71

March 1981
U.K. 60 p
U.S.A. $2

up-to-date electronics for lab and leisure

sound generator

logic analyser
MW receiver

alter page ol i

>ur copy of our 1981

our 320 page catalogue. I
II I am not completely

^outside the UK send £ 1 68
E381

ctronic Supplies Ltd.

P O Box 3. Rayleigh. Essex SS6 8LR
Southend (0702 354155 Sales (0702) 5529 1

Shops

1 59 161 King Sire
284 London Road.
Both shops closed 1

smith. London W6 Telephor
. Sea. Essex Telephone Sou

I (0702) 554000

Catalogue now on sale in all branches of WH SMITH & Price £1.00

1981 - UK 03

selektor 3-01

multichannel TAP 3-02

J. Meyer

Touch switches have been rather neglected lately but their simplicity and
silence makes their use very practical in many applications. The particular
switch in this article is a bank of single pole selectable switches. The
printed circuit board features twelve switches but in principle this may be
any number.

the High Com noise reduction system 3-06

This article covers the complete constructional details of the Elektor High
Com noise reduction system. The High Com modules will be supplied by
Telefunken ready tested and calibrated. The project includes all the extras
needed to construct a very high performance system such as a built in cali-
bration oscillator and LED peak meter. The completed unit can be con-
nected directly between the main amplifier and the cassette or reel to reel
tape recorder.

logic analyser 3-18

The analysis of digital signals is not an easy task without the aid of an
expensive item of test equipment, namely a logic analyser. Unfortunately,
these are invariably priced well beyond the reach of the average enthusiast.

All is not lost however, since an ordinary oscilloscope can be coupled with
the project featured in this article to produce a very respectable logic
analyser.

sound generator 3-22

However strange a sound effect may be, our readers always seem to be able
to find plenty of uses for it. The device featured in this article is able to
imitate many sounds from the twittering of birds to machinegun fire, from
the screeching of brakes as a car runs out of control to the inevitable
crash . . . and all from a single 1C.

movement detector 3-30

The electronic movement detector described in this article not only opens
doors and switches lights on and off, but can also form the basis of a game
- how to sneak out an object from a guarded room.

medium wave receiver 3-32

Many of our readers belong to the younger generation and there is nothing
quite like building your first radio - and getting it to work! - as many of
our more experienced readers will know. The object of this article is to
end up with a neat economical portable radio that fits comfortably inside
a coat pocket.

text display on the Junior Computer 3-36

from an idea by U. Seyffert

The display of the Junior Computer is suitable for displaying both numeri-
cal and hexadecimal data. By utilising a seven segment alphabet it is also
possible to display written texts. If the text is static, a total of six letters
are available. If a longer message is required however, this may 'run' along
the display.

traffic game 3-39

. To sharpen our reflexes for todays traffic situation, this article describes a
game in which everyday problems encountered on the road are simulated
by electronic means.

market 3-42

advertisers index UK-22

contents

©mnitei
(Knum
( m

d

EDITOR.

W. van der Horsl
UK EDITORIAL STAFF
T. Day E. Rogans

TECHNICAL EDITORIAL STAFF

J. Barendrecht

P. Holmes
E. Krempelsauer
G. Nachbar
A. Nachtmann

K. S.M. Walraven

1981 - 301

Hitachi developments

The Hitachi Central Research Labora-
tory and Hitachi Maxell, Ltd. have
recently succeeded in synthesizing a
novel high Li + conductive solid electro-
lyte for use in solid-state batteries. This
development of the new solid electro-
lyte will make possible the production
of ultra-small or ultra-thin (0.7 mm in
thickness or less) batteries having high
energy density and a long shelf life
without any leakage.

The solid electrolyte is a lithium nitride-
iodide-hydroxide type compound. Solid-
state micro batteries using this electro-
lyte show discharge characteristics 100
times or more greater than previous
lithium iodide-alumina solid batteries.
This dramatic improvement will allow
their use in various small electronic
equipment such as wristwatches and
hand-held calculators.

The Central Research Laboratory and
Hitachi Maxell have experimentally
produced ultra-thin batteries with a
thickness of 0.7 mm and diameter of
20 mm, and have shown their feasibility
in practical applications. In the exper-
imental batteries, positive and negative
electrodes are formed on either side of
lithium nitride-iodide-hydroxide electro-
lyte pellet, and this unit cell is enclosed
in a case having a height of 0.7 mm. The
maximum current output of 5 mA is
approximately 100 times greater than
previous solid-state batteries. When a
constant current of two micro amperes
is taken, this battery provides a stable
electric potential of over 1,000 hours.
In the near future, more than 10,000
hours of battery service life can be
expected for an electronic wristwatch.
This new electrolyte will permit
production of batteries which are even
thinner than the present prototype
battery, and consequently will result in
a further increase in performance.
Therefore, layer-comprised batteries
with large capacity and high power
could be manufactured. They are

expected to be eminently suitable in
high voltage applications.

As the electrolyte itself is of solid-state
type, it is envisaged that 1C process
technology will be applied in the
fabrication of the batteries, eliminating
variations in product quality and
opening up wide fields of applications
for high reliability and low cost micro
power sources.

Lightweight portable methanol fuel cells

Hitachi, Ltd. has developed the world's
first compact, portable methanol fuel
cells as a power source for such areas as
home appliances and agricultural and
engineering equipment.

The development of a new type elec-
trode has made possible a three-fold
increase in power output compared to
conventional methanol fuel cells. The
new fuel cells weighing only half that of
an automobile lead-acid battery of the
same size are a high-performance
12 volt, four ampere unit. It can be used
to power VTRs and other home ap-
pliances for use outdoors as well as
pumps and blowers for agriculture and
engineering projects. Demand for the
new cell is increasing as it is a compact,
lightweight power source and can be
used for long hours.

Fuel cells that have been developed in
response to such needs generate energy
through the reaction of an oxidizing
agent with the fuel. It is a kind of direct
current generator that works so long as
there is a supply of fuel.

Hitherto, the methanol fuel cell had
been considered best as a portable
energy source, but its practical appli-
cation had been stymied by the low
voltage per electrode and the small
electric current density.

Hitachi, Ltd. developed an electrode to
activate the reaction using a platinum
ruthenium catalyst in place of the
conventional platinum single catalyst.
As a result, single cell voltage was
increased to 0.4 volts and electric
current density three times to 60 mA/

Furthermore, with the development of a
new battery structure using ion
exchange film as an electrolyte, the

new, compact fuel cells have a generat-
ing capacity of 12 volts, four amperes.
Hitachi will continue research and
development of the methanol fuel
cells for its practical application.

Main specifications:

1. Dimensions: 227 mm (width) x 195
mm (length) x 127 mm (height)

2. Weight: 6.3 kg

3. Rated voltage and ampere: 1 2 V, 4 A

(632 S)

More on noise reduction

The number of noise reduction systems
available on the market nowadays is
bewilderingly high and makes it all the
more difficult to make a suitable choice.
The survey here attempts to clarify the
latest situation with regard to records
and record players.

CBS recently came up with a very
interesting design for reducing noise on
records. It is not available as yet, but
promises to be highly effective. It is
claimed to totally eliminate any noise
caused by the surface structure of the
grooves (surface noise) and improve the
dynamic range considerably. CBS claim
that the new noise suppressor will
enable records to sound as good in
quality as analogue and digital master-
tapes. The system should achieve as
much as 85 dB signal to noise ratio and
could therefore also be used on tape
recorders.

As if that weren't enough, the system
has the added advantage that LP's
recorded with it are compatible to
standard equipment (even without a
corresponding decoder). In other words,
the sound will be improved on an
ordinary record player. This was cer-
tainly not possible using the well-known
Dolby, HighCom, DBX, Anrs, etc
systems. The CBS decoder is expected
to be highly competitive in price.
Naturally, the Dolby laboratories have
not been resting on their laurels. Their
technicians set to work as soon as the
High Com from Telefunken appeared,
for noise reduction has always been this
company's speciality. The recently
introduced Dolby C system seems to be
one step further in the evolution of the
wellknown Dolby B compander. It has
caught up with the competition by
suppressing up to 20 dB of noise. It is,
however, a great deal more complex
than the B compander and it is doubtful
whether the one can be used instead of
the other. Although 'normally' dolby-
sised cassettes may be played on a C
compander system this not true of C
dolbysised cassettes on the 'old' B
version, which may mean having to buy
a new cassette recorder.

Similarly to the B version, the Dolby C
compander is a 'sliding band' type using
an 1C manufactured by a well-known
American company.

(628 SI

3-02 - elektor march 1981

multichannel TAP

J. Meyer

It is high time that Elektor devoted
space and attention to TAP switches.
Their simplicity (and silence) makes
their use very practical in many appli-
cations.

It will be apparent from the heading
that the TAP in question is a multi-
channel set with single point touch
contacts. In other words the switch will
be operated by touching a fixed button
or contact point with a finger. As
mentioned before, there may be any

again. This brings us to the reset circuit:
When one of the contacts (say, contact
no. 1) is touched and the corresponding
output has become '1 ', a voltage of

33

133

(Ub-0.7):

which is about 0.23 x Ub (Ub represen-
ting the supply voltage of the circuit),

muMehannel TAP

the elegant touch

The TAP (Touch Activated
Programmer) switch is 'in' again.
Touch activated switches have
been rather neglected lately, but,
since the rise in popularity of the
electronic clock, more and more
manufacturers are beginning to
include them on all sorts of
devices.

The particular switch in this
article is a bank of simple single
pole selectable switches (SSPSS?).
The standard version features
twelve touch activated switches
but in principle this may be any
number. Only one button at a
time can be 'switched on'. This
means that the system is ideal for
use in tuners, for instance, where
a number of stations are to be
preselected.

number of switches although the twelve
on the printed circuit board of this
version will be sufficient for a variety
of applications. Of course, if more are
required, several boards may be linked
together to construct a 24 or even 36
channel TAP.

The boards are very reasonable, as far as
their constructional cost is concerned,
and the interlocking switch bank uses
very little current. This is due to their
only requiring one CMOS gate each and
after all, a 4071 1C contains four as it is.

The circuit diagram

The circuit involves very little in the
way of sophisticated electronics. On the
face of it, it looks complicated only
because the same circuit has been drawn
twelve times.

The twelve touch contacts are shown
one above the other to the left of the
diagram. Their corresponding outputs
are drawn at the right-hand side. The
switches themselves each consist of four
resistors, two diodes and an OR gate
(N1 . . . N 1 2). At the bottom of the
circuit IC1 and T1 can be seen. These
are included to ensure that when
a contact is touched all the others are
These are included to ensure that when
a contact is touched all the others are
reset thus allowing only one output to
be high at a time.

The circuit works as follows:

When one of the contacts is touched
both the input and the output of the
OR gate corresponding to it will become
logic one. Since the output is coupled
to the other input of the gate via
a resistor (R30 . . . R41), the output
will be latched, that is it will continue
to remain high even when the touch
contact is released. Diodes D13 . . . D24
prevent the output high level from
affecting one of the other gates.

So far so good. The next thing to
achieve is that as soon as another
contact is touched, the output of the
previous contact will become low

will fall across resistor R5. A higher
voltage level, however, will be on the
wiper of PI (as will be explained be-
low), so that the output of comparator
IC1 will be low. Transistor T1 and
diodes D1 . . . D12 will now no longer
conduct. Thus, nothing else will happen.
If after this a second contact is touched
(let's say contact no. 2, for instance),
the corresponding output (i.e. N2) will
again become logic 1. A higher voltage
will therefore fall across R5. In the
formula given above, the value of R42
(100 k) must now be replaced by the
value produced by the parallel circuit
of R42 and R43 (=50 k). This gives

§> ( U„-0„,

thus about 0.37 x Ub-
If the voltage across the wiper of PI is
set between the two values to about
0.3 Ub. the output of the comparator
will go high when the second contact is
touched. T1 will then conduct and reset
all the gates via diodes D1...D12,
except that one belonging to the contact
being touched. In other words, the
output of the last contact to be touched
will be the only one to remain high.
Setting PI to 0.325 Ub can be done
with the aid of a multimeter although
if one is not available, setting PI in the
middle position will probably be close
enough.

There's more than meets the
eye . . .

Readers with sharp eye-sight will have
undoubtedly spotted all sorts of 'extra'
odds and ends on the board which
were not included in figure 1. Well,
since the gates belonging to 1C 4071
must not be overloaded (somewhere in
the region of 0.25 ... 1 .5 mA) room
has been made on the board for adding
twelve buffer stages without making

jltichannel TAP

1981 - 3-03

Table 1.

4049/4050 U b

output 5 V

high 10 V

(Source) 15 V

output 5 V

low 10 V

(sink) 15 V

•out Uout

-2.5 mA 2,5 V

-2.5 mA 9,5 V

-10 mA 13,5 V

6,0 mA 0,4 V

16 mA 0,5 V

40 mA 1,5 V

Table 1 . The output current capability of
each channel when the buffers of inverters
of figure 2a or 2b are used.

