INDIA

Telecommunication
By Optical-Fibre

DIGITAL FREQUENCY COUNTER VDC 18 Smallest SlZe

ever made in INDIA

/ j/7 |u lu {u\ulij Battery cum main* operated wide frequency range 30M

\ (our model VDC 19 la Baaic eenaltlvity of lOmv

V. Bright LED dlaplay 7 digit

MULTI FUNCTION OSCILLATOR

r\JJ~LA/

VASAVI’s VLCR7 SAVES YOU
FROM AGOIMY OF BRIDGES.

Measurement of INDUCTANCE. CAPACITANCE.
RESISTANCE are greatly simplified by
VLCR 7. No balancing, no adjustments.

VLCR 7 gives you directly the digital reading of
value and its loss factor simultaneously.

VLCR 7 is the only instrument in India covering the

widest ranges of 0.1 pf/uH/m ohm

tie. 0.0001 ohm) to 20,000 uf/200H/20 M ohm.

OSCILLOSCOPES

- m@0000

Component tester in our oscilloscopes is
altogether different from other’s. In addition to
all Oscilloscope funtions component tester
provision can test all components passive and
active, in circuit and out of circuit. ONLY OUR
SCOPESCAN recognise NPN/PNP. Distinguish
HIGH/ LOW Gain, Distinguish AUDIO, R.F.,
SWITCHING transistors. J

UiJ&AVI ELECTRONICS

(Marketing division)

630, Alkarirr) Trade Centre. Ranigunj

SECUNDERABAD-500 003.

ph: 70995
gms: VELSCOPE

J

□ Fixing large components on
Printed Circuit Boards.

1 ) To prevent breakage of
soldered ends due to
loads on the components.

2) Prior to use of Wave
soldering equipment.

□ Fixing jumper wires to
avoid hanging wires which
could get pulled off.

.

□ Fixing name/number
identification tags.

□ Repairing cracked P.C.B.s

fevicol-cyn

bonds in 5-10 seconds.
It has high electrical
insulating properties.

Stick with the best

H* .

FEVICOL

India's largest selling brand
of synthetic adhesives

Parekh Marketing Pvt. Ltd.

Mafatlal House. Backbay Reclamation, Bombay-400 020.
Phone: 632 3t 48. Telex. 01 1 5349-PDI-IN. Gram: FEVIFIX.

05

13

mini stereo amplifier

1

good earth. Moreover, all other con- 2
nections should, of course, be kept
as short as possible. This is particu-
larly important in the case of the
supply lines, which should be de-
coupled by Cn as close as possible
to the IC. The negative terminal of
this capacitor should be soldered
direct onto the earth plane; the
positive terminal is soldered in the
normal manner to pin 16.

Finally, the distortion for a power out-
put of about 0.25 W is roughly 0.3 per
cent. (W) _

high dynamic range
mixer

This mini amplifier is based on the
Thomson Type TEA2025. In this
16-pin DIL device hides a stereo
amplifier that with a supply voltage of
9 V will provide 1 watt output per
channel into a 4-ohm loudspeaker. At
full output, the input sensitivity is
about 25 mVpp. If this is too sensitive,
a resistor R may be connected
between pin 6 and C? and between
pin 11 and C>. The sensitivity then
becomes (25 + VzR) mV, provided
R>lkQ, Furthermore, the supply
voltage may lie between 3 V and
12 V.

The operation of the 1C cannot be
discussed here, but for those
interested its internal circuit diagram
is reproduced in Fig. 1. One useful
feature of the TEA202S is that it has a
soft-start circuit on board, thus
obviating annoying plops in the loud-
speaker at switch-on.

Construction of the amplifier is fairly
simple, but has its peculiarities. First,
there is the earth, which in this case
should not be of wire, but rather con-
sist of a metal earth plane (if you
design your own PCB, this would be
of copper). If at all possible, pins 4
and 5 as well as pins 12 and 13,
should be connected to a (copper)
area of not less than 5 cm 2 . The two
areas should be connected in a
suitable manner, and in such a way
that a heat sink is formed under the
IC as shown in Fig. 2. This ensures
both good heat conduction and a

2

A mixer is expected to have low-
noise and high dynamic perform-
ance. Most standard mixers use in-
verting operational amplifiers.
Unfortunately, the noise figure of
many opamps is poor, and opamps
with a good noise figure are normally
not suitable for operating with large
signals.

The noise factor of standard circuits
is often made even poorer because
the source and amplifier are not
properly matched.

The characteristics of a mixer can be
greatly improved, therefore, by the
use of buffers at the input stages, and

the constructing of operational
amplifiers from good-quality tran-
sistors. This has been done in the ac-
companying circuit. The input is
buffered by T. and T?. The input im-
pedance of Ti can be ignored, so that
the source merely needs to be
matched with Pi.

The opamp is formed by transistors
T3 to Ts inch Good-quality RF tran-
sistors have been used in differential
amplifier Ti-Ti-T-. These transistors
have a better noise figure at a greater
bandwidth than AF types.

Vhe proposed circuit has a fre-
quency range (—3 dB points) of 10 Hz

to 80 kHz; third harmonic distortion
of not more than 0.05 per cent at
10 kHz and an output voltage of
9 V PP ; and a signal-to-noise ratio of
100 dB.

The signal-to-noise ratio applies to an
output signal of 9 Vpp with open-
circuit input, and a bandwidth of
20 kHz. The maximum value of the
output signal is about 12 Vpp,
measured across a load impedance
of 560 ohms. If the mixer is ter-
minated by a higher impedance, the
output voltage will be greater.

A further advantage of the circuit is
that the popular valve sound may be

realized in a simple manner. To this
end it is necessary that Ti and T2
commence limiting at a slightly lower
level, i.e: 12 V PP input, than the com-
posite opamp. The supply voltage of
Ti and T2 must then lie between
±6 V and +9 V. Since T2 is connec-
ted as a current source, the exact

supply voltage can be set with the
2k2 preset at the wanted clipping
level.

If desired, the output off set may be
zeroed by inserting a 50 kilo-ohm
preset in the base circuit of Ta . This
base should also be decoupled by a
1 mF, 63 V capacitor.

The current consumption of the
opamp is about 35 mA and that of the
buffer stages not more than 10 mA. If,
therefore, ten buffer stages are used,
the power supply should be capable
of providing 150 mA at + 15 V. (B)

servo-robot driver

Over the past few years, robotics and
cybernetics have become new fields
of interest for many an owner of a
personal micro, equipped with the
necessary add-on boards to effect
peripheral control. However easy it
may seem to write robot control pro-
grams and to build the associated
computer hardware, the construction
of accurately operating mechanical
parts (or, if you like, limbs) often
poses unsurmountable problems,
since a miniature set of gears, ball
bearings, spindles and cog-wheels
are in no way readily made items for
those skilled in programming and
soldering.

Despite their limitations as to pre-
cision of movement, servo-motors

25

§

Free running as well as crystal con-
trolled clock generators in many
digital designs are most frequently
based upon the use of one or more
inverter gates. However easy it may
seem to use these devices for the
construction of reliable oscillators,
the resultant frequency stability is
generally not such as might be ex-
pected from a look at the relevant
quartz crystal data, and this is mainly
on account of the rather poorly de-
fined capacitive and/or inductive
loading of the crystal at resonance.
Stability, however, may be improved
by a factor 3 to 5 by using cascode
type inverters in a symmetrical con-
figuration, as can be seen in the ac-
companying circuit diagram. Two
sets of two n- and p-channel MOS-
FETs, contained in the Type 4007UB
IC. have been connected to form a
highly stable oscillator circuit
capable of operation at frequencies
up to 10 MHz, as determined by
quartz crystal Xi, which should be a
series resonant type.

As the output impedance of the pro-
posed cascode oscillator is relatively
high, buffer stage T< has been added
to minimize drift with low impedance
loads such as (LS)TTL circuits. Fur-
thermore, MOSFET Ti ensures well-
defined logic high and low levels to
interface with (HC)MOS and (LS)TTL.

used in model aircraft or boat con- (20 ms); see the inset timing diagram,

struction may offer an interesting The synchronization interval is

alternative to more complicated generated with Di-Ri-Ci, which reset

mechanical constructions; appli- the Type 4017 counter when no nega-

cations such as robot arms and sort- tive pulse has been received for

ing machines can be made quite about 3 ms. The control inputs of the

easily with the use of cleverly servo-motors may be connected

mounted servo-motors. An example direct to the counter outputs

is shown in the photograph; a simple
robot which is able to walk a few
steps before falling to the ground.

Four computer-controlled servos
have been fitted on the joints.

The servo control circuit for this stag-
gering little creature is driven with a
single computer output signal. A
number of wait loops need to be pro-
grammed to supply active low (0)
pulses lasting about 0.5 ms, while the
interval length between the first and
second pulse determines the pos-
ition of servo 1, while servo 2 is pos-
itioned by means of the interval
between the second and third pulse,
and so on. The repeat rate of the con-
trol process should be about 50 Hz

symmetrical cascode
oscillator

The values of R< and Rs depend on put level swing. In case the oscillator
the supply voltage level (Ub), while is to operate from a 5 V supply, gate
the voltage at gate 2 should be be- 2 of T- must be connected direct to
tween 4 and 6 V to achieve a 5 V out- +Ub. (B)

low noise aerial booster

After having read the design essen-
tials relevant to wideband amplifiers.
RF filtering, intermodulation/cross-
modulation characteristics, etc., as
given in the articles listed at the end
of this article, there would seem to be
little need for us to dwell on func-
tional and electronical aspects of the
present ultra low-noise, wideband
preamplifier incorporating the won-
derful Type BFG65 transistor, which,
although already introduced in (3],
deserves to be put In the RF limelight
as it offers an exceptionally low noise
figure at more than satisfactory
strong signal response, thanks to the
relatively high collector current
(FdB = 0.8 dB at 5 mA, for instance).
Since the important points to ob-
serve in RF construction have been
covered in [1] and [2], the large earth
plane on the component side of the
ready-made PCB Type 86S04 need
not cause any wonder; all parts are
soldered direct onto the relevant
copper fields; the holes merely serve
to aid in locating the parts correctly.
The hole for T. should be drilled to
dia 5 mm for the transistor to be
seated and soldered with the
shortest 'possible lead length.
Additional holes have been provided
to enable the input and output coax
cables to be secured by means of

D1...D4 = 1N4001 | r-5*-] r- t — t " <v)

haaii-

0-K-t-

servo-motor tester

*_h k_

This circuit is of interest to two
categories of readers; first, model air-
craft/boat/train enthusiasts object-
ing to having to leave their radio
control transmitter switched on for
extended periods in order that a
servo-motor and associated mechan-
ical parts can be made to function as
desired; secondly, constructors of
computer-controlled robots incor-
porating servo-motors. The latter
field of interest is a typical combi-
nation of mechanics and electronics
plus software, and it is sometimes
urgently required to be able to keep
them apart as the specific parts of the
robot have been prepared for testing,
which, arguably, should be possible
to do without having to write special
programs to this end on the
computer.

The proposed servo-motor test unit
is, as can be seen from the circuit
diagram, a downright simple design
based on the use of Type 555 or 7555
timer chips connected in a cascade
arrangement which can be ex-
panded further to drive more than
two servos at a time, if desired; it is
readily seen that the second and
third stages of the circuit are
identical.

The first timer, ICi, serves as an
astable multivibrator whose output
pulse period time is determined with
T=0. 693(R- + R2)C. The indicated
values for the relevant timing parts
therefore provide a pulse period
time of about 20 ms at pin 3 of ICi.
The rising edge of this square wave

triggers monostable IC2, whose out-
put pulse width may be set with P ;
the given series connection of Pi
and R« ensures a large enough pulse
width range for most types of servo-
motor, which typically require activa-
tion pulse widths of the order of
1. . .2 ms.

The second servo control stage is
identical to the set-up around IC2,
and up to eight triggered 555 stages,
each with its own pulse width con-
trol, may be cascaded in this way.

It is suggested to fit each of the servo
control potentiometers with a simple
scale in order to have a relative indi-
cation regarding the motor's lateral
or angular position.

The ammeter in series with the

positive supply line offers an indi-
cation about the total current con-
sumption of the servos, and it is thus
readily detected when one or more
have got stuck in their movement.
The test circuit itself does not con-
tribute much to the total current con-
sumption indicated on the meter;
some 3 mA is required for each Type
555 timer in the row, while the use of
low power Type 7555 equivalents
should reduce this figure even
further. It is, therefore, perfectly
feasible to make the tester into a port-
able. battery-operated unit, powered
by four penlight type NiCd batteries.

(W)

We all sometimes wish that some of
the switches around the home were
just a little easier to locate and
operate, notably so in the dark and
with less frequently used light
switches, such as those for the cellar
or garage light. For the physically
handicapped, some switch locations
present a real hindrance to their
mobility in the home; for them, it
would be very convenient to be able
to operate the switch from a distance.
The proposed wireless control
system differs from, say, an IR-based
set-up in that it requires no line-of-
sight path between transmitter and
relevant receiver, while the prac-
ticable operating range is of the
order of a few metres.

The circuit diagram of the control
transmitter shows an oscillator com-
posed of Ti, Ts and T2 , the latter tran-
sistor merely functioning as a
switching device. The oscillator fre-
quency is set at about 30 kHz by
means of Cs , Cs and Li: the latter con-
sists of about 200 turns of 36 SWG
(0 0.2 mm) enamelled copper wire
on a paxolin former to suit the
diameter of 10 to 20 cm long ferrite
rod, which may be salvaged from a
discarded MW/LW pocket radio. The
tap on the coil is made at 20 turns
from the earth connection.

In order to compensate for the rela-
tively low radiation efficiency of the

Just like the associated hand-held
transmitter (see previous article), the
receiver is simple to construct.

As can be seen from the circuit
diagram, parallel tuned circuit L1-C1
receives the transmitter signal, which
is first buffered by means of a dual
gate MOSFET — Ti — in order to pre-
vent excessive loading of the tuned
circuit. Further amplification is per-
formed by T2 , before rectifier circuit
Di-Ce can provide a pulsating voltage
to Ta, which drives PLL detector ICi
with a sawtooth-like signal. The lock
output— pin 8— of ICi controls Rei via
relay driver circuit T4-T5.

As to a few details concerning the
receiver circuit, the PLL chip signals
30 elek.or india Aug/Sepl '986

remote control for light
switches — 1

proposed transmitter aerial, the peak done with a scope; observe the
pulse voltage across C6 amounts to pulsed 30 kHz carrier, which should
some 150 V PP when the oscillator is look as indicated by the inset signal
turned on for 8 ms by Tt , which is waveform drawing; the pufse-on time
driven with an 18 Hz signal from ICi. of 8 ms is determined by Ca-Rj. and
The pulsed mode operation of the their values had better not be
oscillator ensures a relatively low changed, since they are the optimum
mean power consumption of the bat- compromise between transmitter
tery-operated transmitter when a current consumption and output
receiver unit is to be activated. power. (B)

Testing the transmitter is readily

remote control for light

switches — 2

the lock condition by pulling pin 8 L. and L2 should be separated from
low; Ci: is charged and functions as each other with a metal screen to
a buffer device in case the PLL input preclude stray coupling,
voltage disappears because of the The receiver is readily tested and
fact that the transmitter coil is no adjusted by placing an operative
longer held steady for optimum transmitter at a distance of about 4
reception (directive effect of the fer- metres. The optimum position of the
rite rod). At the receiver input, Ri coil on the ferrite rod can now be
should be mounted direct at the rel- found by connecting a scope to the
evant MOSFET gate so as to prevent drain of Ti and sliding Li for maxi-
possible oscillation tendency of T>. mum received signal. In the absence
Like the transmitter coil, Li is wound of an oscilloscope, the signal at the
on a 4 cm long paxolin former, which PLL input (pin 3) may be connected
can be slid over the ferrite rod to find to a loudspeaker to position Li for
the position that gives optimum maximum voice coil movement at
reception. Use 210 turns of 36 SWG 18 Hz. After it has been positioned
( . ' 0.2 mm) enamelled copper wire; correctly, Li may be glued into place
the coil length should be about 3 cm. on the rod.

i m mail'

Adjusting the PLL is done with Pi,
which should be turned carefully
across its travel to establish the points
at which the PLL fails to lock on the
incoming signal (Rei is deactivated
and the lock indication LED, if fitted,

goes out). Now set Pi to the position
in between the no-lock points. Care-
fully manoeuvre the transmitter to a
place where reception is worse, i.e.
where Rei is observed to go off.
Careful adjustment of Pi and further

trial and error will enable the user to
establish the preset position that cor-
responds to optimum receiver sensi-
tivity and reliability under less than
favourable circumstances. (B)

voltage-to-current 1 ^
converter X w

The converter proposed here (also
called voltage-controlled current
source) is based on just one opamp,
and provides to, or draws from,
ground a current that is dependent
on its input voltage. The unit can con-
vert negative as well as positive
voltages into negative currents (from
ground) and positive currents (into
ground) respectively.

When a Type 741 or CA3140 is used
in the Ai position, Rv=lk, and
R = 10k, U,„=±10V max.;

Iom =+ 20 mA max.; and gm = — lmS.
It is, of course, possible to change
any or all of these values as required
by using a different opamp and alter-
ing the values of the resistors. The
maximum output current is always
dependent on the opamp used. To
make such changes, the following
formulas may prove useful.

U + = U- = (Uin — Uout)/2 + Uoui

Uo = 2[(U.n-Uoo.)/2 + Uoui] = Urn + Uou.

Iom = iRv + Ir = Uin/Rv + (Uin — Uout)/2R
If R>>Rv (the usual case),

Iom = Um/Rv. (St) I

elektor mdia Aug/Sept 1986 31

lRv = Um/R„

by R Shankar

This FM band (88-108 MHz) preampli-
fier has been designed to come
round the problems associated with
wideband as well as narrowband
aerial boosters. Most commercially
available boosters are wideband
types with relatively poor selectivity
and adjacent station rejection, while
the (more expensive) narrowband
types are rather impracticable when
it comes to receiving stations well
removed from the (fixed) frequency
of peak amplification.

This proposed design is the best of
both worlds, since it features good
selectivity and strong signal hand-
ling, as well as a relatively low noise
figure and sufficient amplification
over the entire FM band. Tuning the
preamplifier is done in the living
room, by means of a simple poten-
tiometer mounted in an enclosure
which is conveniently located next to
the FM tuner as part of the hifi set.
The unit can also be made to function
as a 2 metres amateur band (144-
146 MHz) preamplifier by modifying
the tuned circuits to suit the higher
frequency.

The circuit diagram of the tuneable

booster— Fig. 1— shows that two re-
mote tuned circuits, along with a
MOSFET tetrode have been incor-
porated to minimize the chances of
running into cross- and/or inter-
modulation caused by strong local
signals. Varicap diodes Di and D2
form the variable capacitance to coils
Li and La respectively. The tuned cir-
cuits are set to the desired frequency
by means of the voltage applied to
the varicap diodes (3 to 24 V. reverse
bias). The RF gain offered by Ti
should be of the order of 25 dB. while
the noise figure is expected to be
about 2 dB.

The amplifier supply/tuning voltage
and superimposed RF output signal
are connected to the coax cable core
which is run to the power supply/tun-
ing unit, shown in Fig. 2. Tuning con-
trol potentiometer Pi constitutes the
feedback loop to the voltage regu-
lator composed of T7. Te and T9 . Turn-
ing Pi thus varies the voltage to the
mast-mounted booster form 15 to 36
volts. Regulator T2-T3-T« (Fig. 1) pro-
vides MOSFET Ti with a fixed
voltage of 11.4 V. irrespective of the
DC level on the coax core. Subtrac-

tion of 12 V from the 15-36 V input
voltage is by means of zener Do and
current source Ts. RF output voltage
and DC supply are coupled to the
downlead cable through Cu and L4
respectively. Cu and Ls (Fig. 2) have
the same function in the PSU. D13
prevents the PSU output voltage from
rising above 37 V in case of any
breakdown in the supply unit, while
D; protects the booster from accept-
ing a reverse voltage in case coax
core and screen are accidentally
reversed. T6 limits the supply short
circuit current to a safe 60 mA.

The following are important points to
observe in constructing the mast-
head amplifier and associated indoor
control unit:

1. Use a copper-clad board of maxi-
mum earth plane surface

2. Mount a metal screen across the
MOSFET case to suppress any

tendency to parasitic oscillation.

3. Keep the source lead as short as
possible; solder it direct to the

copper surface.

4. Keep the leads of G2 decoupling
capacitor Ci as short as possible; a

32 -ikio-in

ceramic disc capacitor is ideal for
this purpose.

5. Keep all coil connections as short
as possible to avoid amplifier

tuning over the wrong frequency
range.

6. Fit To with a small heatsink.

7. Mount a screen between amplifier
and DC supply section.

After the construction of RF head and
PSU has been completed, the latter is
tested by verifying the presence of
the variable (15.6 to 36.6 V) supply
and tuning voltage on the coax cable
core. The voltage across R« should
be lower than 0.4 V with the amplifier
connected at the far end of the cable.
Turning Pi should cause the voltage
at the collector of Ts to vary between
3 and 24 V.

The voltage at the emitter of Tz
should be constant at 11.4 V with
respect to ground, irrespective of the
tuning voltage set with Pi. Drain
resistor Ra should drop between 0.7
and 2 V. Set Pi to the centre of its
travel.

Optimum RF performance of the
booster can be achieved by carefully
stretching or compressing I« for
maximum amplification at about
95 MHz; tune the receiver to a weak
transmission at this frequency and
align for maximum S meter deflec-
tion or optimum audibility of the
signal above the noise level. Do the
same for signals at either extreme
end of the band and set Pi accord-

ingly. Ensure that the tuning poten-
tiometer can be set to give optimum
amplification for every frequency in
the 88 to 108 MHz band and mark the
tuning scale on the indoor unit in
steps of 1 MHz. In case it is not poss-
ible to obtain equal amplification
across the band, L3 may be adapted
carefully by increasing or decreas-
ing the number of turns. The tap.
however, should remain at 3 turns
from ground.

Those constructors striving for
utmost perfection may fit a 40 pF
trimmer capacitor instead of Ci, in
order that the amplifier may be tuned

for optimum (ie. lowest) noise figure,
which is not the same as tuning for
optimum amplification.

