ip-to-date electronics for lab and leisure

elektor july/august 1975 - 703

About this issue D 700

,Cl * t * 753

TV# 1 -TUN list 750

AF nr

amplifier for 1.5 volt supply ig

sniplifier, 6.5 watt 1C 94

aotereo control amplifier 44

■istereo disc preamplifier 45

■stereo power supply 48

■istereo 3 watt amplifier 46

austereo 15-30 watt amplifier 50

diode, ideal 23

filter, active low-pass 101

frequency dependent resistor 4

generator, diac sawtooth- 95

generator, staircase 8

inductance, variable 78

mixer, FET- 7

oscillator, 1 kHz ’ _ 81

overmodulation indicator 36

signal squirt, mini 99

siren, 7400 " 31

tremolo, LED controlled 90

volume control, fingertip g7

volume, voltage-controlled 96

Digital nr

arithmetic teacher 56

capacitance meter gg

clock suppression 28

decimal to 7-segment converter 1

display input selector 51

divider, programmable frequency- 92

division by 2, 5 and 1 0 66

division by 3 ’ 6 7

division by 4 68

division by 6 69

division by 7 70

division by 8 ’ 7j

division by 9. . . 72

EXOR 3

exor ;;;;;;;; 35

four-channel switch 63

frequency reference, universal 12

interference rejector ... 32

NAND-NOR gate . . . 41

NAND-OR gate ; gg

NOR-AND gate g

pulse multiplier ’ 37

reset for TTL, automatic 29

seven segment abc 21

seven segment for MOS-clocks, improved ... 91

time signal simulator 74

the 74121 ' 76

use of 7493 instead of 7490 ’ 73

mvi detector with ceramic filter 10

mono/stereo indicator 26

receiver, sensitive coil-less 22

selective amplifier 25

stereo decoder, CA 3090 AQ ......... . 88

traffic news indicator 33

tv-sound channel, 1C 77

ultrasonic detector ’ 65

Power supplies nf

austereo power supply 48

battery stand-by circuit 58

heatsink shrinker 87

rhythm generator power supplies 40

slow-rise power supply 84

voltage regulator, wide range 5

voltage stabilizer as current source 18

+/0 /— from one winding _ 34

Sundries nr _

afterburner 13

alarm circuit ’ n

calculator doubles as stopwatch ’ 49

capacitance meter g 3

car lights failure detector 1 5

conductance tester, mini siren 82

contrast adapter 199

cuff links 42

curve tracer 30

dark room exposure timer 93

direction pointer 85

economic flashlight 64

economic relay gg

FET £2 adapter 60

'find me' flashlight attachment 83

flasher, LED 7g

flasher, LED, for higher voltages 80

flasher, triac 6

fluorescent lamp dimmer ’’ 62

frequency meter, analogue 61

fuse destroyer ’ 89

hum receiver 54

key contacts for organs, gated 52

LED display dimmer 57

mostronome 47

opto-coupler for mains sync 43

oven, mini- ‘ 24

rhythm generator M 252 38

rhythm generator M 253 39

rhythm generator power supplies 40

running light ’ 17

service meter, auto 20

speed controller, triac 55

temperature monitor 75

tester 14

tester, p.c.board and wiring 53

thermometer 27

timer, kitchen ’ 2

trigger, 555 as a 15

publication’ on many grounds, but the already published, nor is it a preview
most frequent one is that what seems of designs that will be published in the
to be an original idea at first sight turns coming year. Of course the designs are

out to be not so original, or else can not all equally original, and of course

_ ■ • • no * ma< * e t° m eet the required some of the most interesting ones may

flhfMlt this BC CB1 A specifications. be dealt with in greater length as (part

Some of the designs published in this of) a new design in a future issue, but

issue are based on manufacturers’ appli- in general it is the intention that this

cation notes, for which we thank all issue should be a useful collection of

concerned. Originally, it was also the design ideas that will retain its value at

’. . . containing over a hundred ‘Summer intention to publish several designs least until next year’s ‘summer circuits’

Circuits’.’ It doesn’t sound too difficult, using LSI chips (large scale integration). issue is published. We hope that we have

until you start trying to find a sufficient However, a closer analysis of the types succeeded in achieving this goal,

number of interesting designs in time to recommended to us by the major
meet the printers’ deadline. . . . manufacturers showed that most of

The requirements which must be fulfil- them were either too complex to

led are originality, quality and quantity. describe in the space available in this

Finding the requisite 100 circuits starts issue, or too expensive, or not (yet)
to become a real problem when some readily available.

100 circuits have been rejected. A Perhaps it should be made clear that

design can be labelled “unsuitable for this issue is not a review of circuits

710 — elektor july/augint 1975

3 11110 0 1

4 0 1 1 0 0 1 1

5 10 110 11

6 10 11111

7 1 1 1 0 0 0 0

8 1111111

9 11110 11

Cold-cathode numicator tubes are
gradually passing out of use and are
being replaced by various types of
seven-segment displays. As the life of
these tubes is about 1 0 years there are
still many serviceable pieces of digital
equipment around which, although in
good working order, have tubes that
are beginning to fail. Some readers may
have DVM’s or digital clocks, which
they would like to convert for a seven-
segment display.

This article describes two simple
decoders, which will convert the deci-
mal output of the numicator decoder to
a seven-segment code. The circuits are
also useful for demonstrating seven-
segment displays using a ten position
single-pole switch to programme the
numbers.

To work out the simplest decoding cir-
cuit it is first necessary to draw up a

truth table for the seven-segment dis-
play (table 1). A ‘1’ in a column indi-
cates that the particular segment is il-
luminated. A 'O’ indicates that it is
extinguished.

It is apparent that there are a great deal
more 'ones’ in the truth table than there
are ‘noughts’, 49 as against 21 . It is
therefore evident that the simplest
conversion will be achieved via the
‘noughts’, i.e. all segments are normally
illuminated and the ones not required
for a particular digit are suppressed.

The decoder consists basically of a
diode matrix. The rows of the matrix
are the decimal inputs, whilst the
columns are the outputs to the seven
segment display. Where a digit requires
that particular segments are suppressed,
diodes are connected from the appro-
priate row to the appropriate columns.
Two versions of the decoder are shown
in figures 1 and 2. Figure one is
intended for positive logic inputs, i.e.
when a particular digit is enabled, that
row input is ‘high’ and all the others
are low. The TUP’s are all normally
turned on, but when a particular digit
input goes ‘high’ then the diodes
connected to that row hold the bases of
the transistors connected to them to
about +1 1 .4 V, thus turning off the
transistors and extinguishing the appro-
priate display segments.

The version of the decoder shown in

decimal to seven segment

figure 2 is intended for negative logic
inputs such as numicator tube decoder
outputs. The transistors are all normally
turned on, but when any row input goes
‘low’ it will ground the bases of the
transistors connected to it by diodes,

converter using 21 diodes

thus turning them off. either may be used with the other

The two versions of the decoder are supply voltage by substituting the

shown for different supply voltages, but resistor values from the other circuit.

ti

&

cl.

ell

to

‘a

el

(c

hi

U!

SC

S]

di

P 1

di

b

F

1

L c

elektor july /august 1975 — 711

BC140

N4 “CD 4011

mittery or mains-synchronous kitchen
clocks usually have dials large enough to
enable short time intervals of a few
minutes and fractions thereof to be
clearly read (except when they happen
to be steamed up), but few have any
‘alarm’ system to give an audible
warning when a preset time interval has
elapsed. The timers built into electric
(or nowadays gas) cookers supposedly
have this facility, but they are not much
use for short-interval timing because the
scales are either non-existent or very
spall, the setting dial is mechanically
difficult to handle, the termination
point of the timing interval is ill-
defined, and at least one well-known
make is calibrated in quarters of five
minutes — i.e. in 1)4 -minute intervals.
These may be good enough for baking
a cake but they are of little help to the
blancber of twenty successive batches of
runner beans, each to be immersed in
boiling water for 2 minutes 45 seconds.
For this reason it was suggested that an
^ectronic timer should be developed,
possibly using valves. However, the
technical staff felt that the best way to
use these would be to drill holes in them
and fill them with sand.

The time switch described here can be
constructed at very small cost, and is
adjustable between 1 minute and
1 7 minutes. Other times are possible
with small modifications.

Before switching on. capacitor C 1 and
C2 are uncharged. When the device is
turned on with switch S (position 1),
input A of fiipflop N1/N2 remains
briefly at *0’, so that output Q of N2
becomes “O’ and multivibrator N3/N4 is
blocked.

Capacitor C 1 charges through poten-

tiometers PI and P2. When the poten-
tial at B drops below the fiipflop
changeover threshold, the fiipflop
switches over and the multivibrator is
started. The resulting square wave is
amplified by T1 and T2 and reproduced
through loudspeaker L.

On switching off (S to position 2),
capacitor C 1 is discharged rapidly
through resistor R 1 so that, when the
timer is once again switched on, there is
no residual voltage which would shorten
the timing interval.

Calibration:

1. Set PI to the middle of its travel and
P2 to minimum. Then readjust PI to
give a time of 1 minute.

2. Next set P2 to maximum and
measure the time now given by the
circuit.

3. Finally calibrate the scale of P2
linearly between 1 minute

(minimum) and the maximum time
which has been determined.

unequal, the output signal is always
HIGH. Equal input signals give a low
output level because in that case both
transistors are either ‘on’ or ‘off, so that
the output signal is governed by input A.

It is the simplicity of this exclusive OR-
gate that makes its application instead
of TTL-circuits so attractive. It func-
tions as follows. When both inputs are

712 — elektor july/august 1975

in which

T = R 2 x C

R,

R 1 (O 2 T 2

frequency

dependent

resistor

which means that the input impedance

is:

This is a real resistance - with current i
and voltage in phase - but increasing ;
with the square of the frequency.

(Int. J. Electronics)

The main features of the circuit are:
wide range output: 0.1 to 50 volts
good load regulation: 0.005% between
0 and 1 amp

good line regulation: 0.01%
low output noise: better than
250 microvolts.

The wide output range is made possible
by the integrated circuit CA 3130,
which remains effective with a zero volt
input.

Furthermore, upward extension of the
output range is made possible by the
addition of T4 between the IC and the
series pass transistor. The high gain thus
obtained allows a high degree of regu-
lation, and the T1/T2 Darlington pair
gives a sufficiently high current amplifi-
cation. 4

T3 is the output current limiter. When
PI is turned fully anti-clockwise, T3 will
limit at 0.6 amps. The limiting circuit is
inoperative when P2 is turned fully
clockwise.

The regulator circuit proper works as
follows. The integrated circuit CA 3 1 30
compares the output voltage applied to
the non-inverting input with a reference
voltage applied to the inverting input.
The output voltage of the regulator is
stepped down in a potential divider to
prevent damage to the IC. The reference
voltage is set by P2, which should be
a high quality component, as any noise
on its wiper will be passed on to the
regulator output terminals.

wide range

voltage

regulator

A further integrated circuit, CA 3086,
compensates the reference voltage for
temperature changes. The IC contains

four transistors used as diodes or zener t
and one transistor lowering the output
impedance of the reference circuit.

The reference IC also gives a stepped- I
down supply voltage for the CA 3130. *

This function requires the presence of |
both integrated circuits in the regulator; I
omission of IC1 will cause the destruc- I
tion of IC2. All transistors should have
a breakdown voltage of at least 55 volts. I
(RCAappl.)

This circuit makes a standard incan-
descent lamp flash at a frequency that
can be set between 2 and about 10 Hz.
The mains voltage is rectified
by a 1N4004 diode followed by an
adjustable RC network. As soon as the
electrolytic capacitor is charged up to
the breakdown voltage of the ER 900

diac, it will discharge through the diac,
which fires the triac and flashes the
lamp.

After a certain period, set by the 100 k
control, the capacitor will be re-charged
to bring about a further flash.

The 1 k control sets the triac firing cur-
rent.

(Transitron)

triac flasher

elektor july/august 1975 - 713

Inexpensive junction-FETs such as the
E 300 or U 1994 E - already popular in
HF-circuits - can also often be applied
advantageously to LF-circuits.

In a snail audiomixer, for example, the
use of JFETs leads to a great saving in
components because of the relative
simplicity of the biassing arrangements.
The input impedance of each channel is
determined only by the value of the
potentiometer used.

FET- mixer

The number of input channels can be
considerably extended, if required, pro-
vided that the common drain load re-
sistor (Rl) is suitably chosen. Its value
is the standard value closest to ,
where n is the number of input channels.

For other types of JFET it may be
necessary to change the value of source
resistor and/or the supply voltage.

The frequency response of the circuit
shown has -3 dB points at about
20 Hz and 80 kHz.

F. Hebinck

A convenient way of generating a good
‘staircase waveform’ is to use a diode-
transistor pump driven by a square wave
as shown in figure 1 .

Before the square wave is applied to the
input, capacitor C2 is uncharged. Conse-
quently, there is no voltage across it and
transistor T3 is cut off because its
Sjpitter is at +10 V.

During the first positive going pulse, the
diode conducts and capacitor C 1 is
charged to just under the peak value of
the waveform (peak value minus the
voltage across the diode). Because the
diode prevents the emitter of T1 rising
much above ground potential, the
negative-going pulse drives the emitter
negative by an amount equal to the
peak value of the waveform less the
voltage across the diode. Therefore T1
inducts and the voltage on its collector
tails sharply. In other words, Cl is dis-
charged and C2 is charged by an almost
equal amount. (The difference is caused
by the base current of T1 .)

If Cl and C2 are equal, the voltage
change across C 1 will equal the change
across C2 plus twice the drop across the
diode. However, if C2 is much larger
than C 1 , the change across it will be
much smaller.

Capacitor Cl is charged again during the

5... 20 V

staircase

geaerator

next positive-going pulse. As T1 and T3
are not conducting, the charge on C2
cannot leak away and the voltage at T3
emitter remains constant. During the
next positive-going pulse, more charge
flows from C2 to Cl and the voltage on
T3 emitter falls sharply again.

This process is repeated with the voltage
on T3 emitter falling in a staircase
pattern until it is negative. Transistor T3
then conducts, and positive feedback
from T2 to T3 produces an avalanche
effect that rapidly discharges C2. When
C2 is discharged, the whole cycle starts
again from the top of the staircase.

An approximate formula for the
number of steps (n) is

where V s = supply voltage

V w = peak value of the square
wave

Vd = voltage across the diode
Therefore, if Cl is 100 pF and C2
390 pF and the diode voltage 550 mV,
the number of steps produced by the
circuit in figure 1 is

390 x 10

" 100(5-1.1) ’

The staircase waveform generator can be
used in many applications, such as a
D/A convertor, frequency divider, driver
for a transistor curve tracer, as well as to
produce special sound effects and so on.

Ci V s

C, X V w - 2 Vd

Three gates are needed for converting
NOR-gates into an ‘AND’-function. Fig-
ure 1 gives a TTLAND-gate consisting of
three gates of a 7402. When inputs A
and B are ‘1’, so will be input C.

The circuit of figure 1 can be applied to
any ordinary TTL-circuit.