4071

output

(Sink)

t*b •out
5 V -0,25 mA

10 V -0,5 mA

15 V -1,5 mA

5 V 0,25 mA

10 V 0,5 mA

15 V 1,5 mA

Table 2. The output current capability when
the wire links of figure 2c are used to replace
the buffers.

the circuit more expensive. By using
inverters as buffer stages, an inverted
version of the multichannel TAP is ob-
tained. When a contact is touched its
output will now become low, whereas
all the others will be high, instead of
the other way around.

Figure 2 shows what happens in the
buffer stages. The twelve outputs of
figure 1 are linked to the inputs of two
IC's, either the 4050 as standard or
the 4049 for the inverted version. They
each contain 6 gates, so together just
the right number for the 12 channels.
Table 1 indicates the output current of
the 4049 and 4050 ICs at three different
supply voltages. Here it should be noted
that when the outputs are to produce
current (on output high source) the
maximum current will be considerably
less than a current 'sink'.

If buffer stages are not required IC5 and
IC6 can be omitted and the six wire
links indicated in figure 2c are mounted
in the sockets meant for them. The
output current is then as shown in
table 2.

Construction

Figure 3 shows the copper track layout
and component overlay of the 12
channel TAP board. Little need be said
about the construction, as this is simple
and should not be a problem.

1981

Since the points marked A and B in the
diagram have been made easily access-
ible on the board, the number of
channels can be extended without
difficulty. A second board can be
connected to the links and then of
course R1 . . . R5, PI, IC1 and T1 will
no longer be necessary. You don't
necessarily have to have as many as
twelve channels, either. In that case,
the parts for the superfluous channels
are merely left out and the inputs of the
gates that are not used are grounded.
If only 8 or 4 channels are needed, one
or two 4071 ICs may be omitted
altogether.

The touch contacts themselves can be
made in various ways (at least where the
mechanical side is concerned). The most
practical method is to etch a number of
finger-tip-size squares with enough space
between them to avoid mistakes during
operation. Better still buy the contacts
ready-made in the various sizes available.
For real economy, drawing pins will also
dol

The circuit can be fed by means of a
simple, stabilised supply. Current con-
sumption depends on the load at the
outputs. The value of the supply voltage
is not critical; this may range from
5 V ... 15 V. As shown in tables 1 and
2, this value will, however, affect the
output current. It should be noted that
the supply must not be connected to
the mains earth, as this will prevent the
switches from working at all.

A final remark

The circuit can easily be modified to
allow the outputs of all the contacts
touched to remain high until the circuit
is reset by way of another contact. If
D13...D23 and R42...R52 are
omitted and the wiper of PI is turned
clockwise until it can go no farther
(towards R4), contacts 1 ...11 can
each be made high in turn (or low, with
the addition of figure 2b) and then
contact 12 will act as the reset. This
could be very useful. M

L<]J L<jJL<jJ

THJHTHmiGHrjEF

? f f

y y y

1<11<11<I1

If III?

L<H

© ' G -

© 9 ©

T1 = BC547B
D1... D24 = 1N4148
IC1 = CA 3140
IC2.IC3.IC4 = 4071
IC5.IC6 - 4049 or 4050
(or 6 wire links)

Signal /noise i

<02% at 1 kHz

ratio: > 80 dB (DIN input I

> 85 dB (socket input)
noise reduction at 100 kHz: 15dB
at 3 kHz: 20 dB

at 15 kHz: 25 dB
DIN-A: 20 dB

DIN recording: 0.6 mV across 6 kS?

socket recording: 200 mV across 25 k SI

DIN playback: 1 30 mV across 79 kn

socket playback : 200nmV across 100 kfi

output voltage

DIN recording: 1 mV/kSl

socket recording: 600 mV (5k6 output impedance)

DIN and socket playback: 0-1 .5 V (5k6 output impedance)

die High Com noise
reduction system

The previous issue of Elektor included a general discussion on noise reduction systems and the problems
involved for the home constructor. Featured in the article was the Telefunken High Com System. The
present article is a continuation of the above and covers the constructional details of a complete noise
reduction system for the electronics enthusiast. The High Com modules, as supplied by Telefunken, have
been incorporated on a printed circuit board that includes all the extras required to construct a very high
performance system, such as a calibration oscillator and LED peak meter. The completed unit can be
connected directly between the main amplifier and a cassette deck or reel-to-reel tape recorder.

The continued search for better and obviously, and the disadvantage here is
better Hi-fi performance has created the fact that the ICs are only available
important developments in the field of to licensees, in other words, your
audio technology in recent years. Not component retailer will not have them
the least of these are the attempts to in stock. To solve this problem
improve the noise level in cassette Telefunken have made an exclusive
recorder systems. As discussed in the agreement to supply Elektor readers
previous article, a number of companies with completed, tested and calibrated I
have spent a great deal of money and modules which include the elusive
effort to try and obtain the best poss- U401B.

ible performance from this medium. The Elektor system is, of course, based
One of the leaders in this field is Tele- firmly on the Telefunken application
funken with their professional Telcom since this could hardly be improved
C4 system and the consumer oriented upon. However, the 'interface' elec-
version, the High Com system. The tronics between amplifier, compander
excellent specification of the latter and recorder have been subjected to
system has encouraged more and more some fairly extensive modifications,
recorder manufacturers to incorporate Electronic switches now take care of the
it into their cassette decks. entire signal conversion involved during

The High Com system also presents an recording and playback, thereby cutting
interesting possibility for the electronics down the number of long screened
enthusiast since all the principle active cables usually required. In addition, a
components have been integrated onto a calibration oscillator has been included
single chip, the U401B. This, of course, to allow accurate adjustment of the
means that the noise reduction system signal levels between the noise reduction
is now fairly straight forward to build, system and the recorder. A further
Where is the snag? There has to be one, refinement is the inclusion of a LED

the High Cc

arch 1981 - 3-07

peak meter which enables the recording
level to be monitored at all times.

Before going on to describe the com-
plete circuit a closer look at the theor-
etical details, which were mentioned in
passing last month, would help in a
better understanding of the Elektor
noise reduction system.

In theory

- The usable dynamic range of a magnetic
tape recorder is limited by the onset of
overdrive at one end and by the in-
herent noise level at the other end.
When a signal with a dynamic range
greater than the bandwidth of the tape
is recorded, there will be frequent
overdrive excursions at peak signal levels
and important low level signal com-
ponents will be lost in the ever present
noise.

This can be prevented by compressing
the signal level during recording, there-
by reducing the dynamic range, and by
expanding it during playback to bring
it back up to its original value. Provided
that precisely opposite operations are
performed by the expander, the original
dynamic range can be restored. This
ensures that the low level signal com-
ponents remain well above the noise, in
other words, the noise level is effec-
tively reduced. The upper threshold
level can be regulated by the operator
with the aid of (peak) modulation
meters. In any case, the dynamic range
is effectively increased, which is exactly
what is required in a cassette deck.

Figure 1 contains a simple block dia-
gram of a compander (compressor +
expander = compander). Blocks A and
B are voltage controlled amplifiers
where the transfer function can be
expressed as follows:

a™- ft

biu,>.£

From this it can bee seen that the
transfer function of amplifier A is
determined by its output voltage, and

• that of amplifier B by its input voltage.
The output voltage U 4 will have to be
equal to that of the input voltage U, for
the recording to be high fidelity. This
can also be expressed as:

B(U 3 ) = A‘(U 2 ).

In other words, the transfer function
of the expander must be the exact
inverse of that of the compressor. This
should not be a surprise to anyone!

The question now is how to obtain this
'reciprocal' transfer function. Use of
one of the numerous high performance
operational amplifiers that are readily
available these days should make it a
fairly easy matter to solve this problem.
If as in figure 1c, the expander is
included in the negative feedback path

Figure 1. A simplified block diagram of basic compressor expander circuits. The third illus-
tration shows how a compressor can be formed by placing an expander in the negative feedback

Figure 2. An illustration of the principle of three cascaded voltage controlled amplifiers with a
rectifier and a graph showing the characteristics of the circuit.

Figure 3. The block diagram of the High Com compander.

of the opamp, the following transfer
function will be obtained:

A_ 1 + A 0 B “ 1_ + B
A 0

Where A 0 stands for the open loop gain
of the opamp and B represents the
transfer function of the expander.
Assuming the open loop gain of the
opamp to be infinite, the transfer
function expression simplifies to:

This is precisely what we are looking for,
especially seeing that in practice the
open loop gain of a good operational
amplifier will be large enough to come
very close to this equation.

Figure 2 shows a cascade circuit of a
number of identical amplifiers as used
in both the High Com and the Telcom
C4 noise reduction systems. A total of
three VCAs are used in the Telcom
system and two are used in the High
Com system.

The output signal, U Q , of the third
amplifier is converted by a rectifier
into a control signal for all the am-
plifiers. The amplification factor. A,

3-08 - ale

1981

Flyura 5. Tha noita spectrum characteristics of a cassette tape without compander (1) and with
High Com 12 ).

of the amplifiers is controlled so that
the output voltage U 0 of the last
amplifier will remain constant. This
prevents the behaviour of the rectifier
from affecting the circuit in any way.
Thus, the compander characteristic will
not be determined" by the behaviour of
any particular rectifier, but rather by
the cascaded arrangement of the iden-
tical amplifier components.

Figure 2 also features the output
characteristics on a logarithmic scale. It
will be seen that they are remarkably
linear at each output. In the configur-
ation drawn here consisting of three
amplifiers the input signal will be com-
pressed by amplifier A! by a factor of
three. If the output signal has a dynamic
range of 90 dB, it will be reduced to
60 dB (curve U 2 ). The expander works
in exactly the same way as the com-
pressor, although of course, the other
way round.

The block diagram of the High Com
circuit is shown in figure 3. Two am-
plifiers connected in cascade are used.
The High Com system differs from the
Telcom C4 system in that is does not
operate on a number of frequency
bands, but on one band only. For this
reason it is called a broad band
compander.

Each compressor section is preceded
by a pre-emphasis circuit for the higher
frequencies. De-emphasis then takes
place in the expander in order to
compensate for this. To prevent the
tape from being overdriven at high
frequencies two further measures have
been taken. A fixed treble cut is per-
formed at the output of the compressor
together with a corresponding treble
boost at the input to the expander.
This is designed so that a 10 kHz
signal will not be attenuated unless it
has exceeded —8 dB (with respect to
full modulation). The graph in figure
4 shows that compansion does not
cover an unlimited range, but does in
fact have an upper and lower threshold.
As a result of the pre-emphasis and
de-emphasis that takes place, the
characteristics will not be the same
for every frequency.

One important feature of the High
Com system has yet to be mentioned.
The operation of the system is not
affected by idiosyncracies in the
frequency characteristics of the recorder.
Figure 5 illustrates the effect obtained.
Curve number 1 represents the noise
spectrum characteristics of a cassette
tape without the compander and curve
2 shows the characteristics for a cassette
tape with the High Com system in
circuit. As can be seen, the results are
quite impressive.

The High Com 1C

Constructing a circuit from the block
diagram in figure 3 using only discrete
components would be rather a laborious
task. Fortunately, Telefunken have
integrated all the necessary active com-

ponents for the entire compressor/
expander circuit onto a single mono-
lithic chip, the UB4G1. Only a few
capacitors and resistors have to be
added separately.

The circuit diagrams in figures 6 and 7
show the 1C and its internal structure
as given in the block diagram. This
enables us to see what happens to the
signal as it passes through the various
parts of the system. Figure 6 provides
the circuit diagram for playback and
figure 7 for recording. The resistors
and capacitors marked with an asterisk
need to have a tolerance of 2% and 5%
(or better) respectively. The power
supply requirement for the High Com
1C is 15V (pins2and 1).

The voltage gain of the internal low
noise amplifier A is fixed during manu-
facture at 30 dB. This amplifier is only
used during playback. The operational

amplifier following it is connected as a
non-inverting amplifier (also with a
fixed amplification factor). Finally, the 1
gain of opamps C and D is determined
by resistors R7 and R 1 1 . Using the
given values, their gain will be
approximately 5.6.

Expander

Now let us see what happens during
expansion (playback in figure 6). The
output signal of the tape recorder passes
to amplifiers C and D via amplifier B
and the RC network between pins 16
and 17. Resistors R8 and R9 and
capacitor C9, together with amplifier
D constitute an active low pass filter
for expander de-emphasis (block 2 in
figure 3). The integrated electronically
controlled potentiometer between pins
16 and 17 adjust the gain of the filter
stage. The external components R17,

the High Com noise reduction system

R18, Cl 3, Cl 4 and Cl 6 comprise the
treble boost during playback (block 5
in figure 3). The gain between pins 14
and 10 is equal to 1 at low frequencies,
when the total impedance seen between
pins 16 and 17 is 3 k. The capacitors
C11 and Cl 5 prevent DC fluctuations
caused by control resistance changes.

Control voltage

The control voltage for the circuit is
obtained from the output signal at pin
1 6. This signal is fed to amplifiers E and
F via capacitor Cl and resistor Rl.The
gain of the inverting amplifier E is
determined by the ratio of the parallel
combination of R2 and the second
integrated potentiometer to R1. It
follows that the gain will decrease with
the resistance of the electronic potentio-
meter. Subsequently, amplifier F boosts
the signal by a factor of ten. The output
of opamp F (pin 22) is routed to the
rectifier amplifier G via a passive high
pass filter which combines the functions
of the networks (1) and (3) from
figure 3. Since the rectifier operates at
the centre point voltage, pin 24 must
also be tied via R4 to the centre point
voltage at pin 23.