Finally, the coil data for the tuneable
booster are as follows:

Li = 9 turns 22 SWG (0.7 mm dia)
enamelled wire, close wound, coil
diameter 7 mm. Tap at 1 turn from
ground.

Ls = the same, tap at 3 turns from

Lza;hzb = 6 and 3 turns respectively,
26 SWG (0.5 mm dia) enamelled cop-
per wire on dia 10 mm ferrite ring
Type T37-12. (B)

one-chip DC converter

This DC step-up circuit may prove
useful for the incorporation in equip-
ment that requires the presence of a
supply voltage in excess of the nor-
mal circuit supply rail of, for instance.
+ 5 V. Ideal therefore for generating
the necessary + 8. . . 12 V voltage to
feed RS232 transmitter devices, or
the f 25 V programming voltage for
EPROMs, the Type L497 DC con-
verter requires very few additional
passive parts to produce any of the
output voltages listed in the table
below.

As to the components in support of
the converter chip, note Li, which is
a small coil, readily made by winding
about 85 turns of 34 SWG ($ 0.2 mm)
enamelled copper wire on a small
(11x7 mm) pot core having an AJ
rating of 160, e.g. the Siemens Type
6531-L160-A48. The total inductance

of Li should be of the order of
100 pH. Resistor R must be dimen-
sioned as indicated in the table for
any of the no-load output voltages.
Note that the voltage across R:. is
fixed at 1.2 V. and that the value of Ri
may therefore be computed from

11 i n" 111 I R.» I
5 10 I 125 8.8

5 15 I 80 13.8

5 1 20 60 18.8

I 5 25 50 23.8

1 <V> 1 <V> 1 <mA> | <kO> I

' Specifies no-load output voltage.

T Theoretical value:

R’=(Vour—1.2) <kQ>.

Finally, the output current may, of
course, be boosted by means of a
medium power transistor in a
suitable configuration at the Vo
output. HS

33

rms-to-DC

converter

For some obscure reason, estab- ~
lishing the root-mean-square (rms) '
value of an alternating voltage seems
to be among the least familiar pro-
cedures for many an electronics hob-
byist; measuring the alternating volt-
age may be easy, but deciding on the
relevant unit expressing quantity —
rms, mean, or peak-to-peak value —
is quite another matter.

Since the rms value of an alternating
voltage is the most frequently used of
the above mentioned three, some
convenient means of obtaining that
value without calculations may be of
interest in practical measuring tech-
niques.

The rms value of an alternating
voltage U across a resistor R equals
the direct voltage causing the same
dissipation level in R.

Example: a 50% duty factor, 1 Vpp
rectangular voltage across a resistor
R. Find the rms level of this voltage.

The mean dissipation in R, caused by
this periodic signal equals
‘A (Uppf/R = 1/(2R) [W]

The direct voltage causing the same
dissipation has a level of
‘At 2= 0.71 V. since P = ('A / 2)*/R =
I/(2R) (W).

This is also the conversion factor for
obtaining the rms value from the
peak-to-peak value, since
U,ms = t 'AUpp 2 = <A 1 2 U PP =

0.71U P p[V]

therefore U PP * 1.41Uim S in this
example.

Although moving coil meters
measure the mean value of the recti-
fied (pulsating) input voltage, they
are calibrated in terms of rms
voltages. Therefore, the calibration is
only valid for sinusoidal voltages.
The proposed rms-to-DC converter is
a relatively simple circuit as it incor-
porates a dedicated chip, the Type
AD536 by Analog Devices. Alter-
nating voltages applied to input ter-
minal 1 are proportionally converted
into a direct output current, which
causes a direct output voltage across
an internal 25 kohms precision re-
sistor; a buffer opamp outputs the

direct voltage equivalent (i.e. rms
value) of the input alternating voltage.
ICi functions as an input buffer in
view of the relatively low input im-
pedance of the rms converter chip.
The maximum permissible peak-to-
peak value of the input voltage to the
Type AD536 equals the symmetrical
supply voltage level; Di and D 2 have
been added to protect IC: against
accepting input voltage levels in ex-
cess of the ± supply voltage. S 2 func-
tions as a xl/xlO input attenuation
selector to enable high voltage
measurements; the function of Si is
to block any DC components in the

34

input signal to the converter. It is
useful to realize that the rms value of
a composite (AC + DC) signal is
calculated from
Ums = / Udc*+ Uac 2 .

Preset Pi should be turned to obtain
0 V with respect to ground' at ter-

minal 6 of IC2 with no input signal
applied and S2 set to the x 1 position.
The converter achieves an accuracy
of 1% for input voltage levels lower
than 100 mV and input frequencies
up to about 6 kHz. For signals up to
1 V. the bandwidth is expected to be

of the order of 40 kHz, while 100 kHz
may be attained with input voltages
above the 1 V level. Current con-
sumption of the circuit is about 5 mA.

(TW)

rodents deterrent

There are a number of well-founded
arguments against the use of poison
to get rid of mice, rats and other
rodents in and around the home.
From an ecological point of view, the
undesirable side effects are mainly
the disturbance of the natural food
chain of animals we do not wish any
harm whatsoever; most poisonous
substances devised to exterminate
mice are, unfortunately, quite difficult
to break down compounds, which
may, in the end, become manifest as
dangerous to our own health.

The ecologically accepted method
of getting rid of a population of mice
is, therefore, based on the controlled
introduction of such predators as cats
and owls, causing a high degree of
stress on part of the mice, which are
then quite quick to leave the relevant
premises or area.

Another method of bringing about a
high degree of stress is to produce a
high-pitch, frequency-swept signal
just above the audible range for
human beings. The signal is swept
rather than of constant frequency in
order to prevent mice from becom-
ing immune to the sound.

Parts list

Resistors:

R2;R3 15 k

Capacitors:

Ci = 1 n

C2 = 1fi;16 V electrolytic
C3 - 10 n
0 = 220 n

Cs = 1000f«; 16 V electrolytic

Semiconductors:

Dl. . . Da = 1N4001
ICi = 555

Miscellaneous:

Tri =6 V;200 mA.

TDi = piezo horn tweeter.

Fuseholder, PCB type, for Ft.
PCB Type 86490

ABS enclosure tor wall mounting.

The proposed rodents deterrent is
based upon the Type 5S5 timer chip,
which is configured to produce a 20
to 40 kHz output signal, swept at a
50 Hz rate. The latter frequency is ob-
tained from the mains by means of
C« and R3, which pass the
modulating signal to input pin 5. The
output of the swept oscillator is con-
nected direct to a high-efficiency

ensures a sufficiently high sound
pressure level to keep rodents out of
reasonably sized areas, such as attics
and garages.

The completed rodents deterrent cir-
cuit, along with the tweeter, may be
mounted in a simple ABS enclosure,
but care should be taken to observe
the directivity of the loudspeaker
when fitting the unit in its final pos-

35

thermostat-controlled
soil heating

1 I 1

15

Many people with a keen interest in
growing plants insist on the fact that
many of the more exotic species,
such as certain species of orchid and
fungi, will only thrive in warm soil
and relatively high humidity.
Whether or not this is a correct
assumption, this circuit offers the
possibility to keep the soil tempera-
ture in a miniature hot-house at a con-
stant, adjustable level.

The heating element is made of
several loops of plastic covered steel
wire, such as used in gardening. The
wire used in the prototype had a
diameter of 1 mm and a resistivity of
about 0.2 Q per metre.

The circuit diagram of the soil heater
shows that the heating element is
temperature controlled by means of a
triac, driven by a Type TDA1024 elec-
tronic thermostat which gets the
necessary information as to the soil
temperature from R6, an NTC type
sensor.

The circuit is fed from the trans-
former secondary by means of recti-
fier Di and series resistor Rj.
Regulation at 6.5 V is internal to the
1C, and C3 smoothes this voltage. Rj
and R« provide the IC with a mains
synchronizing signal, while Ci
causes a controlled phase shift in
order that the relatively low
operating voltage can still ensure the
correct zero-crossing synchron-
ization.

The temperature sensor circuit is
composed of R->, R6, and Pi. The
sensor proper, R$, must be placed
into the soil at a suitable position,

electrically well isolated, of course.
The optimum soil temperature,
which should be established by trial
and error, is adjustable with preset
Pi; Fig. 2 shows the correlation
between soil temperature, heating
element voltage, and preset tem-
perature.

If necessary, a more powerful

heating element may be dug into the
soil, but the ratings of the fuse, Tri
and Trii should then be changed
accordingly. The transformer sec-
ondary voltage, however, should
remain at 9 V. With the components
as indicated in the circuit diagram,
the heating energy is about 40 watts.

(Sr)

16

This circuit provides motor-cycle
riders with a gear indication to the
foot-operated lever at one side of the
engine block. The proposed indi-
cation unit will be appreciated by
those riders in the habit of forgetting
which gear they have selected when
attempting to drive off at traffic lights
or crossroads and finding that the
engine stalls because it had been
switched to second gear.

The circuit as shown is based on
the use of two gear-lever operated.

motor-cycle

indicator

plunger or roller type microswitches,
along with the neutral gear indication
lamp, which is a standard item on
most types of modem motor-cycle.
Bistables Ni-N and N3-N3 serve
as debouncer circuits for micro-
switches S: (lever down) and S2 (lever
up). If either one switch is actuated.
N. or N; will cause bistable N12-N13
to be set or reset; counter JCs counts
up (U/D - 1) or down (U/D = 0) as a
result of actuating S2 or Si respect-
ively. On release of the relevant

gear

microswitch, AND simulator D1-D2-R6
supplies IC5 with a clock pulse, in-
crementing or decrementing the
gear readout composed of IC6 and
the indication-panel mounted 7-seg-
ment LED display.

Input pin 5 of gate Ns may be wired
to point A, B, or C to suit 4-, 6-, or
5-gear types of motor-cycle respect-
ively. N6 inhibits OR gate Nis from
supplying further clock pulses if S2
is operated when driving in top gear.
N 16 and N11 have the same function

9

■1 £-1

T-- T T -

S-®

JWf

o

u

Sgg,

f

•*• ;.

li

-iZ)^

— K 3

©S>

•F> l§>

- 4 - 4 -

wyn

for the bottom gear, preventing the
counter from decrementing the dis-
play reading at gearing up from neu-
tral to 1. •

If the neutral switch — S — is closed,
ICs supplies the A and B inputs of
IC6 with logic low levels; the level at
the C input need not be forced low,
since the neutral gear is in between

Nowadays, most clocks and watches
are quartz controlled and, therefore,
accurate to within a few seconds a
year. Older type electric clocks, par-
ticularly those used in large groups
in warehouses, department stores,
factories, railway stations, and so
on, were centrally controlled and
synchronized. This synchronization
was effected by pulses derived from
the mains and sent to each clock via
a separate cable network. Many
people have such a clock as a
curiosity, but have not the means of
driving it. The circuit described here
will help. . .

With reference to the diagram, pulse
shaper Ti triggers monostable IC2 at
the mains frequency of 50 Hz.

first and second, both of which pos-
itions cause the most significant bit —
C — to be low anyhow.

Parts RS-C2-N10-N5 have been in-
cluded to prevent an erroneous dis-
play reading at gearing down from 2
to neutral and up again; for two
seconds, N15 is disabled from clock-
ing ICs, so that the lever-up pulse is

Counter IC3 is reset automatically
after every 3000 pulses by IC* and T2 .

not detected.

At power-on, R? and C3 preset
counter ICs to state 1.

In conclusion, it goes without saying
that Si and S2 should be good quality
microswitches, sealed against moist-
ure and dirt. (R)

17

At the same time, bistable ICs
toggles and causes the bridge circuit
composed of T?. . .T» to reverse the
motor polarity every 60 seconds.
Depending on the type of clock you
have, the transformer secondary
voltage may have to be selected to
supply about 0.7 times the normal
operating voltage of the clock motor.
Furthermore, the bridge circuit as
shown should not be made to
operate at voltages in excess of 30 V,
while the maximum current is about
250 mA.

There is only one adjustment point in
the circuit, namely Pi, which should
be set to achieve maximum sup-
pression of mains borne noise; if this
can not be checked, the preset may

industrial-clock controller

be turned to its centre position, resistance, but care should be taken only half the number of 50 Hz periods
Should the clock be slow, Pi may be to avoid setting a monostable time can reach the counter. (Sa)

adjusted to give a slightly lower longer than 20 ms, as in that case

Alan G Hobbs

telephone-bell simulator

This circuit is intended for use in a
small private telephone installation.
The ringing tone sequence is 400 ms
on, 200 ms off, 400 ms on, 2 s off.

In the accompanying diagram, Ni
and Ns form an oscillator that
operates at a frequency of 5 Hz,
which gives a period of 200 ms. The
oscillator signal is fed to two decade
scalers, which are connected in such
a manner (by Nj and Na) that the input
signal is divided by 15.

The second input of Na may be used
to switch the divider on and off by
logic levels. If this facility is not used,
the two inputs of Na should be inter-
connected.

Resistors Rs to R6 incl. form an OR
gate that controls a relay via Ti and T2
which are connected in a darlington
circuit.

Outputs 5 to 9 of IC2 go high sequen-
tially, so that the relay is energized for
400 ms (when 5 and 6 are high), then
off for 200 ms (output 7 is not connec-
ted), and then energized again for
400 ms (when 8 and 9 are high). After
that, the relay is off for 10 periods =
2 s, and then the cycle repeats itself.

38 eleklor india Aug Sept 1986

-**-<§)

This timer automatically switches off
equipment left operating unattended
for more than about thirty minutes.
The circuit operation is readily
understood by following its power-on
and time-out functions. Almost im-
mediately after S< has been de-
pressed. relay contact rela closes to
power Tri and the equipment con-
nected to the mains outlet. This thus
happens because the initial pres-
ence of the + 12 V supply voltage in
the circuit causes counter-oscillator
ICi and set /reset (S/R) bistable N:-Nz
to be reset by means of a short, logic
high pulse at the junction of Ri and
Ci. The outputs of Ni and N3 go high
and low respectively and Ti can
energize Rei. So far for the power-on
automatic hold function of contact
rela.

After being reset, ICi starts counting
down its on-chip generated clock
pulses which have a frequency of
about 2 Hz. LED Di flashes at this
rate to indicate the countdown con-
dition. Note that S2 has been pro-

vided to reset, i.e. disable the timer
permanently, in which case Di lights
steadily. The LED, therefore, has a
threefold indicator function in the
present circuit: timer on (flashing),
timer and equipment off (off) and
timer off while the equipment is on
(steady light).

As long as counter output Q12 re-
mains at logic low level, the voltage
at the collector of 0 inverter T:: can
not cause the relay coil current to be
interrupted by T-s. If, however, some
34 minutes {T(Q:-j ,-x2' 2048s)
have lapsed since IC and N1-N2
were reset, Q:a goes high, causing
the two-gate bistable 10 toggle: the
output of Ni goes low. but Re: re-
mains energized by T;, since the
other input of NOR gate N is still
high, i.e. counter output Q : has not
been set as yet. The selfoscillating
buzzer starts sounding at a 2 Hz rate,
however, since T3 is driven by NOR
gate Ni which receives two logic
low levels at its inputs. The user is
thus notified that the has another

15 seconds or so left to depress Si
for another 34-minute interval. If no
such action is taken to reset the timej^
before Qs goes high, N3 disables the
relay driver transistor, and contact
rela consequently cuts the mains
voltage to Tri and the connected
equipment.

The foregoing outline of the circuit
operation makes clear that depress-
ing Si or switching on S2 is the
only way to keep the buzzer from
sounding and the mains relay from
switching off both equipment and
timer circuit. If desired, push-to-
break switch S3 may be operated to
break the mains supply within the
half hour interval, and without the an-
noying sound of the buzzer.

Finally, the indicated timing intervals
may be changed to suit individual re-
quirements by using other counter
outputs and/or another clock fre-
quency for ICi (adapt the values of
R2-C2).

AS

39

AN OPTICAL-FIBRE
NETWORK

FOR OFFICES, FACTORIES AND HOSPITALS

A very versatile experimental local telecommuni-
cation network based on optical fibres has
recently been installed at the Geidrop Project
Centre — a part of Philips Research. Known as
PHiLAN — Philips Integrated Local Area Network —
this system is used both for studying the
possibilities and difficulties of such a system and
for demonstrating likely applications, which may
not be immediately obvious, to interested parties.

This new network will mainly be used for these
studies; it is not intended as a product prototype.

Computers and their
derivatives have been
used in offices, labora-
tories, hospitals and fac-
tories for years, and in the
future there will be even
more of them. They are
used for word processing,
electronic filing systems,
administration, automated
measurements, processing
X-ray exposures displayed
as television pictures, and,
finally, for monitoring and
control of automated pro-
duction processes. The In-
crease in fhe number of
these machines means
that there is a growing
need for transfer of digi-
tized information (data).
Data transfer via tele-
phone lines, as now often
used, is slow and will not
be able to meet the future
demand. In about 70 per
cent of cases information
is transferred over a short
distance: within an office
block, laboratory, hospital
or factory site. In short,
there is a need for an in-
house telecommunication
network providing ad-
equate capacity (band-

width) for the rapid and
reliable transfer of large
quantities of data. If this
network can also handle

the ordinary telephone
links for the site we have
an integrated local area
network'. Such as system

has the advantage that it
only requires a single net-
work and central facilities.
Work is currently in pro-
gress on the design and
construction of such local
networks at a number of
locations.

In the design of PHILAN,
optical-fibre cable was
selected as a transfer
medium of high band-
width, and a ring network
has been installed. All the
users have their own
branch line on the ring,
enabling them to send in-
formation along the ring
to other users and to re-
ceive messages. Special
plugs and sockets (see
Fig.1) have been designed
for connections to the
ring. Each of the
machines to be connec-
ted has a circuit that first
stores the signals to be
transferred — after digiti-
zation if necessary — and
then delivers them to the
ring at a rate of 20.48 MHz,
the clock frequency
generated by the central
control unit. All data
transfer along the ring

takes place at this rate.

• This circuit also acts as a
regenerative repeater for
all the signals transferred
along the ring so. that a
good signal strength is
always guaranteed,
regardless of the number
ol stations along the ring
or the distances between
them. A Philips-designed
optical relay, shown in
Fig.2, is fitted behind each
| socket to ensure that the
ring is not interrupted
when a plug is pulled out
or a device is switched off.

Time-division

multiplex

Time-division multiplex is
used to employ the band-
width of the optical-fibre
ring as effectively as poss-
ible and to permit the sim-
ultaneous transfer of a
large number of ‘packets'
of information coming
from various stations along
the ring. The continuous
stream of bits passing
along the ring is divided
into frames with a length
of 125 ^s, each containing
320 bytes of 8 bits. A
number of bytes in each
frame are reserved for
messages for the internal
organization of the net-
work. One or more of the
remaining bytes, a 'time
slot', can be allocated for
the transfer of an infor-
mation packet; if
necessary a second time
slot can be used at the
corresponding position in
the next frame, and so on
until the complete packet
has been transferred. An
extreme example is the
use of the network for tele-
phone communication; in
this case a user deter-
mines how long a (narrow)
time slot will be occupied.
A large number of paral-
lel information streams of
varying width, which can
be regarded as so many
parallel communication
channels, travel in this
way. The narrowest chan-
nel, of 1 byte per frame,
has a capacity of
64 Kbit/s, corresponding to
the capacity required for
a digital PCM telephone

Figure j Meander structure in the PHILAN nng To .inn;
the effects on the users of a fault somewhere in the nng.
the ring is subdivided into a number of loops or
meanders, which are linked via the centra! control unit
If there is a fault the meander containing the fault is
removed from the nng and the remainder continues to
operate Besides taking corrective action in the event of
failure, the central control urut synchronizes the signals m
the ring and controls the time multiplexing The time
multiplexing subtiivides the total transfer capacity of the
nng (2048 Mbit/s) into parallel channels of different ca-
pacity as requited

both cases a signal is
transmitted to the centrai
unit and an attempt will
be made to remedy the
failure automatically. In-
tended information
transfer to an inoperative
address will also automati-
cally be blocked to pre-
vent unneccessary
reservation of transfer ca-
pacity. In this way, the net-
work always remains
available to as many of
the user stations as
possible.

Another protective
measure for use with this
transfer of 'packets’ of in-
formation has been incor-
porated: feedback to the
sender when part of a
message is distorted on
receipt. This part of the
message is then sent
again, which means that
the sender can be always
sure his information has
been properly received
without having to check it
himself.

The demon-
stration network

A number of widely differ-
ing machines have been
incorporated in the net-
work that has been con-
structed for research and
demonstration. In addition
to the telephone and in-
tercom facilities usually
found in offices, a word
processor and an elec-
tronic 'archive system'
(MEGADOC) have been
connected, as well as
computers of various sizes
and types with their ter-
minals. The network can
also be used for slow-scan
TV, in which a limited
number of TV pictures are
transmitted pe- minute for
surveillance purposes.

System

protection

In a ring network such as
PHILAN, measures must be \
taken to prevent a failure
at any point in the ring I
from putting the entire net- <

work out of action. In the
PHILAN ring, this can be
done at two levels: a
| single malfunctioning ter-
minal can be short-
circuited or — in the event
of a more extensive failure
— part of the ring can be
short-circuited (cf. Fig.3). In

four-tone siren

This interesting little circuit is par-
ticularly aimed at modellers. Based
on just one IC, it is easy to make and
•it is inexpensive. Moreover, it
operates from a 3 V battery, and con-
sumes only ISO in the quiescent
state, and 28 mA in operation.

The four different tones are selected
by two switches, Si and S*. The table
correlates the switch positions and
the produced sound. (St)

M van Oosten

car lights monitor

t)0-p

Many traffic accidents are caused by
failing car lights. Often, the driver is
not aware of such a malfunction,
because the warning lights provided
on the dashboard do not, strictly
speaking, monitor the relevant lights,
but rather the switch position since
they are almost invariably connected
in parallel with the relevant car lights.
The proposed circuit is intended to
indicate the failure of one light in a
pair: sidelights; headlights (up to
55 W); rear lights; brake lights; or fog
lights. The two lamps must have the
same rating.