Figure 2 shows a COS/MOS AND-gate
consisting of three NOR-gates each with
two inputs. Since the gates N1/N2 in

NOR AND gate

both circuits are connected as inverters,
several types of gate or inverter can be
used, such as 7400 or 40 1 1 gates or
7404 of 4007 inverters, for example.

In this design the fact that ceramic
filters are non-ideal is turned to use. The
impedance at the emitter of T1 is low
enough to have T2 function as an

FIM detector
with

ceramic filter

earthed emitter oscillator. Via emitter
follower T1 the input signal is applied
to the emitter of T2, so that the oscil-
lator is synchronized thanks to the non-
ideal behaviour of the ceramic filter.

The input signal is limited by the

diodes D1 and D2, the value of the
limited voltage being adjustable from 0
to 500 mV by means of PI . The setting
at which the detector functions opti-
mally depends, for instance, on the type
of ceramic filter used. A favourable

property of synchronism detectors is
their superior AM-suppression as com- I
pared with the ratio detector. In ad-
dition, their threshold (smallest signal re-
ceived free from noise) is much better. k

A. Roosink

The circuit described is one more
addition to the variety of warning
signals intended mainly for use in cars.
It causes a warning signal to sound
when the turn indicator control has not
returned to its neutral state. The circuit
is particularly effective for those cars in
which the turn indication is not
accompanied by a clearly audible click.
Further, the circuit sounds a warning
signal when the ignition switch is ‘off
while the headlights are still ‘on’.

The circuit is designed for 12 volt
negative earth systems.

With the turn indicator flashing, the

circuit emits a continuous series of four
tones in random order and rhythm. It
operates as follows.

Point C receives its power from the
ignition switch. Point A receives a series
of pulses from the flashing relay contact
for the warning light. Each of the two
astables generates a signal depending on
the associate RC network; T3/T4 at a
relatively low frequency, the second
astable T1/T2 at variable frequencies
depending on the pilot voltage from D5
or D4.

T5 and T6 generate the delay required
for overriding the flashing intervals
when the warning light is off. The
alarm signal is amplified by T7 ... T1 1
feeding the loudspeaker.

The circuit can also be used as a ‘switch
off your headlamps’ warning. This

operates as follows. Turning the ignition ,
switch ‘off’ removes power from C.

Point B, however, continues to be
powered from the lighting switch. This '
causes the two astables and the ampli-
fier to remain operative, with the
exception of T5. This removes the
control signal from T6 causing its col- I
lector to rise to 1 2 volts and causing the
AND gate D6/D7 to pass the signal on I
to the amplifier. A two-tone warning
signal will then be emitted. - |

Terminals A, B and C, which must be
connected to the flashing circuit, the
headlamp switch and the ignition key \
switch respectively, can be found
behind the fascia board, and there is
no need to tamper extensively with the I
car wiring apart from the three tappings .
mentioned above and the chassis return.

-I

L

alarm

circuit

elektor july/august 1975 — 715

This circuit claims nothing in the way of
originality, but is simply a useful,
general-purpose board that can be used
in many frequency and time measuring
applications. It is particularly suitable
for use - as a gate pulse generator in fre-
quency counters.

The heart of the system is a 1 MHz
crystal oscillator based on two NAND-
gates. The output of this oscillator is
buffered by a third NAND-gate and the
frequency is then divided down by a
series of 7490 decade counters. These
consist of a divide-by-2 stage followed
by a divide-by-5 stage, which means that
in addition to dividing the reference fre-
quency down to 1 Hz in decades, out-
puts of 500 kHz down to 5 Hz are also
available. These outputs are particularly
useful where gate pulses for frequency
counters are required. For example, the
5 Hz output will provide positive pulses
of 100 ms width, so if the frequency of
a 10 MHz signal were being measured a
gate pulse this long would let through
*)00 ,000 cycles of the signal to the
counter, giving a display of 1 000000.

On the other hand, for period measure-
ments the 1 Hz to 1 MHz outputs are
more useful. For example, when
measuring a one second period,

1 ,000,000 cycles of the 1 MHz output
can be counted, giving a display of
1000000.

The p.c. board layout is quite compact
and well laid out. The outputs are avail-
able along the lower edge of the board
h the diagram. There is one spare
NAND-gate in the package used for the
oscillator, and this may be used as the
gate in frequency counter applications.
The connections to it are brought out at
the top right comer of the board.

The oscillator frequency may be
trimmed to exactly 1 MHz by the
trimmer capacitor. The best method of
doing this is to use an oscilloscope to
compare the 100 kHz output with the
200 kHz Droitwich transmissions, using

universal

frequency

reference

rminals

Lissajous figures. The trimmer should,
of course, be adjusted until the

Lissajous figure apparently ceases to
rotate.

W. Ferdinand

afterburner

When starting a car journey after dark it
is useful to have a device which will
keep the interior lighting on for a while
after the doors have been closed, and so
make it easier for the occupants to
fasten safety belts and insert the
ignition key. This can be done with the
simple time-switch circuit shown.

’1N4001or

716 — elektor july/august 1975

Passive testing of an electronic circuit
seems quite a simple job. All you need
is a resistance meter. Unfortunately,
however, using such a piece of equip-
ment for semi-conductors is not always
such a good idea. The output currents
are likely to damage semi-conductor
junctions. The tester described in this
article is easy to build and has the
advantage that no more than about
50 pA can be sent through the circuit
under measurement. So it can be used
for most conventional IC’s and semi-

conductors, including MOS-compo-
nents. The ‘indicator’ is a small loud-
speaker, so that during testing, it is not
necessary to keep looking at the
measuring instrument instead of the
measuring points.

The transistor T1 and T2 form a simple
voltage-controlled LF-oscillator, with
a loudspeaker as the load. The oscillator
frequency is determined by Cl , R1 , R4
and the external resistance between the
measuring probes. Resistor R3 is the
collector resistance of T2 ; C2 serves for

low-frequency decoupling of this
resistor.

As stated above, the tester will not
damage a circuit under test; conversely,
it is advisable to add diodes D 1 and D2
so that the circuit under test will not
damage the tester.

As long as there is no electrical
connection between the measuring
probes, the circuit draws no current.
Battery life is then about equal to its
shelf life.

(Based on Motorola application).

tester

i

Because of the 400 mV drop across Rx,
T1 is normally turned on and T2 cut
off.

If one of the tail lights fails, the current
through Rx is reduced by half, that is
0.84 Amp. The voltage drop across R x
now becomes 0.84 x 0.24 = 0.2 V. This
voltage is too low to turn on T1 , so that
T2 now receives base current via R1 ,
and the LED lights up.

To obtain a reliable indication when the
lamps fail, it is recommended to use one
detector circuit for not more than two
lamps.

It is quite permissible, however, to use
one LED for several detectors: D 1 and
R3 are common to all detectors, and
the collectors of all ‘T2’s’ are connected
together. R3 should be 470 £2 for a
1 2 V system and 220 £2 for a 6 V
system.

car light
failure detector

II O. Dirnhirn

For people who wish to be sure that the
lights on their car are in good order, this
circuit is perhaps the solution. Is is very
simple and provides a reliable indication
when one or more lamps fail.

Depending on the current through
lamp L, a voltage drop occurs across
resistance R x . This voltage drop must
amount to about 400 mV, and this
determines the value of R x .

For example, for the tail lights, where
two lamps of 10 W/12 V are in parallel,
R x is calculated as follows:

The current is ^ = 1 .67 Amp.

Rx then becomes

The IC 555 is normally used as a timer.
The time delay for these applications is
obtained by charging and discharging
a capacitor between two well-defined
voltage limits. For this purpose the IC
has two voltage-sensitive inputs, which
can be used in a trigger circuit.

The input voltage is applied to pin 2;
pin 5 receives a reference voltage from
potentiometer PI . With this poten-
tiometer the trigger threshold can be set
between practically 0 V and half the
supply voltage. The input impedance at
point 2 is about 1 M.

The IC has two outputs, an open-
collector output (pin 7) and an output
(pin 3) which can deliver or sink a maxi-

mum of 100 mA. The output voltage is
‘high’ when the input voltage at pin 2
is below the threshold voltage. If the

threshold voltage is exceeded, the voltage
on pin 3 drops and T1 conducts.

(Electronics)

555 as a trigger

elektor july/august 1975 — 717

R. Kellenbach.

By means of this circuit four lamps can
be made to light up sequentially, so that
‘running-light’ effects are obtained. The
circuit consists of a square-wave gener-
ator (Tl, IC1), a shift register (IC2,
IC3), and the lamp driver stages. With
P 1 the frequency of the square-wave

extinguished and no lamp is burning.
After opening of S2, SI is placed in
position 1 so that the input of the
register receives a ‘1’. After one clock
pulse, the input information of the
flip-flop is transmitted to the output
and the first lamp lights up; SI is now

Integrated voltage stabilizers can also be
used as current sources. The diagram
shows an example of a current regulator
using the IC LM 309 (5 V/l A).

The output current I 3 of the circuit
consists of the output current I2 plus
the quiescent current li of the IC. The
current Ij is determined by the output
voltage V2 of the IC and the resistance

The quiescent current Ii is always speci-
fied by the manufacturer of the IC.
Provided that the quiescent current Ii
can be neglected in comparison with I2
(1 1 <5C I2 ). the output current I3 will be
constant (I3 = I2) because it then
depends only on the (fixed) output

When existing amplifiers lack sufficient
input sensitivity and no space is avail-
able inside their case, separate low
power pre-amplifiers will come in
handy. They should contain a minimum
of components and prefc ably be fed
from a single dry cell.

The self-contained pre-amplifier de-
scribed here consists of a single amplify-
ing transistor preceding an emitter
follower. DC negative feedback stabil-
ises the working point. The gain will be
approximately x 10 to x 20.

When the signal source has an im-

®-± T* I

voltage
stabilizer as a
current source

voltage of the IC and the (fixed) value
of Rl.

For the circuit to work it is essential
that there is a certain minimum voltage

pedance of over 100 k ohms, some
degree of gain control can be achieved
by means of PI.

Some long-term economy can be
obtained by using two 1 .5 volt dry cells

amplifier for
1.5 volt
power supply

(in series) instead of one. When the
power supply drops below 1 volt the
amplifier will cease working. Common
dry cells often run down soon to 1 volt
and must then be discarded, while it will
take more time for each one of two cells

difference between input and output of
the circuit. This difference is
determined by the minimum voltage
difference between the input and out-
put of the IC and by the output voltage
of the IC:

(V, -V 3 ) m i I1 = (V 1 -V 2 ) m in + V 2
The maximum voltage difference
(Vi - V 3 )jnax is equal to the maximum
input voltage of the IC ; similarly, the
maximum output current is equal to
the maximum output current of the IC.
The following values apply to the IC
LM 309 K:

(V, - V 3 ) min = 2.2 V + 5.2 V = 7.4 V
(V, - V 3 ) max = 35 V
hmax = • A

I 3 « j5_ (A, fi) (at I 3 > 100 mA)
V 3 -Ij • Rl“5 • ~(V)

!,5V/350tiA

sumption at 3 volt supply will be about
450 microamps.

Chr. Wiinsche

This handy portable instrument was
designed as an aid to car servicing to be
used in conjunction with a timing light
for setting up a car ignition system for
peak performance. It consists of a
tachometer and dwell meter. The tacho-

auto service
meter

meter is useful for examining the ig-
nition timing at various r.p.m. in con-
junction with a timing light, or when
used as a dwell meter it will measure the
angle over which the contact breaker is
closed and thus provide information
about the setting of the contact breaker
gap.

The circuit is shown in figure 1 and is
intended for use with negative earth
vehicles - which most modem cars are.
It may, however, be adapted for positive
earth vehicles by reversing all diodes and
electrolytic capacitors, and by substitut-
ing TUP for TUN and vice versa. The
circuit derives its power from the car
battery.

The circuit operates as follows: T1 and
T2 form a Schmitt trigger. As long as
the contact breaker is closed T1 is
turned off and T2 is turned on, which
means that T4 is also turned on, so that
a positive voltage equal to the supply
voltage minus the base-emitter voltage
of T4 appears at the emitter. When the
contact breaker is open the input (Rl) is
connected to positive supply through
the primary of the ignition coil, so T1 is
turned on and the Schmitt trigger
switches the other way. T4 is now
turned off so the voltage appearing at its

emitter is zero. The average voltage at
the emitter of T4 is thus proportional to
the ratio of contact-breaker-closed time
to contact-breaker-open time i.e. it is
dependent on the dwell angle. With
switch SI in the ‘a’ position the average
current through the meter is also pro-
portional to the dwell angle, so the
meter may be calibrated linearly in
terms of dwell angle.

With the switch in position ‘b’ the
instrument functions as a tachometer.
Pulses from the collector of T3 are
differentiated by C2 and used to trigger
a monostable consisting of T5 and T6.
The monostable produces pulses of
constant width, but as the r.p.m.
increase the mark -space ratio of the
pulses increases. The average voltage on
the emitter of T7, and hence the average
current through the meter, is dependent
on the ratio of ‘pulse’ to ‘no-pulse’ time.

so as the r.p.m. increase and the pulses I
get closer together, the current through I
the meter increases linearly.

The instrument can be calibrated as
follows: With SI in position ‘a’ earth 1 t S T i
the input (Rl) and adjust PI to give full |
scale deflection of the meter. This corre- p
sponds to a 360° dwell angle and the I
scale can be calibrated linearly from
0 to 360.

The tachometer scale should be cali-
brated with full-scale as the maximum I
required r.p.m. For most applications I
8000 should be sufficient. If the
instrument is to be used with both four I
and six-cylinder engines, then either two I
scales are required, or SI must be re-
placed by a three-position switch and .1 ,

P2 must be duplicated to calibrate one I
scale for different engines. The reason I
for this is that a six cylinder engine
produces more pulses for a given r.p.m. I
The instrument may be calibrated using j
the simple circuit shown. This produces I -
a 100 Hz waveform. This corresponds to I
3000 r.p.m. for a four-cylinder engine I
and 2000 r.p.m. for a six -cylinder
engine. The output of this circuit is con- !
nected to the input of the instrument I
and P2 is adjusted to give the correct C >
deflection of the meter.

bh

I t

E

LEIO

1 5 h
EE

o H

seven

segment

ABC

There are various alphanumeric display
devices available to the electronics
industry, but unfortunately their high
cost places them outside the reach of
most amateur constructors. However,
one of our readers, Mr. P. Geist, has
suggested that it is possible to use
readily available seven-segment LED dis-
plays as alphanumeric indicators, and
has proposed an alphabet using such
displays.

As the display has seven segments, this
means that V - 1 or 127 unique sym-
bols may be displayed. Since 10 of these
are the digits 0 to 9 this leaves 1 1 7 from
which to choose symbols for the letters
of the alphabet, plus other useful sym-
bols such as parentheses. Since the range
of shapes available with seven segments
is limited, some of the characters are

necessarily rather stylised. Nevertheless
they are unambiguous and are easily
read after a little practice. After all, the
figure 4 in a seven segment display is

I

stylised, yet this has gained wide
acceptance, so there is no reason why
the seven-segment alphabet should not j I
prove popular.

dbcdthih uEL
HnoPdRStuddddd

elektor july/august 1975 - 719

This little receiver tunes the medium
and long wave-ranges. It operates with-
out any coils by employing a synchron-
ised oscillator.