Rectifier

The rectifier produces the gain control
voltage according to the full wave
threshhold principle. When the voltage
at pin 24 differs by more than ± 70 mV
DC from the centre point voltage at
pin 23, a current sink circuit is switched
on at pin 6. The current delivered inter-
nally to pin 1 is proportional to the
threshold overshoot at pin 24 until the
maximum value of 2.5 mA is reached,
and is used to discharge the storage
capacitor C7 at pin 6, thus changing the
voltage at this pin. Capacitor C7 is re-
charged towards more positive voltages
from the reference voltage source at
pin 4, via the R6/R7 network. The
equilibrium at pin 6 will be reached
when the integral current flow into pin
6 produced by threshold overshoots is
equal to the charging current delivered
through R6 and R7,

The control voltage throughout the
entire circuit is determined by both the
input voltage of the expander and the
output voltage of the compressor,
since both are related through the gain
of amplifier E to the constant input
voltage at pin 24. The gain of opamp
E is determined by the value of the
second integrated potentiometer (con-
nected between pins 18 and 20), which
in turn is dependent on the control
voltage. As a result, the control voltage
will depend on the input signal of the
circuit, in spite of the constant voltage
to the rectifier. The control voltage
range for amplifiers C and E is designed
to be 30 dB.

The shortest response time of the
rectifier to a positive step change of
input voltage is determined by the value
of the storage capacitor C7 and the

maximum current available from the
current sink at pin 6. This time reaches
0.3 ms for full swing of the gain control
voltage. Since virtually all natural
sound effects have longer transient
times, the compander is thus capable
of accurate processing and reproduction
of such transients.

The gain control voltage decay time
depends on the values of C7, R6 and
R7. Obviously, a short decay time is
preferable for the compander to operate
well, but on the other hand this may
cause signal distortion at low fre-
quencies. For this reason, a delay
circuit has been added in the form
of a retriggerable monostable muiti-

In the passive state, pins 4 and 5 are
linked internally. When ever the signal
exceeds the threshold value of the
rectifier, the MMV is triggered and the
switch between pins 4 and 5 is opened.
Thus, the decay time is determined by
C7 and R7 when a signal is present
and the values of these components
have been chosen to keep distortion
at low frequencies to a minimum.
If the input signal terminates abruptly,
resistors R6 and R7 will be connected
in parallel, thereby reducing the decay
time. If the duration of the input signal
is less than that of the MMV the decay
is further reduced by means of C21.
When the input signal returns, the
charging current continues to flow
through R6 for a short while so that
the decay time does not increase abrupt-
ly. but rather rises slowly through C21 .

Compressor

Little needs to be changed for the 1C
to be used as compressor (see figure 7).
Basically, all that is required is to
incorporate the expansive section
(between pins 15 and 10) within the
negative feedback loop of amplifier B.
Provided the impedance at pin 12 is
sufficiently low, the interval feedback
resistance between pins 12 and 15 will
no longer affect the amplification
factor of opamp B. This condition is
accomplished by ensuring that capa-
citors C8 and Cl 8 have a relatively
large value.

That covers just about all that is re-
quired concerning the High Com 1C
and its associated components. The
circuitry inside the dotted areas of
figures 6 and 7 has been mounted on a
separate board. This also includes four
CMOS switches yet to be mentioned
(ESI, 2, 3 and 4). These serve to switch
between a resistor and two capacitors
(R16, CIO and C12) enabling Dolby
cassettes to be played. Real Dolby
expansion is of course too much to
expect, but the circuit comes very
close to achieving it.

Circuit diagram

Very little needs to be added to the
High Com 1C and its associated com-

elektor march 1981 - 3-11

ponents (see inside the dotted area) in
order to produce an excellent noise
reduction system. For the sake of
clarity, the entire circuit diagram has
been split into two sections in which are
shown the components required for
recording and playback, respectively. In
both instances, the High Com module is,
of course, employed. A choice of input
and output connectors are shown on the
drawings, either DIN or line connectors
are suitable. However, both may not be
used at the same time.

Recording

During recording (figure 7) the signal
enters via the line or DIN connector.
Since the DIN record output of an
amplifier is only a few millivolts in level,
an additional amplifier stage will be
required for the DIN input. This is con-
structed around transistors T1 and T2
and gives a gain of around 70. Poten-
tiometer P3 is used to adjust the record
level and is effective on both inputs.

The signal is then fed to opamp A3
which has a gain of five. The signal is
then passed through CMOS switch ESI 4
to a high pass filter. This consists of a
notch filter (amplifier A4 and corre-
sponding components) and of a passive
6dB per octave filter. Together they
constitute a subsonic filter with a turn-
over frequency of 19 Hz and a decay
time of around 24 dB per octave. This is
included to prevent interference from
low frequency signals during calibration.
The multiplex filter BL30-HR (or HA)
following it suppresses any 19 kHz pilot
tones during the recording of FM broad-
casts. The signal then passes to ampli-
fier B of the High Com 1C via switch
ES6.

Amplifier A is not used during recording
as the input signal level is of sufficient
amplitude to drive amplifier B directly.
As previously mentioned, the U401B
operates as a compressor during re-
cording, since the module connection
B5 is linked directly to module connec-
tion A4 via switches ES10 and ES11
connected in parallel. Bearing in mind
that CMOS switches do have some
resistance, albeit very low, two of them
are connected in parallel to ensure that
the resistance between B5 and A4 is as
low as possible. The compressed signal
from output B6 is then fed to the line
input of the tape recorder via the usual
DIN or line connector.

A calibration oscillator is used to set up
the circuit. This consists of a Wien
bridge oscillator composed of IC6 and
the components around it. Diodes D200
and D201 stabilise the output voltage of
the calibration oscillator. Switch S3 is
used to select either the input signal or
the calibration signal of 400 Hz. The
calibration procedure itself will be dealt
with in greater detail later.

Playback

Figure 6 illustrates the playback circuit
diagram. Compared to the recording

version it is very straight forward. The
output of the tape recorder is connected
to the line or DIN input socket. The
input signal level is preset by poten-
tiometer PI before reaching amplifier A.
The output of this amplifier is connec-
ted to the input of amplifier B by way
of switch ES8. In the playback mode,
the U401B is used as an expander (the
link between module connections B5
and A4 no longer exists). The expanded
output signal then fights its way
through to the output at B5 and is
passed to the buffer amplifier A2. The
signal can now be fed to one of the
main amplifier's line or DIN inputs
(whichever is available) by way of one
of the two output sockets. The tape
sockets in the lower right hand corner
of the circuit diagram are included so
that transfer to a second tape deck is
possible.

The on/off switch for the High Com
system, S4, is shown in both circuit
diagrams. Switch ES12 closes when the
system is switched from recording to
playback and effectively short circuits
the internal feedback resistor of
amplifier B. At the same time, switches
ES10 and ES11 are opened, thus
breaking the connection between B5

and A4. Thus, neither compression nor
expansion will occur.

Switch S2 selects between High Com
and DNR during playback. DNR gives a
similar result to that of DOLBY. In
other words, cassettes taped with
DOLBY can be replayed on the High
Com system. The DNR position is not
functional in the record mode even
though S2 is included in both circuit
diagrams.

Peak meter

The circuit diagram of the peak meter is
given in figure 8. This is a design that
was first published in the January 1978
(E33) issue of Elektor with the printed
circuit board number EPS 9827. The
component values have of course been
modified to suit this particular appli-
cation. The circuit around amplifier A1
constitutes a peak detector which
measures the peak amplitude of the
input signal. The circuit around A2 is
the converter that provides a linearly
increasing voltage with logarithmic
display using LEDs controlled by the
well known UAA 180 1C. Preset poten-
tiometer PI sets the sensitivity of the
modulation meter. The input of the

peak meter is connected to module
output B6, via resistor R41.

Power supply

The power supply (see figure 9) for the
Elektor compander makes use of
integrated voltage regulators. This
circuit although simple is very effective.
The High Com 1C and the peak meter
require +15 V and the remaining op-
amps are fed with ±8 V. LED D5 has
been included as a mains 'on' indicator.

Construction

The printed circuit boards for the
various circuits are shown in figures
10 . . . 14. The system consists of a
main board and two modules. In ad-
dition there is a board for the peak
meter, two boards for the LED display
and finally the power supply board. The
High Com modules include the U401B
and are supplied ready built (more
about this later).

As is usual, all the printed circuit boards
and their components should be
assembled and checked. Then, before
any wiring takes place, the finished
boards, mains transformer, switches and
sockets etc. are fitted into the case.

elektor march 1981 -3-13

Sockets can of course be DIN or phono
to be compatible with the rest of the
audio system. Having made sure that all
the hardware fits together, wiring can
commence. The switches can be wired
as shown in figures 6 and 7 using only
ordinary cable. Since the control
switching is electronic only DC levels
are involved and therefore screened
cable is not necessary. The big advan-
tage with this arrangement is that the
earth loop problems that usually occur
with screened leads do not arise.

The pin numbers for the DIN socket
connections are given in the circuit
diagrams (figures 6 and 7). These refer
to the left hand channel, the right hand
channel being the numbers in brackets.
A total of two DIN or eight phono
sockets will be needed.

The illustration in figure 15 shows the
design of the front panel layout. Due to
practical reasons it is not reproduced
here in full size.

Before connecting the power supply to
the rest of the circuit, it will be as well
to first check to ensure that it operates
correctly. It should supply output
voltages of plus 15 V and plus and
minus 8 V. If all is well the power
supply connections can be made to the
printed circuit boards, after discon-
necting from the mains . . .

The CMOS switches, IC's 3, 4, 7 and 9
can now be inserted and checked (with
the power on). Although they look like
IC's (because, of course, that is exactly
what they are) they really do operate
like a normal switch and can be checked
in the same manner. Practical proof can
be provided by connecting the ohm-
meter probes to the pins of the elec-
tronic switch to be tested and operating
the corresponding panel switch. The
meter reading should change from zero
resistance to infinity (or vice versa). Pin
numbers are again given in figures 6 and
7 with the right hand channel in
brackets. When ES8 and ES21 are
tested, one side of each of these elec-
tronic switches will have to be tempor-
arily grounded with a 10k resistor.

If all goes well, the remaining IC's can
be (turn the power off first!) placed in
their sockets. The voltages at the supply
pins of the IC's should now be measured
just to be on the safe side.

The peak meter and display board
supply can be linked up next and its

9

Figure 9. The circuit diagram for the pov

10

the High Com module:

Resistors:

R1 = 1 k5/2%
R2.R14 =15 k/2%
R3 = 47 k/2%
R4.R9 = 5k6/2%
R6 - 820 k/2%

R8 = 33 k/2%
R11.R12- 10k
R16 = 3k3
R1 7 = 1 k5/2%

R18 = 56 ft

Capacitors:

Cl = 22 M/6V3
C2 = 4p7/16 V
C3.C9 = 3n3/5%

C4.C13 = 1 n
C6 = 680 n/5%

C7 = 220 n/5%

C8.C1 1 ,C1 8 = 47 m/1 6 V
C10= 1n2/5%

Cl 2 = 68 n/5%

C14 = 10 n/5%

Cl 5.C23 = 2p2/1 6 V
Cl 6 = 33 n
Cl 7.C20 =10 p/1 6 V

C19 = 1 50 n
C21 = 15 n/5%

C22= 100 p/16 V

Semiconductors:

IC1 = U401BR
IC2 = MC 1 4066, CD 4066B,
HEF4066B

High Cc

jise reduction syster

elektor

-3-15

C31 !ci 31 .C34.C1 34 = 2<i2/25 V

C32.C132 = 22 m/25 V tantalum
C33.C133 » 100 m/ 25 V tantalum
C35.C1 35.C36.C1 36 = 4m7/1 6 V
Semiconductors:

D200.D201 = DUG
T1.T101 = BC550B
T2.T102 = BC 550C
T3.T103 = BF 256
IC3.IC4.IC7.IC9,

IC10 = MC 14066, CD 4066B.

HEF4066B
IC5.IC8 = RC 4136
IC6 = 741

Miscellaneous:

Fi1,Fi2 = BL30-HA Toko

51 3354 =■ double pole toggle

52 = single pole switch

input connected to the main board.
With a tempory link between B3 and
B6, switch SI to RECORD and S3 to
TEST, the LED's should light. If this is
satisfactory (turning P200 helps), both
the peak meter and calibration oscillator
will be .O.K. The link between B3 and
B6 may now be removed.

Finally, the High Com modules are left
to be dealt with. With the mains supply
disconnected the two modules can be
fitted with their component side facing
the main board. Before fitting the top
panel of the case, the complete unit
must be calibrated.

Calibration

This is a simple matter and does not
even require the use of a meter. The
power is switched on and SI is set to
RECORD, S2 and S4 to High Com and
S3 to TEST. All the potentiometers and
presets are set to the centre position.
The LED's of the peak meter should
now indicate something (if necessary,
adjust the sensitivity of the meter with
PI and P'1 on the peak meter board).
The preset P200 of the calibration
oscillator must now be set so that the
meter gives a clear indication. Switch
the High Com off with S4 and check for
any alteration in the meter reading. By
careful adjustment of P200 in both
positions of S4, a setting should be
arrived at where the meter reading
remains the same for both positions of
the switch. Once this setting has been
achieved, the meter reading can be set
to zero dB with the preset poten-
tiometers PI and P'1 on the peak meter
board.