Counter-wound coils, Li hnd L 2 , carry
the same current when both lamps
are working correctly, so that the
magnetic fields created by these cur-
rents cancel one another. When one
of the lamps fails, the magnetic field
caused by the current through the

other induces a voltage in Lj. This 2
pulse causes the TIC106D to switch
on, and this in turn makes Di light. If
both lamps fail simultaneously (the
probability of which is, however,
minute), the circuit does, of course,
not function.

Because in practice the two lamps do
not come on or go out simul-
taneously, R.-Ca-Ra provide a delay to
enable the magnetic field to stabilize.

Note, however, that Ci must be
matched to the particular lamps
being monitored: increasing its value
makes the circuit less sensitive
(longer delay).

The coil is easily made from an old
(or new) core of a choke or dimmer
switch. First, wind two times 11 turns
SWG22 enamelled copper wire
around the core as shown in the
drawing. Inductor L 3 consists of

twenty turns SWG40 enamelled cop-
per wire (this coil does not carry a
large current). Note that the black
spots in the drawing are the same as
those in the circuit diagram. If the cir-
cuit does not work, it almost certainly
means that the connections of either
Li or La have to be interchanged.

To monitor all the lights of car. the cir-

Since many amateur receivers are fit-
ted with an S meter that functions far
from logarithmically, the proposed
circuit should be a welcome exten-
sion of such receivers.

Although ICs such as the CA3089 or
the CA3189 are not in common use
any more, they serve a useful pur-
pose in the meter circuit, because,
apart from a symmetric limiter, a
coincidence detector, and an AFC
amplifier, they contain a very good
logarithmic amplifier-detector.

As is seen, the circuit is fairly simple,
but remember that these ICs operate
up to about 30 MHz, so that the wir-
ing of the meter, and also its connec-
tions in the receiver, should be kept
as short as possible.

Note further that

■ the input of the CA3189 must be
terminated by 50 Q;

■ the connection to the input of the

cuit will have to be built as many
times as there are pairs of lamps. The
indicator diodes are best fitted in the
dashboard. It is, however, possible to
use only one LED for a number of cir-
cuits: when this lights, it is then, of
course, necessary to walk around the
car to see which lamp has failed.
Once the LED lights, it remains on

until either the thyristor or the
ignition has been switched off.

S meter

CJC-T ^ 0 "*

cable: used between it and the meter cir-

■ if it is not possible to obtain the in- cuit.
put signal from a low-impedance (g)

2

43

23

"Alea iacta est” (the die is cast,
freely) someone said quite a few
years ago, and promptly engaged in
sundry military actions that are gen-
erally reported as having been decis-
ive for global history. Whatever the
relative importance of this notorious
person's decision at that time, he is
not likely to have employed a SMD
die as described here, since he used
the verbal form cast rather than a
clausal construction (in Latin, of
course) to indicate the presence of
clock pulses from a Schmitt-trigger
gate oscillator, at the relevant input of
a Type 4029 binary counter which is
preset to state 9 by means of jam
(preset) inputs ]» . . . J3 while its
Qo. . ,Q2 outputs may represent 1 of 6
pseudo-random states 9... 15 after
removing one's fingers from the
touch-sensitive contacts between os-
cillator and counter clock input.
Counter output states 9 ... 15 were
chosen rather than 1...6 with the
cor resp onding preset 1, in order that
the CO (carry out) could be connec-
ted to PE (preset enable) via inverter
N*. This arrangement causes the
binary value at the Qo . . .Q2 outputs
to vary between 1 and 6, since Q3 is
left unused. CO goes low any time
the counter reaches output state 16,
which can not be represented by
means of the four binary outputs to
the IC (2 4 = 16). Consequently, the
counter loads the preset value 1 (9),
since PE goes high.

LEDs Di. . .D7 are arranged in the
form as usual on the "six" face of a
die, and the random number is, of
course, displayed as an imitation of
the spot(s) seen on the cube faces.
As to the construction of the SMD die,
the tiny parts are fitted onto ready-
made, through plated PCB Type
86454, which comes together with
the Type 86452 (sideway RAM for
BBC and Electron, also a SMD project
in this issue).

It is noted that the 9 V battery is
clipped direct onto the circuit board
to make a compact unit with the
LEDs facing up. The "cast" contacts
are four lengths of stripped wire at
the LED side of the PCB, mounted at
all four sides. Placing your fingers
onto either two of these wires facing
one another causes all seven LEDs to
light, while on release a pseudo-
random value is displayed. (St)

SMD die

44 elekto

fll.; A A
m A D7

[A ? A o A

©-?©

J B88 ;

Ed'dB

85888 ;

ns R5

Parts list (all parts SMOI

Resistors:

R.;Rr 100 k
Rj. Rt -560 B

Capacitor:

C. ^ 12 n

Semiconductors:

O' D; = LED Type CQV231 or LSS210DO
i Siemens!

1C. 4029
ICr - 40106

Miscellaneous:

battery clips lor PCB mounting
PCS Type 86454

9 V battery PP3

PIA for Electron

R v Linden

Despite its neat design and relatively
low cost, the Acorn Electron com-
puter suffers from an unfortunate
lack of I/O support, which is
remarkable, considering the fact that
it is a relatively simple matter to add,
say, two I/O ports to enable the com-
puter to drive a printer, plotter,
modem, or other peripherals by
means of the proposed PIA (periph-
eral interface adapter).

The circuit diagram of the PIA-based
extension shows that address
decoding over the full 64 Kbytes is by
means of two 8-bit magnitude com-
parators Type 74LS688. Address
selection is manual with switches
Si. . .Su, which provide a logic low
level when closed; observe this
when writing out the ones and zeros
to arrive at the desired address in the
I/O map. The PIA chip is enabled
when the preset address matches
that on the computer's address bus;
writing simple I/O drivers is there-
fore mainly a matter of assigning the
relevant address block to control
words and PIA I/O data.

T. has been included to enable the
PIA circuit to generate and forward
interrupt request pulses by means of
the wired-OR arrangement for this
control line.

In case it is desirable to switch
heavier loads than is normally per-
missable with the PIA outputs, it is
suggested to employ power drivers/
inverters such as those in the
ULN2000 series. (St)

45

A B Bradshaw

25 !

13

WP

eil «L ’1

line bar generator

»~r r 1 — | r n _0 " v '

T0

e> <s>

uu jui n 'I

4 i-“-i

The video signal transmitted by most
TV broadcast stations is rather com-
plex. For most tests and experiments,
however, a fairly simple signal will
suffice. The circuit presented here
provides a small, inexpensive source
of line synchronizing pulses and line
bar.

The first of the three timers in the
diagram provides 4.7 ms sync pulses.
It is arranged as an astable multi-
vibrator with a period of 64 (is. The
rising (here: negative-going) edge of
the sync pulse triggers a second
timer. The width of the output pulse
of this timer determines the position

of the line bar. The line bar proper is
provided by the third timer. To obtain
a usable video signal, the sync and
bar signals must be added, which
takes place in R«-Rs-Rs. The resistor
network is followed by a buffer that
ensures an output impedance of
75 ohms. The unit can, therefore, be
connected direct to a standard video
input. The sync and bar signals
occupy 40 per cent and 60 per cent
of the composite signal respectively.
Calibration is carried out by connect-
ing the unit to a monitor or, via a
modulator, to a normal TV receiver.
Presets P>. Pa, and Pa are set to the

centre of their travel. Turn Pi to
obtain a still picture. If the sync pulse
is too wide, it will be visible at the
left-hand side of the picture. The
pulse may be narrowed with the aid
of Pa , after which Pi may need a small
re-adjustment.

Where an oscilloscope is available.
Pa can initially be set to obtain 4.7 \t s
pulses at the output (pin 3) of ICi.
Then, the total period is set to 64 ns
with the aid of Pi.

The line bar is centred with Pa: as its
width is fixed, this completes the
calibration.

2708 alternatives

Thanks to the development of an ever
expanding range of capacious
EPROMs in the 27xxx and 25xxx
series, the Type 2708 has become
completely obsolete. Not only is this
forerunner in EPROM technology
relatively hard to program, it is also
expensive in view of its modest
1 Kbyte holding capacity. Moreover,
it is a more and more difficult to ob-

it stands to reason that replacement
of the 2708 with either the 2716
(2 Kbytes) or the 2732 (4 Kbytes) is
most readily accomplished if the dif-
ferences in pin functions are first
taken into consideration.

The pinning overview and associated
table go to show quite conclusively
that the replacement is no daunting
task, since the former positive and

negative supply pins to the 2708, 19
and 21 respectively, may be hard
wired as suggested for either the
2716 or 2732.

It should be noted that pin 18 (CE for
the 2716 as well as the 2732) is tied to
ground, while pin 20 (OE) is driven
by the computer CS signal. This new
arrangement is of no consequence
for neither EPROM nor computer,
since OE may function as CE if it is

46

realized that the EPROM can not be
switched to its low power standby
state anymore. However, this minor
drawback merely causes an increase
in current consumption, whilst at the
same time offering a faster EPROM
access time, as only the three-state
bus drivers are enabled internally.

rather than the entire chip logic.

As the Type 2716 and 2732 EPROMs
offer double and four times the ca-
pacity of a 2708, respectively, a
manual address block selection may
be added to the circuit; this set-up,
composed of a switch and resistor (to
be constructed double for the 2732)

is marked with an asterisk in the ac-
companying diagram. Wire Aio (and
An, if applicable) to ground if you in-
tend to stick to the 1 Kbyte EPROM
contents, located in the first 1024
bytes block. (Sv)

smart LED selector

In this tiny circuit, for use in, for in-
stance. a two-lights model railway
signal, one of two LEDs may be
selected with either a single pole
switch or a series transistor, as shown
in the circuit diagrams. Note that the
LEDs are fed via a common current
limiter resistor, while a switch is con-
nected in series with one of the
LEDs.

Why do not both light simultaneously
when the switch is closed? Because,
apart from their colours, the two
LEDs also differ as regards their for-
ward voltage drop; when connected
in parallel, therefore, the LED having
the lower voltage drop should be fit-
ted with the series switch; this ar-
rangement causes the high voltage
drop LED to light when the switch is
open and to go out when the switch
is closed, at which moment the other
LED takes over.

Two of the accompanying four small
circuits show the use of a series
switching transistor rather than a real
switch, but the difference hardly re-
quires further detailing, since apply-

ing sufficient drive to the base is in
fact the same as closing the switch.
Two LEDs of identical colour may
also be used as shown, and the ad-
ditional series diode is seen to create
the necessary voltage drop dif-
ference to distinguish between the
LEDs, which, of course, have roughly

the same on/off voltage character-

Finally, the value of R is established
from the supply voltage level and the
typical operating current of the LEDs,
which is usually of the order of
20 mA for maximum allowable
brightness. KD

47

toilet ventilator control

1

Many toilets have a ventilator, which
is energized along with the toilet
light. However, since not every visit
of the toilet requires the ventilator to
start turning, this circuit offers an im-
proved control method, which is still
based upon the use of the light
switch.

The circuit configuration shown in
Fig. 1 may be used in case the toilet
ventilator is powered from the same
mains lines as the light. Bridge recti-
fier Bi and opto-coupler Type TIL113
serve to detect whether or not the
toilet light is on. The ventilator is ar-
ranged to start turning after the light
switch has been operated twice. If
this is the case, the output of Ni will
go high twice; the first time, C« is
charged, the second time will cause
pin 6 of N2 to be logic high, while
the output of this NAND Schmitt trig-
ger gate will supply a logic low pulse
to Na when the voltage at point 3
reaches the logic one level (see
timing diagram Fig. 2). N3, then,
charges Cs which, along with P2 and
R9, determines the ventilator "on" in-
terval, while Pi, C4 and Ra establish
the maximum interval between the
reception of first and second trigger

The circuit option with Tz may be
used if it is less desirable to run an
additional wire to the light for the
purpose of obtaining the trigger
pulses; the LDR should be located as
close as possible to the bulb in order
to preclude erroneous triggering
due to the presence of daylight. The
use of the LDR does not change the
basic operation of the circuit, of
course, and the indirect method of
triggering is in fact to be preferred in
view of the risk associated with
direct mains connection in the case

of the first mentioned circuit option.
Another interesting use of the circuit
option which incorporates Si, T3 and
Ti is a semi-intelligent door bell ar-
rangement; bell 1 will sound any time
Si is depressed, while bell 2 will
only do so if the button is operated
twice within the given interval; it is
not difficult to come up with a
number of useful applications for this
circuit when used in and around the
home. However, note that the timer
parts Cs, Pi and Rs will have to
change places with Cs, P2 and R9 re-

This switch is energized by light and
can, therefore, be used, for instance,
to switch on the aquarium lighting in
the morning. Both the sensitivity and
the hysteresis of the circuit can be
preset; Re is energized in the pres-
ence of sufficient light.

The sensor is an n-p-n phototransistor
Type TIL81 or BP103, which conducts
when light falls upon it. The conse
quent current is divided between Tz
and R4-C1. Since Tz is connected as a
current source, no current will, how-
ever, flow through R4-C1 as long as
the current in T. is smaller than that
through Tz as determined by Pi.

When the current in T. is large
enough, some will flow through Rz
and charge Ci. As soon as the
resulting potential across Ci is
greater than half the supply voltage,
the CA3130 toggles. A current then
flows through Rs. Pz. and Ri, which
will cause a small reduction in the
current through Tz. This means that
even if the current in Ti drops
slightly, the circuit will not revert to
its original state. The magnitude of
this hysteresis is dependent on the
setting of Pz. Note that the hysteresis
prevents the circuit oscillating
around the starting level.

The sensor may also be a photodiode
or light-dependent resistor (LDR), but
a phototransistor gives better per-
formance, particularly when the dif-
ference between the on and off states
of the circuit is small.

Resistor Rz and capacitor Ci could
be omitted, but they augment the
hysteresis by delaying the input
signal from reaching the CA3130.

The current consumption of the cir-
cuit is determined primarily by the
requirements of the relay. Ignoring
the relay, the circuit consumes about
10 mA. which makes it possible to
use a Type 78L12 as voltage regulator.

(R)

desiccator

32

Many of the smaller working areas
available to hobbyists suffer from
humidity, which in no time causes a
number of tools to be covered in a
thin layer of rust. Humidity does not
do most test equipment or books and
the like any good either. The only
solution to this is to try to keep the
area drier by increasing the tem-
perature.

A couple of 100 W light bulbs or a
100-200 W heating element work
wonders in this respect, were it not
for the increases in the electricity
bill. And that is where the present
circuit can help.

With reference to the diagram, the
two HEF4001Bs, in conjunction with
humidity sensor H, generate a volt-

age across Re, that is directly pro-
portional to the degree of humidity.
The function of opamp A. is merely
to present a high impedance to Urs.
The voltage is then applied to the
inverting input of comparator Az,
which has an hysteresis of about
IS per cent. The reference voltage at
the non-inver'ing input of Az can be
set between 0.6 V and 3.0 V with Pi.
which corresponds to a humidity
between 20 per cent and 100 per
cent. As soon as the ambient hu-
midity exceeds the value set by Pi.
the comparator toggles, the triac
conducts and switches on the
heating element. Current consump-
tion of the circuit is a modest 13 mA.
If light bulbs are used, they should

be shielded with a metal hood to pre-
vent the likelihood of a fire.
Calibration is carried out with the aid
of a solution of cooking salt in some
water; placed in a reasonably small,
closed space, this will soon raise the
humidity to 7S per cent. Adjust Ci to
obtain a potential difference across
Rs of 2.25 V. Next, adjust Pi so that the
triac just does not conduct. In prac-
tice, the circuit will then come on at
a humidity of about 80 per cent.

51

MICROELECTRONICS

AND

PHARMACEUTICALS

by Charles Lewis

Public confidence in the
pharmaceutical industry is
such that both doctor and
patient usually accept
without question the qual-
ity and therapeutic re-
liability ot medicinal
products.

However, constant and
stringent precautions and
checks are necessary at
all stages of manufacture
to ensure the quality and
safety of such products.
Since pharmaceuticals
apply directly to human
beings, this is of prime im-
portance and the
manufacturers' responsi-
bilities are that much
more onerous.

Close attention to quality

and safety, of course is in
the interests ot the
manufacturers themselves,
since they have their own
hard earned reputations
to safeguard by ensuring
that their products are
exactly what they are
claimed to be. Conse-
quently there is a
ceaseless search for new
and improved ways of
achieving these ob-
jectives.

The advent of micro-
electronics has proved ot
the greatest importance,
not least to the producers
of manufacturing
machinery. British makers
of such equipment have
been quick to seize the

opportunities presented In
this new and rapidly ad-
vancing field, with the
result that the phar-
maceutical industry is one
ot Ihe most technically
advanced— ensuring a
constant flow ot precision
made, high quality
products.

Perhaps one of the most
outstanding examples of
harnessing modern
technologies to the im-
provement ot quality and
reliability is the Copley^
computerized multi-
parameter measuring,
analysing and data ac-
quisition system known as
DATAPHARM. It is a com-
puterized, multi-parameter
control system for the for-
mulation, in-process and
quality control of tablets
and other dosage forms.

Process

performance

chart

It may be used to assess
any or all of the par-
ameters of weight vari-
ation, friability, thickness,
diameter, and hardness.
The results are presented
as a simple integrated
print-out in the form of a
process performance
chart and a bar
charthlstogram plot. From
this the user is able to
determine instantly
whether the batch being
tested has passed or
tailed. At the same time,
correlations can be drawn
between the various
parameters and an assess-
ment made of any prevail-
ing unwanted trends
DATAPHARM consists of a
central computer supplied
with the appropriate soft-
ware and printer. The com-
puter utilizes the
measured inputs form the

various test instruments,
such as a disintegration
tester and a hardness tes-
ter. It integrates them into
a statistical analysis
program. The user-friendly
system displays simple
operating instructions on
the visual display unit,
and commands are
entered by single key op-
eration. Product specifi-
cation tiles can be built
up and maintained,
together with separate
batch production files,

The system, being
modular, is highly flexible,
so that test instruments
and associated programs
can be added or deleted
to suit individual re-
quirements.

The theory behind the
equipment is that a multi-
parameter control system
not only tells the user
whether a particular
batch of products has
I passed or failed the
chosen parameters, but
allows those parameters to
be interrelated. The
parameters associated
with tablet production are
in many cases correlative.
For example, high hard-
ness values may increase
disintegration times and
decrease dissolutions
values. On the other hand,
if hardness is too low,
friability and the percent-
age defective could well
be too high.

Disintegration
and dissolution

A range of values could
therefore be chosen so
that relatively high hard-
ness would produce ad-
equate disintegration and
dissolution values, while
maintaining low friability
and percentage defective
values. Similarly, high pro-
portions of values relating

| to weight variation could
| increase hardness and /or
| decrease dissolution
values.

I By exploiting the obvious
correlations between
hardness, disintegration,
j dissolution, friability,

; percentage defective, and
j weight variations, the
parameters can be ad-
justed to produce the best
dosage form.

I In operation, a new prod-
uct file is created or, if one
already exists, is recalled.
The parameters to be
tested— weight variation,
hardness, disintegration,
dissolution, and so
on— are chosen to suit the
test program. A represen-
tative sample of the prod-
I uct is then divided into an
[ appropriate number of
parameter batches and
the test proceeds, sequen-
tially, to build up a full
case file.

Electronics are also being
increasingly incorporated
into pharmaceutical
check-weighing machines.
One example is the Best B>
rotary unit for check-
weighing cylindrical con-
tainers such as miniature
aerosol cans. This is done
with the cans in a vertical
position and to very high
accuracies.

The machine satisfies the
need for both accuracy
and speed. It can handle
more than 100 containers
a minute with a zone of in-
decision of ±75 mg.

Starwheel

control

It incorporates its own
feed scroll and starwheel
to eliminate the need for
prior pitching of the prod-
uct. Cans fed at random
along a conveyor are
picked up by the scroll
and fed into the first star-
wheel, which controls the
flow and pitch for check-
weighing. After this has
taken place, a second
starwheel collects the con-
tainers from the weigh-
plate and feeds them
‘ back on the same con-
veyor, eliminating the
i need to side transfer.
Incorporated into the

i machine is a microproces-
sor system employing
pushbutton set-up facilities
| for full data acquisition,
j Between each weighing,
l the machine is automati-
I cally zeroed, and the con-
| sole incorporates a digital
| display for instant weight
information.

The mean weight control
causes the target weight
I of the check-weigher and
I associated control limits to
J adjust automatically, fol-
lowing a trend of increas-
ing or decreasing weight
of the packs being passed
over the weigh-cell.

I A running mean weight is
calculated over a block of
weights. The size of each
block is adjustable at the
keyboard and the block
size can be set at any
number between two and
64.

Audible alarm

j The four control limits (two
j upper and two lower)

| move with the moving '

J target weight. Should a
j pack be rejected, its
j weight is not included in
the calculation of the
| target weight.

Two fixed upper and lower
limits, representing the
limit to which the target
weight is permitted to
move, are set at the
I keyboard and this
eliminates the possibility
of the control system
becoming unstable. An
audible alarm sounds if
this should occur.
Manesty*®, which makes
tablet manufacturing
| equipment, has
| developed what is termed
a total system installation
option. This includes tablet
presses and coaling,
machines, powder filling
systems, capsule filling
machines, and blister
packing and cartonirg
machines.

An example of the
company's commitment to
constant development is
the new Rotapress Mk 4
high speed rotary tablet
press for large batch pro-
duction. It is capable of
outputs of up to 600 000
tablets an hour.

The double-sided press

i has a single pre-
! compression station on
each side. Both are strain
| gauged and linked to
| monitors in the main con-
j trol panel, allowing for the
] precise monitoring and
control of the pre-
compression force.

Motorized
\ adjuster

Mean tablet weight con-
trol is through a motorized
adjuster which automati-
cally compensates for any
; change that may occur.

I The compaction force and
the force on individual
tablets are also monitored,
I and the operator warned
j if a change exceeds the
pre-determined upper or
lower values.

I Also available is a pro-
j grammable tablet sam-
| pling device, which
provides a relatively
| simple method of sam-
pling individual tablets
from an individual station
| of tooling. Its also allows
tablets to be sampled
sequentially for monitoring
purposes. A group sample,
with one tablet from every
station of tooling, can be
taken to represent output

at any one time.