When the oscillator is barely able to
maintain oscillations, it will synchronise
to an incoming signal frequency close to
its free-running value. The amplitude of

the oscillation follows, more or less
linearly, the modulation of the
incoming signal. The circuit can
synchronise itself to a signal of some
tens of microvolts, so that its effective
sensitivity is very high.

P4 sets the level of oscillation; the
stereo-ganged pair P2 and P3 form the
‘tuning’ control. Since this stereo-
potentiometer covers a very wide range
it may be necessary to add a ‘fine’
control consisting of a low-value stereo-
potentiometer in series (typically 1 k or

500 ohm).

The modulated oscillator signal is passed
to the amplifier stage with T4, via the
full-wave ‘detector’ with D1 and D2.

The output level is sufficiently high for
driving most amplifiers.

A short aerial, such as a piece of wire, is
enough to provide quite reasonable
reception. If a good aerial is to be used,

P 1 can be inserted to prevent over-
driving by too strong a signal.

sensitive
coiMess
synchrodyne
receiver for
MW and LW

F.G. Hebinck

Every semiconductor diode has a
threshold-voltage, below which there is
no useful conduction, so that it cannot
rectify small signals. This threshold
voltage can however be made very small,
with the aid of an op-amp.

The op-amp input signal is the differ-
ence between the available input
voltage (Vj) and the output voltage
(V 0 ). The diode will conduct whenever
this differential voltage, multiplied by
the amplifier’s open-loop gain, more

ideal

diode

than equals the diode threshold voltage.
The effect at the input of the circuit is a
reduction in the apparent diode
threshold voltage by a factor equal to
the open loop gain of the op-amp.

In practice this apparent threshold will
be far smaller than the offset voltage of

the op-amp, so that it is this offset
voltage that determines the apparent
threshold. It will usually be a few milli-
volts. (Lit. Sescosem application note)

720 - elektor july/august 1975

This circuit arrangement will operate as
a low-power temperature stabiliser, suit-
able for ‘ovening’ quartz crystals.

If the NTC resistor R3 and the transis-
tor T1 (heating element!) are mounted
together with the crystal on a metal
block, the circuit will hold the tempera-
ture of the block constant within 0.5°C.
A suitable operating point would be in
the range 60 ... 70°C; this can be
maintained for ambient temperature
0 ... 40°C with or without supply volt-
age variations within the range 8 ... 16 V.

The NTC resistor forms one arm of a
Wheatstone bridge (Rl, R2, R3 and PI ).
The unbalance voltage from the bridge
is applied to the differential input of the
TBA 221 op-amp, which operates as a
low-hysteresis voltage comparator. The
op-amp quickly saturates in one direc-

mini-oven

tion or the other, so that the current
through T1 is effectively turned on or
off. It is the heat dissipation of this
transistor which maintains the block
temperature above ambient (as re-
quired). For optimum performance it is
necessary to mount the transistor and
the NTC as close as possible to each
other.

The values of Rl, R2. R3 and PI have
been so chosen that the bridge arms
have roughly equal resistance values at
the typical operating temperature-range
of 60... 70° C.

(Milliard)

This selective amplifier can be continu-
ously tuned, by means of R la and Rib
(stereo-gang potentiometer), from
100 kHz to 2 MHz. The same relative
tuning range, about a lower frequency,
can be obtained by increasing the values
of all the capacitors.

The circuit contains two phase-shifters
and an amplifying stage. The first of
these phase-shifters is built up around
Tl, using Cl and Rla. The voltage gain
of this stage approaches the ratio
Rc/Re, which is unity, so that the
signal is almost unattenuated. The phase
shift varies slowly from 0° tot 180° as
the frequency is increased, equalling 90°
at a frequency determined by Cl and
the setting of Rla.

The second phase shifter, built up with
T2, C2 and Rib, is identical to the first.
The amplifying stage, with T3, produces
a phase reversal, i.e. a further 180°. At
the critical frequency set by R 1 a+b the
total phase shift going around the feed-
back loop will be 360°, which means:
positive feedback (‘reaction’). R2 sets
the amount of this feedback. If the total
gain inside the loop exceeds unity the
circuit will oscillate. When the gain is
just marginally below unity the circuit

H. Benzel

Those who may find it fun, or even use-
ful in some respect, can easily extend
the single lamp stereo indication of an
FM-receiver to give an automatic stereo-
mono indication. Experiments with the
well-known IC stereo decoder
MC 1 3 1 OP resulted in the circuit given
here. The green LED (D1 ) is the normal
stereo indication; lighting up of the red
LED (D3) means that we are listening
to a mono broadcast.

behaves as a high-Q tuned circuit.

To use the design as a selective amplifier
the ‘tuning’ control Rl (stereo potentio-
meter) is set to give the required fre-
quency. As the feedback control R2 is
now advanced from zero the circuit will
become increasingly selective with
respect to the frequency of the input
signal at A-B. As the oscillation
threshold is approached, where the
highest Q values are obtained, it may be
necessary to set Rl more precisely. If
oscillations once start it will be
necessary to back off R2 some distance
(log potentiometer!) and try again.

A second application of the circuit is as
synchronised oscillator. The tuning
range in this case is the same as for the
selective amplifier (and can once again
be lowered by using suitably increased

C-values), only now the feedback
control is advanced to obtain oscillation '
and then slightly backed off so that this?' ,
oscillation is only just maintained.

If a signal very close in frequency to the L
circuit oscillations is now applied to the '
input, the oscillator will be pulled into I
synchronism. The beating that occurs |
just before ‘sync’ is achieved can be
observed at T3 collector.

It is also possible to synchronise the
oscillator to an input at one of its sub- I
harmonics. Once again the trick is to set I
R2 for barely-maintained oscillations, I
then use an input signal just strong
enough to secure synchronisation. The
circuit will latch without difficulty onto I
a frequency l/20th of that of its own
oscillation.

During mono broadcasts, the red LED I i
is connected direct to the supply volt- f
age via load resistor R2 (and diode D4 I
in series). However, as soon as the 1C | 1
switches on D1 , signifying a stereo
broadcast, the voltage across D2 will
be considerably less than the drop
across D3/D4, with the result that
LED D3 will extinguish.

mono-stereo

indicator

elektor july/august 1975 —

In this circuit, a silicon diode is used as a
temperature sensor. The junction poten-
tial of a silicon diode decreases by about
two millivolts per degree centigrade, so
that the temperature of the diode can
be ascertained by measuring the voltage
across it. As a temperature sensor, a
diode has the advantages of high lin-
earity and a low time constant. It can
moreover be used over a wide tempera-
ture range, from -50 up to 200 C.

As the diode voltage has to be measured
very precisely, a stable reference source
is needed. A good choice is the 723 volt-
age stabiliser. Although the absolute
value of the zener voltage in this IC
varies from one 723 to another, the
temperature coefficient is very small
(typically 0.003%/°C). Furthermore,
the 723 can be used to stabilize the
1 2-volt supply for the rest of the circuit.
Note that the pin numbers in the circuit
diagram are only correct for the dual-in-
line (DIL) version of the 723; the
pinning of the metal-can version is
shown in the IC list elsewhere in this
issue.

The second IC, the 3900, contains four
amplifiers of which only two are used.
These op-amps do not function in the
usual way; they are current -driven
rather than voltage-driven. An input can
best be regarded as the base of a transis-
tor in a common-emitter configuration.
Consequently, the input voltage is
always about 0.6 volt.

R1 is connected to the reference voltage
and a constant current therefore flows
through this resistor. By virtue of its
high open-loop gain, the op-amp will
adjust its own output so that the same
current flows into its inverting input,
and the current through the
temperature-sensing diode (Dl) there-
fore remains constant.

This arrangement is necessary because
the diode is, in effect, a voltage source
with a finite internal resistance, and any
variation in the current flowing through
it would therefore produce a change in
the voltage which would be incorrectly
interpreted as a change in temperature.
The output voltage at pin 4 is thus equal
to the voltage at the inverting input plus
the voltage across the diode (the latter
varying with temperature). C3 prevents
oscillation.

Pin 1 of IC 2B is connected to the fixed
reference potential and a constant

current therefore flows into the non-
inverting input. The inverting input of
IC 2B is connected via R2 to the output
of IC 2A (pin 4), so that it is driven by a
temperature-dependent current. IC 2B
amplifies the difference between its
input currents to such an extent that
the voltage variation at its output
(pin 5) can easily be measured with a 5-
to 10-volt f.s.d. voltmeter.

If a panel meter is used, Ohm’s law will
have to be invoked to calculate the
series resistance. If a 1 00-/iA f.s.d.
instrument with an internal resistance of
1200 £2 is used, the total resistance for
10 V full-scale deflection must be

R5 should therefore be
100 k - lk2 = 98k8. The nearest
standard value (100 k) can be used.
Calibration proceeds as follows: the
zero point is first set by PI with the
temperature sensor dipped in a cup of
melting ice. Full-scale deflection can
then be set with P2 ; for this the diode
can be immersed in a hot liquid whose
temperature is known (e.g. boiling water
or water which has been checked with a
reliable thermometer as being at 50 ).

dock

suppression

E. Krause

Digital clocks in which the timing pulses
are derived from the 50 Hz mains fre-
quency are very sensitive to mains inter-
ference. An excessive amount of inter-
ference can often cause the clock to
gain several minutes per hour.

An adequate solution is offered by the
circuit given here; if it is used in series
with the clock, the latter will always
receive a ‘clean’ 50 Hz pulse train.

After the 50 Hz supply has been
positively rectified, the signal is fed to
the trigger input of a monostable multi-
vibrator 74121. The values of Cl , R 1 ,
and PI are chosen so that the pulse
width of the monostable is just under
20 ms.

Adjustment is simple: after the clock
has been connected, PI is first set to its
maximum value; the clock should then
run at half speed, because alternate
50 Hz input pulses occur whilst the
monostable is still triggered and have no
effect. Then PI is turned back until the
clock begins to run normally.

The value of R x depends on the supply
voltage. It must be chosen so that a

voltage of no more than 5 V p .p occurs
on pin 5 of the IC. If the voltage across
the smoothing capacitor (Cfi) is

measured, R x can be calculated by
means of the following formula:
R x = 100(V C B-5)(J2).

722 — elektor july/august 1975

W. Hollevoet

When the supply voltage is switched off,
TTL counters and memories lose their
information. To prevent them from
assuming a random position when the
supply voltage is switched on again, we
can use the circuit described in this
article for a complete automatic reset-
ting or presetting to a certain position.
The auto reset circuit functions as
follows:

When the supply voltage is switched on,
the output of N1 will become logic ‘1’,
because capacitor Cl is not yet charged.

Since C2 will not be charged either, the
output of N2 will also be logic *1’.

The time constants are chosen so that
the charging time of C 1 is shorter than
that of C2. This means that at a certain
moment Cl is sufficiently charged for
both inputs of N1 to be logic T’, so that
the output will be ‘O’.

A short time later, when C2 has charged
sufficiently, the output of N2 will
become logic ‘O’, with the result that
the output of N 1 will immediately
return to logic ‘1*. From then on the

circuit is stable.

The resistors R1 and R2 serve to dis-
charge the electrolytic capacitors Cl
and C2 when the supply voltage cuts
out. The moment of switching, and also
the pulse duration, can be modified by
experimenting with the values of the
two resistors and capacitors; the
charging time of Cl must, however,
always be shorter than that of C2! The
values given here will be suitable in most
cases.

Switch S2 serves to set the value of the
base-current. In the position drawn this
bias current can be varied from

0 ... 1 00 nA by means of the

1 M potentiometer; in the other
position this bias is non-adjustable and is

modulated - according to the instan-
taneous value of the incoming supply -
between 10 and 100 //A (at 6 V supply).
The circuit is only suitable for use with
low-power devices.

H. v. Hooydonk

Those who have access to an
oscilloscope can use this simple
accessory-circuit to display several of
the characteristic curves of diodes and
transistors. The table lists the possi-
bilities.

A normal bell-transformer can be used
for the power supply to the circuit. The
voltage delivered by this transformer
(which is invariably higher than the on-
load value marked on the housing!) may
vary between 3 and 6 volts. At 6 volts
the highest DC level in the circuit is
about 1 5 volts.

With switch SI in the position drawn
the circuit is correctly poled for testing
PNP transistors; the other switch pos-
ition reverses all polarities, for testing
NPN devices.

Function

X-input to:

Point M to:

Y-input to

Remark

curve

l c =f<'b>

4

1

3

S2 in position 1

A

l c =f<Vce>

1

3

5

S2 in position 2

B

lb= f <Vbe>

4(1)

2

1(4)

S2 in position 1

C

ld=f(Vdl

1(4)

2

4(1)

S2 in position 1

C

elektor july/august 1975 - 723

M. Mergel

The electronic siren described here is
easy and cheap to build.

The circuit consists of two astable
multivibrators, N1/N2 and N3/N4. The
0.2 Hz square-wave signal from the
latter oscillator is integrated by R3 and

7400-siren

C3 ; this voltage swings the frequency of
the other AMV (N1/N2) up and down
at 0.2 Hz.

The output level is about 2 V p .p,

^ sufficient to drive a power amplifier
directly.

Parts list

Resistors:

R1,R2 = 4k7
R3= 10 k

R4 = preset potentiometer 4k7
R5 = 5k6
R6 = 1 k

Capacitors:

* C1,C2 = IOOO/t/6 V

C3 = 500 #1/6 V
C4.C5 = 470 n
C6 = 150 n

Semiconductor:

N1 . . . N4 = 7400

B. Krone

Many TTL circuits have relatively low
noise immunity and consequently
suffer from spurious pulses upsetting
*» the logic process.

This interference can be greatly reduced
by using a circuit that recognises logic
4, pulses by their pre-determined length
and level.

Some type of RC network (figure 1)
could be used for this purpose. It
would, however, affect the TTL levels,
narrowing the margin for the R and C
values.

A better performance is given by the
circuit of figure 2. It compares the dur-
ation of the incoming signal with the
length of the ‘set’ state of a monostable
multivibrator. The incoming pulse will
have effect only when its length outlasts
this state. Shorter pulses will not only
be rejected but also, at their trailing
edges, ‘reset’ the MMV to prepare it for
the next pulse recognition.

Components shown in figure 3 deter-
mine the delay characteristic of the cir-
cuit (connected to pins 14 and 15).
Figure 4 shows the logic pulse diagram.
The MMV is ‘reset’ by a logic *0’
whereas a logic ‘1’ of pre-determined
length is admitted.

interference rejector

l »J

MVQ | ty j

i n n

724 - elektor july/august 1975

Control PI adjusts the correct oscillator
frequency to 57 kHz: as soon as the
pilot tone is detected, the LED will
light. The LED series resistor R4

H. Emmerl

For frequent commercial or holiday
drivers by car in W-Germany it is
interesting to know that a number of
West German FM radio stations trans-
mit, from time to time, useful traffic
information. The transmitters in ques-
tion also carry a continuous 57 kHz
pilot tone, enabling a pilot tone indi-
cator in the car to show that the car set
has been tuned to the traffic news
transmitter.