For the idealist, the calibration may be
made more accurate by connecting a
millivoltmeter to pin B6 (careful, AC
here) and then switching S4 to and fro.
The meter will then give a clear indi-
cation of the difference in readings
between when the High Com is switched
on and off.

At last the moment has arrived for the
recorder to be connected to allow a test
recording to be made. The switches will
be in the following positions: SI to
RECORD, S2 to High Com, S3 to TEST
and S4 to High Com. The relevent
connecting cables between the tape
recorder and the noise reduction unit
are fitted. The recorder is switched to
record and its recording levels are
adjusted to give a reading of 0 dB on the
record level meters. From now on, the
recording controls of the tape/cassette
deck should not be moved at all. The
potentiometers on the noise reduction
unit will be used for any recording level
adjustments that are required.

Now a test tone can be recorded on
tape, two minutes should be long
enough. The noise reduction unit is then
switched to playback (SI to PLAY, S3
off) and the recorded test tone played
back. Using the preset PI on the main
board (seperately for each channel) the
reading of the peak meter must again be
set to 0 dB. The potentiometer P2

enables the output level to be adjusted
to suit the input sensitivity of the
amplifier in use.

And that is it, the unit is calibrated! The
lid can finally be fitted on the case.

A 'normal' recording can be made and
listened to. Again, don't forget to use
the controls of the noise reduction unit
for any level adjustments and not those
of the record deck itself otherwise the

calibration procedure will have to be
repeated. Make a test tone recording . . .

If a second tape/cassette deck is to be
used, the recording level controls of this
deck will have to be adjusted in the
same manner with the aid of the cali-
bration oscillator. The test tone is again
recorded and PI on the main board is
calibrated during playback.

The noise reduction system in use.

Once calibration has been completed,
the unit is very simple to use. The
various switches operate as follows:

SI: RECORD/PLAY.

S2: High Com DNR switch. W.hen DNR
is selected, Dolby cassettes or tapes
can be played back.

S3: TEST. Always switched off during
normal use.

S4: High Com on/off switch. In the off
position for playing 'standard'
recordings (also required during
calibration).

S5: Mains on/off switch. Should be in
the 'on' position for optimum

The recording controls set the levels
during recording with the aid of the
built in peak meter. Only SI has to be
switched for playback.

The High Com modules

To allow the High Com system to
become a viable project for our readers,
the Elektor editorial staff have reached
an agreement with Telefunken. The
High Com system will be available only
from Elektor in the form of completed
and calibrated modules. All components,
IC's, precision resistors and capacitors,
inside the broken line in figures 6 and 7
are included. The main board together

with two modules and a self adhesive
front panel will be available through the
Elektor print service (EPS). Additional
items required are the components for
the main board and the necessary
hardware. The peak meter and power
supply printed circuit boards and their
corresponding components must of
course also be added for the High Com
noise reduction system to be complete.

M

play /highcom nr%

•o o*

compander

elektor

r? ^

power — • test

3-18 - elektor

1981

logic analyser

logic JlllilljMT

indispensable aid to digital fault finding

The analysis and comparison of digital signals is not an easy task without
the aid of an expensive item of test equipment, namely a logic analyser.
Unfortunately, these are invariably priced well beyond the reach of the
average enthusiast with the result that fault finding, especially on
computer systems, can result in a succession of inspired guesses.

All is not lost however, since an ordinary oscilloscope can be coupled
with the project featured in this article to produce a reasonable logic
analyser.

Photograph 1. An enlarged view of the data stored in RAM. The small spots on the lines are
the indication of the logic level at that point.

Readers who work with digital circuits
regularly, and especially with micro-
processors, know that their oscilloscope
is absolutely essential if reliable infor-
mation is to be obtained. However,
complicated circuits require a lot more
than the one or two channels that
the average oscilloscope has to offer.
A microprocessor with eight data and
sixteen address lines would need a
whole bank of oscilloscopes since fault
tracing in this area would require that
all the lines are monitored at the same
time. After all, processors operate on
bytes and each one consists of eight
'bits' worth of parallel information
(disregarding the new 1 6 bit processors
for the moment).

It would be fairly simple to design an
eight channel trace 'switcher' for the
oscilloscope to provide eight lines on
the screen simultaneously, but this
would be rather pointless. Pointless
because, with data continually changing
at high speed the information is lost
within microseconds. Thus, a memory
is required where the series of digital
signals can be stored before they are
read out on the scope.

We also have to know which infor-
mation the computer is to record at the
data inputs. If, for instance, not more
than twenty bytes can be stored in
memory, whereas the entire program
consists of over a thousand, tracing the
bytes in question will be like looking
for a needle in a haystack, unless there
is some sort of device to help us. Test
equipment manufacturers realised this
long before now and came up with the
logic analyser. This instrument is a
combination of an oscilloscope and a
digital test and memory circuit. Unfor-
tunately, it can cost as much as £ 2000,
well beyond the reach of most of us.
Elektor's design team felt it was time
something was done to remedy the
situation. Their logic analyser is a trace-
switching circuit that can be connected
to an ordinary oscilloscope. The Elektor
version is by no means simple and the
components can total up to quite a few
pounds, but this is still cheap, con-
sidering the quality and potential
offered.

Since the circuit diagram is rather com-
plicated, the discussion of the practical
design, its possibilities and operation
will be left until a following issue and
this article will concentrate upon the
principle it is based on.

To start with the simple block diagram
in figure 1 meets the requirements
which have been dealt with so far.
Firstly, there has to be a memory to
store a certain amount of eight bit
parallel information, a few hundred
bytes, for instance. Next the circuit has
to be informed when to start reading in
and this is carried out as follows. The
memory stores data continuously. An
eight bit word can be preset in the
trigger unit. As soon as this word is
recognised in the input signal, the
trigger unit will generate a pulse. This
activates a counter which will produce

logic analys

1981 -3-19

a 'stop' signal after a certain period.
When this occurs the reading-in stops
and the data stored in memory can now
be shown on the screen of an oscillo-
scope. This requires a control device to
process the digital signals so that all the
bytes appear on the screen in a clearly
visible manner.

Reading the display is made easier with
a cursor. This moves across the screen
and will always indicate the eight bits
belonging to one byte at a time in order
to avoid any reading errors.

It will be obvious that the block diagram
in figure 1 is a very much simplified
version of the original. A great deal
more is involved in actual fact, as the
full block diagram of the logic analyser
will later show.

How does it work?

In order to discuss the basic principles
of the logic analyser in greater detail
a more elaborate block diagram is given
in figure 2.

Initially the two flipflops FF1 and FF2
are reset so that their Q outputs are at
logic zero. A clock oscillator combined
with a presettable divider generates the
clock pulses for an eight bit counter A,
the outputs of which provide the
address code for a 256 x 8 RAM. The
digital signals that are to be sampled,
DO . . . D7 are written into memory at
the clock frequency by means of an
8 bit latch.

After the 255th pulse, the counter is
reset, starts to count from scratch again,
and the memory once more starts to fill
with incoming data. When a trigger
pulse is generated, FF1 changes state
causing counter B to start counting. The
initial state of this counter can be preset
with the 'trigger mode' switch. In the
'post trigger' position, the initial state of
counter B will be zero. In the 'centre
trigger' and 'pre trigger' positions it will
be 126 and 255 respectively. The pos-
ition of this switch will determine
whether the final contents of the RAM
will consist of data stored after, before
and after, or before the trigger pulse was
generated.

Depending on the initial setting of
counter B, a certain number of clock
pulses will be required to 'fill' it and
generate a carry pulse. The carry out
from the counter at that point will set
FF2 thereby preventing any further new
data from being read into the memory.
For example, when the trigger switch is
in the 'post trigger' position, the write
operation of incoming data into the
memory will continue for 256 clock
cycles before the write cycle is halted.
In other words, the 255 bytes of data to
enter after the trigger pulse is generated
will be stored in the memory. In the
'centre trigger' position, 126 bytes be-
fore the trigger pulse and 129 after it
will be stored, and in the 'pre trigger'
position the 255 bytes before the trigger
pulse will be stored. This facility is of
course very useful and well worth the
few components it requires.

1

Figure 1. A very simple block diagram of the logic analyser showing the relationship of the
memory, triggering, clock and oscilloscope control.

It may be as well to make a small di-
gression and see where the trigger pulse
originates from. There are in fact three
methods which can be used to produce
it. The first, and probably simplest
method is by means of an external
trigger signal. Perhaps some point in the
circuit under test can be used to furnish
a pulse at the correct point in time that
it is required.

Secondly a trigger pulse can be derived
from incoming data which as expected
proves to be somewhat more compli-
cated. The final possibility is a combi-
nation of both of the foregoing choices.
For the last two options a 'word recog-
niser' will be required. This, as its name
suggests, is a circuit that can recognise a
(preset) ten bit word when (or if in some
cases) it occurs in the incoming data.
Since all the input data is entered into
the eight bit latch initially, it is a fairly
simple matter to generate a trigger pulse
when the contentsof the latch equals the
preset word in the word recogniser.

Back to the memory input. Now the
contents of the RAM have to be read
out and displayed on the oscilloscope in
a legible manner.

When FF2 was set it simultaneously
caused switch S2 to be activated, as a
result of which the entire system was
switched from the preset clock fre-
quency to a fixed scan frequency. For
this the monostable multivibrator MMV
will be pulsed on every carry out signal
generated by counter B. This ensures
that the clock oscillator is inhibited for
the time period of the MMV in order to
prepare the time base of the oscillo-
scope for a new line trigger. Once this
has passed, the contents of counter C
are incremented by one and at the same
time, a trigger signal is issued to the
oscilloscope. A line is then displayed on
the screen with a vertical position deter-

mined by the state of counter C. The
outputs of this three bit counter control
a D/A converter which is connected
directly to the Y input of the oscillo-
scope.

After the trigger pulse, counter A con-
tinues counting and the data now stored
in the RAM are passed to a multiplexer.
Provided the contents of counter C
remain unchanged, the multiplexer will
pass on a single bit of each byte at a
time to the LSB input of the D/A con-
verter. In this way all the 256 bits of
data in a single input line are transfered
to the D/A converter and then written
onto the screen. In the event of a logic
one, the voltage level at the Y input of
the oscilloscope will be increased
slightly, whereas a logic zero cause the
level to remain at a constant level
depending on the contents of counter C.
This therefore enables the digital infor-
mation of a complete data line to appear
on the screen.

Now, how is the line height determined?
If the contents of counter C are '000'
the voltage output level of the D/A con-
verter will therefore be 0 V and the line
will be traced along the bottom edge of
the screen. The multiplexer will then
switch data line D7 of the RAM through
to the D/A converter for the period of
time that it takes for the data line to be
read out.

After 256 clock pulses all the infor-
mation on line D7 will have been writ-
ten out and counter B will then produce
a carry signal and the MMV will be trig-
gered. At the end of its time period,
the contents of counter C will be in-
cremented by one and at the same time
the oscilloscope will be triggered. The
line now being written on the screen
will be slightly higher up than the pre-
vious line (as the contents of counter C
are now 001 ). As before the multiplexer

connects the RAM data output line to
the D/A converter only this time it is
data line D6. All the information on this
line will now appear on the screen on
the second line up.

The above operation is repeated until
there are eight lines on the screen corre-
sponding to the eight data lines of the
memory. This is the whole of the
memory in fact, eight lines each having
256 bits. After the entire read out cycle
is completed, it starts from the begin-
ning again and is repeated.

Figure 3 shows the voltage levels that
are fed to the screen from the D/A con-
verter. The upper signal is the trigger
input and provides a pulse for each new
line to be written. The Y input wave-
form consists of stepped voltage levels.

one step for each 256 bit data line from
the memory. The lower illustration
shows the corresponding oscilloscope
screen, each line corresponding to a step
at the Y input. In this example only a
few of the data bits are shown, in actual
fact each line can contain up to 256 bits
across the screen.

Once the reader becomes familiar with
the above operation it begins to lose its
apparent complexity. However, we still
have a long way to go.

So much for the logic analyser. Now let
us look at the auxiliary devices which
have been added for the user's benefit.

The cursor

Although the screen now contains all

the data from the RAM, eight lines with
256 bits each add up to a considerable
amount of information for such a tiny
scope. You don't need to be short
sighted to find reading an 8 bit word
from amongst that lot a strain on the
eyes! And after all that's what it's all
about. For this reason the logic analyser
includes a useful aid which is shown in
figure 4: the cursor.

The cursor acts both as a pointer and as
a hexadecimal display of the byte being
indicated. It consists of a pair of LED
displays each connected, via a seven
segment converter and buffer, to four of
the eight data output lines of the RAM.
The information appears on the display
in hexadecimal form, the first display
corresponding to data on lines D4 . . . D7

logic analyser

rch 1981 -3-21

and the second to lines DO . . . D3. When
data is being read into the RAM the dis-
plays are switched off by the Q output
of FF2, which will be at logic zero. The
displays will not be lit until the data is
written onto the screen.