Another useful feature is
the detachable control
panel, which can be
remotely sited, and this is
particularly beneficial
when tabletting toxic or
dangerous products. All
power and sensor connec-
tions between the panel
and the machine are
housed in non-toxic flex-
ible conduits.

The obvious next step is
the application of the
latest video capabilities to
inspection processes, and
already some British
companies are well ad-
vanced in this direction.
But since pharmaceuticals
apply directly to human
beings, each step has to
be thoroughly and
meticulously evaluated
and tested before final in-
tegration into the overall
production system. (LPS)

1. Copley Instruments (Not-
tingham) Ltd, Private Road
No 7, Cot wick Industrial
Estate, Nottingham,

! England, NG4 2ER.

2. Best Inspection Ltd, 3
Fleming Road, Newbury,
Berkshire, England, RG13
2DE.

3. Manesty Machines Ltd,
Speke, Liverpool, England,
L24 9LQ.

gure 2. The Best .

I

heart beat monitor

The proposed circuit is based on the
fact that the degree of translucence
of parts of a mammal’s body de-
pends, among others, on the flow of
blood. Because the blood supply
pulsates at the frequency of the
heartbeat, this may be monitored in a
simple way without the need for an
electrical connection between the
mammal and the measuring
equipment.

In the proposed circuit, the flow of
blood through a finger is monitored.
To obviate errors caused by the pos-
ition of the finger, the receiver diode
is included in a loop.

The positive input (terminal 3) of IC1
is held at about 2.5 V. The gain of the
device is determined by the ratio
Rs:R<. Network Re-D2 ensures that the
circuit stabilizes rapidly. The ampli-
fied signal is rectified by IC?. Time
constants Re-Ca and Rr-Ca are chosen
such that the potential at pin 2 of IC2
has a sawtooth shape. The CA3130 in
the 1C 3 position functions as a trig-
ger. The output signal may, for in-
stance, be applied to the input port of
a computer.

If a computer is not available or
deemed necessary, the beat is made
audible by a piezo-electric buzzer
operated by gates N> and N2.

Circuit ICs provides a WAIT indi-
cation that shows when the circuit

2

54 elekio

has stabilized and is ready for use.
The programme is compiled as
follows: wait for a trailing edge, then
count until the next trailing edge ap-
pears. The count is converted into a
number per minute, and this is dis-
played on the monitor screen. How-
ever, the heart beat is not constant,
which is quite clear from listening to
the buzzer or observing the monitor
screen. It is, therefore, advisable to
calculate an average over, say, sixty
seconds. It is then possible to display
the instantaneous value, the average
value over 60 seconds, and the trend
(rise or fall).

Once the programme is known to
work satisfactorily, it becomes in-
teresting to display the actual signal
on the screen. If the computer used
has an analogue-to-digital converter,
the output signal of IC> may be used
for the display.

Fig. 2 shows a possible construction
of the heart beat monitor in an ABS
enclosure; the measurement may
simply be taken by gently pressing
one’s finger onto the photodiode. (B)

quartz-controlled ^ ;
tuning fork O

Musical instruments are tuned with
the aid of a signal source that
generates a signal at a frequency of
440 kHz. An electronic tuning fork is
superior to its mechanical counter-
part as far as dimensions, weight, and
stability with temperature are con-
cerned. The stability is obtained by
controlling the signal source by a
quartz oscillator. The output of the

oscillator is frequency-divided and
then amplified. The output may be
made audible by, for instance, a small
loudspeaker.

In the accompanying diagram, N.,
Na, and the quartz crystal form the
oscillator. The precise frequency,
measured at the 0 terminal of FFs
with a calibrated frequency meter, is
set with Ci. Divider Type 4059 is eas-

ily programmed to a different divisor.
A duty factor of 50 per cent is en-
sured by Fp2.

The transducer is shunted by a
100 nanofarad capacitor, because
most transducers have a much better
high- than low-frequency response,
which causes very shrill sounds.

(B)

battery guard

37

by P C M
Verhoosel

This protective circuit is readily in-
corporated in battery-powered
equipment which is typically in-
tended to operate for less than about
a minute; possible applications that
come to mind include IR remote con-
trol units, calculators, etc. Forgetting
to switch off such devices irrevo-
cably causes the built-in batteries to
be exhausted after a while, however
’Tow” the standby current.

The proposed battery guard auto-
matically switches off the supply cur-
rent to the circuit, either after about
one minute has lapsed after power-
on, or when the battery voltage has
fallen below the acceptable level for
normal operation.

Series regulator FET Ti can pass a
maximum current of 150 mA in the
circuit as shown, and it is advisable
to use a more powerful type than the
BS250 in case more than about
100 mA is expected to be consumed
by the equipment connected to the
output terminals. The Type BS250
FET drops about 0.5 V at a drain cur-
rent of 100 mA, and 0.8 V at 150 mA,
whence the foregoing consideration.
As Ti is a p-channel FET, it conducts
and powers the equipment when the
output of Schmitt trigger NAND gate
N3 is low, i.e. when both gate inputs
are high. This is so at power-up, since
C2 is still discharged and the inputs
of Ns are kept at logic low level via
Rfi. Consequently Ti is enabled and
causes C2 to be charged via Rs. After
about one minute (R-C time), the
voltage across R3 is low enough for
N3 to recognize a logic low level at
pin 1, thereby turning off Ti. N2 pro-
vides a hold function of this state,
since otherwise Ns might oscillate
owing to the slowly varying voltage
across Rs.

At power-on, the output of N* is
pulsed high by means of R-C network
Ri-Rj-Ci, whereby any residual
charge in C2 is cleared; the circuit
may, therefore, be switched on with
Ss-Si immediately after automatic
power down.

Battery voltage monitoring is ac-
complished by Ds, Rs, Re and Ns. The
latter’s trigger threshold level is. as
with all Schmitt trigger gates, in di-
rect proportion with the supply volt-
age level to the IC. As long as the
supply (i.e. battery) voltage is suf-
ficiently high, N4 will recognize a
logic low level at junction Rs-Rs-Ns.
However, if the battery voltage falls,
D3 keeps the input voltage to Ns at a

Semiconductors:

D.iDj - 1N4148
Di zener diode 6V8/.00 m

Miscellaneous:

Si = single pole mi

fixed level, causing the gate to
supply a logic low level to Ns, which
consequently turns off the series
regulator FET.

It should be noted that the exact
values of R3, C2, Rs and Rs may have
to be adapted to suit operation with
certain makes of the Type 4093. Also
note that the interval time of one mi-
nute may be changed to individual
requirements by suitable redimen-
sioning of timing elements R3-C2.
Adjustment of the battery guard is
carried out by temporarily exchang-
ing Rs and Rs with a 100 k preset to
determine the correct resistor values

for a given switch-off level. Current
consumption of the proposed circuit
is mainly determined by the zener di-
ode. which has been biased to pass
only 1 mA. After automatic power-
down, the guard circuit draws a (neg-
ligible) current of less than 1 jsA.

38

Computers and computer-driven
peripherals are notorious sources of
RF interference, and receiver jam-
ming may occur at frequencies well
above 100 MHz, even though the
computer is said to run at a mere
16 MHz or so. The cause of this prob-
lem lies in the very fast pulse rise
time of the switching and timing
signals internal and/or external to the
computer system and its peripherals,
which are often located well away
from one another (printer, modem,
mass storage).

Much of the interference originating
from long peripheral wiring systems
may be suppressed quite effectively
by inserting simple low-pass filters in
the signal lines for data and hand-
shaking. The proposed L-C filters are
composed of small (3 mm) ferrite
beads with 10 turns of 0.2 mm (36
SWG) enamelled copper wire, plus a
ceramic 1 nF capacitor: the coil
inductance is about 80 jiH, which
gives a cut-off frequency of about
60 kHz (120 Kbaud).

The filters are mounted on a small
piece of veroboard which may be cut
and filed to fit into a standard D-
connector housing. Other cut-off fre-
quencies may be defined by mod-

39

Unfortunately, we are all well aware
that the annual holiday season is an
anxious time for many people, since
they worry about leaving the home
unattended and therefore liable to be
visited by burglars and/or hooligans.
Right now is, therefore, an ideal time
to construct this circuit before you
leave your home and all of your
highly-valued property.

It goes without saying that simulating
one's presence in the home may be
accomplished by having some elec-
tronic or mechanical timer device
switch on a number of lights when it
grows dark, merely keeping them on
until a fixed time interval has lapsed.
The potential housebreaker, how-
ever, may soon detect the regular
pattern that occurs every evening,
encouraging him to embark on his
nefarious activities, since he realizes

58 elektor India Aug/Sapt 1986

filtered connector

ifying the small coils: inductance is
proportional to the square of the
number of turns, while constructors
boasting of good (near) eyesight and
lots of patience may endeavour to
use thin (0.05 mm) copper wire to run
through the beads. However, the L-C
ratio as given should not be mod-
ified.

In conclusion, it should be noted that
a filtered connector dimensioned for,
say, 10 kHz, should not be connected
to a high frequency (20 MHz) com-
puter output, since the excessively
high capacitive load may cause
damage to the line driver IC. (JB)

random lights controller

he is dealing with a harmless timer
rather than persons in the home.
This circuit, while also being a timer,
offers a better simulation of human
activity, since it automatically ar-
ranges for a number of lights to be
switched on and off in an apparently
random manner, which gives the
burglar the impression that there are
people at home. In actual fact, the
lights pattern is pseudo-random, but
16 possible configurations are bound
to ensure sufficient diversity to keep
your mind at ease and that of the at-
tentive burglar quite puzzled for at
least a few weeks.

And now for the operational prin-
ciples of this easy-to-build circuit.
The evening's specific lights con-
figuration is determined by the four-
bit logic code supplied by counter
IC: at the moment it becomes dark.

Since this never happens at precisely
the same time every evening, IC:
may be considered as a four-bit (1 of
16) random code "generator. When-
ever the LDR fails to detect the
presence of daylight, the output of
N: goes high, and D; charges C..
Meanwhile, Ni constantly applies
100 Hz pulses to the input of counter
IC>. When the voltage across C: and
R: has risen to a level, sufficiently
high to be recognized as a logic one
by the clock input of quad latch IC3,
the four-bit counter code is latched
and transferred to the Qe . . .0: out-
puts of IC3. In addition, N3
simultaneously enables IC4 to start
counting and dividing its on-chip
generated clock signal.

The latch (IC3) and counter (IC4) out-
puts are combined in AND gates
Ns. . ,Ni6. The oscillator parts to IC4

59

the circuit output; the former may be
wide enough to make a digital circuit
go completely haywire!

Further sources for possible inter-
ference are mainly separation jitter
and crosstalk caused by the colour
and/or sound carrier; the former
is not always suppressed during
synchronization intervals, while the
latter is, of course, continuously pres-
ent in the demodulated input signal,
which fact necessitates additional
filtering at the sound carrier fre-
quency.

The foregoing considerations have

41

The objective of this circuit is to
obtain a synthesizer-controlled
equivalent sound as produced by
such metal indefinite pitch per-
cussion instruments as cymbals,
gong, and anvil. Fig, 1 shows that the
generator comprises four indepen-
dently tuneable VCOs which supply
rectangular output signals to a com-
bination of XOR gates.

One of four identical KOV (keyboard
output voltage) driven VCOs is
shown in Fig. 2. The use of fast
opamp types ensures linear VCO
operation well up to 4 kHz. while FET
Ti improves upon the linearity of the
voltage-frequency curve relevant to
the combination of integrator and
comparator. With the VCO con-
structed four times over and connec-
ted as shown in Fig. 1, drive controls
Pi. . . Pa allow the user to set the out-
put sound as desired.

The outputs of buffer opamps
A-...A4 (ICi. Type TL084) should
measure 0 V offset with the KOV rail
grounded. If this can not be attained,
the IC will have to be exchanged
with a more stable type.

Linearity of each of the VCO circuits
is set with the preset at the drain of
the FET, Ps and Ti respectively in
Fig. 2. Use a scope to check whether
the rectangular VCO output signal
has a 50% duty factor; if not, adjust
the relevant preset.

As the four VCOs lack a linear to
exponential KOV converter at their
inputs, it is not possible to use the
present circuit with a keyboard of the
1 V per octave type. However, many
keyboards provide an exponential
KOV signal whose frequency
doubles with every octave and which
are, therefore, suitable for use with
this generator. (KD)

led to a circuit which effectively make the opamp toggle, providing a
removes the colour carrier from the negative (active-low) output signal,
demodulator output signal, before composed of the TV sync signals,
separating sync from video. Note that the circuit output signals

A combination of series (L2-C3) and can be made positive (active-high) by
parallel (L:-C2) tuned circuits is interchanging the opamp inputs,
peaked at the colour carrier fre- As to coils Li and La, these may be
quency (4 43 MHz), while D; serves replaced with fixed value (3.9 mH)
as a clamping device for a direct out- chokes, provided both C2 and Ct are
put voltage of about 0.7 V. The invert- consequently replaced with a 270 pF
ing (— ) input of opamp Type LM311 is capacitor and a 60 pF trimmer in
arranged to be at 0.75 V with respect parallel. If necessary, similar filters
to ground in order that signals caus- may be added to remove the sound
ing the non-inverting (+) input to be carrier. (D)

at a voltage lower than 0.75 V can

metal percussion
generator

1

call counter

-TX

d>^d> 4) d) d> d) d> d> f-M”) «i *

T-Jj

“ f ? f ? ?> r

A telephone has the unfortunate dis- 2
advantage that you have to be near it
to be able to make use of its com-
munication possibilities. If you have a
telephone answering unit, you know
at least who has called and how many
callers there were. If you cannot, or
do not want to, hire or buy such a
unit, the present low-budget one may
be of interest. Low-budget involves
the limitation, however, that the in-
coming calls are merely counted:
who has called, or what the message
was, can only be guessed. Moreover,
to avoid problems with British
Telecom (or whoever your PTT auth-
ority is), the unit is acoustically
coupled to the telephone. Such a
design must, of course, have ex-
cellent pulse suppression, since ex-
traneous sounds must not be inter-
preted as an incoming call. Finally,
the counter's current consumption

A1\A2',A3\ = J /,IC1' = LM324

should be (very) low to enable its op-
eration from a battery.

A small, inexpensive loudspeaker is
used as the detector, the output of
which is applied to window com-
parator A1-A2. In the absence of a
signal from the detector, the output at
interconnected pins 1 and 7 is logic
high. When the loudspeaker picks
up a sound from the telephone, the
output consists of negative-going
pulses. Monostable MMV, is trig-
gered by the leading edge of the first
pulse, and suppresses the pulse.
Only after a time lapse of 0.4 s does
MMVi enable a second monostable,
MMV?. If the sound is still being
detected by the loudspeaker, MMV.:
is then also triggered. This arrange-
ment ensures that noise pulses of
less than 0.4 s duration are effectively
suppressed. Since MMV; is retrig-

43

This circuit is not really a technical
novelty, but it has its practical uses. If,
for instance, it is desired to connect a
video recorder AND a computer per-

gerable, and its mono time is about 5
seconds, the intermittent ringing of
the telephone is converted into a
single pulse.

The decimal point of the display is
switched on via R:& and T. indicating
that the circuit is in a triggered state.
The remainder of the circuit is
straightforward: a decimal counter,
IC3. with switch-on reset (R7 and C3 ,
and a BCD 7-segment decoder. IC4.
In the quiescent state, the display is
not energized in order to keep the
current consumption low. Pressing
S2 will indicate how many telephone
calls there were. The circuit is reset
by briefly switching it off, and then
on again, but could be arranged by a
simple switch across C3. Current
consumption in the quiescent state
amounts to about 0.6 mA. so that a
reasonably long life may be ex-

manently to the SCART socket at the
back of a modem television set, it
will be found that that is impossible.
All that can be done is to connect

pected from the PP3 battery.

If the input sensitivity is poor, it may
be improved by lowering the value of
Ri and R2 to 10 Q. If this is still not
sufficient, a simple input amplifier as
shown in Fig*. 2 should be added.
The LM393 is then replaced by an
LM324, which has four suitable
opamps. One of these is then used as
input amplifier, and two of the re-
maining three as the window com-
parator. Diodes Di and D2 are
necessary in this case, because the
outputs of the LM324, in contrast to
those of the LM393, are not open-
collector. The value of R21 is estab-
lished by trial and error to find opti-
mum input sensitivity. Adding the
input amplifier has the small disad-
vantage of increasing the current
consumption to around 1 mA.

(TW)

either the video recorder or the com-
puter. But the proposed SCART
switch offers a solution .to this
problem.

SCART switch

«

62 elekti

The switch is constructed from a
small (110x60x30 mm) metal case, a
six-pole change-over switch, two
SCART sockets, one SCART plug,
and a length of screened coaxial
cable. Suitable holes should be pro-
vided in the case to receive the two
sockets, a cable outlet, and the
switch. The various components are
connected together as shown in th
accompanying diagram. The SCAR,
plug is connected at the free end of
the cable, which should not be
longer than 1 metre. The connections
to the sockets and plug are also
identified in the table.

The completed switch should find a
home beside, under, or on top of the
TV set: the SCART plug is inserted
into the SCART socket at the back of
the set. The two SCART sockets on
the case are then used to receive the
computer and video recorder
respectively. From then on, it is a
simple matter of switching between
recorder and computer! (Sv)

to

12

13

15

16

20

21

Audio output (right-hand)

Audio output (right-hand)

or channel 2

Audio output (left-hand)

Audio earth
Blue earth

Audio input (left-hand)
Blue component

Switching voltage:
0 - TV reception

Green earth
Not used
Green component

0.5 V for input
impedances <10 kQ
0.5 V for output
impedances <1 kQ

impedances i 10 kQ

signal level 0.7 V,
load impedance 75 Q;
superimposed direct
voltage - 0 .2 V

0 0...2V

1 9.5. 12 V

Identical to 7

Not used
Red component
Blanking signal
1 blanking

Video input

Housing screen and/or

earth

Identical to 7
0 0...0.4 V

Load resistance 75 Q

Difference between
peak white level and

voltage = 0...2V
Synchronization signal
only = 0.3 Vpp

Identical to 19
Connected to chassis

fast voltage-controlled A A
pulse generator 'I'l

Certain measuring and process con-
trol applications require pulse
generator sections which are to
operate over a large frequency range
and must, therefore, produce a signal
with very low pulse width. It is for
this reason that the proposed circuit
uses high-speed complementary
MOS (HCMOS) type gates; the proto-
type typically produced an output
pulse width of 20 ns over the fre-
quency range of several hundred
hertz to 25 MHz.

The combination IC 1 -T 1 is a voltage-
controlled current source which
discharges Cz. The fast charging of
this capacitor is effected through the
voltage at the output of Schmitt trig-
ger N 1 -R 3 -D 1 . The lower frequency
limit of the proposed circuit mainly
depends on the offset voltage of
opamp ICi. In order to enable set-
ting the lower frequency limit, Ti
must be arranged so as not to draw

GED— r

any current at an input voltage of 0V;
to this end, offset preset P. should
be correctly adjusted. Finally, the
output pulse width may be widened
by increasing the capacitance of C 2 ;

this will not alter the attainable
sweep range.

Literature: E Abbel, Electronic
Design 18 (1984), pp 270-271. j B

45

guitar fuzz unit

The fuzzbox, fuzzer, tube screamer,
or whatever other name there may
exist for the controlled guitar sound d
distortion unit, is a well-known item 11
in the electrophonic field, which is of
common interest to both musicians
and electronics enthusiasts.

The majority of fuzz units are simply
opamp configurations with some
form of maximum input level control,
which determines the degree of
overdrive by the guitar input signal,
and, consequently, the amount of
audible distortion, generally referred
to as the object "sound" the player
has in mind as his very own musical
visiting card.

This is probably one of the few fuzz
units to feature controllable sym-
metrical clipping facilities, which
means that the limit for distortion-free
amplification may be separately
defined for both the negative and
positive portions of the input
smewave(s), the peaks of which may
be clipped by means of shunt tran-
sistors T> and T? respectively, each
with its own clipping level control
potentiometer (Pi; P2). The tran-
sistors, when driven, pass the signal
from input opamp IC> to the positive
supply or to the ground rail, before
buffer IC2 can pass the "fuzzy" guitar
sound to the connected amplifier.
Preset P3 determines the minimum
gain of the fuzz unit; the desired level
may be set with P4 turned to its
minimum resistance position. Next,
Pa is adjusted to suit the maximum in-
put level that can be expected from
the guitar. P3 and Pa may then be
alternately adjusted to hit the correct
compromise between these two sig-
nal levels.

46

' WMI i

"Watf j...

'Em

feB

y- o-u-o o_o Os oy
<o 4 — k> , L -“.

isfir^o t

oT^lo a< j°

Ofi.fi I •

Finally, note the three-pole simultaneously switching it off I

changeover switch which allows preserve battery power.

easy bypassing of the fuzzer while j

two-gate bistable

Probably unequalled as to its
simplicity given the digital function,
this circuit may serve as a single-
button on/off control for incorpor-
ation in a wide variety of electronic
designs. The operation of the pro-
posed bistable is best understood if
it assumed that the input of Schmitt-
trigger inverter Ni is at logic high
level; the output of N2 will therefore
be high as well. It is seen that the ca-

21 , depression of the button pulls the in-

i' ciT ' M N 1 ...H2 = i/3 ici = 40106 put of N ' ,0 lo 9' c low level, causing
.^1 ! ~ the bistable to toggle; the capacitor

0 is charged via the 1 M resistor, and

t — I 0 the circuit will change state again at

the next switch action. The indicated

t . I — 1 resistor values have been found to of-

***”'' fer optimum stability of the bistable,

while the use of Schmitt-trigger
pacitor is discharged because of the CMOS inverters is essential to the
low output level of Ni. Therefore, correct operation. HS

mains-based remote ^*7
controller * X

This combination of transmitter and
receiver is based upon the use of the
mains network in the home for re-
mote control of mains-operated dom-
estic appliances.

Figure 1 shows the transmitter, which
merely superimposes a 36 kHz signal
on the 50 Hz mains voltage if Si is op-
erated. It is noted that IC. is fed
direct off the mains voltage by means
of a rectifier circuit composed of D; ,
D2, zener diodes Da, D«, and
smoothing capacitor C;; the pro-
posed configuration is to supply
+ 20.V with respect to the mains
neutral (0) line. The 36 kHz output
signal of the opamp is fed to the
mains by means of coupling capaci-
tor C3. R2 is a bleeder resistor to dis-
charge Ci and C2 after the circuit
has been unplugged from the mains
outlet.