Readers will, no doubt, recognise the
stereo decoder integrated circuit
MC 1310 P(SN 76 11 5) put to a some-
what unorthodox use. Due to the fact
that the frequency of the traffic news
pilot tone differs from that of the stereo
pilot tone, the PLL oscillator used for
stereo decoding must be adapted. The
diagram shows the components R1 , PI
and Cl values that determine the VCO
frequency.

R2 controls the circuit sensitivity, C4
(which should preferably be a tantalum
type) gives a slight switch-on and
switch-off delay.

For a 1 2 V battery voltage and 20 milli-
amps LED current this gives
R4 = 500 ohms; the nearest standard
value (470 J2) can be used.

It is recommended to incorporate the
circuit in the car set itself. The

de-emphasising network in the detector
output circuit must be removed; it
would impair the 57 kHz pilot signal.
The treble boost caused by the removal
can easily be compensated for in the car
set audio amplifier.

F.H. Burmester

It is often necessary to provide a nega-
tive supply in circuits where the main
supply voltage is positive. For instance,
stabilised power supplies may require a
negative reference voltage, so that out-
puts close to zero volts may be obtained,
or it may be necessary to use operational
amplifiers requiring + and — supplies in
logic systems having a single positive
supply rail. For these applications a
second transformer winding is an ex-
pense that can be avoided using this
simple circuit.

The circuit operates on a voltage doubler
type of principle. The positive supply is
obtained from a conventional bridge
rectifier. On the positive half-cycle (with
respect to the end of the transformer
winding marked with a dot) Cl charges
through D1 and D2. C2 charges through
D5. During the negative half-cycle Cl
charges through D3 and D4.

D5...D6=1N4002

Whilst D3 is conducting, the potential
on the positive end of C2 is held at just
below zero volts (due to the voltage
drop across D3). This means that the
negative end of C2 is at minus the
supply voltage. C3 therefore charges
from the voltage across the transformer
secondary and C2 through D4, D6 and
Cl. To ensure that (almost) equal volt-
ages appear across Cl and C3, D3 must
remain in conduction.

This means that the current drawn by
the positive supply must be greater than
that drawn by the negative supply, so
that the difference current will flow
back along the zero volt rail, through
D3. If the positive and negative supply
currents are equal then an indicator
lamp may be connected across the
positive supply to produce an im-
balance. This will also serve as a mains
indicator.

The circuit described here is
input 'exclusive or’ gate.
When all inputs are logic ‘ 1 ’,

a four
the output

of N 1 is logic *0’. One input of each of
the following gates is now driven to
logic ‘0’ from N1 and the other input

of each gate is driven to logic ‘1’ by the
inputs. The outputs of N2, N3, N4 and
N5, and thus also the inputs of N6 are
now high, so that output Y is logic ‘O’.
Similarly, when the inputs are all logic
0’ output Y will be 'O’. If only one of
the input signals is low, say B, the out-
put of N1 (and with it the common
inputs of the following gates) becomes
logic ‘1’. The outputs of gates N3, N4
and N5 now all produce a *0’, so that
Y becomes logic ‘ 1 ’. For all other
combinations ( 14 in total) the output
will always be logic '!’.

elektor july /august 1975 — 725

During recording on tape, it is necessary
to ensure that the signal voltage at some
point does not exceed a predetermined
level. A similar situation can occur with
(pre-)amplifiers and measuring instru-
ments.

In such cases a meter-type indication is
£ot the only possibility - and indeed it
is invariably not the best. An overmodu-
lation indicator that lights a lamp is
cheaper - and gives a more distinct
warning. The accompanying circuit is
designed for this purpose.

Transistor T1 is an emitter follower,
providing a high input impedance
(about 100 kJ2) to minimize the load on
the signal source. The trimmer R3 sets
the voltage at which the lamp will just
light up (overmodulation level).

The circuit around T2 is a x 100 ampli-

fier which enables the threshold to be
set as low as 5 mV. When this high
sensitivity is not needed, i.e. when the
threshold is 0.5 volt or higher, the stage
can be omitted. The points A and B are
then bridged. If the high input im-
pedance is also unnecessary, as for
instance when a loudspeaker-connection
is being monitored, it is obviously per-
missible to omit the input stage also.
Figure 2 shows how the input is made
to point B in this case.

The circuit following point B is the
indicator proper. The current through
R6 normally ‘bottoms’ T3, so that T4 is
cut off. Alternating signal voltage at
point B however, rectified by the action
of D 1 , D2 , C3 and C4, will cause a nega-
tive drive to be applied to T3 base.

When this AC voltage exceeds about

overmodulation

indicator

0.5 volts, T3 will no longer be
bottomed , so that T4 will start to con-
duct. ‘Monoflop’ action via C4 will now
ensure that even short signal peaks are
clearly indicated by the lamp.

When selecting the type of lamp, one
should note that the maximum available
current is about 100 mA. With a supply
voltage of 7 V as shown, the lamp
should be a 6 ... 7 volt type. If circum-
stances dictate, a resistor can be inserted
in series with the lamp. Given the
supply voltage (Vb), the lamp volt-
age (Vl) and the lamp current (II) in
amps, the series resistor (R) value re-
quired is:

n _ vb - vl
II

To take an example, suppose that a
6 volt 50 mA (=» 0.05 A) lamp is to be
used on a 9 volt supply :

for which the nearest lower standard
value of 56 would be taken.

G. Tajbl

This circuit uses a start/stop oscillator
and a monostable multivibrator. The
monostable is triggered by the trailing
edge of the pulse to be multiplied (see
pulse train 1 in the diagram). This sets
the Schmitt trigger SI oscillating (2) for
'• period that can be set by PI . P2
controls the oscillator frequency. In this
way, one single incoming pulse
generates a well-defined number of out-
going pulses. A further Schmitt trigger
S2 is used as a buffer stage and produces
the pulse train shown at (3) in the dia-
gram.

PI must be adjusted to a value permit-
ting the highest number of incoming
pulses to be passed on correctly. If the

monostable period is too long, the
74122 will act as an unwanted fre-
quency divider. After PI is set to the
correct value, P2 is used to set the
desired multiplication coefficient.

The pulse multiplier will find a number
of different applications. It will, for
instance, permit a low frequency such a:
heart beats to be counted with reason-
able accuracy.

® _r

©j

L_*

ijtjtj-ltl

726 — elektor july /august 1975

Electronics is beginning to establish a
firm foothold even in the world of
music, as is evident from the introduc-
tion by SGS-Ates of two ‘rhythm
generator’ ICs.

The simplest of these ICs is the M 252,
which can be obtained in a number of
different versions. In addition to an IC
with standard programming, ICs with
rhythm patterns tailored to the
customer’s requirements can also be
supplied.

The M 252 B1 AA or D1 AA has fifteen

integrated

standard rhythms and facilities for
driving eight generators (‘instruments’).
With this IC, the choice of rhythm is
determined by four programmable
binary inputs, as shown in the block dia-
gram. This makes it possible to have a
rhythm generator of very simple design,
because four single-pole programming
switches, operated as shown in Table 1,
are all that is needed for selecting the
rhythms.

As the circuit given in figure 3 shows,
only a few additional components are

N1 ...N3 = IC1=CD4011i
N4 . N6=IC2=CD4011-

elektor july/august 1975 — 727

needed for a complete rhythm gener- Besides the programmed inputs, the IC width) is lengthened by a monostable

ator. Each of the outputs of the M 252 has also a reset input with which the multivibrator in order to obtain a clear

can be connected to an appropriate rhythm pattern can be interrupted. This indication.

electronic percussion instrument, the input also serves as an output for the The stop-switch (SI) provides for reset-

output at pin 12 is used for both the so-called ‘down beat’ indication, because ting the IC and at the same time blocks

snaredmm and the claves. The snare- the internal reset signal is routed via the the tempo oscillator,

drum is generally used with rhythms 1 reset input to an output connection. Figure 4 shows the p.c.board and the

to 9 inclusive and the claves with By adding the isolating diode D2, the component layout for this circuit,

rhythms 1 0 to 15 inclusive (the South reset connection can be used as both an

American rhythms). This interlinking input and an output. In this way it is (SGS-A TES applj

does not occur automatically, so either possible to give an optical indication of
the claves or the snaredrum can be used the beginning of each beat directly from
with any rhythm. the reset pulse. The reset pulse (5 ps

W

N The M 253 exhibits a number of basic
^ differences in comparison with the

1 ^1 252. In the first place the chip is
simpler, as the binary programming
facility has been omitted. Secondly,
only twelve standard rhythms are avail-
able with this IC. As the block diagram
(figure 1 ) shows, the mode of operation
is in other respects practically the same
as the M 252.

As a result of the omission of binary
inputs, each rhythm must be selected
with a separate switch. In some ways
this is a disadvantage, but on the other
j hand it opens up the possibility of
making up a greater number of rhythms.
As each rhythm must be selected with a
separate switch, a 1 6-pin package is not
enough, so a 24-pin dual-in-line package
is used to accomodate the chip.

The design of the oscillator and the
down-beat indicator is the same as for
the M 252 circuit.

Pin connection data for this IC are
given in figure 2, and the complete cir-
cuit is shown in figure 3. It should be
noted that, as with the M 252, a single
output serves for both snare drum and
•laves. For the South American
rhythms, the claves are used in place of
the snare drums. The one exception to
this is the tango.

SGS-ATES appl.)

integrated rhytlim generator M 253

728 — elektor july/august 1975

D1 ... D4= 1N4002

,BFX34

IBC140

The rhythm-generator ICs which are
described elsewhere in this issue need
power supplies at several voltages, and
the circuit given here is one example of
the way in which these can be provided.
The IC itself uses a +5 V and a - 1 2 V
supply ; the associated instruments and
the preamplifier, however, need a poten-
tial of +12 V. The circuit shown here
can also provide this.

If the +1 2 V supply is not used, the
+5 V stabilisation section can be
omitted, and a different stabiliser IC,
the TBA 625 A, can be used. This
provides +5 V instead of the +1 2 V
from the TBA 625 B.

(SGS-ATES appl.)

BFX41

BC160

rhythm-generator power supplies

Figure 1 shows a NOR-gate with two
inputs (A and B) built from a NAND-
gate. The NAND-gate inputs are con-
nected together and the new inputs (A
and BJ are provided with two diodes.
The input and output of this NOR-gate
are TTL compatible.

Figure 2 gives a NOR-gate configuration

similar to figure 1 , but this time for j
COS/MOS circuits. The input resistance 1 •
at the points A and B in figure 2 is .1
nearly equal to Rl.

easily soldered to the transistor housing
Those who have the facilities can then
silver-plate the completed accessory (for
example in a ‘spent’ fixing bath!). The
least expensive approach is to use a tran-
sistor with a chromium-plated housing;
those with more surplus cash can have
the jeweller do a gold-plating job.

The photograph illustrates a possible

D. Jacobs

Here is a suggestion for fanatical
electronics-amateurs who wish to be
immediately distinguishable from any
uninitiated persons with whom they
might be consorting. It is certainly
guaranteed to attract the attention of
any other initiate who might be in the
vicinity - with the object of starting a
suitable conversation.
Power-semiconductor victims of
accidental death - a particularly suit-
able type is called ‘2N3055’ - can be

final result. This might even be an idea
for somebody dreaming up a present
particularly suitable for the intended
recipient !

easily turned into very exclusive cuff-
links that any initiate will recognise at
great distance. The necessary clips are
readily available in many shops that
cater for hobbyists - and they can be

opto-coupler
for mains
synchronised
pulse generator

Many different methods are followed
to generate mains synchronised pulses.
This circuit employs an opto-coupler to
give complete isolation from the mains.
The coupler is shown in the diagram ; it
consists of a LED, a phototransistor
and a few standard components.

A resistor is used to connect the LED
direct to the mains, or to a transformer.
The resistor value is chosen for a LED
current of about 1 5 m A. The reverse-
parallel diode D 1 by-passes negative
peaks across the LED. The light emitted
by the LED causes a current to flow
through T 1 , which is amplified by T2

and T3. The high gain causes a (TTL
compatible) square wave to appear at
T3 collector.

The phototransistor is made from a
BC 107, by carefully filing or sawing
down the top of the metal case, which
exposes the base.

elektor july/august 1975 — 729

Resistors:

R1 = 2M7
R2 - 4M7

R3,R4,R5,R12 = 1 k
R6,R9,R13 = 4k7
R7 = 39 k
R8 - 5k6
RIO -47 k
R11 = 220 k
R14= 100 k

Capacitors:

Cl = 1 /i, 6 V tantalum
C2 = 470 /i, 6 V electrolytic
C3=100/i, 16 V
C4 = 100 /i, 25V
C5.C6 = 2n2

^C8,C9,C12 -25/i. 16 V
C10= 1 n
C11 =50ji,6V

Sundries:

’ PI = preset potentiometer

P2 = potentiometer
4k7 log. stereo
P3.P4 ** potentiometer
100 k lin. stereo
P5 = potentiometer 10 k lin.

Semico nductors :

.T1 ,T3 = TUN
T2 = TUP

ouster eo

control

amplifier

Transistors T1 and T2 form a voltage
amplifier with a high input impedance
and a low output impedance. When the
slider of preset potentiometer PI is set
to give the full value of 1 k, the input
sensitivity in combination with the
3-watt amplifier is about 150 mV for
the 1 2-volt version working into a 4-12
load, or 200 mV for the 17-volt version

PI can be set to a value lower than 1 k.
If switching to different values of
input sensitivity is needed, fixed re-
sistors can be used in place of PI , with
values determined according to the
formula:

Rx = ^ ohms ^

where Vj n is the RMS input voltage in
mV. The formula holds good for input
voltages from 5 mV to 250 mV. T3 is
used in a standard Baxandall tone con-
trol circuit. The 1 n capacitor between
the collector and earth is to prevent
oscillation.

730 — elektor july/august 1975

The disc preamplifier, of which only one
channel is shown in the circuit, incor-
porates equalisation to correct the out-
put of a magnetic cartridge according to
the RIAA playback curve, and also
amplifies the signal to a level sufficient
to drive the control amplifier. It consists
of a two-stage voltage amplifier, T4 and
T5, with the RIAA feedback network
R18, R19, CIS and C16 connected
from the collector of T5 to the emitter
of T4. DC feedback and biasing of T4 is

provided by R15. The disc preamplifier
board should preferably be mounted in-
side the turntable box as otherwise the
capacitance of the screened lead be-
tween the cartridge and the disc pre-
amplifier can form a resonant circuit
with the self-inductance of the car-
tridge. If this resonance lies within the
audio spectrum it may cause a peak in
the frequency response. Of course some
cartridge manufacturers quote a rec-
ommended load capacitance and if this

is so their recommendations should be
adhered to. Another good reason for
mounting the disc preamplifier inside
the turntable is to keep it away from the
hum fields of the amplifier’s mains
transformer. Turntable motors usually
have much less stray field then the aver-
age mains transformer!

It can be seen that the layout for the
two channels is symmetrical.

Resistors:
R15 = 47 k
R16 = 1k5
R17 = 18k
R18= 12 k
R19 = 120 k
R20 = 2k7
R21 = 10 k
R22 = 4k7
•R23 = 100 k

Capacitors:
C13.C18 =
C14 = 100,
C15 = 27 n
C16 = 6n8
C17 = 47 ..