The actual pointer is made up by the
cursor control, counter D and an 8 bit
comparator. The position of the cursor
can be controlled as required by coun-
ter D which is in turn controlled by the
left/right buttons. The comparator
compares the contents of counter D to
that of counter A (which provides the
RAM address code). When the two are
equal, the comparator generates a pulse
which is fed to the Z modulation input
of the oscilloscope and a spot will
appear on each line on the screen.

The comparator output pulse is also
used to latch the eight data lines of the
address concerned into the seven seg-
ment converters and the data will
appear on the two displays in hexa-
decimal form. If the oscilloscope does
not feature a Z modulation input, the
cursor will appear as a 'dimple' in the
data line.

After every 256 clock pulses the con-
tents of counter A will again equal that
of counter D and the comparator will
generate another pulse.

The comparator output is also used to
alternate the displays, a simple form of

) multiplexing, in order to maintain the
current passing through the display at a
safe level.

, In this way a very useful indication is
obtained. A vertical row of eight spots
appear on the screen, one for each data
line, and at the same time the byte
indicated is shown in hexadecimal
format on the display. It is a simple
matter then to move the 'cursor' left or
right on the screen with the two buttons
until the byte in question has been
found.

More to come . . .

It will be readily apparent now why a
block diagram is so important to the
description of circuit operation in a
complicated circuit such as that in this

Figure 4. The block diegrem of the cursor control. This includes the two LED displays for the
hexadecimal raadout. The RAM. FF2 and counter A sections are the same as those shown in
figure 2.

logic analyser. It must be noted that,
as yet, we have still only covered basic
principles in this article. However, this
has the advantage that the in depth
details of the circuit diagram in the
following article can be started with a
good idea of what should be happening
as we progress through the various

Perhaps it would be as well to mention
that this particular design is fairly
extensive and as such may not be an

ideal project for beginners, but where
there's a will there's a way ... A logic
analyser must, if it is to be of any
significant use, operate at fairly high
frequencies and therefore a great deal
of care must be taken during its con-
struction. Readers who are familiar with
microprocessors should not experience
too many problems in assembly. For
those who are still a little nervous we
will include some constructional tips in
the next article. M

multiple
sound effects
generator

a single 1C full of sound surprises

Yes, a single 1C is all (well practically all) that is
required to produce a cacophony of sound effects.

The device is able to imitate virtually every sound
under the sun, from the twittering of birds to
machinegun fire, from the sound of a plane
overhead to the high pitched squeal as it plummets
to the ground out of control, from the screeching

multiple sound effects generator

A great many letters have been received
in response to the various sound effects
generators that have been published in
Elektor over the years, which shows
that such circuits are still very popular.
However strange or frivolous the sound
may be, people always seem to be able
to find plenty of uses for it. Sound
freaks will be delighted to know, there-
fore, that Texas Instruments have devel-
oped an 1C specifically for them, the
'complex sound generator' SN 76477N.

It comes in a 28 pin DIL package and *
contains all the ingredients required to
serve up a whole menu of interesting .
and refreshing sound effects. A resistor .
here, a capacitor there and a couple of
transistors constitute the only other
components required. The latter simply
amplify the output signals to a level
sufficient to drive the loudspeaker.

This 1C is by no means new. In fact it
was first described in the September
1978 edition of Elektor under the
heading Applikator. We were then
mainly interested in the theoretical and
technical aspects of the device. Now,
however, we shall concern ourselves
more with the practical side of things.

The 1C was found to produce a total of
seven basic (and different) sound effects.
There are, therefore, a minimum of
seven different circuits involved, but as
they all have certain things in common,
it was possible to design a single printed
circuit board on which they can all be
incorporated. Each individual effect
requires a slight modification to the
combination of components used.

of brakes as a car runs out of road to the inevitable
crash ... All in all an amazing repertoire of
everyday (?) sound effects.

tiple sound effects gar

1981 - 3-23

The 1C

Since the 'internals' of the 1C have
already been discussed in detail in the
September 1978 issue, it will suffice to
give a brief survey of the most import-
ant features. Readers who wish to delve
a little deeper into the technology
involved are referred to the article
mentioned above. It includes all the
various formulae and tables that are
of interest.

4 The block diagram of the 'complex

[ sound generator' 1C is given in figure 1,
together with a few external com-
ponents that are required. A closer
examination reveals that there are three
fundamental signals produced. These
signals are obtained from: the Super
Low Frequency oscillator (SLF), the
Voltage Controlled Oscillator (VCO)
and the noise generator.

The SLF section contains an oscillator
which covers the super low frequency
range between 0.1 Hz to 30 Hz. Under
special circumstances it can also be used
at higher frequencies. The frequency of
oscillation is determined by the values
of Resistor R s and capacitor C s (con-
nected to pins 20 and 21 respectively).
Also shown is the fact that the SLF
oscillator provides two output signals.
The first is a squarewave signal which
is processed by the mixer stage and the
second is a triangular waveform which
can be used to control the VCO by way
of the external VCO/SLF select
section.

The VCO block consists of an oscillator
whose frequency is totally dependent

on the input voltage. This can be either
the SLF oscillator output signal or an
external signal applied to the U p input
at pin 16 of the 1C. Which one it is to be
is determined by the logic level at the
VCO select input (pin 22). In addition,
a signal at the U p input enables the
VCO output to be frequency modulated.
The voltage at U v (pin 19), on the other
hand, affects the duty cycle of the
squarewave produced by the VCO and
thus the timbre of the resulting audio
signal. The free-running frequency of
the voltage controlled oscillator is
determined by the external com-
ponents R v and C v (pins 18 and 17
respectively).

The pseudo-random white noise gener-
ator is triggered by the noise clock
whose internal current level is deter-
mined by the value of R c (pin 4). The
resultant signal is then passed through
the noise filter. The turnover point of
this low pass filter can be altered by
selecting different values for the com-
ponents R n and C n (pins 5 and 6 res-
pectively). Alternatively, the nature of
the noise signal can be changed by
applying an external clock signal to the
input at pin 3.

So far so good. Now for the remaining
sections. A logic 'one' level at pin 9
causes the system enable logic to
suppress the output signal from pin 13
of the 1C . . . When this same input is

SN76477

ANALOGUE LEVELS

Figure 1 . The block diagram of the 'complex sound generator'.

taken low, the one shot (monostable
multivibrator) is triggered. This is used
to generate 'single' sounds such as that
of gunfire. The duration of the output
pulse of the one shot is determined by
the values of resistor Rt and capacitor
Ct (pins 24 and 23) and can be anything
up to a maximum of ten seconds.

The output signals from the SLF
oscillator, the VCO and the noise gener-
ator are all fed to the mixer stage.
According to the logic levels presented
to the mixer select inputs (pins 25, 26
and 27) one, or a combination of, the
three signals are passed on to the next
section, this being the envelope gener-
ator/modulator. Here the mixer output
signal is amplitude modulated with
either the VCO output signal or the one
shot depending on the logic level
applied to the envelope select inputs
(pins 1 and 28). If the VCO signal is
selected, both amplitude and frequency
modulation are possible.

Finally, we come to the output am-
plifier. The gain of this stage is deter-
mined by the values of resistors Rf
and Rg (pins 12/13 and 1 1 respectively).
In the given circuit examples this am-
plifier is immediately followed by a
symmetrical output stage made up from
a complementary pair of transistors.
These in turn drive the loudspeaker.

The sound effects

Clearly, the SN 76477N 1C is extremely
versatile. The signals that can be gener-
ated by component combinations and

modulation possibilities are more than
sufficient to produce a whole host of
aural phenomena.

Seven different basic variations were
constructed and a universal printed
circuit board was designed accordingly.
There is no need for readers to restrict
themselves to the examples given here
however. Feel free to experiment with
different values of resistors and capaci-
tors to modify and in some instances
even improve on the effects.

One major advantage is that all the cir-
cuits can be supplied from a single 9 V
battery or from an unstabilised power
supply. This is because the 1C contains
its own internal regulator (not shown
in figure 1, so you'll have to take our
word for it!!) which derives a stable
5 V from the original input voltage
(pin 14 : 9 V in, pin 15 : 5 V out). Ob-
viously, the current consumption will
depend on the volume of the output
signal, but should not exceed more than
about 20 mA.

The various circuits will now be re-
viewed briefly, regarding the 1C itself
more or less as a ‘black box', so as not
to have to go into too much detail for
each variation.

Siren (figure 2)

This is a nice straightforward circuit
for a start. Only three resistors and two
capacitors are involved, in addition to
the loudspeaker amplifier which is the
same for all the examples and is connec-
ted to pins 11... 13. The VCO is

controlled by the triangular waveform
from the SLF oscillator. This results in
a characteristic siren effect provided
the frequency of the SLF oscillator is
kept low (by means of potentiometer
PI). If, however, the pitch of the SLF
oscillator is increased, the effect is
similar to that of a science fiction space
rocket 'zipping' through unknown
galaxies.

The effect can be modified by changing
the frequency of the VCO by means of
R4andC4and that of the SLF oscillator
with C3, R3 and PI.

Gun shot (figure 3)

The sections we have just mentioned
(VCO and SLF oscillator) are not re-
quired for this effect. White noise,
together with a sharp attack, is what is
needed for the gun shot. When switch
S2 is pressed the one shot is activated
via a negative going pulse on pin 9. Low
frequency noise is fed to the output via
the envelope generator/modulator. The
attack and decay time constants are
determined by the components RIO,
R1 1 and C6. The sound of the gun shot
is made more realistic as the noise gener-
ator actually produces a sound similar
to that of thunder. By repeatedly de-
pressing S2 the sound of machinegun
fire can be simulated.

Explosion (figure 4)

Basically the circuit is identical to that
of the gun shot, only this time the noise

R3.R4.R12 = 100 k
R7.R9 = 470 k

R13 = 47 k
R14 = 3k9
PI = 1 M preset
P3 = 4M7 preset

C4 - 2n2

C5 * 100 m/10 V

C7 = 100 m/10 V

T1 =■ BC 547
T2 = BC 557
IC1 = SN 76477 (Texas)

miscellaneous:

S4 = single pole switch
loudspeaker 8 S7/0.2 W
wire links: B2. B3. B5. B6, B8

is a little lower in frequency and some-
what drawn out due to the longer time
constants that are used for the attack,
decay and one shot.

Steam train and whistle (figure 5)
Since the sound produced by a steam
train consists mainly of white noise, this
effect will present no problems to our
1C. As soon as the on/off switch S4 is
closed, an intermittent white noise
signal will appear at the output. The
rhythm of the signal will be determined

by the frequency of the SLF oscillator
and the resulting sound will resemble
a steam train (if any of you can
remember what they used to sound
like!) The speed of the train can be
altered by adjusting the frequency of
the SLF oscillator by means of poten-
tiometer PI .

The whistle is generated by the VCO.
When switch S3 is depressed the VCO
is switched through to the output. The
frequency of the whistle is determined
by the values of R4, R5, R6 and C4. To
be honest, the whistle sounds a lot less
authentic than the steam train. Unfor-
tunately, the only way to improve on it
would be to include many more external
components, which we feel would
defeat the object of the exercise, this
being to keep the circuit as simple as
possible.

If this particular sound effect is to be
used for a model railway, the frequency
of the SLF oscillator should be directly
related to the speed of the locomotive.
This can be done by replacing potentio-
meter PI with a light dependant resistor
(LDR) which is optically coupled to a
6 V/50 mA lamp that is connected
across the rails (with a 1 k preset poten-
tiometer wired in series). When the vol-
tage on the rails (speed) is increased, the
lamp will burn brighter, so that the
resistance of the LDR will drop and the
frequency of the SLF oscillator will
therefore rise (QED).

Aeroplane (figure 6)

Since an aeroplane makes a sound very

similar to that of a fast steam train, the
circuit diagram is very much the same as
that of figure 5. The only differences
being that the SLF oscillator will have
a much higher frequency and the VCO
is not required, we have yet to hear an
aircraft with a steam whistle!!

Racing car plus crash (figure 7)

The roar of the engine is provided by
the VCO whose frequency is determined
by potentiometer P2. Depressing switch
SI will interrupt the motion of the car
with a shattering 'crash' effect. For this,
capacitor Cl is fully charged and the
attack/decay system is activated. During
this period low frequency white noise is
passed on to the output. When SI is
released the car (or its ghost!) will start
up once more after the short delay
required for Cl to discharge.

Dawn chorus (figure 8)

Again, this mainly involves the VCO
and the SLF oscillator. The VCO is
controlled by the negative going edge
of the triangular waveform provided
by the SLF oscillator. As a result, the
signal produced will slowly drop in fre-
quency. Nothing happens during the
positive period of the triangle, so that
there is a short break. The frequency
of the VCO is determined by the setting
of potentiometer PI.

Since the effect thus obtained is rather
monotonous, the noise generator has
been called in to help and R7 . . .R9,
P3 and C5 have been included. Poten-

tiometer P3 is adjusted so that the noise
generator produces a low frequency
output signal. Then the random saw-
tooth signal generated across C5 is fed
to the frequency determining input of
the SLF oscillator. This livens up the
birds considerably.

The setting of P3 is quite critical, as it
has to be turned very slowly and care-
fully until the dawn chorus becomes
sufficiently lively. If required, the values
of R8 and P3 can be altered to give a
somewhat less critical control.

