected to substantial
current drains. Power rating is 3 W maximum,
with maximum voltage and current ratings of
200 V and 500 mA DC.

Impectron Ltd.,

Foundry Lane.

Horsham,

W. Sussex RH13 5PX.

Telephone: 0403-5011 1

(1510 M)

Economy wire stripper and cutter

A new simple-to-operate wire stripper and
cutter has been introduced by AB Engineering
Company. Known as the ABMK001, it
features a knurled knob adjustment to control
the stripping depth, a retaining clip to ensure
it remains in the closed position in the tool
box or pocket and a curved cutting edge
which provides a secateur-like action for clean

Based on the well proven AB MK 100, the
new MK 001 has an improved locking device
and is priced at £1 .85.

AB Engineering Company,

Timber Lane,

Woburn. Beds. MK 17 9 PL.

Telephone: (052525) 322/3/4/ 5.

(1511 M)

Weston 6000 autoranging
multimeter

This compact, rugged, lightweight instrument
combines the accuracy and convenience of
digital readout and measurement hold with
the broad range coverage of conventional
VOM’s. 3 VS digit 54" high reflective liquid

crystal display gives excellent readability over

a wide range of ambient light levels. The

autoranging facility means that 26 ranges of
measurement are available with only two

switches to select functions. Two disposable

9 V transistor batteries offer more than

handle serves as a tilt stand and a display

window protector. The unit is supplied with

Tool range L td.,

Upton Road,

Reading RG3 41 A.

Telephone : (0734) 29446 or 22245

(1526 M)

Wire wrapping kit

contains the tool plus the JUW-1 unwrapping
tool and four 50ft wire refill cartridges one
each in red, white, blue and yellow, all packed
in a sturdy, re-usable clear plastic box.

OK Machine & Tool (UK) Ltd,

Dutton Lane.

Eastleigh,

Hants S05 4AA
Telephone: 0703 610944

(1516 M)

r ELEKTOR BOOK SERVICE

300 CIRCUITS

"FOR THE HOME CONSTRUCTOR"

This book of more than 250 pages contains clear descriptions of
300 projects ranging from the basic to the sophisticated

PRICE £1.95 + 50p postage & packing (UK only)

OVERSEAS SURFACE MAIL £3.70

BOOK 75

PRICE £3.00 + 50p postage & packing (UK only)
OVERSEAS SURFACE MAIL £3.70

A selection of some of the most interesting and popular construction
projects that were originally published in Elektor issues 1 to 8.

30 projects are contained in this book, plus a 'DATA' section which
includes a chart of pin connections and performance for common-
anode LED displays, valuable information on MOS and TTL-ICs,
opamps, transistors and our tup-tun- dug-dus code system for trans-
istors and diodes. With over 100 pages, the wide variety of projects
in this book stimulates the professional designer to up-date his know-
ledge and even the beginning amateur should be able to build most
of the projects.

FORMANT

COMPLETE CONSTRUCTIONAL DETAILS OF
THE ELEKTOR FORMANT SYNTHESISER

If you are even mildly interested in synthesisers then this is the book
for you. It's all here - setting up procedures, hints and tips and more.

We have taken the trouble to include a FREE cassette of sounds that
the FORMANT is capable of together with suggestions on how to
achieve them.

PRICE £4.00 + 50p postage & packing (UK only) OVERSEAS SURFACE MAIL £5.70

DIGIBOOK

Provides a simple step-by-step introduction to the basic theory and
application of digital electronics. Written in Elektor's typical style,
there is no need to memorise dry, abstract formulae - instead you
will find clear explanations of the fundamentals of digital circuitry,
backed up by experiments designed to reinforce this newly acquired
knowledge. For this reason DIGIBOOK is accompanied by an
experimenter's PCB which will facilitate practical circuit construction.

PRICE £4.50 + 50p postages, packing (UK) OVERSEAS SURFACE MAIL £5.20

When ordering please use the

ELEKTOR READERS ORDER CARD .

in this issue