Sem iconductors :
T4 = BC 109 C
T5 = TUN

Transistors T 1 and T2 form a direct-
coupled voltage amplifier. Resistor R6
and diodes D1/D2 determine the quiesc-
ent current of the quasi-complementary
driver stage T3/T4 and the output
stage T5/T6. The values of resistors R7
and R8 are chosen so that the output
transistors are either just biased on or

JN1613

Capacitors:

Cl =2,2 At. 16 V
C2 = 100*1, 16 V
C3 = 10 n
C4 = see table
C5.C6 = 47 n

■ 2N 1613

Semico nductors :
T1 ,T3 = TUN
T2.T4 = TUP
T5.T6 = 2N1613
D1,D2 = DUS
heatsinks forTO-5

austereo

3-watt

hifi amplifier

IOI01GIO1OIOIO1

elektor july/august 1975 - 731

just cut off depending on the gain of the
transistors used. C3, C5, C6 and R3 help
to maintain stability. The input sensi-
tivity of the amplifier is about 400 mV
for 12-volt operation with a 4-£2 load,
and 600 mV for 17-volt operation with
an 8-£2 load. The gain may be increased
by reducing R4 but this is not rec-
ommended as instability may occur
and distortion is increased.

The following layout precautions
should be noted when assembling the

2. Separate leads must be run from the
supply to the supply points on each
board.

3. Outputs of any board should be kept
well away from inputs of other
boards (except of course where the
output of a stage is connected to the
input of the succeeding stage).

4. Care should be taken to avoid earth
loops. Each section of the amplifier
should have only one connection to
supply common.

completed board onto a chassis:

1 . Loudspeaker common must be con-
nected directly to the power supply
common and should be kept well
away from the boards.

N2. With potentiometer PI the fre-
quency can be varied between about
40 x per minute and 10 x per second.
Point 3 of gate N1 drives transistor T1
which is connected as a buffer stage. Via
resistor R4, the emitter voltage of this
transistor periodically drives transis-
tor T3 on and off. The gates N3 and N4
form a square wave oscillator with a fre-
quency around 500 Hz. The output of
this oscillator (point 1 1 of gate N4)
switches transistor T4 via T2 and R8.

T4 does not pass the oscillator signal
continuously because the base of this
transistor is periodically cut off by T3.

mostronome

The short whistle (abt. 0,15 s) period-
ically available at the emitter of T4
arrives at the base of T5 via R9 . Via R 1 0
and volume control P2, the collector of
T5 drives a loudspeaker.

The impedance of this loudspeaker may
lie between 4 £2 and 16 £2.

Since the average current consqmption
is low because the whistle is very short,
a 6 V battery (powerpack) is quite suf-
ficient to serve as the power supply.

— © 6...15V/50mA

The mostronome is a metronome which
differs from all other types in that it
does not produce a ‘tap’, but a brief
whistle. This is pleasant to the ear and
at the same time far more penetrating
than the sound produced by the con-
ventional ‘tap’ generators.

The repetition frequency of the mostro-
nome is generated by the gates N1 and

12 V

17 V

R12

680 £2

1 k

C4

LS

4700 H

4 £2

2200 H

8£2

732 - elektor july/august 1975

Transistors T1 and T2 form a Darlington
pair acting as a compound emitter-
follower with a reference voltage pro-
vided by Zl. Z1 is chosen as a 13 or
18 volt zener for a 12 or 17 volt supply
respectively. Since T2 dissipates only a
small amount of power a heatsink is not
required.

ouster eo
power supply

Semiconductors:
T1 = 2N3055
T2 = see table

B40C2200

Trafo:

Tr = 2 A :

BC107I

on/off switch

is derived from the 50 Hz mains: the
output from the bridge rectifier
(100 Hz) is divided by 10. The resulta: 4
1 0 Hz signal is fed to gate N3 ;
depending on the state of the set/reset .
flip-flop (N1/N2) it can be passed on to
the output transistor. This transistor car
be used to drive a (reed) relay; the relay,
contacts are connected in parallel witlf^
the ± key.

Of course, if enough is known about the
calculator circuit the relay can usually
be omitted. For instance, if the ± key
connects a positive voltage to supply
common, the relay is omitted, the out- J
put transistor can be a TUN, and its ']
collector is connected directly to the
positive terminal of the ± key. The
supply commons are interconnected.

surgery on the calculator can be reduced
to a minimum: all that is needed is an
external connection to the contacts of
the ± key (figure 2).

The stopwatch adapter itself is simply
a standard frequency generator that can
be switched on and off. The frequency

Many pocket calculators have a constant
factor memory. If a 1 is stored in this
memory, depressing the ± key adds 1 to
the number already displayed.

If the original number displayed is 0 and
the ± key is depressed at regular
intervals, the number displayed at any
given time is equal to the number of
times the key has been depressed, which
in turn is proportional to time. This
means that the calculator can be used as
a stopwatch.

Figure 1 shows a practical circuit. It can
be built as a separate unit, so that

doubles

elektor july/august 1975 - 733

I The ‘austereo’ 3-watt amplifier is used
| as a drive amplifier for the 2N3055 out-
put transistors, and very few changes in
the circuit or the component values are
needed. Capacitor C7 is introduced to
compensate for the phase shift due to
the output transistors. The value of R1
is reduced to 56 k, and additional de-
coupling, in the form of a 47-k resistor
and a 10-/t capacitor, is inserted be-
tween the high-potential end of R1 and
supply positive. The output impedance
is very low, as T5/T7 and T6/T8 form
i power darlingtons. The ‘austereo’

1 control amplifier is well capable of
supplying the 1-V RMS input voltage
1 -needed.

Because of the low input sensitivity, the
amplifier has good stability and its
sensitivity to hum is low. Substantial
negative feedback via R4 and R5 ensures
low distortion.

» Maximum permissible supply voltage is
I 42 V. The power supply circuit is de-
veloped from the stabilised power
supply unit for the ‘austereo’ ampli-
fier, with circuit modifications and also
changes of component ratings to suit the
higher working voltages.

Lp addition to the heat sinks shown in
the amplifier and power supply circuits,
the three 2N3055 transistors should be
cooled by mounting them on the ampli-
fier or power supply boxes (as appli-
cable) using mica insulating washers.

Jhe power supply table is worked out
fr stereo. Power for the control ampli-
fier is drawn from ~ *

Semiconductors:
T1.T3 = BC107
T2.T4 = BC177
T5,T6 = 2N1613
T7.T8 ■ 2N3055
D1.D2-DUS
5 heatsinks for T05

.2N1613

BC177L,

J2N3055

2N1613 with its
*^ase potential held at half the main
^supply voltage.

'austereo

15-30 watt amplifier

2N1613.

Parts list for 15-30 W power supply

B40C2200

amplifier

■2N1613

Transformer:
see table

Sundries:

B = Bridge rectifier; see table
Z = Zener diode, see table
N = Neon lamp
S = On/off switch
Heat sink for T05

Power o

utput (w) with:

8ft

4ft

2 ft (R13.R14 = 0.1 ft

power 30

9.5

19

35

25 C4

supply 36

15

30

55

25 working

voltage (V) 42

20

40

70

35 voltage (V)

2200

4700

10,000

C4 capacity 1//F)

Output power

Transformer

secondary

V | A

B

B40C...

Cl x 100 /If

Z

R1 ft

1 watt

9.5-19-35

30

1-2-4

1000 - 2200 - 5000

22 - 47 - 100
50 V

ZD 33

680

15-30-55

36

1 .2 - 2.4 - 4.8

2200 - 3200 - 5000

22-47-100
60 V

ZD 39

820

20 - 40 - 70

42

1 .2 - 2.5 - 5

2200 - 3200 - 5000

22-47-100
70 V

ZD 43

1 k

t]

J-BL* i» ^IT)b‘ ci j7^

T"

y x r n Bb ° 7

7N1613

BCKJ7 Too

D i^Ldus in " -M

)

i dsi

2N3055

734 - elektor july/august 1975

H. Bense

The relatively high price of seven-
segment displays or nixies is not
economically justified when the display
is used for reading out a single instru-
ment only. A selector switch at the
display input end gives a ‘display-
sharing system’ that is more economical.
The logic Boolean OR function is
written as follows

x = as + bs

in which ‘s’ stands for:
switch open = M’,
s for switch closed = ‘O’.

In order to handle this function by
NAND gates, it is converted, using
de Morgan’s theorem, into

x = (as) (bs)

Selection in accordance with the given
function must be carried out for each
bit to be chosen, and for one bit the
function is performed by the part of the
circuit diagram in dotted lines.

The switching part (S, Rl, N13) is
common to all bit selectors.

The gates operate as follows. The TUN
emitter receives the output signal from
the tone source. Striking the key will
open KS and admit the tone, via Rb, to
the 2', 4’ and 8' stops S2', S4' and S8'.
Points VI in figures 2 and 3 are inter-
connected. Point Vb is the power
supply. Its voltage must amply cover the,
largest signal amplitude from the tone ’*’/
sources.

R 1 and R2 bias the stop rails to prevent
switching clicks; their resistance can be
high, as the electronic gates, when
closed, do not load the stop rails. PI is J
used to minimise the clicks.

G. Knieling

Most electronic organs use multiple
mechanical key contact assemblies,
which can be eliminated by multiple
electronic gates operated by a single key
contact.

Figure 1 shows the principle. KS is the
normally closed contact. Figure 3
shows, as an example, three keyboard
contacts each feeding a set of three
organ stops. Resistors Rb define the
tone balance and are determined when
the organ stops are voiced.

for organs

Parts list.

P. Rampelbergh

When testing continuity by means of an
ohmmeter, there is always the danger
that resistors, semi-conductors etc. are
involved in the measurement, giving
false readings. Furthermore the current
or voltage of the meter can sometimes
damage the circuit.

With the design described in this article,
these disadvantages are avoided; the
tester does not ‘see’ a resistance greater
than 1 ohm as a connection.

The measuring voltage is no more than
2 mV, so that no diode, IC or any other
element is included in the measurement.
The maximum measuring current into
the circuit under test is 200 pA.

The indication is by way of a LED; if it
i8 fitted in one of the two measuring
probes, the tester becomes quite a
handy instrument. The supply can be
via two 9 V power packs.

The voltage offset of the IC can be
compensated by means of poten-
tiometer PI (which can compensate
about 8 mV). For adjustment, the
measuring points are short-circuited and
P 1 is adjusted so that the LED just
lights. When the measuring points are
opened again, the LED goes out.

A printed circuit board was designed for
this very inexpensive and yet
extremely handy measuring instrument.
By virtue of its modest dimensions
(18 x 68 mm) it can be fitted in practi-
cally any suitable housing.

resistors:

R1.R3 = 22 k
R2 = 10 £2
R4.R5 = 1 k
R6 = 470 k
R7 = 470 fi

PI = preset potentiometer
lOklin

IC ■ 709
D1 = DUS
D2 = LED

p.c. board
and wiring
tester

The hum receiver consists of one
COS/MOS IC comprising four NAND-
gates (CD 4011). The four gates are
series-connected as amplifier elements.
The first gate (N 1 ) picks up the hum
radiated by the mains. The inputs of
this gate should not be in the close
vicinity of other sources of inter-
ference (amplifier outputs, etc.). A
copper wire 2 to 3 cm long is sufficient

hum receiver

to pick up the hum and to produce a
square wave with a frequency of 50 Hz
and a risetime of about 20 ns at the
output of gate N4. Depending on the
conditions, one or more gates can

K) 3 ... 1 5 V/2 n A

sometimes be omitted.

The current consumption of the entire
CD 401 1 is so low that if a 4.5 V bat-
tery is used as the power supply, its life
will be essentially shelf life.

This circuit is intended for continuous
speed control of small series wound
motors, as used in most electric hand

drills, etc.

The speed is adjusted by potentiometer
P 1 . The setting of this potentiometer

determines the moment at which the
triac is triggered. If the motor speed
drops below the preset value (under
load), the back EMF of the motor
reduces. Consequently, the voltage

triac speed
controller

across R 1 , P 1 and C5 increases so that
the triac is triggered sooner, and the
motor speed will be increased. In this
way, a certain amount of speed stabilis-
ation is pbtained.

(Transitron)

iHH

736 — elektor july/august 1975

The MM 5780 is an integrated arith-
metic teacher. When a pupil has to solve
a problem involving basic addition,
subtraction, multiplication and/or
division, he can enter the problem into
the teacher while he himself is also
solving it, and then enter his solution.
The teacher tests the pupil’s answer, and
if it is correct LED ‘R’ (= right) lights.

If the answer was wrong, LED ‘W’

(= wrong) lights, but the correct answer
is not given . . .

Before starting a new calculation, key
C’ (= clear) must be depressed.

The circuit is quite simple, as all necess-
ary functions (contact-bounce sup-

arithmetic

teacher

pressor, calculator, comparator and
LED drivers) are integrated on the chip.
Singlepole pushbuttons can be used for
the keys; it is also possible to use the
keys (and case) from a cheap pocket
calculator - the original calculator chip
and LED displays can be used for
various other purposes, e.g. as a counter
or stopwatch.

A 9 volt battery can be used as power
supply — it should last for
10 ... 30 hours of operation.
Alternatively, a mains supply can be
used, delivering 7 . . . 9.5 volts.

(National appl./

LED display
dimmer

Segment currents for the commercial
range of LED displays must be limited
to some 25 mA, which is usually done
by means of series resistors of which a
six -digit display will, then, require a
number of 42 (disregarding the decimal
point). Once installed, the LED lumi-
nance can no longer be adjusted.

The circuit shown, on the other hand,
feeds the segments from an adjustable
voltage stabiliser. The luminance can be
varied over a wide range and, moreover,
the pcb lay-out is simplified by avoiding

the use of series resistors.

LED brightness depends on the setting
of voltage controls PI (coarse) and P2

(fine), somewhere between 0 and
4.3 volts, the exact adjustment being
rather critical due to the diode charac-
teristic of the LED. When setting up the
display, the voltage is first set to the
lowest value, then carefully raised to
obtain the correct brightness.

The total current for a six-digit display
should not exceed about 1 amp for a
safe segment current of 25 mA (7 seg-
ments at 25 mA for 6 digits).

The choice of the series transistor (Tl)
is governed by its safe dissipation.

This simple circuit will find many
applications as a battery eliminator for
low power requirements. It consists of a
transformer, a bridge rectifier and an

electrolytic capacitor followed by a
zener controlled series pass transistor.
The output is stabilised at 7.5 volts. The
stand-by battery, 7.5 volts, in series

with D2, floats across the output ter-
minals, ready to take over in case of
mains failure. The voltage drop across
D2 will then reduce the power supply to
about 7 volts.

Resistor R2 has an additional function:
when working off the mains it will
trickle charge the dry cells or storage
battery. Since not many accumulators,
and very few dry batteries, will stand
prolonged overcharging, R2 must not
allow for more than just the self-leakage.
Its correct resistance can be found by
dividing the voltage potential difference
between the zener and the battery by
the safe trickle current, which may
amount to some 0.7 milliamps.

elektor july/august 1975 — 737

R.L. Rappenecker

The supply voltage for a relay is chosen
so that the relay will pull-in reliably.