Construction

As mentioned previously, the various
sound effects circuits require very few
components (one 1C and a handful of
external components). Once these have
been mounted on the printed circuit
board (see figure 10), all that remains
is to connect up a small loudspeaker and
a battery - then the fun can begin!

The printed circuit board can be used
for all of the circuits described in this
article (and doubtless a few more
besides). In each case it is simply fitted
with the components indicated in the
parts list. Any parts marked on the com-
ponent overlay, but not mentioned in
the particular parts list can be omitted.
For those of you who would like to do
a little experimenting, figure 9 provides
a general survey of all the components
for which there is room on the board.
This makes it a lot easier to 'translate'
the circuit diagram in relation to the
board and vice versa. M

Before it reaches the amplifier the out-
put of the tuned circuit first has to pass
through the two sections (R4, C5 and
R5, C6) of the low-pass filter. In ad-
dition, the negative feedback of A1
has been made to be frequency de-
pendent, so that the amplifier's fre-
quency curve will also feature a low-pass
characteristic.

PI presets the sensitivity of the circuit
in such a manner that the amplifier A2
will not produce an output signal until
the voltage level at its input has ex-
ceeded a certain value.

Opamp A3 is connected as a re-
triggerable monostable multivibrator.
If A2 produces only one output trigger

detector

Inspite of the fact that an
electronic movement detector is
used in many department stores
and office blocks nowadays, a
great number of people still think
it something of a mystery when
the doors slide open automatically
as they approach the exit with a
loaded trolley. In vain they
look for the 'thing' that noticed
their presence. In some cases it
is a shaft of light catching passing
legs, but usually the method of
detection will be invisible and
therefore all the more mysterious.
The system described here not
only opens doors but also switches
lights on and off and can in fact
even be used for a game, the
purpose of which is to sneak out
an object from a guarded room
(good practice for burglars!)

What is the principle behind the move-
ment detector? Well, when an electri-
cally charged object happens to be
inside an electrical field, the latter will
be disturbed. The interference can only
be detected the moment it occurs,
in other words, not once the situation
has returned to normal. Similarly, a
moving electrical field will affect a
conductor that is present.

Generally speaking, human beings are
surrounded by a weak electrical field
created, mainly, by friction. Every
physical movement of a conductor
causes charge carriers inside it to be
displaced and this can be detected.

How the circuit works
Together with the FET T1, resistors
R1 . . . R3, capacitors Cl, C2 and C4
and the 'sensor plate', the input stage of
the circuit acts as an LC tuned circuit
followed by an amplifier. The input
circuit, consisting of R1, R2 and Cl,
functions mainly as the inductor. In
parallel with the capacitor formed by
the sensor plate, the 'inductor' produces
a parallel tuned circuit with a centre
frequency just below the mains fre-
quency. Any change in the surrounding
electrical field will affect the coil,
causing the tuned circuit to oscillate.

pulse, the output of A3 will be 'high'
for a short while and then drop back
to zero. Further pulses will hold the
output high for as long as they con-

The output high of A3 reverse biases
diode D5 therefore the noninverting
input of opamp A4 will also become
logic one. This opamp then generates
a square wave output with a frequency
in the 400 Hz range, which, via transis-
tor T2, will be heard from the loud-
speaker. The volume can be regulated
with P2 as required. To prevent the
generator from oscillating continually
opamp A4 is switched off by diodes
D4 and D5 whenever the output of A3
becomes low again.

In addition to the audible signal, the
monostable multivibrator A3 provides
the base drive current for transistor T3
by way of the voltage divider R19 and
R22. The collector current then oper-
ates the relay Re which may be used for
various switching purposes. When the
relay is released, diode D2 protects T3
by eliminating any induced voltage.
Diode D3 makes sure T3's collector
will remain higher than the emitter
voltage by more than 1 V.

It is important that the earth of the
circuit is connected to that of mains, if
it is to work well. H

Resistors:

R1 - 12M

R2 ' 1 M

R3.R15- 10 k

R4 = 15 k

R5.R6 = 47 k

R7.R21 =470 0

R8 = 33 k

R9.R10 = 4k7

R1 1,R16 = 470 k

R12,R13,R14.= 100 k

R17,R18 = 22k

R19 = 2k7

R20 = 1 k2

R22 = 1 k

PI = 220 k preset

P2 = 100 fi/1 W, linear

Capacitors:

Cl = 560 n
C2.C7 = 330 n
C3 * 10 m/16 V
C4 - 10 n
C5 = 390 n
C6.C12 = 100 n
C8 = 47 m/10 V
C9 = 220 m/16 V
C10= 1 M
C11 - 10m/10 V
C13 = 3n3
C14 - 47 m/25 V
Cl 5 = 1 000 m/25 V

Semiconductors:

IC1 - LM 324
IC2- 7812
T1 * BF 256C
T2 = BO 139
T3 = BC 547B
D1.D2.D3.D4 = 1N4148
D5= AA 119
D6.D7.D8.D9 « 1N4001

Miscellaneous:

Tr - 12 V/0,5 A transformer
Re = 1 2 V relay
LS = 8 n loudspeaker

3-32 - ele

1981

MW receiver

MW

receiver

a straightforward 'straight-through' receiver

Why should there be a need to construct your own medium waveband
receiver? Surely it is far cheaper to buy one at the local supermarket?
This may well indeed be true, but it is far more fun to actually build
one yourself. After all, many of our readers belong to the younger
generation and there is nothing quite like building your first radio —
and getting it to work! — as many of our more experienced readers will
testify.

Amateur constructors often feel like
magicians. It is quite amazing what can
be accomplished with very few com-
ponents. Take the design for this
receiver for instance; an RF amplifier
and a couple of transistors to bring
music to your ears! In any case, many
readers felt that it was high time that a
simple receiver circuit found its way
into the EPS list once more.

The object of the exercise is to end up
with a neat, economical portable
radio. One that fits comfortably inside
a coat pocket and can keep you up to
date with the latest news and pop
music, as you travel around town.
Another important factor, of course,
is that a single 9 V battery should last
as long as possible (a few months at
least).

When designing such a project, the first
choice has to be between AM and FM.
Nowadays, FM is favourite, but the
problem here is that is not so easy for
the novice to build, especially if the
finished unit is to be really small,

Elektor does have printed circuit boards
available for something a little larger
than that described here, but by no
means one that requires so few com-
ponents, it can be virtually put together
with your eyes closed. Which is our
main objective, remember.

We therefore came to the conclusion
that there is nothing wrong with the
medium waveband. It certainly has not
run out of stations yet and what is
more, the set will be much simpler (and
cheaper) to build than an FM radio. It
can be far smaller in size and last, but
by no means least, it needs no finnicky
aerial. In other words, it really is a
pocket radio.

Superhet or superreg?

Now that we have decided upon medium
wave and the main requirements are
that it be small, simple to build and
conservative on batteries, we need to
work out a few more design parameters.
The majority of manufactured radio
receivers operate on the superhetero-
dyne principle. However, most single
waveband receivers utilise the super-
regenerative principle. This is, in fact,
the recipe for a reliable receiver if it is
to have a fairly high performance and
feature reasonable sensitivity in spite
of its compact size. Nevertheless, if
such considerations as simplicity of
construction and ease of calibration
are involved, the 'super' part is best
omitted. This is further illustrated in
figure 1. All the most common AM
receiver principles are shown there.

First the 'straight-through' receiver
(a). This is comprised of an adjustable
LC tuned circuit, a high frequency
amplifier, a detector, an audio amplifier,
a detector, an audio amplifier and a
loudspeaker. The RF stage could even
be left out, so that the set would then
be a 'sophisticated' crystal receiver. If
it is to be sufficiently sensitive, how-
ever, rather a lot of RF amplification

la

will be necessary. This is why the RF
amplifier usually incorporates an ad-
justable feedback network (see dotted
line) which enables the set to be ad-
justed to the point of oscillation (maxi-
mum sensitivity) for every station.

The reflex receiver in figure 1b also
offers a reasonable degree of sensitivity.
Here the RF amplifier stage is not only
used in the conventional manner, but
it also amplifies the audio signal. This
type of receiver used to be very popular
in the days when transistors were rather
expensive and difficult to obtain.

Figure 1c shows that even the 'simple'
superhet can be quite complicated. The
aerial signal is now 'added' to that of an
oscillator in the mixing stage. The
oscillator produces a somewhat higher
(or lower) frequency than the input
signal and is varied simultaneously with
the tuning capacitor. This generates a
constant 'sum' or 'difference' frequency
regardless of the actual frequency of the
input signal. This 'intermediate' (IF)
frequency is filtered at the output of
the mixer and is further amplified. If
necessary, the signal can be filtered and
amplified several times to improve the
selectivity. This is because the constant
frequency of the IF signal makes the
tuning of the LC circuits for each
station superfluous. Obviously, the
receiver will be rather complicated to
set up.

Straight-through

Seeing that practically all the receivers
that have been published in Elektor
over the past few years were either
superheterodyne or superregenerative,
our design staff thought that it was time
that a simple version was produced. In
any event, an 1C exists which will fit
the bill perfectly, but more about this
later. Thus, after due consideration, the
recipe illustrated in figure la was
chosen for the miniature MW receiver.

2

Figure 2. What could be simpler? The
straight-through receiver using the 2N 414 1C.

albeit without the feedback stage. The
latter, even in the version shown in
figure 1b, will make any receiver a lot
less portable. Also, construction be-
comes a critical task, the set is likely to
'whistle' and more often than not the
receiver will have to be operated with
both hands as the amount of feedback
has to be adjusted for each individual
station. If an ordinary 'straight-through'
receiver can be built to incorporate
enough RF amplification for feedback
to become superfluous, without causing
it to oscillate, it will have many practi-
cal advantages. The miniature integrated
circuit that we have in mind does just
this and furthermore features other
useful characteristics, as will be seen
later.

Compared to more usual sets, a simple
single tuned circuit receiver (such as
this one) will be much less selective
and therefore not so sensitive. Since MW
receivers, especially pocket-sized ones,
are more often than not used for
reception of a limited number of local
radio stations, this disadvantage will not

be so noticeable. It is amply com-
pensated by the following advantages
over other types:

• it is much easier to build

• it does not require any alignment

• it does not include an oscillator,
thereby avoiding stability problems

• no mixing is involved, reducing
'whistle' considerably

• its sound is of superior quality to
that of the average superhet

The ZN 414

By far the easiest method of con-
structing a straight-through medium,
waveband receiver is to use the ZIM 414
integrated circuit from Ferranti which
was designed specifically for this
purpose.

Having only three pins, it looks more
like a transistor than a 'proper' 1C.
Although it has been around for quite
some time, it continues to provide the
best solution for a receiver where a
minimum number of components is
required. This is clearly illustrated in
the diagram in figure 2. It shows the
complete MW receiver constructed
around the ZN 414. All that is required
is a single transistor amplifier stage to
provide a first class matchbox receiver.
Certain items immediately catch the
eye. First, the low supply voltage. The
ZN 414 is designed to be powered from
a single battery. Its supply voltage range
is between 1 .2 and 1 .6 V and the
current consumption is in order of
0.3 mA. This device could hardly be
more economical.

Also remarkable is the fact that the coil
(LI ) consists of a single winding, instead
of the usual double-wound or tapped
inductor, and that the detector diode
which one would usually expect is

The double wound coil is superfluous as
the 1C features an extremely high input
impedance (4 Mfl) which is only a very

1981

| AGO

40 ^

r. The entire section inside the dotted

slight load for the parallel tuned circuit.
Not only does this make the coil that
much easier to wind, but also it helps
to prevent interference from short wave
transmitters. As far as the detector
diode is concerned, this is already inte-
grated in the 1C in the form of a transis-
tor detector which uses capacitor C3
as the only external component.

The block diagram of the medium
waveband receiver, see figure 3, shows
just what goes on inside the case of
the ZN414 (see inside the dotted
area). It is comprised of a high im-
pedance input stage (drawn here as an
emitter follower), a (three stage) RF
amplifier with a frequency range of
1 50 kHz ... 3 MHz and a gain of
72 dB, an AM detector and, finally, an
automatic gain control (AGC).

Too much should not be expected from
the latter, as its range is about 20 dB,
just enough to smooth out any slight
differences in amplitude between the
various radio stations. As soon as the
unit is in close proximity to a powerful

transmitter, however, the automatic
gain control will be unable to adjust
to the particular station required.
Nevertheless, it is far better to have
20 dB than none at all, as is the case
in some elementary receivers.

Circuit diagram

From the block diagram in figure 3
it can be seen how straightforward the
complete pocket sized medium wave
receiver is. Apart from the parallel
tuned circuit, the ZN 414 and the audio
amplifier, all it needs is a suitable
circuit to derive the 1.3 V required by
the ZN 414 from the power supply for
the audio amplifier. A simple bleed
resistor and zener diode would have
been more than adequate, but a far
better method has been employed here.
The complete circuit diagram of the
MW receiver is shown in figure 4. The
actual receiver section is constituted
by IC1 and the surrounding components
and is, of course, identical to the

diagram shown in figure 2. The only
difference between the two is that
the values of R2 and C3 have been
slightly modified. This is because they
are based on the ideal supply voltage
for the ZN 414, which is between
1.3 and 1.4 V.