The receiver, shown in Fig. 2, is fed
with an inexpensive door bell trans-
former, although any other type sup-
plying 6 to 8 V AC at about 300 mA
should do just as well.

Apart from being used to power Tri,
the mains voltage with the 36 kHz
carrier is filtered by parallel tuned
circuit Iii-Ce to detect the presence
of the s'uperimposed 36 kHz carrier,
which is passed to amplifier IC; via
R7. Subsequent rectification by Da
enables the relay driver circuit com-
posed of Ti and Ts to energize Rei.
Preset Pi is adjusted to find the right

compromise between receiver sensi- with a small hole drilled into it for Si .
tivity and noise immunity. Ru should The receiver ABS enclosure is likely
be dimensioned to suit the relay coil to be of larger size if a mains socket
current. is incorporated for easy connection

As to the construction of the receiver to the appliance to be controlled,
and transmitter, it should be made The contact rating of Rei should be
quite clear that the presence of the duly observed in case heavy loads,
mains voltage necessitates the use such as a coffee machine (4 A), are to
of sound and safe con-ABS enclos- be switched. /yy\

ures to prevent accidental contact
with the live wires Do not take any
risk in this respect, neither while ex-
perimenting with the circuits as
shown nor while setting up and
testing.

The transmitter, then, is readily fitted
in a salvaged mains adaptor case

65

subwoofer filter

The filter described here is intended
primarily for experimenting with a
(central) subwoofer (see Active Sub-
woofer March 1986. p.5-18. As the
human ear cannot sense direction in
a standing wave, directional sensi-
tivity is generally poor at low fre-
quencies, so that is would seem
superfluous to use a stereo set-up
below about 200 Hz. Therefore, the
low frequencies can be concen-
trated on one good bass enclosure,
which, of course, keeps the cost of
the overall system down. The satellite
loudspeakers, (see may 1986.

p. 5-46 will then have to cope with
the higher frequencies only.

The requisite cross-over network de-
scribed here is based on 24 dB/
octave Bessel filters: the cross-over
frequency lies around 200 Hz. With
reference to the circuit diagram. Ai
and A2 buffer the left-hand and
right-hand signals respectively. The
high-pass filters for the two channels
are formed by A3-A4 and A9-A10 re-
spectively.

At the same time, the two channels
are combined in As, and the
resulting signal is passed through

low-pass filter Ae-Pn. The amplifi-
cation of As can be varied with Pi, so
that the level of the low-frequency
signal can be matched to that of the
high-frequency signals. Note that the
component values given in parenth-
eses are the calculated values, wich
perfectionists may try to approach.
The power supply is a symmetrical
design with short-circuit protection,
which also prevents annoying
"plops" at on and off switching.

If a different cross-over frequency is
required, refer to Active Cross-over
Network in the October 1984

50 !

Before any analogue voltage can be
measured and subsequently pro-
cessed by a computer, a converter
device with the necessary precision
is required to provide the computer
with the digital n-bit equivalent of the
voltage as applied to the DAC circuit.
Obviously, the higher n, the more
steps involved in the conversion pro-
cess, but also the higher the accu-
racy that can be obtained.

This 8-bit ADC circuit works with
very few parts; yet it is versatile, fast,
and sufficiently accurate for most
purposes. The maximum input volt-
age to the circuit is arranged at 5V, as
determined by the resistor network
connected to the Am terminal of the
Type ZN427 ADC chip. Given this up-
per limit for Vm, the conversion ac-
curacy equals 5V/(2*-l)= 19.6mV/step.
Other input voltage levels may be ac-
commodated by appropriate re-
dimensioning of the input voltage
divider.

Since the proposed ADC chip
features an analogue-to-digital con-
version time of only KVs (typical
value), alternating voltages may be
measured (digitalized) and pro-
cessed under machine language
control; just as with the above DAC
circuit, BASIC is usually not very
suitable for this purpose, and its use
is restricted to applications where
timing requirements are less
stringent. It will be understood that
fast and therefore smooth computer
response to, say, joystick movement
is only feasible if the ADC reading
subroutine is written in machine
code.

A low SOC (start of conversion) pulse
at the WR input of the chip triggers
the internal voltage conversion pro-

51

This simple-looking circuit enables
the arbitrary programming of seven
outputs in a series of not more than
2048 (2») steps. The step length may
be set as required. The time base is
derived from the mains voltage. Tran-
sistor Ti produces a square wave
from the mains voltage applied to its
base. This square-wave voltage is div-

8-bit ADC

cess and the BUSY output is activated computer CPU.

(i.e. pulled low); this, in tum, enables Calibration of the present circuit is
Schmitt trigger gate N. to generate straightforward, since this merely in-
the ADC clock frequency of about volves setting two presets. First, a
900kHz. On completion o f the clock- simple test loop may be written in
controlled conversion, BUSY goes machine language; next, adjust Pi
high, and the CPU may read the 8-bit (offset) for a computer reading of 0
value contained in the ADC latch by with no input voltage applied to the
activating the read line. Note that the circuit; Pz is set to give a reading of
SOC and read signals must be 255 (FFhex) with the maximum input
decoded with suitable circuitry as re- voltage at Vm, i.e. 5V. Finally, test the
quired by the type of computer or ADC linearity by applying 2.5V from
CPU. Provision has been made in the a sufficiently accurate source; the
ADC circuit to select either the BUSY computer should read 128 (80hex).
or BUSY signal in order to Bag the HS

conversion condition to the host

versatile timer

ided by 10 in ICi, so that the fre- would enable the use of a Type 2732

quency of the signal at the clock (4096 steps), but, on practical and

input of IC is 5 Hz. Circuit IC 2 financial grounds, a Type 2716 is

serves as address counter for the used here since 2048 steps are nor-

Type 2716 EPROM. This means that mally quite sufficient.

IC 2 . after a reset, counts upwards The outputs of the EPROM are buf-
from 0 and runs over the successive fered by a Darlington array, ICs, so
addresses of the EPROM. that seven switch outputs are

Circuit IC 2 has twelve outputs which available with a sink capacity of

©*

500 mA at a maximum voltage of 50 V.
The eighth output contains the stop-
bit that provides the facility of stop-
ping the programme if this is shorter
than 2Q48 steps.

The start-stop circuit is based on
bistable N3-N4. When the supply is
switched on, IC2 ensures that the
bistable resets from the stop state.
This means that both divider IC; and
counter IC2 are in position ’’zero".
The first address in the EPROM must,
therefore, have a neutral content,
because it is addressed in the stop
state and thus appears at the output.

The bistable is set. and both resets
cleared, when the start button is
pressed. Circuit ICi then com-
mences to divide, and IC2 starts to
count. With the present time base,
the programmed content of suc-
cessive addresses will appear at the
output of the buffers at 0.2 s intervals.
Counting continues until a stop-bit
appears at pin D7 of the EPROM, or
stop button Si is pressed. If re-
quired. a HOLD function may be ob-
tained by connecting a switch across
capacitor Ci, which enables the time
base to be switched off.

Switching on a specific output a. . .g
merely requires the corresponding
bit position in the EPROM to be left
unprogrammed (logic high); pro-
gramming a 0 disables the relevant
output. The stop-bit operates with
negative logic: a 0 therefore causes a
stop.

Finally, the time base may be
adapted for the setting of the re-
quired step frequency and accuracy.

(Sv)

analogue & digital

Leafing through some electronics
magazines published over the past
few years, it is surprising how fast
and vigorous digital techniques have
come to the fore. 1 Even audio, until
recently virtually untouched, is now
becoming digitalized at a rapid pace.
What are the consequences of these
changes to us engineers, tech-
nicians, and hobbyists alike?

As long as a circuit is totally analogue
or totally digital, all is weil. But as
soon as these two techniques

become mixed strange things some-
times happen. Well-known examples
are analogue-to-digital converters
that will not give a stable reading: the
last few digits do not match and it
appears as if there is a certain
regularity in the deviations. Another
example is an otherwise good ampli-
fier that generates whistles in perfect
rhythm with the digital clock oscil-
lator. And soon...

Often, these flaws can be traced to
faulty earth connections, i.e. the zero

supply line, or common ground.
Because of that, here are a few tips
that may prevent these annoying
defects.

■ Avoid earth loops.

■ Keep the analogue and the digital
earths separated.

■ Interconnect the analogue and
digital earths at one point only, for

instance, at the analogue-to-digital
converter, but NOT at the power
supply.

■ If there are more earths, connect

these to the same common point.
■ At high frequencies, the im-
pedances of earth lines are not
negligible: short, thick wires should,
therefore, be used.

An example that gives good results is
shown in the accompanying draw-
ing. All sensitive parts of the circuits
have been isolated from those parts
that carry (large) earth currents. Most
converters have, therefore, two earth
terminals, or an earth terminal and a
differential input (which is the same
thing).

In audio amplifiers most of us do not
dream of wiring the power supply to
the output amplifier via the preampli-
fier. In mixed analogue-digital cir-
cuits, such considerations are not so
self-evident, although the principle is
the same.

Note that in the accompanying draw-
ing the system needs several elec-
trically isolated power supplies: that
is unfortunately the price often to be
paid for new techniques. (W)

53

Sooner or later, most types of
frequently used multi-way rotary
switches develop contact resistance
instability or other malfunctions,
either caused by internal oxidation or
wear and tear of the rotary mech-
anism. Broadly speaking, the same
goes for multi-contact relays. It is,
therefore, hardly surprising to en-
counter the electronic, free-of-wear
equivalents of the above devices; n-
way electronic switches and solid-
state relays are at present available in
a wide variety of contact arrange-
ments.

The circuit diagram shows the elec-
tronic counterpart of a 16-way rotary
switch whose pole is connected to
earth. Two push buttons have been
provided to enable the switch to be
"turned” clockwise (up) or anticlock-
wise (down).

Debouncing bistables N5-N6 and
Nj-Ns supply a stable low logic level
to monostables N1-N2 and N3-4
respectively in order that these can
output approximately 3.5 j<s long
pulses to the relevant input of
up/down counter ICi. The rising
edges of the up/down pulse(s) cause
this IC to generate the correspond-
ing binary code at its Qa. . Qd out-
puts, which are connected direct to
the Di. . .D4 inputs of latching 4-to-16
decoder IC2 which, in turn, activates

electronic rotary switch

the next lower or higher output correct active-low outputs to block
Se. . .Sis if the relevant control button gates N2 and N4 when the desired
was activated. Provision has been stop positions are reached,
made to ’’stop" the switch if this Finally, push button S3 resets the
reaches its first or sixteenth position, counter IC and consequently causes
which conditions cause the down or IC2 to activate its Se ouput, which is
up monostable respectively to be also the default switch position at
disabled. Other switch configur- power-on.

ations may be defined by using the CD

selex is

Measuring

Techniques

| Chapter 1

The measured values of any
parameters in a circuit are
the most important for an
Electronics Engineer or
Hobbyist. They give the
most reliable information
about what is going on
inside a working circuit.

Even in case of a circuit
| which has stopped
functioning, the measured
values are of great
importance because they
tell us what is wrong with

To understand more about
I the measuring techniques
used in Electronics, we are
starting a new series on
Measuring Techniques",
from this issue of Elektor
i India. This and the
subsequent articles in this
series will explain how the
measurements are carried

1 instruments.

The expectation that a
! measuring device should
work as accurately as
possible is obvious
However the measuring
instruments used by
Electronics Engineers and
Hobbyists show different
types of tolerance ranges
An average multimeter has
a tolerance value of a few
percent. This leads to an
expectation that the
measured values would
deviate by just a few
percent from those
mentioned in the circuit
diagrams, or from the
theoretically calculated
values. Especially the
beginner is very much
surprised to see the
measured values deviate
considerably from the

see that the circuit still
functions correctly

deviations is mainly the
tcierance on the
components used in the

circuit Resistors can have
tolerances upto ± 10%.
Condensors can have even
greater tolerances. The
current gain of a transistor
can vary by 100% or more,
and semiconductor
threshold voltages can
deviate from the expected
value by 10 to 15%
depending on the current
and temperature.

Measuring instruments
available in the market can
be divided into two groups
Measuring instruments (or
meters) with a fixed range
for mounting permanently
on the panels of various
equipments, and measuring
instruments with multiple
ranges (like multimeters)

Meters with a fixed
range

These type of meters have
built in shunts or
compensating resistors and
their scales are calibrated
with a fixed range. The
important technical
specification about these
meters is printed in a corner
of the scale provided with
symbols.

I Most frequently

I encountered symbols are

illustrated and explained in
! Table 1 . The last entry
"Class" indicates the
accuracy of the meter The
figure shown indicates the
maximum deviation from
] the actual value as a
percentage of the full scale
reading For example, a 10-
V moving coil meter of class
2.5 will give a maximum
reading error of i 2 5% of
the full scale value. That
works out to an accuracy of
* 0.25 V. This deviation can
occur at any measured
| value. A reading of 1 V on
I the scale means that the
j actual voltage is anywhere
| between 0.75 to 1 25
Volts. Thus the accuracy as
I a percentage of the actual
| reading at this point is t
25V To stretch the example
| a little further we can see
[ that if we are trying to

I ! measure a voltage of the
order of 0.25 Volts with
such a meter, the deviation
would be i 100%.

High class laboratory grade
meters generally have
accuracies of the order of
±0.1% to ± 0.5%. but they
are very expensive. On the
other hand inexpensive
meters can have accuracies

| of the order of £ 1 % to i
5%.

! Most of the commonly used
I meters are of the Moving
I Coil' type. 'Moving Iron' type
I meters are occasionally
| used, as they are suitable
for both AC/DC, however
[ their sensitivity is less

0

Moving Coil Type
Moving Iron Type

— Direct Current (DC
^ Alternating Current (AC)
^ AC/DC
_L Rated position Vertical
i | — | Rated position

o

☆

■£r

Rated position Angular
(for example 60 °)

Zero Adjust
Test Votlage 500 V
Test Voltage 2 KV

0.5 Class (for example 0.5)

selex

. Meters with multiple
ranges (Multimeters):

\ Multimeters contain a
moving coil meter, but they

| shunts and compensating
resistors which can be
I switched. A rectifier and a
I battery is also included in
the housing of the
[ multimeter. By switching
j the ranges or by replugging
the measuring leads into
different sockets, one can
, use the multimeter as a
Voltmeter. Ammeter or an
| Ohmmeter. The rectifier
allows for measurement of
AC voltages and the battery
is required for measurement
| of resistances. The dial of a
multimeter is marked with
| various scales for all the
different ranges, however it
I is marked only with
abbreviations due to lack of
space on the dial. These
| abbreviations are explained
| in Table 2

I Operation of multimeters is
quite simple but to avoid
| damaging the meter or
J getting incorrect readings, it
| is always essential that the
instructions given in the
operating manual should
be followed.

Position of meter
Any meter must be used
only in its rated position -
i.e. Vertical, Horizontal or
Angular. In any other

position the measurements
j may give incorrect readings
| This will happen even if the
meter is set to zero with the
| zero adjust screw before
starting the measurements,
j The spring is so delicate
that even the imbalance
! caused by the incorrect
I positioning of the meter
I gives rise to measuring

I Polarity

| When measuring the DC
| Voltages or Currents, the
polarity must always be
observed. The black lead is
always used for minus and
I the red lead is always used
; for plus pole. On a
| multimeter, the minus pole
is marked as COM'
meaning the COMMON
I lead. This is so because DC
| voltages are mostly
measured with reference to
the common ground line.

In case the polarity is
unknown, the safest way to
confirm the polarity is to
switch the multimeter in the
highest voltage (or current)
range and touch the leads
I to the test points only
| momentarily. If the needle
i deflects in the reverse
direction, remove the lead
I from the test point quickly,
j the leads must be
_ interchanged to get the
I correct polarity.

- | Even in case the polarity is
1 j known but the voltage or

| current value is unknown it
“ I should be first measured on
the highest range so as to
si | get a rough idea about the
! order of magnitude Then a
s I suitable range should be
| selected to read the actual
j value. Generally the range
‘ ' ' for reading should be so
s [ selected as to get the needle
in the upper one third of the
:s scale. The measuring

accuracy has lesser effect in
| the upper portion of the

— j scale. Another important
,s point to note is that the
i range switch should never

| be changed during
j measurement.

' Direct Current
| In some cases, the polarity
| of the DC current being
j measured is doubtful. A

2

p©-0-©-]

_©J

T Storage

EUmVnator B *"<"Y

5

3 Mutlim

j

simple example is shown in I
figure 2. A storage battery
is being charged by a
battery eliminator.

In this case the plus pole of
the eliminator is connected
to the plus pole of the
storage battery and it may
seem doubtful whether the
plus pole of the meter
should be connected to the
plus pole of the battery
eliminator or the storage
battery.

In such cases, the plus pole
of the meter should be
connected in tne direction
of the plus pole of the
voltage s&jrce in the circuit,
i In the above example the
battery eliminator is the
voltage source and hence
we must connect the plus
pole of the meter to the plus
pole of the battery
eliminator.

Resistance

Measurements

Before carrying out
resistance measurements, it
may be necessary to do the
zero adjustment. For this,
the test leads are shorted
1 and the setting knob is
: turned so as to get the
needle to real 0 If This zero
i setting for resistance. is*

necessary for every range,
and must be done when a
new range is selected. If it
becomes difficult to do the
0 11 adjustment even with
the knob in extreme
position, it means that the
battery voltage is very low.
The battery must be
replaced with a new one.
During the resistance
measurements, the polarity
of the test leads gets
inverted in case of almost
all multimeters. This is due
to the resistance measuring
circuit of the multimeter,
which has the internal
battery as the voltage
source. The meter measures
the current drawn by the
test resistance from the
internal battery. As the
multimeter itself is acting as
tha voltage source in this
case, the polarity of the test
leads naturally gets
reversed. The basic circuit
of a multimeter for
resistance measurement is
shown in figure 3.

The polarity reversal of test

resistance measurement but
it must be observed
correctly while testing
semiconductor devices. The
'COM' terminal acts as the
plus pole in this case.

selex

Loudspeakers

No audio system can work
without Loudspeakers. You
will find at least one in
every radio set and many of
them in a stereo system
How does this funnel
shaped structure of paper
and metal convert electrical
| oscillations into sound ?

A loudspeaker is a sound-
transformer. Most
loudspeakers are electro
dynamic. The electrical
current supplied by the
I amplifier to the loudspeaker
| flows through a coil. The
! magnetic field set up by this
current interacts with the
magnetic field of the
I permanent magnet inside

i the loudspeaker and sets
the coil in motion. This is
due to the fact that the
current is oscillating at the
rate of several cycles per
second. The coil is
connected to the membrane
in such a way that the coil
I sets the membrane also in
| motion. The sizes and
shapes of these
j loudspeakers can be quite
| different, as shown in figure
I 1 Figure 2 shows the basic
i construction of a dynamic
j loudspeaker. The
! construction is quite simple
1 as can be seen from the
I illustration,
j The Most important

I loudspeaker is the
permanent magnet. The
magnet is circular in shape I
and like every other magnet. |
has a north and a south
pole I

Loudspeakers with other
j shapes of magnets are also
j available but they are very

The so-called pole plates
are mounted on both sides
of the magnet to improve
| the flux distribution of the !

magnetic field. The upper I

; pole plate has a circular
J opening where as the lower j
I pole plate has cylindrical j

! core There is a very small
| airgap between the core

and the walls. The smaller j
the air gap, the stronger is '
the magnetic field in the air I
gap. Figure 3 shows the
schematic representation of I
the path of the lines .
magnetic field. The nu, !

and the pole plates are
drawn in section so as to
make the inside view as
clear as possible.

A moving coil wound on a
paper tube is placed in the
airgap and rigidity attached j
to the conical membrane.

The membrane is funnel
shaped and made of thin I

paper or plastic. The upper |
edge of the membrane is .
fixed to the metallic frame of I

selex

Ihe loudspeaker through a
rubber or special fabric
suspension. This prevents
the horizontal movement of
the core and provides the
necessary elasticity for
vertical movement. The
] lower end of the core is
held in place concentrically
by a web. which ensures that
the oscillating coil never
touches the magnet core or
the pole plates. The web
also keeps the membrane in
| a stable position when no
signal is fed to the
loudspeaker The alignment
of the coil is very important.

| because the airgap available
is just about a few tenths of
a millimeter wide.

The terminals of the
oscillating coil are brought
| out to two soldering lugs,
j The amplifier output is
connected to these two
j lugs. Whenever a voltage is
I applied across these two
lugs, the membrane is set

1 current flowing through the

selex

current. The amplifier must
be capable of supplying
sufficient current to produce
the movement of the
membrance.

The movement of the coil in
the airgap can be explained
as follows:

An electrical conductor
carrying a current produces
a magnetic field around
itself. In case of the loud
speaker coil the signal
current flowing through it is
responsible for producing a
magnetic field. This field
interacts with the field of
the permanent magnet. As
the opposite poles of two
magnets attract each other
and similar poles repel each

coil. If the current is AC.
then a sound is produced
depending on the frequency
of the AC current.

If the current is DC.
then the membrane moves
forward or backward only
once depending on the
direction of DC current and
produces just a click.lf you
can get hold of an old
loudspeaker which is still
working, this experiment
can be easily carried out
using a 1.5V battery cell. It
can be observed that the
membrane moves forward
or backward depending on
how the cell is connected to
the terminals. The
movement increases with
increase in the supplied

amplitude of the sound
produced. The frequency of
oscillations of the cone is
decided by the frequency of
the signal current, and this
also decides the frequency
of the sound produced. The
movement of the membrane
is transmitted to the
surrounding air and thus
the output signal coming
from the amplifier is finally
converted into sound waves
by the loudspeaker.

Most of the loudspeaker
boxes used in stereo
'■/stems contain many
i.-'diviriual loudspeakers of
different shapes and sizes.

my single loudspeaker can
not reproduce all the
frequencies equally well.

The loudspeakers with large
diameters and large, heavy
membranes can reproduce
low frequencies much
better. A light membrane of
| small size is much more
mobile and can move
I rapidly with high frequency
I sounds.