The holding voltage is, however, much
lower: often even less than half the
pull-in voltage.

Consequently most relays function satis-
factorily at a reduced voltage provided
it is ensured that at the moment of
activation the voltage is increased suf-
ficiently. The circuit given here is
suitable for relays drawing up to
100 mA, whilst the supply voltage must
be lower than 25 V.

Using this circuit has two advantages:
firstly the relay draws considerably less
current ; at half the supply voltage the

power consumption has already been
reduced to one quarter! Secondly, a
given relay can be used at voltages
which so far were too low to ensure
reliable functioning. (For example a 6 V
relay that must function on the 5 V
from a TTL supply). The circuit is con-
nected to a supply voltage which is
certain to hold the relay. As long as SI
is open. Cl is charged via R2 to the
supply voltage. R 1 is connected to the
+ terminal and T 1 will not conduct.

As soon as SI is closed, the base of T1
is connected to supply common via R1 ,
so that it conducts and drives the relay.
Via SI the positive terminal of Cl is
connected to supply common, and since
this capacitor was charged to the supply
voltage its - terminal is now at a nega-
tive potential. The voltage across the
coil of the relay is now equal to double

economic relay

the supply voltage, and the relay will
pull in.

Switch S 1 can, of course, be replaced by
a transistor (TUP or TUN) which is then
switched on or off.

Since the price of a universal meter is
usually more than proportional to its
‘ohms-per-volt’, it is useful to have an
adapter which makes the input
impedance sufficiently great to measure
high-impedance circuits, such as pre-
amplifiers and MOS circuits. The
adapter shown here can be used with
rfi'eters having an impedance exceeding
about 2 kfi/V. The circuit consists of a
source follower with an input
impedance of about 1 0 Mohm where
the pinch-off voltage is used to obtain
a zero setting for the instrument. The
voltage at T1 source is higher than the
artificial zero (via R3 and R4) and the
voltage at T2 drain is below the zero,
provided sufficient current flows

through T1 and T2. The choice was in
favour of a FET current source because
of its very low temperature coefficient.
PI is used to set the current through T1

FET fT adopter

and T2 to about 1 mA (4.7 V across
P3). Next, P3 is used for a coarse zero
adjustment, and finally P2 is used for
fine adjustment of the zero setting of
the meter. The values of P 1 , R 1 , P2 and
P3 apply for a meter with a sensitivity
better than 1 0 kJ2/ V. When using a less
sensitive instrument, the current
through the FETs should be increased
accordingly and the values of the above-
mentioned components reduced. The
supply can be two series-connected 9 V
batteries, or one 1 8 V mains supply
which need not be stabilised. The
maximum permissible supply voltage is
30 V. At 18 V the current consumption
is 10 mA, the maximum input voltage
is then ± 6 V. If higher voltages must
be measured, an input attenuator is
needed.

The frequency meter is suitable for any
sinusoidal or pulse input signal. The
minimum input voltage amplitude at
which the circuit still functions satisfac-

torily is 25 mV r.m.s. The measuring
range of the circuit is from 10 Hz to
1 00 kHz depending on the setting of
switch SI . Each setting is calibrated
with a 20 k potentiometer. The total
current consumption is about 10 mA.
The values of R 1 and Cl are selected
according to the type of micro- or
milliammeter. The table gives the

f.s.d, R1 Cl

100 HA 39 k 2/i, 10 V

500 /iA 6k8 15ji, 10 V

1mA 3k9 25 fl, 10 V

various values of R 1 and C 1 for meters
with different values of full-scale deflec-
tion.

aaalogue

frequeacy

meter

(rev.couater)

738 — elektor july/august 1975

The luminous intensity of fluorescent
lamps cannot be controlled by means of
conventional light dimmers, unless
certain modifications are carried out. In
the circuit described here the heaters of
the fluorescent lamp are pre-heated by
means of a heater transformer with two
separate windings. The starter is

omitted, while the choke (LI) can
remain in the circuit.

The (conventional) triac control is con-
nected via the choke with a 33 k/2 W
‘bleed’ resistor across the tube and
choke to supply current to the dimmer
whilst the tube is extinguished.
Alternatively, three resistors of 100 k/

fluorescent
lamp dimmer

'A W can be connected in parallel. Any
suppression networks originally in the
triac dimmer should be removed; the
large self-inductance of LI will limit the
interference caused by the dimmer to a
minimum. If the range of control is
considered insufficient, it is possible to
experiment with the value of capacitor
Cl.

Standard precautions should, of course,
be taken: the circuit must be mounted
in an insulating box, PI should have a
plastic spindle, and Cl must be a 400 V
type.

This four-channel switch is an excellent
aid when comparing digital signals with
an oscilloscope. In the version described
here the switch works as a chopper.

With a small addition it can also work in
‘alternate’ mode. The chopper fre-
quency is generated by a Schmitt trigger
(N 1 ) connected as an oscillator. With
the given component values the oscil-
lator frequency is about 2.5 MHz. This
means that the chopper cycle frequency
will be around 600 kHz, for the four-
channel version described here.

The oscillator signal is fed to a divide-
by-four counter, from which the chan-
nel selection signals are derived. These
signals are used to drive the gates N2,
N3, N4 and N6 sequentially. The signal
from the selected gate is fed to the
emitter of T1 via N5. This transistor
works as a current source, where the
current is determined by the state of the
divide-by-four counter and the input
signal level The current gives a voltage

digital

four-chaaael

switch

drop across R7, which in turn deter-
mines the position of the trace on the
oscilloscope tube. With no input signal
on the four inputs four horizontal lines
are displayed. The position of each line
is determined by the setting of the
current source, which in tum is deter-
mined by R2 to R6.

With the given component values the
DC level of the output voltage is about
4.6 V. The logic signals have an ampli-
tude of about 50 mV.

With some oscilloscopes the range of the

vertical position control will be insuf-
ficient to bring all the signals onto the
screen. In this case resistor R7 can be
replaced by a preset potentiometer of
470 J2, so that the 70 mV now available I
per channel can be expanded to a maxi-
mum of about 250 mV. However, this
in turn leads to a reduced switching
speed of the circuit. T1 should certainly I
never be driven into saturation, because j
this would have a very unfavourable
effect on both the switching time and
display of the signals.

When digital signals of a very high fre-
quency must be compared, the oscil-
lator frequency must be reduced by
increasing Cl to 10 n or even 100 n.

The signals are then no longer chopped,
but displayed sequentially on the
oscilloscope tube (‘alternate mode’).

To maintain the proper phase relation, I
one of the signals must be used as an
external triggering source for the
oscilloscope.

elektor july/august 1975 - 7:

economic

flashlight

Since the full brightness of a torch is
only rarely needed, a suitable dimmer
will be a welcome energy saver.

The system is designed around an
astable multivibrator whose duty-cycle
can be varied by means of poten-
tiometer PI . The diode has been added
to improve the rise time. Via T3 the

AMV switches transistor T4, which in
turn switches the lamp. T4 requires no
extra cooling.

The control range is such that the lamp
can be adjusted to bum at about one-
third of its normal light intensity; this
means that the batteries will then last
three times as long.

The application of the circuit is, of
course, not limited to torches; it can
also be used for panel lights, radio scale
illumination, etc. If PI is replaced by
an LDR, it is also possible to obtain
an automatic dimmer that adapts the
brightness of the lamp according to the
ambient light conditions.

ear is limited. The purpose of the ultra-
sonic detector if to overcome this

limitation by changing the frequency of
high-pitched sounds such as dog

whistles, hardly perceptable gas leaks,
bat’s bleeps, and various man-made

ultrasonic noises such as softly fingering
a newspaper (which also produces a
sound in the audible range).

The “ultrasound’ picked up by the trans-
ducer is amplified and applied to a
product detector. An astable multi-
vibrator is used, since the BFO stability
is of minor importance. Apart from the
required difference signal, the circuit
also produces the BFO signal itself and
the sum signal, the latter two being
rejected in a low pass filter set at 4 kHz.

ultrasonic

detector

The resulting signal is once more ampli-
fied to drive a pair of headphones. The
circuit draws about 8 milliamps, so it
can be run from a dry battery.

740 - elektor july/august 1975

The 7490 contains two independent
(except for reset function) counters,
a divide-by-two (flip-flop) and a divide-
by-five counter. These may be used
separately, or cascaded to form a divide-
by-ten counter.

For use as a divide-by-two counter, the
input count is fed into input A (pin 14)
and the output is taken from output A
(pin 12). For a divide-by-five counter
the input is to pin 1 , and a binary
output sequence is obtained at pins 8,

9 and 1 1 .

There are two methods of connecting
the IC as a divide-by-ten counter. The
divide-by-two counter may be connec-

ted before the divide-by-five counter
(pins 12 and 1 linked, and the input
connected to pin 14). In that case a
BCD count sequence is obtained at
pins 8,9,11 and 1 2, in accordance with I
the truth table. If a symmetrical square- I
wave output is required for frequency I
synthesizer or other applications then I
the divide-by-two stage is connected
after the divide-by-five stage. In that
case input BD (pin 1 ) receives the input I
count. The D output (pin 1 1 ) is connec- |
ted to input A (pin 14) and the sym-
metrical square-wave output is obtained I
from the A output (pin 12).

In order for the counter to function the I
reset 0 and reset 9 inputs (pins 2, 3, 6 I
and 7) must be grounded.

For division by 3 the counter must reset
to 0 when the BCD output is 3
(i.e. 0011). The counter is therefore
connected in the BCD (non-symmetri-
cal) divide-by-ten mode, but the reset 0

division by 3

inputs (pins 2 and 3) are connected to
the A and B outputs, pins 1 2 and 9
respectively. The reset 9 inputs remain
grounded. Asymmetric divide-by-three
output, pin 9.

Division by 4 requires the counter to be
reset to 0 when the output reaches 4
(BCD 0100). The reset 0 inputs are thus
both connected to the C output (pin 8).

As with division by 3 and by 4, the
DCBA output of a 7490 can be
exploited to give a reset to 0000 at
every sixth count. The binary output at
this count is DCBA = 01 10; i.e. C and B

Since in BCD code 7 is 01 1 1 it is
impossible to provide a reset every 7 in-
put pulses using only the 2 reset 0
inputs, since using only 2 of the bits in
the code would lead to confusion with
3, 6 or 5. It is, however, possible to
achieve a divide-by-seven function,
though not with a BCD output se-
quence. Use is made of the reset 9 in-

On the fourth count pulse output C will
momentarily become ‘1’, but the
counter will then immediately reset to
zero. Asymmetric divide-by-four out-
put, pin 9.

division by 4

9

puts, and the reset 0 inputs are
grounded. The C and B outputs are con-
nected to pins 6 and 7 respectively. The
counter will now count 9, 0, 1, 2, 3, 4,

5, 9, 0 and so on. An asymmetric

divide-by-seven output may be obtained
from either output C or output D.

division by 7

are both ‘1’ for the first time.

Outputs B and C are thus connected to
pins 2 and 3 respectively. Asymmetric
divide-by-six output, pin 8.

division by 6

This simply entails connecting output D
(pin 1 1 ) to the reset 0 inputs. Asym-
metric divide-by-eight output from out-
put C.

division by 8

elektor july/august 1975 — 741

division by 9

Since the BCD code for 9 is 1001 out-
puts A and D must be connected to
pins 2 and 3 respectively. Asymmetric
divide-by-nine output from output D.

The 7493 may be use for division ratios
up to 16 since it contains a divide-by-
two and a divide-by-eight counter. The
pin connections are identical to those of
the 7490 except that no reset 9 inputs
are provided. The following additional
division ratios can be obtained with no
external gating.

Division by 12

Outputs C and D connected to the reset
inputs, pins 2 and 3 respectively. Asym-
metric divide-by-twelve output, pin 11.

Division by 16

Counter connected as for 7490 in
divide-by-10 mode.

use of 7493
instead of 7490

9

na .[~F~

-

°?N , C D © SJ

in 7493

H ^

This circuit generates a ‘six pips’ signal
indicating the exact hour.

At the instant the first ‘pip’ is to start,
an eight-input NAND gate, which
receives BCD-coded minute and second
marks from a TTL clock, transmits a
negative-going edge to MFI (at A). Each
digit requires not more than two inputs,
the logic ‘ 1 ’ level only being used for
decoding. The code shown in the dia-
gram corresponds to the moment
59' 55".

The period of MF 1 monostable is
about 6 seconds. During this period
monostable MF 2 is triggered by a 1 Hz
signal at (C), as shown in the pulse
diagram, its B input being logic ‘1’ (B).
The MF 2 monostable period has been
set to about 100 millisec, which permits
MF 2 to be triggered every second. The
result is a signal at (D).

Until the turn of the hour, the minute-
unit mark remains logic ‘1’ (E). N1
inverts this level, causing the signal at
(F) also to be ‘1’.

At the instant 00' 00", N2 by-passes
MF 2. This means that MF 2 only
determines the length of the first five
‘pips’. The length of the sixth ‘pip’ is
determined by the remainder of the
MF 1 period (G). The new hour is, thus,
marked by a ‘pip’ of a length of about
1 second.

The tone itself appears at the input of
N4. If the clock is crystal controlled it
may be possible to derive a frequency of

around 1 kHz from the divider circuitry
in the clock. Otherwise a simple 1 kHz
multivibrator may be used.

742 — elektor july/august 1975

The integrated circuit LX 5700 is a
temperature sensor with a working
range extending from —55° C to
+125° C. The package includes the
sensor, an op-amp and an active
reference potential source (figure 1).
Two transistors operating under similar
conditions but at different collector
currents will have different base-emitter
potentials. The sensor works according
to this principle, the transistors being
matched in such a way that the output
voltage (measured between points A and
B in figure 1) rises linearly by 10 mV
per degree Kelvin. The inverting input
of the internal op-amp is brought out
to pin 2, so that a wide choice of output
characteristics and working ranges can
be preset.

If the op-amp is used as a comparator,
its output condition changes when the
sensor output voltage Vs (which is
proportional to sensed temperature)
exceeds the extemally-preset threshold
voltage. The LX 5700 thus acts as a
temperature monitor, which can be used
to switch warning lights or equipment
on and off.

In the circuit shown in figure 2, the
working temperature range can be set
by the voltage divider R 1 , P 1 , R2. With
the values given, a switch-over point
between + 1 5° C and +60° C can be set
with PI . Other working ranges are
possible, if new values are calculated
for the voltage divider. It should be
noted that the voltages are in relation to
pin 3 of the LX 5700, not to supply
common!

The ceramic capacitor C2 serves to
suppress interference voltages. R3 , R4
and D6 determine the hysteresis, which
is 1° C with the values given. Reducing

temperature

monitor

the value of R4 increases the hysteresis.
The external Darlington pair T1 and T2
drives a relay with which, for example,
switching operations can be initiated or
an alarm set off. An LED, in parallel
with the relay coil, gives an optical indi-
cation. The relay and the LED are fed
with pulsed DC; Cl only smooths the

supply to the LX 5700. R5 determines
the dissipation in the internal voltage
stabilizer; it should be chosen to give
a current consumption of about 3 mA.
Among the possible applications of the
temperature monitor are: monitoring
room temperature or the water
temperature in a bath, remote control
of heating installations, or giving an
alarm if the temperature rises above
or drops below a preset level. Electronic
thermometers can also be constructed
with the LX 5700 IC.