When calculating the value of R2 three
parameters have to be taken into
account. First, the ratio R1/R2 will
affect the automatic gain control. Since
the value of R1 must be 100 kS2, only
the value of R2 can be altered. There-
fore, the value of the latter will also
affect the gain of the ZN 414 and, as
the voltage supply to the 1C must be at
a constant level, the gain will be reduced
if a relatively high value is chosen for
R2. Moreover, it is important that the
values of R2 and C3 constitute a low
pass filter with a turnover frequency
of around 4 kHz, which is necessary
for the detector included in the 1C.

The solution, therefore, is to select the
best compromise value for R2 and to
find an effective method of regulating
the supply voltage for the ZN 414. This
is why the voltage source constructed
around transistor T1 has been added to
the circuit. The voltage on the emitter
of T1 can be adjusted between approxi-
mately 1.2 and 1.45 V by means of the
preset potentiometer P2. This may not
seem to be very much, but it affects the
gain of the 1C somewhat considerably.
This is an advantage as the sensitivity
of the receiver can now be adapted to
specific circumstances by presetting the
gain of the amplifier as required. Ob-
viously, this will be at a maximum in
isolated areas and lower in the vicinity
of powerful local transmitters so that
the set is not overdriven, which could

cause distortion and poor selectivity.
Batteries spend a very brief period of
their lifespan at their rated nominal
voltage and for this reason, together
with the fact that the supply voltage for
IC1 is critical, the voltage source T1 is
not fed directly from the battery.
Instead, the voltage is first regulated by
a zener diode (D1) to iron out any
fluctuations in the battery voltage.
Since the receiver also has to be econ-
omical, the current supply to the zener
diode is limited by means of a fairly
large series resistor (R6). As the current
consumption of the ZN 414 is very low
the zener diode will operate very well,
even at a low - ' voltage than that which
it is rated at (about 3.9 V in this case).
So much for the receiver section. Now
for the audio amplifier. Initially, it was
proposed that one of the well-known
amplifier ICs should be used. These,
however, turned out to rapidly exhaust
the small 9 V battery's current supply.
Instead, it was decided to combine two
pnp transistors and two npn transistors
to form a discrete amplifier. Very little
can be said about this, as it is con-
structed entirely according to the 'four
transistor recipe'. It requires very little

current and there is virtually no quiesc-
ent current for the output transistors
T4 and T5. Thus, when there is no input
signal the entire amplifier will only
consume around 2.5 mA.

Since the current requirement for the
ZN414 is also fairly modest, the
receiver will only consume a total of
4 mA. A reasonable battery will there-
fore last a considerable time, provided
the volume is not turned up to an
ear-splitting level, that is.

The maximum output power of the
audio amplifier is in the region of
250 mW. In theory it will produce
more from a 9 V battery (about 1 W
maximum into 8 ft). but the voltage
gain of the audio stage is limited so that
no more than 4V pp is available across
the loudspeaker output even when the
output signal from the ZN 414 is at a
maximum (approximately 30 mV e ff).
This maintains the current consumption
at a level acceptable to the 9 V battery
and also eliminates the need for heat-
sinks on the two output transistors.

Construction

The printed circuit board and com-

ponent overlay for the medium wave-
band receiver are shown in figure 5. The
only components which are not actually
mounted on the board are the variable
capacitor Cl, potentiometer PI and the
loudspeaker. The leads connecting the
capacitor to the board should, ob-
viously, be as short as possible.

As you probably already know, the
performance of a parallel tuned circuit
is largely dependent on the Q of the
tuning inductor. For this reason, the
aerial coil, LI, will have to be wound
with the utmost care and attention. It
is best to use the parameters specified,
that is, 48 turns of 0.3 mm diameter
enamelled copper wire on a ferrite
rod with a diameter of 10 mm and a
length of 10 cm. The ferrite rod can be
mounted onto the board by means of
two short pieces of string. Holes have
already been drilled in the printed
circuit board for this very purpose.

It is a good idea to wind LI around a
paper or cardboard tube so that it can
be moved up and down on the ferrite
rod later; the permeability of ferrite and
ferroxcube material tends to vary, so
therefore it may be necessary to 'trim'
the receiver if the stations are not at

3-36 - ele

arch 1981

displa

TiorCc

the correct places on the waveband.
Further remarks. Firstly, something that
probably does not need mentioning. As
the ferrite rod coil is in fact an aerial, it
would be unwise to mount the com-
pleted receiver in a metal easel
Secondly, the zener diode D1 must be
either a 250 mW or a 400 mW type, as
stated, as otherwise the input level for
the voltage source T1 (3.9 V) will not
be correct. This is because the current
flowing through D1 is far lower than
normal in order to keep the current
consumption of the circuit to a mini-
mum. Thirdly, as the output transis-
tors do not require any quiescent
current, the value of resistors R13 and
R14 are fairly critical. If the stated
values are not adhered to the chances
are that the output transistors will start
to draw current after all and, as there
is no temperature compensation net-
work, this could well have a detrimental
effect on them. Using the values given in
figure 4, transistors T4 and T5 will not
have to be cooled. They can be ordinary
types without any need for heatsinks.

Results

In practice, the miniature medium
waveband receiver was found to per-
form very satisfactorily. Being a single
coil type, it may require constant re-
tuning due to the set 'drifting' off
frequency, especially where distant
stations are concerned. Even so, it is
eminently suitable as a 'stand-by'
receiver for news bulletins etc. which
is quite often all that is required any-
way. It is only when the owner wishes
to listen to a weak station in the neigh-
bourhood of a powerful one that the
MW receiver is going to have problems.
This can often be remedied by turning
the receiver towards the weaker station
thereby eliminating the stronger one.
Local stations can be received very well.
In unfavourable circumstances, an ex-
ternal aerial may be experimented with.
This should be connected to the top of
the tuning coil via a small value capacitor
(4p7). This, however, should hardly
ever be necessary. If the input signal is
clean enough, the sound quality of the
receiver will be suprisingly good. In this
respect it really stands out amongst
similar commercial radios.

Finally, the receiver is remarkably
inexpensive. If, like countless other
constructors, you have a 'junk' box full
of ferrite rods, tuning capacitors and
transistors, it will only cost a few pence.

M

This particular topic receives full
attention in Junior Computer Book 2
(to be available shortly), but there is
no harm in whetting the appetites of
our readers even if it is a little prema-

How can the Junior Computer display
words? Normally speaking, data and
address information is displayed with
the aid of the monitor routine
SCANDS. This involves one of the
hexadecimal numbers, 0 . . . F, in each
display. Where texts are concerned,

text display
on the
J unio r
Computer

As we know, the display of the
Junior Computer is suitable for
displaying both numerical and
hexadecimal data. By utilising a
seven segment alphabet it is also
possible to display written texts.

If the text is to be static, a total of
six letters are available. If,
however a longer message is
required, this may 'run' along the
display rather like the electronic
news display at the top of tall
buildings (dynamic text).

from an idea by U . Seyffert

however, the monitor routines are no
good. What is needed is the subroutine
SHOW with the addition of a special
look-up table which contains the
corresponding seven segment pattern
for each individual letter.

Table 1 provides a survey of letters and
figures together with the corresponding
data which has to be entered into
port A for them to be displayed. This
table has been partly based on sugges-
tions made to us from one of our
readers. Obviously, letters which include
diagonal lines (such as K, M, N, Q, V,
W, X and Y| will have to be adapted to
the horizontal and vertical set up of
the display segments. Experience has
shown, however, that the eye and the
brain soon become accustomed to
this.

Now for a short program that will allow
a six letter word to appear on perma-
nent display. A good example would be
the word 'Junior' as indicated on the
prototype of the Junior Computer in
the front cover photograph of the May
1980 issue of Elektor and Book 1. The
program, JUNIOR, is listed in table 2.
Here the modified SHOW routine will
be called SHOWDS and the look-up
table that holds the information relating
to the display of any particular
character is called TXT (text table).
The Y index register acts as the display
counter and text index. The value con-
tained in the Y register increases from
00 tc 05 as an index for the particular
character to be displayed. As soon as
the value in the Y register becomes 06,

00 (jump to DISMPX to begin another
round). During the subroutine SHOWDS
the Y register contains a delay value
which determines the length of time
that each display is actually lit. For
this reason the previous value contained
in the Y register (display counter/text
index) must be saved in the address
location TEMPY (0004) before the
jump to the SHOWDS subroutine takes
place.

The function of the X index register, on
the other hand, is the same as it was for
the SHOW routine: it acts as a display
digit switch by way of port B. In other
words, the information contained in the

X register (08, 0A, 0C, 0E, 10 and 12
consecutively) is passed to port B data
register to turn each of the displays on

Text on the run . . .

A stationary text is all very well, but it
does tend to get a little monotonous
after a while. A much more interesting
possibility would be to update the

displayed text every few moments. In
this manner whole sentences could be
displayed instead of just single words.
This can be accomplished with the aid
of the program JUNTXT shown in
table 3. The effect is very similar to that
of an electronic news display. It is an
expanded version of the earlier program
JUNIOR (table 2). Page 03 is used to
store the actual text which can, there-
fore, be up to 256 characters in length

Table 3.

JUNTXT 0200
0202
0205

0207

0208
020A

BEGIN 020C
020E

DSTIME 0210
0212

DISMPX 0214
0216

ONEDIS 0218
021 A

021 C
021 E
021 F
0222

0224

0225
0227
0229

TMECHK 022B
022D
022F
0231

0235

SHOWDS 0239
023C
023F
0242

DELAY 0244
0245
0247
024A
024C
024 F

0250

0251

— enough for the average length
paragraph!

Again, this program uses the subroutine
SHOWDIS, only this time the text table
(TXT) is located at address 0300 and
although the Y register is still used as a
display counter it is no longer used as a
text index directly, Instead, the particu-
lar section of the text to be displayed is
calculated by adding the instantaneous
value in the Y register to the contents
of address location NUMVAR (0001).
The value contained in NUMVAR will
be constant for the period of time a
certain text is on display (the actual
duration can be adjusted by modifying
the contents of location 021 1 ). As soon
as that period of time is over the con-
tents of NUMVAR are incremented by
one: the entire text shifts one location
to the left and the right hand display
shows a new character. When the
contents of NUMVAR are greater
than the contents of location NUMCOR,
we will have arrived back at the be-
ginning, as this means that the entire

A9 7F
8D81 1 A
A5 00

38 SEC C = 1

E9 05 SBC * 05

85 02 STAZ-NUMCOR NUMCOR = NUM minus 05

A9 00 LDA = 00

85 01 STAZ-NUMVAR first display text

A9 6F LDA ft 6F

85 03 STAZ-DISCNT establish text display time

A2 08 LDX * 08 start from Di 1

A0 00 LDY = 00 display counter (Y) - 00

84 04 STYZ-TEMPY store display counter

98 TYA Y to accumulator

18

65 01
A8

20 39 02
A4 04
C8

CO 06
F0 02
DO ED
C6 03
DO E5
E6 01
A5 02
C5 01
BO D9
90 D3
B9 00 03
8D 80 1 A
8E 821 A
A0 7F

10 FD
8C 80 1 A
A0 06
8C 82 1A
E8
E8
60

CLC

ADCZ-NUMVAR

TAY

JSR-SHOWDS

LDYZ-TEMPY

INY

CPY =06
BEQ TMECHK
BNE ONEDIS
DECZ-DISCNT
BNE DISMPX
INCZ -NUMVAR
LDAZ-NUMCOR
CMPZ-NUMVAR
BCS DSTIME
BCC BEGIN

LDA-TXT, Y
STA-PAD
STX-PBD
LDY = 7F
DEY

BPL DELAY
STY -PAD
LDY =06
STY-PBD

RTS

C - 0

A «-Y ♦ contents NUMVAR (00011
accumulator to Y
display first/next character
retrieve state of display counter

have all 6 display been accessed?

if not, next display
time up?

if yes. update text
end of text?

see 'JUNIOR' program
TXT = 0300 (table 4)
text index = Y + contents NUMVAR

0 1 23456789ABCDEF
0300 7F 7F 7F 7F 7F 7F 07 0B 2F 23 01 7F 20 7F 02 7F
0310 01 6F 07 0B 7F 07 0B 06 7F 61 63 2B 6F 23 2F 7F
0320 46 23 48 0C 63 07 06 2F 7F 3F 7F 03 63 1 1 7F 03
0330 23 23 0A 7F 24 xx xx xx xx xx xx xx xx xx xx xx
0000 INUMI = 34

text will have been displayed. This is Table 4 provides a sample text which
because the contents of NUMCOR are can be displayed on the Junior Com-
05 less than those of location NUM. The puter with the aid of the program
latter (location 0000) is where the user JUNTXT as given in table 3. The text
must store the low order byte of the last contains a message for Junior Computer
memory location of the text table. In Book 1 owners. A text should always be
other words, if the last character of the preceded by at least six blank spaces
text message is stored in location (7F), so that the beginning and end of
0332, the value 32 is stored in location the message are clearly separated from
0000 (NUM). each other. M

0

A

R. de Boer

□ ®

l raffle game

No, this is not a game to play in the middle of the M2. Since it is
somewhat safer to simulate traffic situations by electronic means,
this game can be used to familiarise inexperienced drivers with the
various perils to which they are exposed every day on the road. It
enables them to cope without the risk of an accident. In addition,
it's fun!