I In addition to the
loudspeaker principle
I described above, there are a
I few other types also
j available, such as the
Electrostatic loudspeaker,
Strip loudspeakers and
Piezo loudspeakers. The
most well known and most
frequently used type
however is the
Electrodynamic loudspeaker.

An oscillating current
through the coil produces
an oscillating magnetic field

oscillating motion. Along
with the coil, the membrane
is also set in oscillating
motion. However the
suspension of the
membrane at the top

deflection. The strength of
current flowing through the
coil decides the deflection of

other, the coil is also
repelled or attracted by the
permanent magnet
depending on the polarity of
the magnetic field produced

The

Digilex-PCB

is now available!

Price:

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

Send full amount by DD/MO/PO.
Available from:

precious

ELECTRONICS CORPORATION

Journal Division

1 1 , Shamrao Vrthal Marg (Kiln Lane)

Off Lamington Rood. Bombay-400 007.

selex

Magnetic Flux Photograms

Magnetic force does not
[ need a conductor like the
I Electric force needs in form
of cables and wires
Magnetic force can act
independently through
| space, without a visible
j connection. However, we
can make the lines of
magnetic force visible using
an interesting technique.

j One or more magnets are
J kept on a flat surface and a
paper is placed on top. Now

scattered on the paper.
These iron filings align
j themselves with the
I magnetic lines of force of
the magnets lying below the
| paper, thus making the
magnetic lines of force
visible. While scattering the
filings on paper they should
| be rubbed between the
I fingers like a pinch of salt,
so that they fall uniformly
on the paper.

All you need to do this
simple experiment is a few
magnets, a paper and some
iron filings. If you are

planning to buy the iron
filings from a shop, you will
be very much disappointed

You will have to get them
either from a small
workshop, or file a bolt from
the junk box into iron filings
yourself, and collect the
filings with a magnet.

The patterns generated by

fascinating as can be seen [
from the magnetic flux
photograms reproduced
here. These are nothing but
the images of the magnetic
flux lines captured
photographically using iron I
filings All we have to do is |
replace the ordinary paper

from the previous
experiment by a
photosensitive paper and
expose it using an enlarger
or an ordinary white light
bulb and then develop the
paper and fix the image
Obviously, this has to be
done in a darkroom. The
shadows of the iron filings
can be seen in the
photogram as white spots,
and the remaining area
which gets exposed to the
light, becomes totally black
after developing the paper.
If an enlarger is not
available, a 25 W bulb or a
table lamp can be used for
exposing the paper.
However, if the enlarger is
not used, the result will not
be very sharp.

Amateur photographers will
need no more instructions,
and can go ahead on their
own without reading
further. But for those who
are new to this subject and
would like to try their hand
at photography on this
occasion, here are few
guide-lines.

7 i

%

%

selex

How does a
transformer work ?

How does a transformer work ?"

"A transformer increases or decreases the AC voltage

"I know that, we give the mains voltage at one end of
| the transformer and a lower voltage comes out at the
other end of the transformer".

I 'The one end where we give the voltage is the input and
the other end is the output. In technical language it is
called primary and secondary. The input terminals are
the primary terminals and the output terminals are the
secondary terminals."

"But I don't still understand how the current comes from
the primary to the secondary? I have checked the
connections of a bell transformer using a multimeter
There is no continuity between the primary and the
secondary side"

"That is right. The transformer is intentionally so
I designed that there is no connection between the
primary and secondary We don't want the 230 V mains
voltage to be connected to the bell head"

“But how will the current flow to the bell without a
connection ?"

"This is done by magnetisml The input terminals lead to
the primary winding and the output terminals come out
{ from the secondary winding. When a current flows
j through a winding, which is nothing but a coil, it
produces a magnetic field."

"Quite clear, that means a transformer is just an
electromagnet I"

"No, not just that, because the transformer winding is
| connected to an AC Voltage, and an AC current
J continuously changes direction"

"That means the magnetism in the transformer also
| 'Continuously changes its direction."

| "Yes, the magnetic field is also an alternating field"

"But what mak6S a current come out of the secondary

"The alternating field produced by the current flowing in
the primary coil also passes through the secondary coil
and the secondary coil transforms this magnetic field
again into a voltage across itself."

"Oh, I see, then this is the reason why we call it a
transformer I"

And as the magnetic field needs no wn es for conduction,
there is no electrical connection between the primary
and the secondary "

"What is this thick black line between the two coils of
the transformer symbol ?""

"It is represents the Iron Core of the transformer"

"Is this same as the block of plates in the transformer ?"
"Yes. these plates are the iron laminations built into a
bundle. The magnetism travels, better through iron than
through air. Very little magnetic energy is lost when
passing through iron."

"And that is why we get a lower voltage at the
secondary, is it ?"

"No. it has nothing to do with the loss of magnetic
energy. The value of the voltage at the secondary is
decided by the ratio of the number of turns in the
primary and secondary windings"

"I still don't understand how the voltage can be
increased by the transformer"

"That is obtained by winding the secondary coil to have
more number of turns compared to the primary. This
however, reduces the current output at the secondary "

"Because the total energy remains same at the input
and output, so a higher voltage must be accompanied by
a lower current and a lower voltage with higher

"Just like the gears of a carl In the first gear, one can
drive at a lower speed but with greater force."

"Yes, that is very interesting comparison"

"Tell me. is there also a reverse gear in electronics ?'"

up/down clock generator

Various designs of clock generators
have appeared in previous Summer
Circuits ' issues of Elektor Elec-
tronics, and this tradition is kept up
with the present design which, un-
like the other circuits, outputs an
up/down indication as well as a rec-
tangular signal over a wide fre-
quency range; 0 Hz to several JcHz.
The output signal and the U/D indi-
cation are both controlled by a single
potentiometer. If this is set to the
centre of its travel, nothing happens;
turning the potentiometer in the
clockwise direction causes the U/D
output to be at logic high level, and
the frequency of the output signal
rises with turning Pi further in this
direction. The same goes for turning
it anti-clockwise, U/D being at low
logic level.

The basic operation of the circuit is
as follows. Operational amplifiers A:
and A2 together constitute a
sawtooth/square wave generator.
The falling edge of the sawtooth
voltage has a fixed duration of about
200 ns, as defined by the current
through D4. The rising edge time,
however, depends on the voltage at
the wiper of Pi. The wiper of P2 is ar-
ranged to be at a slightly higher
voltage than that at the wiper of Pi,
when this is set to the centre of its
travel. The STOP LED will light in this

condition. If Pi is turned in either
direction, the voltage across Ri rises
and causes a low current to flow
through R2. This current, and
therefore the output frequency, is
proportional to the position of the
wiper of Pi, but this only goes for a
limited frequency range. If the
voltage across R2 exceeds about
0.6 V, Di conducts and connects R3
in parallel to R2. D2 and D3 do the
same for R4 at about 1.2 V; this
method causes the oscillator fre-
quency to be an exponential function
of the voltage, set with P ; the ar-
rangement ensures a considerable
output frequency range for the oscil-
lator A1-A2.

Together with one or more universal
counter modules the proposed

2

O

clock generator may offer a neat re-
placement of the well-known BCD
coded thumbwheel switches; the
potentiometer-set value is present at
the O' • • Q‘ outputs of IC2, as well as
visible on the seven-segment display.
The U/D and clock output of the
present generator are connected to
the relevant points on the modules,
as explained in the above mentioned
article, but remember to observe the
different supply voltages of clock
generator and counter module; keep
all points marked + 5V at that voltage,
except the supply pin of the LM324
and Ri4 and R15, which are connec-
ted to the counter module + 12 V
supply. Current consumption of the
present up/down clock generator is
modest at about 10mA.

PT

71

551

Since good crystal filters are expens-
ive, there is a constant search for (less
expensive) alternatives. One of these
is the ceramic filter, now widely used
as IF filter in short-wave receivers.
The somewhat poorer temperature
characteristics of ceramic filters (as
compared with those of crystal fil-
ters) are normally not of much conse-
quence.

Numerous experiments have finally
led to the circuit of Fig. 1, which uses
five 45S kHz ceramic filters. As com-
puter crystals can be obtained
cheaply nowadays, it would also be
possible to construct a similar filter
with a number of such crystals.

The result of our experiments is a

narrow-band IF filter

rl -i ii- .11- .

Q ? tt tf

FF

3 dB filter bandwidth of about 800 Hz;
the attenuation outside the pass-band
is of the order of 60 dB.

A possible application is its use in a
receiver with variable bandwidth for
SSB, AM, and FM operation.

Another application is as input Biter

in a receiver whose dynamic fre-
quency range is inadequate (but the
IF should then not be 4SS kHz).
Finally, note that correct matching of
both the input and output impedance
(330 ohms) is imperative.

(B)

G Schumann

current indicator

72

The non-standard resistors, R4 to R7
incl., "measure” the actual current
consumption. Between point A and
the positive output terminal there
exists a potential difference. When
this p.d. reaches a value of 0.6 V, T2 .
and consequently T>,, switch on and
this causes D2 to light. In the same
way, when the p.d. between points B,
C, and D respectively, and the posi-
tive output terminal reaches about
0.6 V, transistor pairs Tj-T?; Tj-Tb; and
Ts-Tn switch on, and the associated
LED will light.

Resistor R2 and capacitor C? provide
a soft start facility at switch-on. Tran-
sistor Tt provides an emergency
switch-off facility, which in practice
has proved very useful
The input section (not shown) should
consist of a mains transformer with
24 V; 2.8 A secondary; a bridge recti-
fier (e.g. B80C2200/3300); and a
4700 >iF; 40 V smoothing capacitor.
The L200 regulator should be
mounted on a suitable heat sink. This
device has internal short-circuit and
overload protection; its pin assign-

is given in Fig. 2. (R)

R7

alternating flasher A R Kambach

The proposed circuit is intended for
use by modellers at railway cross-
ings, work in progress, advertising
boards, and many others. It can be
built quite quickly from but a handful
of components.

In the accompanying diagram, Ai
determines the flashing frequency,
which may be altered by changing
the values of R2 and, particularly, R3.
The latter may be replaced by a
suitable preset if the frequency
needs to be varied often.

Inverters A2 and A3 function as fixed
delay elements, while Aj inverts the
drive to D2. The alternately flashing
LEDs may, of course, be of any colour
suited to the application.

low-drop
voltage regulator

58

Integrated 3-pin voltage regulators
are not suitable for use where the in-
put and output voltages are nearly
equal. In fact, with most such
regulators, the input voltage is
typically 3 V higher than the output
potential. To cater for situations
where the two voltages are nearly
equal, it is necessary to use discrete
components. The series transistor is
then connected in a common emitter
circuit, so that the output voltage is
lower than the input voltage only by
the saturation voltage of the transis-
tor. However, it is then difficult to pro-
vide short-circuit protection as is the
case in integrated regulators. But,
where there is a will, there is a way.
In Fig. 1, the series transistor obtains
its base current from T2, which
together with Ti forms a differential
amplifier. This arrangement ensures
that the junction of voltage divider
R^-Rs has the same potential as the
cathode of zener D2. The crux of the
circuit is that T3 has a certain current
amplification, but T2 can only pro-
vide it with as much base current as
R2 allows. The potential difference
across R2 has a maximum value of
the zener voltage minus the base-
emitter voltage, Ube, of T2, which in
practice is about 4 V. The maximum
current through R2 is, therefore,
about 11 mA, so that, assuming that
T3 has a current amplification of 50,
the maximum output current is
0.55 A. If a higher current is drawn,
the output voltage will drop. If it
drops below the zener voltage of D2,
the p.d. across R2 will drop also. The
result is that the output current will
behave as shown by the fold-back

159!

Crafty designers are forever trying to
use ICs for applications they were
never intended for. In this circuit a
member of the newish HCMOS fam-
ily is used as a voltage-controlled
oscillator (VCO). This is achieved by
using the characteristic of the
HCMOS family of operating from a 2
to 6 volt supply. However, at 6 V these
ICs are faster than at 2 V.

In the present circuit, a "supply
voltage" variable between 1.5 and

characteristic in Fig. 2. It is clear,
therefore, that the series transistor is
protected against high (short-circuit)
currents.

Diode Di and resistor R: provide a
soft start, because the voltage across
the diode, which is connected to the
output of the regulator, is nought at
switch-on. Since the circuit, because
of the high gain, has a tendency to
oscillate, capacitor Ci is included to
improve the stability.

The output voltage level, Uo, can be
freely selected, within the limits of
the series transistor, by D2, R3, and
R<, and is determined from
Uo=Uz(Rs + R4)/R 5 .

Resistor Rz must be matched with
the actual current amplification of the
transistor used. The maximum dissi-
pation of a well-cooled BD140 is of
the order of 5 W. If a noise-free out-
put is required, an additional 10 *jF

5 V is used as the input signal of the
oscillator, which consists of three
cascaded NAND gates. The VCO
operates as follows: a logic 1 at pin 2
causes a logic 0 at pin 3; this
becomes a 1 at pin 6. and a 0 at pin 8.
Pin 8 is, however, connected to pin 2.
which, therefore, is no longer 1 but
becomes 0. This 0, because of the
delay times of the gates, appears a
little later at pin 2 as a logic I. And so
on: the oscillator works! Gate N«

2

Uo^-' rs " 8 u,

electrolytic capacitor should be con-
nected in parallel with D2. The cir-
cuit will then have a real soft start:
there will be no output for about 0.2 s
after switch-on. (R)

functions as a buffer for the oscillator
output.

Since the peak output voltage cannot
be greater than the supply voltage,
i.e. the input voltage to the oscillator,
its level must be adapted to those at
the remainder of the circuit, which
normally will be 5 V. This is ensured
by inverter Ns, which is powered by
a genuine 5 V supply. Because of
feedback resistor Ri, the inverter is
arranged as a linear amplifier. It is,

HCMOS VCO

74

therefore, sufficiently sensitive to
amplify positive signals between 2
and 5 V adequately.

The characteristic in Fig. 2 shows that
the VCO is reasonably linear. Other

output frequencies are not possible
with the circuit of Fig. 1, unless the
number of gates in the oscillator
proper is extended by an even
number of identical gates, which

increases the total delay times, so
that the frequency is lowered. It is
also possible to add dividers to the
output circuit. (W)

super dimmer

Most dimmers use a silicon-con-
trolled rectifier (triac or thyristor)
which is triggered at a fixed phase
angle and then conducts until the
next zero crossing of the mains
voltage. This method is simple, but at
the same time it gives problems in
controlling small or inductive loads
(hysteresis; flickering). The cause of
these problems lies in the fact that
owing to the small load the current
supplied to the bases is insufficient
to allow conduction to continue. This
means that a region of the control
characteristic is not used. The effect
is even worse when the load is in-
ductive.

The proposed circuit offers a sol-
ution by providing the SCR continu-
ously with gate current, so that even
loads of 1 watt can be controlled. To
keep the circuit as small and simple
as possible, it makes use of the well-
known timer-buffer Type S55.

The output of the 555, which is nor-
mally active high, is made active low
with the aid of a negative supply
voltage. The supply is provided by
network C1-R3, rectifier D1-D2, and
stabilizer D3-C2. Transistors Ti to T3
provide a start pulse at the trigger in-
put of the 555 during the zero cross-
ings of the mains. For a period deter-
mined by the setting of Pi and P 2 , the

output of the timer is high, and there
is, therefore, virtually no potential dif-
ference between pins 3 and 8, i.e. the
SCR is turned off. When the set
period has lapsed, pin 3 goes low
and the SCR is triggered. For the re-
mainder of the half period, a gate
current flows which keeps the SCR in
conduction. The minimum position at
which, for instance, a light bulb just
should not light, is set with Pi.

Filter R7-C5-L1 provides the requisite
decoupling of the SCR.

Finally, note that the maximum power
that can be controlled is of the order
of 600 watts.

( SV lb

car radio alarm

It is an unfortunate as well as a gener-
ally acknowledged fact that the car
radio (plus cassette recorder) ranges
among the most desirable and often
surprisingly easy to steal objects on
many a burglar's "shopping list".
This circuit may help to prematurely
end the criminal practice by sound-
ing the horn if it is attempted to
remove the radio set; cutting or
unplugging an additional ground
wire, which has been hidden in the
cable for connection to the battery
and loudspeakers), causes the alarm
to be set off, since the connection to
the car chassis (ground) is inter-
rupted.

The circuit for the car radio alarm is

composed of a single timer, the well-
known Type S55, surrounded by a
few additional odds and ends to
make an astable multivibrator, whose
on-time is determined with Ci. Horn
relay Re should have a coil resistance
to enable the timer chip to energize
it direct by means of the voltage at
output pin 3.

It is seen that the multivibrator is in
the reset state as long as point M is
connected to earth, ie. when the set
is in the place where it should be.
Removing the car radio inevitably
causes the voltage at M to rise to
nearly 12 V. ending the reset state of
IC : . which responds with activating
Re, i.e. the car horn, since this is

energized via the relay contacts in
parallel with the horn switch in the
steering wheel.

Note that Re is a PCB-mount type,
e.g. the Siemens Type V23127-A0002-
A101; where this is not available, any
other type of small changeover relay
haying a 12 V coil may be 'wired to
the circuit, provided the S5S is
capable of handling the coil current;

2

76

many motorists’ and car repair shops
can, no doubt, supply you with a
suitable relay for the alarm circuit.
The sense wire to point M should be
hidden in the multi-wire cable to the

radio set, while the circuit itself must
be fitted in an out of the way position,
somewhere behind the dashboard.
In order that not even an attempt is
made to break into your car, it is, as

will be readily understood, prudent
to stick adhesives to the car side win-
dows, warning of the presence of the
radio alarm. (Sv)

solid-state O
dark-room light WU

Light-emitting diodes are perfectly
suitable for dark-room light, because
they (a) obviate the need of filters; (b)
emit cold light; (c) have a life that is
not shortened by continuous on-off
switching; and (d) do not radiate
infra-rays. The types used must, of
course, have a high light output; for-
tunately. there are nowadays LEDs
with a luminous intensity of hundreds
of millicandela.

The sensitivity of photographic
paper lies between wavelengths
300 nm and about 550 nm, whereas
the wavelength of the light emitted
by green LEDs is about 565 nm; that
by amber types around 585 nm; and
that by red LEDs about 640 nm. From
this, it is clear that all three types of
LED may be used with impunity.
None the less, in practice, it is best
not to use green ones. Because of the
special composition and high sensi-
tivity of colour negative paper, only
yellow LEDs with reduced light out-
put should be used when processing
this paper. The proposed light,
therefore, has provision for reducing
the emitted light. Note that since
colour reversal paper is sensitive to
all colours, it can only be processed
in total darkness. When working with

orthochromatic paper, only red LEDs
should be used.

With reference to the diagram, each
group of three LEDs is fed from a cur-
rent source, T> to To respectively. The
current level, and consequently the
light output of the LEDs, is deter-
mined by the setting of Pi. Zener
diode Dis provides the reference
voltage for the current sources,
ensuring that the light output of the
lighting unit remains virtually con-
stant over the life of the PP3 battery.
Maximum light output is set with the
aid of P2. To this end, both Pi and P2

are first set to maximum resistance;
after this, P2 is adjusted until a poten-
tial of 0.2 V is measured at point A.
The maximum current through the
LEDs is then about 20 mA.

As the photograph shows, the unit
has been constructed so that Si is
easily operated. Since this switch is a
press-to-make type, the light will
switch off as soon as it is put aside,
thus preserving the battery. It is poss-
ible to have the light on continuously
by connecting an external battery to
Uext. In that case, Rio must be
matched to this source according to

Rio = (Uexi— 9)10 [kQ]

but only if NiCd batteries are used. If
standard cells are used, D20 and Rio
must be omitted.

If a variety of photographic paper is
processed, it may be useful to be
able to switch between red and
amber LEDs. For that purpose, each
of the eighteen original yellow LEDs
is duplicated by a red LED, shown in
dashed lines. Switch S2 may be used
to select the relevant bank of LEDs
(red or yellow) as required for the
specific application. (R)

Mandelbrot graphics

The computer-based implementation
of certain iterative types of calcul-
ation may offer highly attractive
graphics screen representations, as
we got to know when keying in a
program to crunch a few numbers in
the Mandelbrot series, and found
that doing so with the support of the
computer’s graphics facilities took us
through a regular graphics adven-

On further investigation, it was found
that the degree of complexity of the
resultant graphics image is in direct
proportion with the number of
iterative steps the control program is
arranged to perform. However, since
the necessary calculations to obtain a
Mandelbrot series become the more
complex, and therefore time con-
suming, as the computer crunches
through its approximations and
evaluations, it should not strike the
programmer as odd that obtaining a
nicely detailed graphics image may
take as long as 12 to 24 hours, even
with the fastest types of personal or
semi-professional types of computer,
such as the BBC equipped with a
second processor.

The Mandelbrot series of numbers is
basically obtained with the use of
complex numbers, in a calculation
that converges rather than diverges
the intermediary results according to
the equation Z=Z‘+C, where C is
the complex number constant having
a real part between —2 and 1, while
the imaginary part ranges between
—1.5/' and 1.5/; Z is the result of the
preceding calculation.

Stepping through a section of the
series is possible by assigning start
values and/or differently dimen-
sioned step rates to either the real or
the imaginary part of C. It goes with-
out saying that calculation time and
image resolution increase with the
number of iterations used for obtain-
ing results in accordance with the set
requirements; the calculations may
be stopped when the result is larger
than 2. The colour assigned to any
pixel on the screen depends on the
number of iterative steps required to
satisfy the Mandelbrot equation; if
this is not the case, the iteration loop
is consequently aborted.

The program shown in Listing 1 has
been written for the Electron or BBC
computer, and arranges for 15
iterative steps; the screendump of
Figure 1 shows the result. Figure 2
illustrates how a section of the

78 elektor india Aug/Sept 1986

graphics image is enlarged by means
of relevant redefinition of the
equation variables, as outlined
above. Obviously, the suggested
program allows a good deal of
further patching and experimenting
to arrive at even more attractively
styled graphics designs, but it should
be pointed out that producing Fig. 2
took our BBC no less than. . .2 days!

(SO

car radio alarm

The purpose of this one-chip circuit
is to give an audible alarm in case a
thief attempts to steal the car radio,
which is generally considered an
item of prime importance to the
motorist's well-being during any trip
with his vehicle.