(National apptf

The pin connections of the 74121
(monostable multivibrator) are the
same for ‘dual in line’ and ‘flat package’.
A1 , A2, and B are control points. In-
puts A1 and A2 are intended for rapid
triggering edges (1 V/ps). Input B can
be used for relatively slow variations in
level ( 1 V/s). The conditions under
which the monostable produces a pulse
are given in the tables.

The arrow indicates what variation is
required in the given situation to make
the monostable function.

Together with the internal capacitance
and resistance, resistor R1 and
capacitor Cl govern the pulse width.
The graph shows the output pulse width
as a function of the value of external
capacitor Cl , with a fixed value for R1 .
The straight line ‘R1 =0’ on the graph
is related to the configuration in which

the 74121

B IAi old I A 2 old | Aj new A 2 new

T~t I o I

11 i -*■ i o

1 I 1 I It-*’ 0 I 0

pin 9 of the IC is connected direct to
the supply. R1 may have a maximum
value of 40 k.

elektor july/august 1975

743

The IC TDA 1 190 (SGS) comprises all
elements required for a complete TV-
sound channel. Consequently, this 1C
and a minimum of external components
are sufficient to build a circuit that
amplifies and detects the intercarrier
sound of the TV and drives a loud-
speaker directly.

The TDA 1 1 90 consists of the following
parts:

• IF limiter/amplifier

• active low-pass filter

• FM demodulator

• * AF volume control

• AF preamplifier

• AF output amplifier.

Since the specifications of the IC are
practically constant for the range
between 4.5 and 6 MHz, it can be used
in all types of TV set.

The diagram shows a circuit designed
around a TDA 1 190 that, in spite of its
simplicity, is the complete audio section
of a TV-receiver. The circuit can be
driven by a supply voltage of 1 2 V or
24 V ; the respective LF output powers
are then 1 .5 W (into 8 ft) and 4.2 W

(into 1 6 ft). Owing to the high input
impedance of the IC, the required selec-
tivity can be obtained by means of a
ceramic filter (for instance the Murata
type SFE 6.0 MA). The value of R x
determines the gain of the AF amplifier.
This offers the possibility of setting the
gain in accordance with the maximum
frequency deviation. Capacitor C 1 0
determines the upper audio cut-off
frequency. C8 and an internal resistor
together form the de-emphasis network.
The input sensitivity is about 30 /iV.

The AM suppression is 55 dB and the

IC tv-sound
channel

signal-to-noise ratio about 70 dB. The
input impedance is about 30 kohm.

At a supply voltage of 24 V, a loud-
speaker impedance of 16 ft and a value
of 18 ft for R x the frequency range is
from 50 to 1 2 000 Hz (within 3 dB); at
R x = 10 ft the range is from 50 to
9000 Hz.

At a supply voltage of 24 V and an
output power of 50 mW distortion is
0.55%; at full power this percentage
increases to about 10%.

The IC is provided with two cooling fins
which can be mounted on a suitable
heat sink. One of the fins must be
connected to supply common.

It is possible to imitate an inductor,
using a capacitor and a gyrator. This
arrangement uses a Miller integrator and
a passive differentiator instead.

The component values must be so
chosen that R1C1 = R2C2 = t. In this
case the impedance developed between
the points A and B will be jwPlr. This is

variable

inductance

an inductance of value L = PI • t!

The value of R1 (= R2) should be taken
somewhere between 5k6 and 22k. In
some applications the voltage follower
(IC2) can be omitted.

If R2 is made adjustable it will be
possible to achieve maximum quality
factor by precisely balancing the circuit.
When the dashed components are added
the circuit becomes a simple hum-rejec-
tion filter. Careful adjustment of PI and
R2 can enable 50 dB of attenuation to
be obtained at 50 Hz, with 3 dB attenu-
ation at 40 Hz and 62 Hz.

744 - elektor july/august 1975

The integrated circuit LM 3909 used in
this device is a relatively inexpensive
monolithic oscillator in an 8-pin DIL
package. The original concept is in-
tended for low current LED flashers,
powered by a 1.5 volt source. The IC
internal circuitry (figure 1 ) shows a cur-
rent limiting resistor for the LED
(100 SI to pin 8) and two re-
sistors R2 + R3 in series, which can be
used for the delay function. The whole
flashing circuitry of figure 2 needs only
two components plus a single 1.5 volt
cell. Capacitor C sets the flashing fre-
quency, which is about 1 Hz in this case.
The current consumption is so low that
one battery should last a year. The
LED flasher is a very handy device for
locating, in the dark, appliances such as
fire extinguishers, key lockers and the
like. It can also be used as a warning
light for model railways.

As a general rule, the flashing frequency
depends to a large degree on the accu-
rate capacitance and quality of the
electrolytic capacitor. It will, therefore,
take some experimenting to arrive at a

LED flasher

specific frequency. It has been found
that capacitors with a high leakage cur-
rent impair the operation. Best results

are obtained with tantalum electrolytics, I
which have the added advantage that I
they are very small.

(National Semiconductor) I

Many electronic devices have no pro-
visions for the 1.5 V power supply for
the LED flasher. Two resistors can, in
that case, be added enabling the circuit
to operate at higher voltages. The resist-
ance values depend on the voltage avail-
able.

The following table shows the values of
the components required to operate the
circuit at different voltages. The last
column indicates the range of voltages
over which the circuit will work satisfac-
torily.

Table

Sup-

Flash-

C

Rv

Rf

Oper-

ply

ing

fre-

range

age

(V)

quency

(Hz)

<mf>

(ft)

(£2)

(V)

+ 6

1.6

470

1 k

1 k5

+ 5...
25

+ 15

1.6

220

3k9

1 k

+13...

50

+100

1.3

220

43 k /
1 W

Ik

+85 ...
200

LED flasher for higher voltages

circuit is a 1 kHz square-wave oscil-

lator.

PI is used for making the square-wave
symmetrical P2 sets the amplitude of
the output signal; the maximum is
about 1 . 1 volts (peak-to-peak).

1 kilohertz
oscillator

This miniature device consists of only
four components. It can be used for
continuity tests in wiring harnesses and
on p.c. boards, using suitable test
prods connected to points A and B.
After some experience it is possible
to estimate contact resistance by
interpreting differences in pitch and
sound level.

A further application is, to use the
mini-siren as a morse code practice
set by connecting a morse key between
A and B.

mini-siren conductance tester

pin 4 is connected to the spring contact
of the case.

The current for the IC flows through
the torch bulb.

attachment
for 'find me'
flashlights

Many amplifier designs, both do-it-
yourself efforts and commercial cir-
cuits, have the bad habit of
delivering a more or less hefty switch-on
‘thump’. This is usually due to too-rapid
charging of the output-electrolytic. If
this fault occurs in a high-power design

slow-rise
power supply

there is every chance of loudspeakers
departing for Valhalla.

An alternative to redesigning the offend-
ing amplifier is given by this power-
supply circuit. It is essentially a simple
regulator with a slow-start character-
istic.

The raw supply is delivered by recti-
fier B and reservoir capacitor C v . Zener-
diode D 1 provides the reference voltage,
the output voltage being about 600 mv’
lower than this. If necessary the desired
voltage can be made up using two zeners
in series. The total zener voltage may be
chosen between 28 V and 63 V
(roughly). Switch SI turns the supply
on and off (coupled to the mains
switch). When it is closed the voltage
on Cl rises in about 1 second to its

operating value. The output voltage
follows it up until the point is reached
at which the zener becomes conductive.
When SI is opened the voltage on Cl
falls away in about 5 seconds, due to
base current supplied to T1 . If the
amplifier shows no serious switch-off
transient, so that well-defined turn-off is
not required, it is permissible to omit
SI.

The unregulated voltage at Cl should
not exceed 80 V. It should be chosen so
that there is just enough voltage drop
across T3 to take care of regulation
requirements. Too high a drop is a waste
of power and - much more important -
a waste of expensive heat sink. The idea
is that with the supply fully loaded and
the incoming mains voltage at its lowest

(expected) value, there should be about 'j
2 volts across the series transistors at the
troughs in the ripple waveform.

A reasonable rule of thumb, alterna-
tively, is to allow about 10 volts across
T3 (no load) - and hope! T3 will in any
case need a modest heat sink (e.g. 2 mm '

thick bright aluminium, about 10 cm by 1

10 cm). In extreme cases it may also be
necessary to provide T2 with a cooling
vane or clip.

The value of 1000 nF given for C v is
only intended as an illustration. If one
is going to also build the raw supply,
matched to a known maximum load,
there is not much else for it but a little
mathematical exercise with Q = CV
(remembering that the rectifier delivers
one hundred wallops-per-second).

746 — elektor july/august 1975

W. Lewald

The direction is indicated by means of
four lamps that flash on in sequence :
one-two-three-four-out. T] and T2 form
a multivibrator whose period determines
the rate at which the lamps 'flash for-
ward’.

At switch-on all the thyristors are non-
conductive. The first positive-going
flank at the collector of T1 will trigger
thyristor Thy 1 , causing lamp LI to light
up. The same trigger-pulse is also
applied to Thy3 - but this will
cause no anode current to flow, because
Thy2 is still non-conductive. The next
change of state of the multivibrator
produces a positive-going flank at the
collector of T2, and this will trigger
Thy2 and light L2 . L3 and L4 come on
during the second cycle of the multi-
vibrator. Now all the lamps are burning,
having come on one-by-one. There is

direction pointer

now a positive voltage at the collector
of T3, so that the next pulse will cause
the relay to attract. This will interrupt

the feed to the thyristors, quenching all
the lamps, so that the sequence can start
again.

One must take care that the thyristors
are adequately rated for the lamp load’;
Bear in mind that Thy 1 has to carry the
current for all four lamps; Thy2 the
current for three, etc.

In principle an OR-gate can consist of
cathode-connected diodes with one re-
sistor to the supply neutral. A drawback
of such a circuit is that its properties are
not the same as those of TTL or
COS/MOS. If such is required, the cir-
cuit of figure 1 can be applied with
TTI^circuits; the one of figure 2 with
COS/MOS. The number of inputs of the

circuits of figure 1 or 2 can, of course,
be extended by adding a diode for each
additional input.

J.W. Gnodde

The designer of a heavy-duty, variable-
voltage, regulated power supply unit
always meets the problem of having to
allow for the very high dissipation that
can occur in the series regulator transis-
tor. When a high output current is
drawn at low voltage (TTL!) it may
even be necessary to use a cooling fan at
the heatsink.

Perhaps an extreme example is the case
of a supply unit capable of delivering
5 amps at between 5 and 50 volts. Such
a unit will have typically a 60 volt un-
regulated supply. Suppose this unit is to
supply TTL circuits at its full rated
current. The series element in the circuit
must now dissipate 275 watts! The cost
of providing adequate cooling is likely
to be exceeded only by the cost of the
series transistor needed.

If the voltage drop across the regulator
transistor could be limited to 5 .5 volts,
independently of the selected output
voltage, the dissipation would be drasti-
cally reduced - in the above example to

heatsink

shrinker

1 0% of its original value.

This can be achieved by using three
semiconductor devices and two resistors
(figure 1 ). y ]

This is what happens: thyristor Thy is
normally kept conductive by means of
R1 . However, when the voltage drop
across T2 - the series regulator -
exceeds 5.5 volts, T1 will start to con-
duct, causing the thyristor to ‘open’ at
the next zero-crossing of the rectifier
output. This arrangement continuously
controls the charge supplied to Cl - the
reservoir capacitor - so that the unregu-
lated supply is held at 5 .5 volts above
the regulated output voltage.

The value required for R 1 is found as
follows:

r, _ 1 4 x v sec ~ ( v min +

where V sec is the RMS secondary volt-
age of the transformer and V m ; n the
minimum value of the regulated output.
The thyristor must be capable of carry-
ing the peak ripple current, and its
working voltage must be at least

1.5 Vjec- The series regulator transistor
must be rated to carry the maximum
output current, I ma x , and be provided
with a heatsink on which it can dissipate

5.5 x I sec watts.

... ./ fuse destroyer

elektor july/august 1975 - 747

The CA 3090 Q, an integrated PLL
stereo decoder, has been quite popular
for the last four or five years. Not long
ago RCA introduced a new, improved
version: the CA 3090 AQ. The most
important improvement is a reduction
in harmonic distortion from 0.35%
down to 0.2%.

A complete circuit with the new IC is
shown in figure 1 : figures 2a and 2b
show the two basic ways of connecting
pins 3 and 4. If these pins are connected
to supply common (figure 2a), the cir-
cuit performs as a normal stereo
decoder; however, one or two minor
additions (figure 2b) give a stereo-defeat
option : if the voltage on pin 4 is less
than 0.9V the decoder will not operate,
Sa

CA3090 AQ

so that the output remains mono.

A voltage greater than 1.6 V on pin 4
switches the decoder back to normal
operation. This option can be used for
an automatic stereo-defeat, driven by a
carrier-level dependent reference
voltage, or simply as a manual switch as
shown in figure 2b.

The p.c. board (figure 3) allows for both
possible connections to pins 3 and 4.
Mounting the (dotted) wire links gives

the circuit of figure 2a; if R3 is
mounted, the stereo/mono control
voltage can be connected to V(\

The trimming procedure is as follows:

1. Connect the decoder to the audio
output of an FM receiver. NB. The
original de-emphasis network in the
receiver should be removed, of course!

2. Tune in to a stereo broadcast - if
necessary, this can be checked with the
aid of a second receiver.

3. Screw the ferrite core of L2 out as
far as possible.

4. Now screw the core down until the
LED (D1 ) lights up; screw the core
down still further, counting the number
of turns, until the LED goes out again;
finally, set the core midway between
these two extreme positions.

Of course, if one has access to a good
oscillator and frequency counter, the
coil can simply be tuned to 76 kHz.

748 — elektor july/august 1975

■

Existing tremolo controls using an
incandescent lamp and a light-depen-
dent resistor have a response too slow to
follow a fast and deep tremolo. New
possibilities are opened by the avail-
ability of light-emitting diodes and rela-

LED controlled
tremolo

tively fast responding light-dependent
resistors (mostly of Japanese origin).
Obviously, a LED permits luminance
variations much faster than a filament
lamp.

The circuit diagram shows an RC oscil-
lator T1 with control P2 setting the
tremolo frequency between 2 and 9 Hz.

PI controls the tremolo depth. The
tremolo signal is admitted to the buffer
stage T2 via S 1 , which switches the
tremolo on/off, and C5. T2 emitter
passes the tremolo signal plus a standing
current via R8 on to the LED.

The LED is optically coupled to the
LDR. The optical path is screened from

external light.