By means of light and sound signals, the player is confronted with all
sorts of problems which he has to deal with efficiently in order to cover
as many miles as possible within an a lotted time. Obviously, this does
not mean he can Dreak any traffic regulations on the way.

elektor march 1981 — 3-39

Present-day traffic is no joke. There are
so many vehicles on the roads nowadays,
that for safety's sake children should
know the rules and regulations practi-
cally as soon as they can walk.

Although technological progress has
literally taken the weight off our feet,
our reflexes are very much overbur-
dened. However, it is impossible to
turn back the clock, so we will just have
to sharpen our reflexes.

Which is exactly what this traffic game
proposes to do. The player learns to
react instinctively, automatically after
having to face similar situations again
and again. What's more he/she can come
to grips with unpleasant reality without
leaving the safe confines of the living
room. In short, an excellent method to
teach children.

The rules of the game

Figure 1 shows how the game can be
built. A potentiometer with a scale divi-
sion in miles per hour constitutes the
accelerator. A clock generator is con-
trolled by the pot to provide the sound
of the engine. The clock frequency is
divided, counted, decoded and shown
on the display as the number of miles
covered.

The main purpose of the game is to
drive for as many miles as possible
within a certain time. This is not merely
a question of putting the foot down on
the accelerator, or here, turning the
speed up. LEDs are used to represent
all sorts of obstacles which the driver

Figure 1. An example showing how the traflic game can be built. The speed at which the 'car'
keep to all the traffic regulations in order to save time.

is likely to encounter. These light at
varying intervals and may symbolise a
sharp bend, a slow car ahead (30 miles/
hour), a 40 mile hour speed limit and a
bumpy road surface.

If the driver is taking it too fast, an
'angry' note will be emitted by the
loudspeaker. Whenever something goes
wrong, the mile counter will come to a
halt, but time will continue to run

At the beginning of the game, the push-
button in the top left-hand corner is
depressed to introduce a handicap into
the game. The more often this is de-
pressed, the more obstacles the driver
will meet en route.

The player is allowed to overtake if he is
stuck behind a slow vehicle (30 miles
per hour). This makes the game a lot
more realistic. An oncoming car will
then be imitated by a running light (to
the left of the panel). The tenth and last
LED in the series indicates that over-
taking will inevitably lead to a crash. If
the driver is stupid enough to risk it
anyway, by putting the switch on 'over-
take', this will cost him several seconds.
As an oncoming vehicle is difficult to
see from a distance, the running light
LEDs increase in brightness as they
approach the player. Exceeding the
maximum speed limit will also cost the
driver quite a few seconds.

If for some reason or other the game has
to be interrupted, depressing the reset
key will be enough to restart it.

The electronics

Since the circuit diagram of the traffic
game (figure 3) looks rather compli-
cated, it is a good idea to consider the
block diagram first. This is shown in
figure 2.

At the heart of the circuit the clock
generator 1 provides the sound of the
engine. In addition, the number of miles
covered is obtained from the clock
signal. A monostable multivibrator
(MMV2) determines the duration of the
game. The handicaps are caused by a
shift register. Lighting LEDs represent
various regulations and other difficul-
ties the motorist might encounter. A
comparator checks wether he/she is
keeping to the traffic regulations. This
it does by comparing two voltages: one
depends on the driving speed and is
therefore linearly related to the clock
generator's frequency; the other is ob-
tained from the obstacle shift register
and determines the maximum speed
that can be preset for each obstacle by
means of the potentiometers.

If the maximum speed is exceeded for
too long, the driver is elektorcuted, that

is, the voltage at the output of the com-
parator will become low. This starts the
monostable multivibrator MMV1. As a
result, the clock generator (the engine)
will stop for a few seconds (penalty)
and at the same time generators 2 and 3
will produce an 'angry' note. During the
penalty period, generator 4 will also be
inactive, since the car has stopped and
so there will not be any obstacles.

If an error is corrected quickly, no
penalty time will have to be paid. The
period during which this correction is
to be carried out is determined by the
network C5/P8.

The maximum speed limit on motor-
ways (70 miles per hour) can be preset
with P9. Putting S2 in the b position
will enable the driver to overtake. The
'accelerator' is then depressed a little
further, gaining more speed and there-
fore mileage. Before the last LED of
the running light lights, however,
(oncoming car) S2 will have to be in the
a position as otherwise the car will be in
for trouble (penalty time).

Now that the block diagram has been
dealt with it won't be necessary to dwell
on figure 3 at great length. Since the 1C
numbers are indicated in the various
sections (figure 2) the components can
easily be spotted in the circuit diagram.
Here and there a few matters still have
to be cleared up. Five of the outputs be-

longing to the running light (IC1) are of their purpose and corresponding P9: 70 mile speed limit on motor-

buffered with inverters. The other five functions: ways.

are not buffered, since these only have PlOa + b: driving speed (accelerator),

to produce a low current. Potentio- pi : speed at which the various ob- P1 1 : volume control of sound of engine

meters PlOa and PlOb are mechanically stacles follow each other. and 'angry' tone,

coupled (a stereo potentiometer). This P2: speed of oncoming traffic. P12: duration of game,

is a simple way to convert the clock P3: penalty time. SI: obstacle entry (before game is

frequency into a direct voltage (accor- P4: speed of slow car ahead (30 miles/ started),

ding to the driving speed). Potentio- per hour). S2: overtake (position b).

meter PI Ob should be connected so P5: maximum speed on poor road S3: start,

that it has minimum resistance when surface. S4: reset,

the wiper of PlOa is turned towards P6: speed limit. S5: on/off.

resistor R 20. P7 : sharp bend to the right; maximum

There are quite a few switches and speed 35 miles/hour,

calibration points in the circuit diagram, P8: time within which error can be
so that it is a good idea to have a survey put right without being penalised.

Wiltshire SN2 68 N.

Telephone: 107931 693681-7.

Telex: 444166.

(1849 M)

with a soft-spoken operator. The equipment
meets Home Office specification MPT 1302.
Dymar Electronics Limited.

Colonial Way,

Radlett Road,

Watford,

WD2 4LA

Mobile radiotelephone system

The new Lynx 2000 range of radiotelephones
from Dymar Electronics Ltd. is a British
designed and manufactured system which

The TP 600 can
PFM200 hand-he
Thandar TF 200

ivledium-size enclosure

1981 — 3-43

Bug

Multimeter with automotive
functions

A new multimeter, the LNIM-400, recently
introduced by Lascar Electronics has two
Bxtra functions for automotive applications.

25°c will Self-adhesive temperature
r cent per indicators

strument features AC/DC Voltage Walton Road,

output are claimed. An important feature,
thought to be unique, is that brushes and
commutators are guaranteed for 100,000

Normand Electrical Company Ltd..

easurement from 0.1 mV to 1 K
C/DC current range up to 20 Amps,
he resistance measuring range is
1 Ohms to 20 M Ohms.

cs**ES>

Eastern Road,

Cosham,

Hampshire P06 ISZ,
Telephone: 0705 370988

Thermindex Chen

RCA launches £ 138 single-board P.O.Box 112.
computer WaHord^'

The latest addition to the RCA Solid State Telephone: Watfo

range of Microboard single-board computers

is the CDP18S604. which costs only £138

(plus V.A.T.). Featuring very low power

consumption and simplicity of use, it is a

complete computer system on a 4.5 x 7.5-inch

a CDP1802 C-MOS - Rotary < electronic c

Lascar Electronics Limited.

Unit 1. Thomasin Road,

Burnt Mills. Basildon,

Essex SS13 1LH,

Telephone: Basildon 10268) 727383

New DC tachometer generator
from NECO

tages of lowpower static C-MOS circuitry and
the COSMAC microprocessor architecture.
Powered from a single 5 V supply, it requires
only 4 mA when populated with a C-MOS
read-only memory.

The CDP 18S604 is designed to provide the
key hardware for a variety of low-cost micro-
computer applications, enabling the designer
to concentrate on software development and
the special requirements of his specific
application.

Two CDP 1 852‘s are used to provide eight
parallel input and eight parallel output lines.
It has four flag inputs and a Q serial data
output. All the input/output lines are on edge

The new low-priced Microboard computer
is compatible with both the CDP 1 8S005
!’ design of DC tachometer and CDP18S007 COSMAC Development
> the range of high-precision Systems, facilitating prototype design and the
n Neco Electronics (Europe) debugging of both hardware and software.
’**• Like all the other RCA Microboard products,

per 100 rpm, (12,000 rpm jt js expandable by use of the 44-pin
(ref ST-7447A-2I is more COSMAC Microboard universal backplane.

4 . *

An anti-dazzle mask is included among the
range of accessories - which makes reading of
possible built-in LED desplays easier. The
system includes; Case, Rear panel/rear panel
with opening, 2 feet, 2 knurled screws and

,he 4 adhesive feet. Standard ci

CONDITIONS OF ACCEPTANCE OF CLASSIFIED ADVERTISEMENTS

1. Advertisements are accepted
subject to the conditions appearing
on our current rate card and on the
express understanding that the
Advertiser warrants that the advert-
isement does not contravene any
Act of Parliament nor is it an in-
fringement of the British Code of
Advertising Practice.

2. The Publishers reserve the right
to refuse or withdraw any advert-
isement.

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

PRINTED CIRCUITS. Make your
own simply, cheaply and quickly!
Golden Fotolak Light Sensitive
Lacquer — now greatly improved
and very much faster. Aerosol cans
with full instructions - E2.25. Dev-
eloper - 35p. Ferric Chloride - 55p.
Clear Acetate sheet for master 14p.
Copper-clad Fibre-glass Board
approx. 1mm thick - £1.75 sq. ft.
Post & Packing 60p.

WHITE HOUSE ELECTRONICS,
P.O. BOX 19,

PENZANCE, CORNWALL.

CLEARANCE PARCELS :
TRANSISTORS, RESISTORS,
BOARDS, HARDWARE, 101 lbs. -
only £5.80!

1000 RESISTORS £4.25

500 CAPACITORS £3.75

BC108, BC171, BC204, BC230.
2N5061, CV7497 TRANSISTORS
10 . . . . 70p 100 .. . .£5.80

2N3055, 10 for £3.50. S.A.E. for
lists : W.V.E. (4),

15 HIGH STREET, LYDNEY,
GLOUCESTER.

TELEPHONE answering machines
from £75.00, FLIPPHONES £25.00
WATCHES, gents and ladies, from
£5.75. CAR RADIO/CASSETTES,
BOOSTERS, SPEAKERS, AER-
IALS, WALKIE-TALKIES down to
earth prices. Send SAE for details:
H.P. SUPPLIES,

162 LONDON ROAD,

WEMBLEY, MIDDX.

4. The Advertiser's full name and
address must accompany each
advertisement submitted.

The prepaid rate for Classified
Advertisements is 20 pence per
word (minimum 12 words). Semi-
display setting £5.50 per single
column centimetre (minimum 2.5
cms.). All cheques, postal orders,
etc. to be made payable to
Elektor Publishers Ltd. Treasury
notes should always be sent by
registered post. Advertisements,
together with remittance, should be
sent to the Classified Advertisement
Manager, Elektor Publishers Ltd.,
10 Longport, Canterbury, Kent
CT1 1PE.

MODERN BARGAINS S.A.E.
Early radios.

SOLE ELECTRONICS.

37 STANLEY STREET.
ORMSKIRK, LANCS. L39 2DH.

Chartpak UK 09

Coiltronic UK 08

Doram Electronics UK 07

Electronics and Music Maker . . UK 17
Electrovalue UK 08

Gailan UK 15

Keytronics UK18

Maplin UK 02

Marshall's UK 13

Microcircuits UK12

G.F. Milward UK 15

Phonosonics UK 09

T. Powell UK 24

Ramar UK 08

BALLARD'S, Tunbridge Wells, have
moved to 54 GROSVENOR ROAD,
TUNBRIDGE WELLS, KENT.

Phone 0892 31803.

DISK DRIVES £ 135

5% inch double-density, 40-track
disk drive giving you 250k bytes
of unformatted storage. Shugart
compatible. Removed from brand
new equipment, fully tested and
aligned. Three months' warranty.
£135 each, post & packing £3.
Data sheets 50p. Manual & circuit
£9.50.

DISKETTES £2.95
Maxell 5 '/• inch double-density
floppy disks. Compatible with all
35 and 40 track single-sided, soft-
sectored drives.

£255 each, post 8i packing 30p.
£28 for box of ten, post & packing

80p.

DOUBLE-SIDED DISKETTES
£3.50

Maxell double-sided, double-
density disks. £3.50 each, post 8i
packing 30p.

£33 for box of ten, post &
packing 80p. Please add VAT to
all prices. Send cheque or P.O. to:

Helistar Systems Ltd.,

150 Weston Rd., Aston Clinton,
Aylesbury, Bucks, HP22 5EP.
Tel: Aylesbury (0296) 630364

Science of Cambridge .UK 10/UK 11

Suretron Systems UK 08

Technomatic UK 23

TK Electronics UK19