Since removing the car radio necess-
arily involves cutting or unplugging
the supply cables, the present circuit
detects disconnection of an extra
earth lead, which has been fitted to
the rear side of the car radio (metal)
housing. In the circuit diagram, this
point is marked as M. If M is at earth

potential, Ti is off (high collector
voltage); if the earth connection is cut
or unplugged, the voltage at M rises
to a positive level, Ti conducts, and a
negative-going pulse triggers timer
ICi, which has been arranged to
provide a 30-second timing interval
as defined with R6-C3. The second
timer contained in IC 1 functions as a
0,5 Hz (R7-R8-C4) oscillator section
with an output duty factor of 50%
(D3). Note that the Type 556 dual
timer chip directly energizes a 12 V,
low-power relay, whose contacts are
connected in parallel with the horn

switch in the car's steering wheel.

If it is attempted to steal the car radio,
the alarm intermittently sounds the
horn for 30 seconds. It is, of course,
imperative that constructors of this
car radio alarm locate the additional
earth connection on the radio set in
such a way as to necessitate discon-
nection at an early stage of attempted
theft, otherwise the alarm would
come on too late, enabling the thief to
get off at his leisure.

Sv

79

deceptive lock

This circuit offers a means to fool all
but the cleverest burglar, but,
although it is a clever design, it has
been kept simple, as a glance at the
diagram will show. On the surface it
looks like a simple operating panel
with ten push buttons. However,
anyone trying to open it illegally is in
for a surprise! It is not just a matter of
keying in the correct code, it is also
necessary to keep one of the keys
depressed for about 10 to 15 seconds.
The circuit is based on a single Type
CD4093, which contains four NAND
gates with Schmitt trigger inputs.
Gates Ni and Nz form a bistable that
contains the status of the lock.
Assuming that the circuit has been
off for some time, switching it on
causes network R1-C2 to set the
bistable to the "lock" position, that is,
the output of N? is logic low. Capaci-
tor Ci is discharged, and the only
way the circuit can toggle is by
recharging this capacitor. This is
done by pressing key Sx long
enough for the trigger threshold of
Ni to be reached. When that hap-
pens, the bistable is set to the "open"

position, that is, the output of N2 is
logic high.

Capacitor Ci remains charged via
R4-D2. even after Sx has been
released. In other words, the bistable
remains in the "open" position. The
lock is closed again when one of the
other keys is pressed, or, if required,
by means of a special lock key. This
causes Ci to discharge rapidly via

D1-R3, which returns the circuit to the
"lock" condition.

When the lock is "open”, relay Re
will also be open in the present cir-
cuit. It is, however, possible to have
the relay energized in this condition
by connecting the remaining free
gate in the 4093 in series with Rs as an
inverter. • (Sv)

from an idea
by

A Biihlmeier

Noise on an audio signal becomes
more troublesome as the signal itself
becomes smaller. When a mixer is
connected to a number of signal
sources, it becomes particularly
disturbing when one or more of
these sources produce only noise. In
these situations, a noise gate is a real
help. Such a gate continuously
monitors the level of the audio signal
and switches it off, after a predeter-
mined period, if the level drops
below a preset value.

The circuit consists of two parts: a
control section and a regulator sec-
tion. The control section, based on
opamps Ai to A4 incl., derives a
voltage from the audio signal that is
used to drive the regulator.

The regulator is a voltage-controlled
amplifier, for which one of the two
operational transconductance

amplifiers contained in a Type
LM13600 or LM13700 is used. For a
stereo system, one control section

noise gate

and two regulator sections are re-
quired. For a double mono version,
two control sections and two
regulators are needed. One LM13600
or LM 13700 will thus suffice for all
these requirements.

Opamps A: and Az form a full-wave
rectifying circuit. Opamp A3 com-
pares the peak value of the signal
with the direct voltage set by Pz. If
the peak value is larger, capacitor Cr
is charged via T: : the attack time is
set by P3. The time lapse after which
the audio signal is switched off is de-
termined with P;. The control of the
voltage-controlled amplifier (VCA)
and the LED indicating whether
there is a signal present is effected
by A-j. Diode D4 ensures that the am-
plification of the VCA is really zero
when the output of A* is low (i.e. less
than -15 V).

The input of the regulator section has
an impedance of about 10 kQ and is
designed for audio signals of 1 Vims.

However, even for a 12 dB higher in-
put signal, the distortion is still not
greater than 1 per cent. Where
higher input voltages are the norm,
the value of Ri should be altered ac-
cordingly. Where lower inputs are
the norm, a preamplifier should be
used.

It is, therefore, seen that the noise
gate should preferably be connected
between the preamplifier and power
amplifier.

The output level is set with Rs, while
Pi enables the circuit to be adjusted
for minimum switching noises. To
this end, the drive input is switched
on and off by Si, while the audio in-
put remains open-circuit.

It is best to use a 3.5 mm chassis
socket with break contact for the
drive input: the break contact then
replaces Si. As soon as the jack is in-
serted into the socket, the connec-
tion between the audio input and the
regulator is broken.

-0

This type of drive input affords a
number of special effects, such as
the switching in of, say, an echo unit
at the command (sufficiently high

signal level) of a given instrument
(e.g. a snare drum). The command in-
strument is plugged into the drive in-
put for this purpose, while the regu-

lator is connected into the effects

VIP-bleeper

i

Here is a circuit that enables you to
leave a boring meeting at any time
you like. When you have had quite
enough of the exasperating, seem-
ingly endless discussions and would-
be interesting speeches about trivial
subjects, just fumble in your breast
pocket to take out a pen or a note-
book, stealthily press a relief button
to activate the timer function of your
VIP bleeper, and pretend to be very
attentive and busy so as to be more
surprised when, after 20 seconds,
you are called up by a loud bleep
from your very personal tracer
system. Now reset the VIP bleeper
by pressing the switch once more,
pretend to be quite disturbed, gather
your papers, and leave the meeting
after having informed the chairman
that your presence is urgently re-
quested elsewhere.

The VIP bleeper functions with only
very few components, and will be
operative for a whole year if powered
from a 9V alkaline battery; a normal
9V battery enables use for about 230
days, so it is not even necessary to fit

an on/off switch. Gates N-, and Na
have been arranged in a bistable
setup which toggles after Si is
pressed; as long as the circuit is not
activated, the output of N2 will be at
logic high level. Some 20 seconds
after depression of Si, N3 will enable
to low-frequency oscillator gate N« to

activate the bleep generator proper,
Ns. The resulting intermittent bleep
from the piezo element resembles
the ringing sound of a cheap, hand-
held telephone set. The bleep vol-
ume may be set with S2, depending
on the feigned urgency and relative
importance of the VIP. GD

81

68

Despite the many laudable qualities
of the BBC microcomputer as to
speed and ease of peripheral inter-
facing, many users are slightly disap-
pointed with the sound quality of the
standard version as manufactured by
the Acom company. An investigation
into this matter has revealed that
Acom have disregarded the optional
connection of an external audio
amplifier to the computer; this is the
more surprising since special holes
have been provided to this purpose
on the main PCB. The result of this
omission manifests itself in a very
poor sound quality, caused by the
small loudspeaker in the cabinet, the
high noise level of the improperly
driven audio amplifier chip, and the
rather coarse volume setting. How-
ever, a minor modification to the BBC
computer is sufficient to boost its
sound production by means of an ex-
ternal, more powerful audio amplifier
which may be connected to a sound
output socket on the computer. Pro-
ceed as follows:

1. Open up the computer, remove the
keyboard and the main PCB.

2. Locate the PCB holes for plug 16, to
the left of IC7, the Type LM386

audio amplifier chip.

3. Use desoldering braid to open up
the holes for plug 16, if these are

filled with solder.

4. Cut off the centre pin of a three-
pin, 0.1 inch pitch single row PCB

header, and solder it in the holes pro-
vided for plug 16.

69

Halogen lamps are, unfortunately,
rather prone to burn out when they
are switched on, and this is mainly
owing to the high current consump-
tion of these devices during the in-
itial stage of heating up to the normal
operating temperature of the filament
in haloid gas.

A typical value for the cold resist-
ance of a 6 V - 4 W halogen lamp is
about 0.3 ohm, demanding a turn-on
current of 20 A. In view of the rela-
tively low internal resistance of car
and motor-cycle batteries, such a
current surge is not at all to be

improved sound for the
BBC micro

* * M

W 1L

77 ,

5. Mount a 3.5 mm jack-type audio
socket with a break contact at the

rear side of the computer, and wire
Pie, Pis, and the internal loudspeaker
as shown in Fig. 1.

6. Reassemble the computer and test
the new audio output by connect-
ing an external amplifier set to the
jack socket. Insertion of the jack
plug should silence the internal loud-
speaker.

Now that we are on the subject of the
BBC computer, it is just as well to

give a few hints concerning re-
duction of the total power c.onsump-
tion of the computer. The Type 6522
VIA chips may be replaced with
their new CMOS equivalents 65C22
to reduce the total current consump-
tion by some 240mA. The 6850 chip
may also be replaced with a 6350, but
this is a riskier matter since the
former chip is soldered direct onto
the PCB.

HS

halogen lamp protector

dismissed as purely theoretical, and
it is easily seen that the ehsuing rapid
heating inside the lamp is a prime
cause for the thin filament to melt at
the sudden temperature effect. What
is required, therefore, is a series
regulator system to limit the current
during the heat-up phase; in other
words, a soft tum-on facility.

The circuit diagram shows that Ci is
charged to the battery voltage by
means of Ri and R2, causing FET Ti to
become slowly conductive after Si
has been closed. The Type BUZ10(A)
power FET is used in view of its low

82

drain-source resistance in the fully
conductive state; a typical value for
Rdgon) is 0.19 ohm, which ensures a
low voltage drop across the FET, and,
therefore, a sufficiently high
operating voltage for the halogen
lamp. Parts Di and R3 discharge Ci

after opening Si, so that the power-on
delay functions correctly any time
the lamp is turned on.

Lamp voltages other then 6 V require
R? to be modified according to
R 2 =200,000/(Vt*r - 2) [Q].

In case the BUZ 10(A) proves hard to

obtain, other types of n-channel
power MOSFET may be used in the
circuit. The minimum requirements
are: drain-source voltage Vds = 50 V,
drain current Id = 19 A, and drain-
source on resistance Rdstom = 0.2 Q.

AR

simpie NiCd charger

Dry batteries have one major disad-
vantage: they go flat. Rechargeable
types, such as NiCd cells, also suffer
from this drawback, but they can at
least be recharged. Sometimes even
a fifteen minute charge is sufficient to
give enough life to, say, an electronic
flash battery.

A NiCd charger is, in essence,
nothing but a sophisticated current
source. The present design contains
four such sources with a common
control switch, but each with a
separate LED that lights as soon as a
battery is connected to it.

In position 1, the sources each pro-
vide a current of about 90 mA; in pos-
itions 2 and 3 values of between 100
and 300 mA as required. Note, how-
ever, that with charging currents
above about 200 mA the transislors
must be'fitted on suitable heat sinks.
On stability considerations, it is
advisable to mount diodes Di to D*
incl. in good thermal contact with the
relevant transistors.

Terminals + B and — B enable the cir-
cuit to operate from a 12 V DC
source, such as a car battery, in situ-
ations where a mains supply is not
available.

1

The present unit can charge Type
AA ( = HP7 = R6) cells in about 8 hours
(position l = 90mA); Type C
( = HP11 = R14) cells in 10-14 hours (pos-
ition 2 = 180 mA); and Type D

83

by

P Sicherman

two-tone chime

This electronic chime is easily built
from commonly available, inexpen-
sive parts.

Depression of the door bell button,
S2, causes inverter Ti to pass a logic
low level to NAND gate Ni, which
responds with a logic high level at
its output, enabling the oscillator
composed of N? and N3 to toggle at
about 1 Hz. Since buffer capacitor C 1
remains charged for some time after
S2 has been released, the oscillator

72

This circuit offers impeccable
monochrome images when driven by
the digital RGB and sync signals of a
high-resolution graphics card such
as the one featured in Etektor India
issues from December 1985
to April 1986

Transistor T2 in the video combiner/
buffer ensures a short-circuit proof,
75Q impedance monochrome and
composite video signal with an rms
value of about IV, as usual for con-
nection to a monitor. The combi-
nation of a PNP and an NPN transis-
tor, Ti and T2 respectively, for ampli-
fication and combination of video
and sync typically exhibits a good
response to the fast rise and fall times

will continue to provide 1 Hz pulses
to Ct and Cs. as well as to a second
oscillator section, ' composed of N<
and associated parts via Re.

A logic high level at pin 10 of inverter
Ns enables T? to connect preset Pz in
parallel with frequency determining
parts R P , which arrange the fre-
quency of Ns to toggle at a few hertz.
The two superimposed frequencies
may be adjusted to individual taste
with Pi and P2.

of these signals, and thus enables
sufficient picture definition in the
case of, for instance, text presen-
tation at 80 characters per line, or
hlgh-resolution graphics appli-
cations.

The input circuit arrangement with
Ti and the mixer resistors is a D/A
converter in its most rudimentary
form; the R, G B, and I signals are ap-
plied to resistors at values that corre-
spond to the luminance percentage
of each basic colour, to the effect that
any colour shade is represented as
one of 16 shades of grey on the
monochrome monitor. Where this is
desirable, the intensity bit resistor
may be replaced with a 2k5 preset;

In addition to controlling the tone fre-
quencies of the chime, the 1 Hz
pulses also determine the envelope
shape of the resultant chime sound
by means of Tt-Ts and associated
parts. Preset P3 is used to define the
desired decay characteristic, while
emitter follower T6 functions as
a very simple voltage-controlled
amplifier, driving one-chip AF output
amplifier Type LM386.

KD

this enables setting the intensity
ratio. In case the present combiner is
used with a video interface that
merely supplies a video and sync
signal, or merely RGB signals, the
unused inputs may simply be left

The sync signals are combined with
the video signal at the base of T2; de-
pending on the system setup, the
sync input may have to be slightly
modified. Where an inverted CSYNC
(composite sync) signal is available,
as in the case of the Elektor graphics
card, all parts relevant to the sync in-
puts may be dispensed with, except
for Di. Resistors Rs and R6 determine
the black level of the composite

RG B-to-monochrome
video combiner

video signal, to which the sync com-
ponent is added by Di, which is
capable of lowering the base voltage
of T2 to a value, lower than the black
level. Diode D2 allows separate, in-
verted sync signals to be applied to
the combiner, and T3, Rio, R12 and Rn
may be added for video systems that
output a positive-going, separate
sync.

The combiner may also be used with
the video-interface incorporated in
the IBM PC if this computer is
equipped with a monochrome video
board, the sync mixer only requires
D2, Rn, R12, and T3. The IBM colour
board requires the addition of Rio,
whereas D2 may be left out. Figure 2
shows how the video combiner is
typically connected to the IBM
colour board. Note the use of the
‘reserved’ line (7) for the + SV supply
voltage of the combiner circuit.

CD

I = intensity -J"L

supply protection

The use of external mains voltage
adaptors for cassette recorders, port-
able radios, home computers, pocket
calculators, and so on, is common
practice since the typical enclosure
sizes of this type of electronic equip-
ment either does not allow the incor-
poration of a mains supply, or the
device has been primarily intended
for battery operation.

Unfortunately, the degree of idio-
syncrasy among manufacturers of
adaptors is rather high; a standard for
adaptor output voltage and output
polarity is definitely hard to find. It
may, therefore, be quite risky to
power, say, the home computer from
an adaptor which is not tailored to
this (expensive) piece of electronic
equipment.

Here is a simple circuit to prevent a
lot of trouble; its extremely low cost
fully justifies incorporation in any
equipment operated from an exter-
nal DC supply. The supply protection
consists of a mere four parts and a

fuse, which may already be incor-
porated in the equipment. The zener
diode is selected to have a zener
voltage of about one volt higher than
the equipment supply voltage.

In case the input voltage to the cir-

cuit has the wrong polarity, the zener
diode conducts and causes the triac
to fire, since its gate is driven positive
with respect to MT2; the current flow
through the triac is sufficiently high
to look like a short-circuit to the fuse,
which duly melts and breaks the
supply voltage, before damage is in-
flicted upon equipment parts.
Operation of the circuit in an over-
voltage condition is even simpler,
since in that case the zener also sup-
plies the gate of the triac with a firing
voltage.

Obviously, if everything is in perfect
order, the protection circuit is as if
non-existent to the equipment it is
part of, because it introduces no ad-
ditional voltage drop. Finally, the only
modification to the circuit for use in
positive-earth equipment involves in-
sertion of the fuse in the negative-
supply line.

HS

1986 85

H J Walter

electronic toss-up

The electronic version of the well-
known coin to toss up prior to com-
mencing a footbal match— or any
other sports event where this a gen-
erally established formality on part of
the referee— consists of a row of
seven LEDs, the centre one being
green, the others red. After having
reset the circuit, the odds are exactly
equal for either one red LED located
next to the green one to light when
the toss-up key is pressed; we have,
therefore, a left/right decision circuit
operating on the basis of pure ar-
bitrariness.

As to the operation of the circuit, but-
ton Si may be pressed at any time to
preset counter ICi, which responds
with outputting the binary code for 0
at its Q», O' and O2 outputs, causing
BCD-to-decimal decoder ICs to light
the corresponding LED, i.e. the green
one— D4— at the centre of the row.
The preset code for the initial state of
the circuit is determined with preset
inputs P....P3 being tied to ground,
causing ICi to load 0000 as the binary
start-up value when Si is pressed.
Depression of button S2 causes the
bistable composed of Ni and Nj to
toggle, providing a single pulse tran-
sition at the clock input of ICi.
Depending on the logic level at the
UP/ DOWN input of ICi, the one-of-
eight decoder will light either Ds
(right) or Dj (left), since counting up
from 0000 gives the next higher
binary code 0001, while counting
down gives 1111. The latter value
causes IC» to light D3 at the 0? out-
put, since the most significant bit
input— D— has been tied to ground.
The arbitrariness of the toss-up cir-
cuit is ensured by the speed at which
oscillator N3-N4 applies pulses to the

counter UP/DOWN input. The odds
are 1 to 1. theoretically, while the cir-
cuit can not be bribed. .

Seven of the eight active-high outputs
of IC2 have been wired direct to the
corresponding LED. while 0- serves
to inhibi t the counter via the
CARRY IN terminal. It is readily seen
that counter inhibiting occurs auto-
matically when ICi counts up from
output state 3, or down from state 5;
both conditions cause Q* and

therefore CARRY IN. to go high,
disabling further counting until the
reset button is pressed.

Finally, repeatedly depressing S2
without resetting the circuit will
cause any other, random, LED in the
row to light, and this facility may be
put to good use in any other, random
decision based game or serious
application you have in mind.

GN

B Vogel

tuning

AF

power stages

Simple, economically priced audio
output stages, such as, for instance,
those using the hybrid ICs in the STK
series, may be improved in a simple
manner as regards distortion, noise,
and off-set voltage. To this end, the
output amplifier is included in the
feedback loop of an op-amp. Fig. 1
shows the set-up for inverting output

amplifiers, and Fig. 2 that for non-
inverting ones (the normal situation).
In the calculations to arrive at the
new gain of the output amplifier,
determined by Ri and R;, it is
assumed that the LF356 provides an
undistorted signal of 5 Vmvs; note
also that this type of op-amp must
work into a load of not less than

5 kilo-ohms to prevent distortion.

For an output power of SO W into
4 ohms, the output stage must pro-
vide a voltage, U= PR = 14.2 Vrms. If
the amplification of the stage is 3, the
op-amp should deliver 4.73 V. For the
set-up in Fig. 1, the value of R? is then
R? = 3Ri, while for that in Fig. 2, R2=
2Ri. Note that in both versions only

the value of R> should be altered. The
total amplification may be calculated
from the ratio of Ra and Rb as follows:
A=(Ra+Rb)/Rb. Furthermore, be-
cause of the load impedance of the
op-amp, Ri>10 k (Fig. 1); Rj>10k
(Fig. 2); Ra> 10 Q; and Rc>10 S (Fig.
1 and 2).

To compensate for the off-set voltage
of the output amplifier, the input
capacitor should be replaced by a
wire link. The capacitor in series
with Ri in Fig. 2 should also be short-
circuited. The lower frequency limit
of the complete circuit is then deter-
mined by Cb = l/2nfumRB. The off-set
voltage is then smaller than 3 mV,
provided both Ra and Rc are equal
to, or greater than, 100 kQ. Where
greater accuracy is required, Pi can
be used to set the off-set to exactly
0 V.

To ensure that there is no direct
voltage at the new input of the ampli-
fier, capacitor Cc should have a
value of Cc=l/fiimRc.

Since the amplification of the output
stage has been reduced to 3, its feed-
back factor has gone up, and the dis-
tortion has gone down. The ad-
ditional feedback of the LF356
reduces the distortion even further.
An overall reduction in the distortion
from 1 per cent to 0.1 per cent is fairly
typical.

The altered feedback unfortunately
results in a change in stability. If
there is a tendency to oscillate, the
first thing to do is to bring the upper
frequency limit back to its previous
value with the aid of Cv = l/2nfi,mRA.
If the tendency persists, capacitors
Cx must be used: their value lies
between 100 pF and 1 nF. Our proto-
type (using STK ICs) worked satisfac-
torily without either Cx or Cy. (W)

electronic fuse

It is perfectly understandable why
many a constructor is in a cold sweat
when faced with the decision
whether or not power may be ap-
plied to his circuit board or ex-
perimental set-up at a more or less
advanced stage of completion. Even
after the closest inspection of the rel-
evant circuit, sheer experience has
taught many of us to be wary of
calmly curling smoke signals from
quite unexpected PCB locations,
along with a sudden and violent ac-
tion on part -of the ammeter in the
supply line, often followed by rapid

® —

overheating of the power supply
series regulator and blowing of a
fuse, all of which typically take place
before the mains switch can be
reached to prevent further damage
from being inflicted upon power
supply and/or expensive parts in the
circuit under test. Murphy strikes
again!

Some of the above-mentioned plight
may be alleviated by this circuit,
which functions as a resettable, non-
destructible fuse replacement for
series connection in the positive
supply line to DC operated circuits

consuming up to about S00 mA at an
operating voltage between 5 and
30 V.