The tone signal entering at A is modu-
lated by the tremolo controlled LDR
when it arrives at the output end B. P3
pre-sets the range of the tremolo depth.
Transistors of an equivalent type can __
be used forTl and T2 (e.g. BC 107B,
108B, 109B).

c

b — *

R-

2 improved
.C 7-segment
S for

3 MOS-docks

In a previous issue (Elektor 2) the
principle of improved readability for
7-segment displays was explained. The
idea is to add an extra stroke at the top
of a six and at the bottom of a nine
(figures 1 , 2 and 3).

The circuits given there were based on
the TTL decoder driver SN 7447, and
derived the necessary ‘logic’ information
from the BCD inputs. In many mono-
lithic integrated circuits, such as the
MOS-clock IC MM5314, only segment-
drive outputs are accessible however —
and moreover these outputs deliver a
sequentially multiplexed signal for all
four (or six) digits. In this case, a
different approach to ‘improved
readability’ is required.

Programmable frequency dividers are
particularly useful for producing a large
number of frequencies from one
reference frequency. An example of the
type of circuit in which such dividers
can be used is a frequency synthesiser.
Programmable dividers are generally
rather expensive. The circuit given here
offers a less pricey solution. In this
instance, two 7490s connected in cas-
cade are made programmable by provid-
ing a switchable reset point. This is
achieved by connecting the BCD out-
puts of the two dividers to the reset
inputs through diodes, and presetting
the required division ratio by closing
one or more switches. When, and only
when all the outputs connected to
closed switches go ‘1’ simultaneously,
the diode anodes all become free to rise

The circuit in figure 4 can be used with
common-anode displays, as in the MOS-
clock. The points A, D, E and G are
connected to the collectors of the

in potential and the reset operates.

For example; the circuit divides by 75
when the switches for IC 1 are set to
DCBA (5=0101) and the switches for
IC2 are set to T)CBA (7 = 0111). Each
bar indicates an open switch; the other

corresponding segment-drive transistors
(T1 , T4, T5 and T7 in the MOS-clock) %
For common-cathode displays the
circuit of figure 5 can be used.

switches should be closed. When the
divider count reaches the number which
has been preset, both Roi inputs go to
logic ‘1 and the dividers are reset. Both
dividers then start again from the
beginning.

llll-

IC2= 74 90

b 90) GN0 R 0I»

8* DUG

m

Mi

IC1 =74 90

m««i GN0 "■

0

r=

— ©

programmable frequency divider

An automatic exposure timer makes it
quite simple to obtain correctly-
exposed prints from negatives, and the
usual experimenting with test strips can
be almost completely done away with.
Operation is as follows: with the
switches in the positions as drawn, the
lamp in the enlarger is off. When
switch SI is operated, this lamp is
turned on continuously for focussing
adjustment. After a red filter has been
inserted the unexposed printing paper
can be set in the frame under the en-
larger. The lamp is then turned off with
SI, and the red filter removed. Finally,

. when S2 is changed over, the paper is
automatically exposed correctly. The
automatic exposure control circuit is
quite simple, in the neutral position (S2
in the position as drawn) Cl is dis-
charged. At the moment when S2 is
changed over, the voltage on the non-
inverting input of IC1 (pin 3) is there-
fore initially +1 8 V. The voltage on the
inverting input (pin 2) is lower, so that
the output voltage of the IC (pin 6)
is about +18 V. Transistor T1 is driven
into saturation via resistor R7, so that
the relay is energised and the lamp
switches on. The LDR measures the
quantity of light falling on
the printing paper. The lighter the
negative, the lower the LDR resistance,
and therefore the faster the rate at
which Cl is charged. As Cl charges, the
voltage on the non-inverting input of
the. IC drops. As soon as this voltage
becomes less than the voltage on the
inverting input (adjustable by PI) the
output voltage of the IC becomes
zero. As a result T1 is cut off, so that
the relay cuts out and the lamp is
extinguished.

There are some important practical

elektor juty/august 1975 - 749

points to be kept in mind in the con-
struction. The LDR must measure the
average quantity of light incident on the
printing paper, but no extraneous light.
It is therefore best to use the arrange-
ment shown in figure 2, in which a light-
tight tube and a single lens are fitted in
front of the LDR. The position of the
LDR relative to the printing paper
should not be altered after calibration.
This is best done with a negative having

a uniform distribution of mid-grey
tones. The correct exposure time for
this negative is determined as usual with
test strips, after which PI is adjusted so
that the timer gives the same exposure
time for this negative. If every grade of
the printing paper used has the same
speed, this setting need only be carried
out once. If, however, the speeds differ,
PI must be provided with a calibration ’
scale.

The SGS integrated circuit TCA 940E is
a monolithic audio amplifier in a plastic
12-pin quad-in-line encapsulation. When
power requirements are moderate this
IC can be used to construct an excellent
little amplifier.

The integrated circuit has very low non-

linearity and crossover-distortion; it is
also effectively protected against output
short-circuiting.

The application circuit given here will
deliver 6.5 W into an 8 load, with a
20 V supply; at 1 8 V the output be-
comes 5.4 W.

6.5 Watt
IC-amplifier

The amplitude-frequency response is flat
(within 3 dB) from 40 Hz to 20 kHz,
while the harmonic distortion is only
0.2% for output power levels between
50 mW and 3.5 W, The input has an im-
pedance of 100 k (the volume control
potentiometer) and can be fully driven
with 110 mV.

The IC can be cooled by mounting the
two protruding ‘feet’ on a suitable heat-
sink, or by soldering them to pads on
the PC board. If the PC board has a
copper cladding of about 35 nm thick-
ness, each pad should have an area of
about 1 0 cm 2 . One of the ‘feet’ has in
any case to be connected to the ‘earth’
rail of the amplifier wiring.

750 — elektor july/august 1975

The use of a unijunction transistor in a
sawtooth generator is well known. By
way of variation, this circuit uses a diac.
One specific property of the diac is its
high breakdown voltage. While a normal
diode will conduct at a forward voltage
as low as about 0.5 V, a diac's threshold
can be as high as 30 V. A diac shunted
across a steadily charging capacitor will
abruptly discharge it every time the
breakdown voltage of the device is
reached.

If the charging of the capacitor is done
by means of a current-source, the volt-
age wave across it will be a sawtooth
with an amplitude equal to the break-

diac sawtooth-
generator

down voltage. A resistor in series with
the diac protects it against excessive
peak current.

Any load connected directly in parallel
with the capacitor will spoil the lin-
earity of the sawtooth. The voltage-
follower combination of JFET and bi-
polar transistor solves this problem by
providing an extremely high input im-
pedance and a very low output
impedance.

The specified E 300 FET’s can actually
be replaced by almost any other type;
the circuit merely uses fundamental
FET properties.

A good sawtooth contains all even and
uneven harmonics of the fundamental
frequency - so that this generator can
produce interference far into the
VHF range. It must be screened!

Ch. Voss

This volume-control circuit offers an
unusual approach to the well-known
problem of distortion in active-device
attenuators. The zero-output in this case
is obtained by allowing equal signals of
opposite phase to cancel each other.

The input transistor operates as a
‘concertina’ phase-splitter, producing
equal-amplitude opposite-phase voltages
at its collector and emitter. The two
signals can be brought into complete
cancellation, at the summing point, by
means of preset PI . The harmonic
content of the two signals is very small
- but not quite identical. There will
therefore actually be an - even smaller -
distortion output at the nominally
‘zero’ point.

If something now happens to the ampli-
tude ratio of the signals at the summing
point, there will be output passed to the
buffer-stage. The necessary unbalance is
achieved by means of the JFET and
capacitor C2. The gate bias on the FET
is set by the DC control voltage applied
to point A. With this voltage close to

zero the FET will be cut off, so that the
above-mentioned cancellation takes
place. As this voltage is increased there
will come a point at which the channel
starts to ‘bleed ofF AC collector current
from the splitter; this will upset the
balance and so cause an output signal to
appear. The more conductive the FET,
the more output. Unfortunately, the
more channel current there flows the
lower will be the negative gate bias - and
so the greater will be the distortion of
the ‘regulated’ summing component.
The trick is now te employ only a

moderate degree of unbalance - so that
the FET operates at low distortion per-
centages. The process is helped also by
the always-present ‘clean’ summing
component. The buffer stage provides
gain, so that a sufficient output level is
obtained.

The circuit’s frequency response
extends from 50 Hz to 35 kHz (-3 dB
points). The input voltage should be
limited to 100 mV p-p; the output can
be varied from ‘0’ to 1 V p-p (by using
the appropriate range for the control
voltage at A).

R. Meschenmoser

The circuit consists of an ‘electronic’
attenuator, arranged to respond to a
fingertip touching one of its contact
plates.

Transistor T1 (a type E300 or similar
JFET) operates as the variable resistor in
a voltage divider (Tl, Rl). The channel
resistance of Tl depends on the negative
voltage across C 1 .

If one touches the contact-pair associ-
ated with the negative supply a current
through D2, R2 and D3 will charge the
capacitor. The charging time is deter-
mined by the values of R2 and Cl .

When the FET gate is biassed suf-
ficiently negative the device will no

longer conduct, so that the audio signal *
passes unattenuated.

To reduce the audio volume one simply '
touches the ‘positive’ contact-pair. The
discharge of Cl causes the FET-bias to
be reduced, so that the channel becomes
conductive again and diverts signal
current to ‘earth’. The level will con-
tinue to slowly change as long as one of
the contact-pairs is being touched.

The FET channel is only approximately
a linear resistor, due to the audio signal
modulating the gate bias. At an input
level not exceeding 30 mV there will
nonetheless be no audible distortion.

fingertip
volume control

elektor july/august 19 75 - 751

This instrament has 14 capacitance
measuring ranges, from 5 pF f.s.d. to
1 5 ixF f.s.d., and a linear scale. SI is the
main range switch, and works in combi-
nation with S4 (xl/xlO) and S3 (xl)or
S2 (x3).

The 7413 works as an AMV, with R1
and Cl ... C6 as frequency-determining
components. This in turn triggers the
74121 (a monostable multivibrator) so
that it produces an asymmetric square-
wave, the repetition frequency of which
is determined by R1 and Cl ... C6 and
with a pulse width determined by R2
(or R3) and C x . The average value of

WWW a

this square-wave voltage is a linear func-
tion of the duty cycle, which in turn is
linearly dependent on the value of C x ,
the value of R2/R3 (x 1 0/x 1 ) and the ’
frequency (determined by the position
of SI). The final range selector switches
S3i_x 1 ) and S2 (x3) simply add a re-
sistor in series with the meter.

The wiring to pins 10 and 1 1 of the
74121 and to C x should be as short and
rigid as possible, so that the parasitic
capacitance at this point is small and
constant. P5 and P4 are used for separ-
ate zero calibration in the low capaci-
tance ranges; for all higher ranges one
calibration (with P3) suffices.

F.s.d. calibration is fairly simple. Do not

By means of the IC TAA 131 and a few
external components a simple square-
wave generator can be built which is
eminently suitable for testing amplifiers,
for instance.

The circuit operates well on a voltage of
1.5V, but it is still reliable with a supply
voltage of about 1 V. Consequently, it
is possible to build the entire circuit
plus the battery (mercury cell!) into a
small case.

capacitance

meter

solder C6 in circuit but connect it across
the C x terminals. Set SI in position 3,
S4 in position xl and S2 closed (x3);
this corresponds to 1 500 pF f.s.d., so
that C6 can be used as a calibration
standard, and PI is adjusted so that the
meter reads 2/3 of f.s.d. After this, S4
can be set in position ‘xlO’, S2 is
opened and S3 is closed (x 1 ); this

15V

corresponds to 5000 pF f.s.d., so using
C6 as C x will give 1/5 of f.s.d. Alterna-
tively a range of standard capacitors
may be incorporated into the instru-
ment for calibration purposes.

Editorial comment; although the author
specifies 100/r/l%, 10/r/l% and 1 /r/1%
electrolytics for Cl , C2 and C3, we fear
that these will not be readily obtain-
able.... ForC2 and C3, 5% poly-
carbonate types can be used instead; if
the highest range is also required C 1 can
be individually calibrated (with extra
padding capacitors) using C3 as cali-
bration standard. Of course, the
temperature stability of an electrolytic
will always remain questionable.

The frequency generated is about
1000 Hz but, because the output signal
is a good square-wave, harmonics as high
as 1 MHz will be present.

Choosing a larger value for Cl or R1
gives a lower frequency, and vice versa.

mini

signal squirt

The amplitude of the square wave at
the output is practically equal to the
supply voltage. The symmetry is some-
what dependent on the supply voltage.

752 - elektor july /august 1975

This circuit adjusts the brightness of an
incandescent lamp to suit ambient
lighting conditions. This is useful for
instrument panel lights, bedroom clock
lighting and similar applications.

The circuit is designed for 6-24 V lamps;
the total current must not exceed
1 amp.

The adapter functions as follows.

LDR 1 senses the ambient light. LDR 2
is optically coupled with an incan-
descent lamp. The circuit balances when
both LDR 1 and LDR 2 receive the
same amount of light. The circuit must,
however, cause the external lamp(s) to
be brighter than the ambient light. For
this reason LI should have a lower
current rating than L2, L3 etc; or,
failing this, a small screen (small sheet
of paper) can be placed between the
lamp (LI) and the LDR inside the opto-
coupler.

The 0.68 ohm resistor limits the lamp
current; the 1 nF capacitor prevents
oscillation of the circuit. The circuit
must be powered by at least 8.5 volts;

contrast adapter

lower voltages affect the operation of
the IC. It is recommended to use a
supply that exceeds the lamp voltage by j
about 3 volts.

The zener (Zl) is chosen to match the
lamp voltage; for 6 V lamps the internal J
zener in the IC can be used by
grounding terminal 9.

i@i

The circuit diagram shows an active-low-
pass filter design which can be given any
desired cut-off point - over a wide
range — simply by calculating two
values for four capacitors. The filter
consists of an RC-network and an
NPN/PNP transistor-pair. The transistor
types shown can be directly replaced by
many other types without spoiling the
circuit’s performance. The supply
voltage used should be between 6 and
12 V.

active

low-pass filter

The capacitor values chosen for
Cl ... C4 determine the cut-off fre-
quency. These values can be obtained
from the following two formulae:

in which ‘fc’ is the desired cut-off fre-
quency (in Hertz), where the amplitude-
response is down 3 dB, and the values
for Cl ... C4 are obtained in micro-
farads. (Inserting kilohertz will deliver
nanofarad values and inserting mega-
hertz will produce picofarad values.)

As an illustration the measured result is
given for a filter made up with
Cl = C2 = C3 = 5n6 and C4 = 3n3. The
‘—3 dB point’ in this case occurs at
1350 Hz. One octave higher, at
2700 Hz, the attenuation is already
19 dB.

754 — elektor july /august 1975

elektor july/augutt 1975 — 755

OPAMPS. COMPARATORS

VOLTAGE REGULATORS

766 - elektor july/august 1975

TUP

TUP

Wherever possible in Elektor circuits,
tors and diodes are simply marked 'TUP*
(Transistor, Universal PNP), 'TUN' (Transistor,
Universal NPN), 'DUG' (Diode, Universal Ger-
manium) or 'DUS' (Diode, Universal Silicon).
This indicates that a large group of similar
devices can be used, provided they meet the
minimum specifications listed above.

Motorola LED sentries never die.