The circuit diagram shows series
resistor Ri, which drops a voltage in
proportion to the current it passes to
the load at the circuit output.
Whenever this voltage exceeds
about 0.5 V, series-connected tran-
sistor T2 is turned off and LED Di
lights to indicate the overload con-
dition.

Si is pressed to restore the supply
current to the circuit under test, but
only after the cause of the "fuse” ac-
tion has been investigated and cor-
rected.

Whenever the current consumption
of the circuit under test is lower than
about 500 mA (determined by Ri). Ti
remains off. However, when te load at
the circuit output consumes more
than 500 mA, Ti is arranged to act as
a relatively low resistance between

77

This simple circuit enables computer
users to generate analogue voltages
under software control, which, no
doubt, offers interesting possibilities
for intelligent control of, for example,
volume adjustment of audio equip-
ment, light dimmer circuits, etc. It is
also possible to write machine
language algorithms for the gener-
ation of several different, complex
periodic output voltages, in short, to
construct a computer-controlled
function generator using a minimum
amount of hardware.

The circuit is based on the Type
ZN426 digital-to-analogue converter
(DAC), which is an 8-bit resolution
(255-step), high conversion speed
(1 jjs) device for direct microproces-
sor interfacing. The circuit may be
connected to an 8-bit output port
which provides TTL or CMOS com-
patible digital levels; most computers
currently on the market have such a
port, or the manufacturer has made
provision to add one or more of these
in the form of an expansion. The con-
version time of the DAC chip allows
the use of machine code for high fre-
quency output voltages; BASIC is
usually too slow for this purpose. The
DAC output voltage is buffered with
an BIFET opamp, which can be ad-
justed for a step response of
15mV/step, which means that the
maximum output voltage of the pres-
ent circuit is 3.825V, since 8 bits

the base and emitter of T2, which is
consequently turned off, interrupting
the current flow to the load, just like
a fast acting fuse. Resistors R3. . .Re
have been provided to maintain the
necessary base current for Ti once
the circuit has disabled the load,
since that condition implies the
absence of a sufficiently high voltage
across Ri. Therefore, it is seen that
pressing Si is the only way of restor-
ing the current flow to the circuit
under test.

The circuit as shown is readily
modified to suit a wide range of
selectable output currents by replac-
ing R: with a 12-way, single pole
switch and a set of 12 suitable series
resistors, each calculated from
R=0.4Inu* for operating voltages
below about 20 V. The indicated tran-
sistor types, however, do not allow a
current consumption in excess of
500 mA, while the voltage drop

caused by the circuit is of the order
of IV.

In case the proposed electronic fuse
is to monitor the current consump-
tion of digital circuits, notably those
incorporating ICs from the well-
known TTL families, the fuse action
LED may light after power has been
applied, despite the fact that the rel-
evant digital circuit is known to func-
tion all right. The explanation of this
phenomenon lies in the relatively
high peak power consumption of
these chips, which draw the more
current as their switch rate increases.
In order to render it impossible for
the electronic fuse to react to such
current peaks, smoothing capacitor
Ci may be added; its value should
be 10 to 100 jiF. determined on a trial-
and-error basis.

TW

8-bit DAC

represent 255 steps (2*-l).

Adjustment of the circuit is straight-
forward: connect a DVM to the out-
put and adjust Pi for an indication of
0.00V with nought (0) written to the
DAC; next, write 255 (FFhex) and ad-
just P2 for the maximum voltage indi-
cation of 3.825V.

The circuit is also very suitable as an
D-to-A converter driven by 8-bit I/O
port (EE, December 1985) as part of

the universal I/O bus. It should be
noted, however, that writing FFhex to
this port gives an analogue output
voltage of 0V, since the ULN2003
buffer IC in the 8-bit output port is an
inverting device: moreover, the eight
data lines to the DAC chip should be
fitted with pull-up resistors as shown
in the circuit diagram.

HS

Some popular computer games re-
quire the joystick to be turned 45° in
order to get the correct cursor move-
ment on the screen. Obviously, this
presents problems in case the
joystick is desk mounted or of the
type that is ergonomically styled and
hand-held.

The electronic solution to this in-
convenience starts from a redefi-
nition of the joystick axes, as shown
in Fig. 2. Direction A is defined as in
between the positive X and Y axes;
direction D as in between the nega-
tive X and positive Y axes. Directions
C and B are opposite to A and B re-
spectively. Table 1 summarizes the
old and new direction assignments
and associated activated outputs.
The circuit diagram of the adaptor
circuit — Fig. 2 — shows that the out-
put levels to the computer are active
low rather than high as in the un-
modified joystick connection; this
necessitates the use of inverter gates
between adaptor and computer in-
put. A Type 74LS04 hex inverter may
be used to this end, and the trigger
(fire) functions) can also be inverted
at the same time, since this IC con-
tains six inverters.

Table 1

D irection Contacl

C • - X and - Y
D • — X and • V

B and C • V
C and D - X

The double trigger function enables
the turned joystick to be connected
to MSX types of computer as well.
Table 2 lists the relevant connections
for both the C64 and the MSX com-
puter type.

The adaptor input and output signals
may be visualized with red and green
LEDs, clearly indicating the elec-
tronic signal turn over 45 °. When the
joystick is moved into direction .<4, for
instance, input LED +Y lights, as well
as output LEDs +Y and +X. Current
consumption of the adaptor circuit is
about 75 mA.

89

!79l

One of the major problems in the
design of switch mode power sup-
plies is that most available (and
suitable) ICs only offer the absolutely
necessary facilities and not, for in-
stance, thermal or short-circuit pro-
tection.

Linear Technology offers a solution
to these problems with their LT1070
range of switching ICs. These
devices are as easy to use as the
familiar 3-pin regulator ICs. All steps
have been taken to make the design
of a simple, yet efficient, switch
mode supply as easy as possible.
The peak output current is 5 to 9 A,
and a current-limiting circuit is
provided.

The diagram shows a switch mode
DC-to-DC converter, whose output
voltage may lie between 12 and 48 V,
provided the input voltage is greater
than 3 V. The input voltage of the cir-
cuit as shown must always be lower
than the output voltage. It is, how-
ever, possible to modify the circuit to
obtain an output voltage that is lower
than the input voltage. One of the
modifications is replacing Li by a
suitable transformer.

The output current is dependent on
the value of the input voltage. For an
input voltage of 3 V, the output power
is 10 watts maximum. Our prototype,
operating from 3 V, delivered about
50 mA at 48 V, while at an input
voltage of 24 V the output current
was over 1 A.

In the construction, account must be

The proposition that television
"brings the world into your living
room" has gathered strength with the
latest addition to the colour vision-
stereo sound-three dimensional TV
line of development in "tube tech-
nology": the programme-controlled
odour generator, which is expected
to herald a new era of real life to tele-
vision, since broadcasters will no
doubt avail themselves of the oppor-
tunity to delight us viewers with the
appropriate smell to go with a certain
action in the relevant film or docu-
mentary.

90 elektor india Aufl/Sepl 1986

switch mode power
supply

taken of high peak currents: all con-
nections should, therefore, be short,
and the input and output lines should
be SWG20 (0.8 mm . ) or thicker.
This also applies to the earth con-
nections.

It should also be noted that spikes
are present on the output voltage. If

necessary, these may be eliminated
by an LC filter, the inductance of
which has the same value as Li, and
the capacitance is between 10 and
100 nF. The quality of the capacitor is
of importance because of the re-
quired low series resistance for RF
signals. (B)

programmable odour
generator

Odorant Elektronik GmbH of Col-
ogne, Germany have recently in-
troduced their Type CD4711 4-bit
programmable, bio-electronics
based odour generator, suitable for
digitally controlled TV sets. Jour-
nalists were regaled on a number of
sample odours from the new chip at
a rushed press conference, during
which a spokesman of Odorant
claimed that the Type CD4711 is
capable of producing 1 of 15 odours
as selected by the bit pattern on the
four CMOS- and TTL compatible data
inputs (see Fig. 3), while an 8-bit ver-
sion with odour intensity modulation
will shortly be announced.

3. . .15V / 2mA

-e~e

Mo - 1

Jf'S

speech processor with O l
background suppression O X

A speech processor is commonly
used in public-address installations
and in utility transmitters. It augments
the average value of the speech
signal, so that in spite of a high level
of background noise or, in the case
of a radio transmission, a lot of inter-
ference, speech recognition remains
possible. In many cases it is, how-
ever, undesirable that this back-
ground noise or interference is
enhanced together with the wanted
signal. A possible remedy, as out-

lined here, is to provide an adjustable
threshold at which the speech pro-
cessor becomes active.

With reference to the diagram, the
signal from the microphone is
amplified in Ti (a low-noise ampli-
fier) and in A;. Limiting (or clipping)
of the signal takes place in A 3 .

The signal (taken from the output of
Ai) is also amplified in A2. When the
output of this opamp reaches a cer-
tain level, electronic switch ESi is ac-
tuated. Consequently, the mono-

stable formed by ES2 changes state,
and this closes ES3, whereupon ES*
is opened, which in its turn increases
the amplification of A3. When ES* is
closed, the amplification of A3 is de-
termined by the ratio Pi:Rs; when
the switch is open, by the ratio
(P2+Rs):R5.

The mono-time, determined by the
time-constant R20-C19, has been
chosen such that speech is not
clipped. The low-pass filter between
A3 and A* ensures that frequencies

above 3 kHz are severely attenuated.
The required output level is set by

P3.

Calibration is somewhat unorthodox:
a signal source with a continuous out-
put of speech by trained speakers is
used. The microphone is positioned
in front of the loudspeaker at normal
speaking distance and the sound
level adjusted to roughly the level of
the user. Next, connect a pair of
headphones to the output of the pro-

cessor and make sure that only the
output of these phones can be heard.
Adjust P< for maximum resistance,
and then set the clipping level with
P2 (which is a matter of personal
taste). At maximum clipping level, in-
telligibility of the speech will remain
good in the presence of interference,
but it will have a somewhat harsh,
metallic character. Then, adjust P
for maximum resistance, and P4 till
all background noise disappears. Fi-

nally, set the ratio signal: background
noise with Pi; this is best done by
making a recording of the user’s
speech via the microphone and the
processor. When the processor is ac-
tive, i.e. clips, D4 lights.

Li to L4 incl. are 6 turns 36 SWG CuL
through 3 mm ferrite beads. (B)

82

electronic bell-pull

The simplest circuit in this issue of .

Elektor Electronics consists of a 1
single transistor and resistor, which,
when put together as shown, con-
stitute the electronic equivalent of an

old fashioned, stylish bell-pull used R _ Ub

in conjunction with a chime or bell

circuit of any relative complexity of- 1—

fering whirling melodies, buzzing or ^

ringing sounds, or chime imitations F*

to prompt the houseowner to open lr~^ ' BC ,„, |
the front door.

The bell-pull is made from a
T039-style NPN transistor which is
taped or isolated by means of a
length of heat shrink sleeving, after
the emitter and collector leads have
been fitted with wires for the elec-
trical connection to the bell circuit. A

small, conductive plate is secured
onto the isolated transistor head, and be cast into epoxy
the base lead is joined to this plate nice compact unit
over a series resistor which is dimen- the treatment of ev
sioned according to R-Ub/5 [kQ]. caller at the door!
The completed assembly may also

■sin to make a
lat can handle
t the roughest

92

The environmental nuisance value of
discos is in direct proportion to their
sound level. The circuit proposed
here cannot be disabled by the disc
jockey, since it is built into the output
amplifiers used in the disco. Its oper-
ation is amazingly effective: if the
preset sound level is exceeded, the
input of the amplifier is short-
circuited for a few seconds. Any disc
jockey whom that has happened to a
couple of times soon gives up trying
to break the sound barrier.

The power amplifier output is con-
nected to the metering input of the
present circuit (Ci). This signal is ap-
plied to low-pass filter R4-C2 via Pi
(which sets the maximum volume)
and buffer ICi. In case of line inputs,
this opamp can be given a gain of
20 dB by the omission of the wire link
across R2.

The signal from the low-pass filter is
rectified (half wave) by IC2 and IC3.
The resulting direct voltage is ap-
plied to A 1 and A2 which compare it
with two reference voltages derived

2

from potential divider Ra-RrRio. R14 and Ris, and capacitor Ce, obviate
When the first threshold is ex- any "plops” from the loudspeakers,
ceeded, Ds lights to warn that maxi- Power for the present circuit may be
mum sound level has almost been derived from the output amplifier,
reached. When the sound level then The normally quite high supply
increases by 6dB, At also toggles, voltage there is reduced to + 15 V by
which triggers monostable ICs. two complementary power tran-

The input signal to the power ampli- sistors. Current consumption of the
fier (via Cn, R>e. and P2) is then short- circuit is about 40 mA.
circuited to ground via T>. Resistors

93

sideway RAM for BBC
and Electron Plus One

As already reported on numerous oc-
casions in this magazine, the BBC
micro ranges among the most widely
used types of personal computer
currently available. To newcomers in
the computer field, the amount of
commercially available ROM-sup-
plied software is truly staggering,
and there seem to be programs to
suit almost any requirement and
budget.

However, the number of ROMs that
may be located in the BBC computer
is limited to four in the basic version
and sixteen when it is equipped with
a sideway ROM expansion card.
Users in posession of a good many
ROMs and EPROMs are, therefore,
often forced to exchange these
before a program can be run; a
method that is both cumbersome and
possibly bad for the ICs and their
sockets.

A way of getting round this problem
is to install RAM rather than ROM or
EPROM chips on the sideway board,
so that software may be readily
moved about between ROMs, direct
access memory, disk and RAM, since
many of the originally ROM-based
programs may also be run from RAM.
it has appeared.

Since it was thought convenient to
plug 16 Kbytes of static RAM into any
one vacant ROM socket, the circuit
was constructed in all-SMD tech-
nology on a ready-made PCB of very
small size.

The circuit diagram shows two
8 Kbyte, low-power static RAMs Type
6264FP-15 as a replacement for a
16 Kbyte EPROM Type 27128; a single
inverter selects the relevant 8 Kbyte
block when the (formerly) ROM
socket is addressed.

Working with SMD parts to achieve a
truly miniature ROM replacement
should be based on the necessary
skills in soldering and handling
these new parts, and the construc-
tion of the proposed extension
therefore requires to be done as
follows.

It should be noted that the through-
plated PCB for this project comes
together with the SMD die, de-
scribed elsewhere in this issue.

Fit 28 short (1 cm) pins at the sides of
the PCB to enable it to be received in
an IC socket— see the accompanying
photograph.

The SMD RAMs are mounted piggy-
back onto the PCB, with the excep-
tion of pin 26 of the top mounted

2

ICi ;IC2 = 6264FP 15
IC 3 = 74HC04
PCB 86425

RAM; this terminal should be wired
to socket pin 28. The SMD parts
74HC04 (IC3) and Ri may now be fit-
ted to conclude the PCB construc-
tion. Once the unit has been plugged
into a ROM socket, a short wire is run
from pin 8 of IC77 on the BBC main

board to the NWDS input on the SMD
board.

Finally, although not mentioned so
far, the Electron Plus One computer
may also benefit from the proposed
sideway RAM circuit which, as will
be readily understood, need not

necessarily be constructed with SMD
parts; a veroboard and normal sized
components, along with a bit of wir-
ing, will also do in many cases,
although it may be hard to surpass
the elegance of the plug-in unit.

HS

long interval timer

<♦> to.tzv

This low-cost timer circuit can offer at any time during the interval causes Setting the exact duration of the

switching intervals up to about 24 the timer to be reset and Rei to be timing interval is readily

hours and may, therefore, be useful deactivated. accomplished by temporarily using

for a variety of domestic as well as The funtion of FFi is that of a counter output Q3 rather than Qu to

electronic applications. debouncer circuit for Si which, when reset FF2; with the component values

Depression of Si causes Rei to be actuated, causes FFi to apply a logic as indicated, the interval should be

energized and the timer to be high pulse to the clock input of FF2. adjustable between 3 and 45

started; the position of Pi determines which toggles. IC2 starts counting, seconds. Divide the desired relay-on

the duration of the timing interval — since its reset condition is ended. At time by 1024 and set Pi accordingly;

the given value for C2 allows a maxi- the same time, Ti is driven with a connect the FF2 R input to Qu again,

mum of 12 hours. Doubling the positive logic level, and Rei is ener- depress Si and have Rei power the

capacitance of C2 lengthens the gized. After the timing interval has relevant equipment for as long as set.

timing interval accordingly; the timer lapsed, i.e. when counter output Qu (St)

may thus be employed to control a goes high, FF2 is reset and Rei deac-
NiCd battery charger. Depressing Si tivated in consequence.

95

from an idea
by R Conell

digital volume control

This digital potentiometer circuit is a
hybrid analogue and digital design
offering push-button controlled pro-
grammable attenuation as well as
high to low impedance conversion
by means of a single active device.
Digital noise is eliminated as effec-
tively as possible through galvanic
isolation of digital and analogue parts
in the input attenuator.

At the heart of the digital control sec-
tion is a Type 2716 EPROM, which
can be programmed either as shown
in Table 1 or to individual re-
quirements, as will be detailed
below. At power-on, debouncer
bistables N1-N2 and N3-N4 force logic
low levels onto EPROM address lines
As and At respectively, selecting a
programmed address range that
supplies the digitally coded, initial
volume settng. R-C network R16-C2
causes gates N7 and Ns to generate a
clock pulse for IC2, which latches
the 8-bit word from ICi, passes this

information to driver ICs, and thus
determines which relay(s) is/are en-
ergized, thereby fixing the attenu-
ation before the AF signal is applied
to opamp 1C;. Depression of Si (up)
or S2 (down) causes the correspond-
ing address line As or A6 to go low,
selecting a certain address range in
the EPROM. The exact address lo-
cation is determined by the value last
latched into IC2 after either key has
been released. It is readily seen that
the five available databits at the
Oi. . .05 outputs of IC2 allow 32 (2 5 )
simulated potentiometer settings.
The digital control section has been
designed to offer an auto-repeat
function when either one of the step
control keys is kept depressed; oscil-
lator gate Ne then provides a clock
pulse train to N?-Ne. and so causes
successive addresses in ICi to be
scanned automatically, until either
the lowest or highest possible
volume setting is reached, at which

moment the circuit forces itself to a
hold state, which can also be
selected at any time by simul-
taneously depressing the up and
down key.

S3 enables the user to select a
further address block, programmed
with another set of volume steps; the
circuit as shown, along with the data
from Table 1 arranges for 3 dB steps.
The analogue section of the circuit is
basically a four-section, relay-con-
trolled attenuator composed of re-
sistor networks to achieve a signal
attenuation in 3 dB increments, as
defined with the relevant bit pattern
at the Q1...Q5 outputs of IC2. Ret
(Q5), if deactivated, enables IC6 to
amplify its input signal by 3 dB. The
inset resistor and preset combination
may be used take over the function of
C10, since the latter should be a high
stability foil type, which may be a
rather difficult to obtain part. Both cir-
cuit alternatives function as click

suppressors when stepping through
the available range of volume set-
tings. The preset, if used, should be
set for zero offset voltage at pin 6 of
the opamp; replace Cio with a wire
link.

It is suggested to use miniature DIL
relay types in the Re i... Res pos-
itions, while all resistors in the at-
tenuator are preferably close toler-
ance (1%), high stability types. Also
observe that the supply voltages to
analogue and digital section are kept
well apart and decoupled so as to
preclude introducing switching
pulses and digital interference in the
sensitive attenuator sections as well
as the opamp output stage.

Finally, Table 1 offers a suggestion
for programming the EPROM with
data to achieve circuit operation as
set out above. (SvJ

Table 1

0000 00 01 02 03 04 05 06
0010 10 11 12 131415 16
0020 00 00 01 02 03 04 05
0030 0F 10 11 12 13 14 15
0040 01 02 03 04 05 06 07
0050 1 1 12 13 14 15 16 17
0060 0E 0E 0E 0E 0E 0E 0E
0070 0E 0E 0E 0E 0E 0E 0E
0080 00 01 02 03 04 05 06
0090 10 II 12 13 14 15 16
0080 00 00 00 01 02 03 04
00B0 0E 0F 10 11 12 13 14
00C0 02 03 04 05 06 07 08
0000 12 13 14 15 16 17 18
00E0 0E 0E 0E 0E 0E 0E 0E
00F0 0E 0E 0E 0E 0E 0E 0E

courtesy light delay

Ever been groping about for the
safety belt, ignition slot, choke con-
trol or a map while in utter darkness
and happy to have closed the car
door(s) because of the cold, or foul
weather? Wouldn’t it be convenient
to have the courtesy light on for a few
more instants in order to get the ve-
hicle started and ready to move off?
Figure 1 shows a courtesy light delay
circuit for easy incorporation in
almost any type of car. The courtesy
light is switched by power MOSFET
T* , which is a Type BUZ72A ensuring
a low voltage drop (0.2 V typ.) across
drain and source and therefore the
lowest possible power loss. The door
contact, connected to terminals B
and C, is normally a push to break
type. T. is therefore off and Ci
discharged when the door is closed;
MOSFET T2 does not conduct, so
that the courtesy light remains
quenched. Opening the door, how-
ever, causes Ti to charge Ci, and the
courtesy bulb will therefore light in a
gradual manner. Although closing
the door turns Ti off again. C. con-
tinues to supply gate drive to T2 for a
few more seconds; the courtesy light
will be dimmed slowly. The
suggested MOSFET type should not
switch more than about 10 W, which
is the usual power rating for the
courtesy light.

Figure 2 shows how the circuit may
be modified to enable the courtesy
light to go out immediately after the
ignition key is turned. The terminal
numbers refer to the wiring code

1

convention as relevant to most types
of European car:

15 = +Vban - ignition on.

30 = +Vi»h - unswitched.

31 = ground.

31b = door contact (connects to
ground).

50 = + Vban - starter motor on.

Figure 3 clearly shows the circuit
connections in accordance with the
foregoing convention.

In case the suggested MOSFET Type
BUZ72A (Siemens) is a difficult to ob-
tain item, any equivalent n-channel
power MOSFET to the following
specifications will also do ad-
equately: VdsZ 100 V: Idz 9 A;

Pd >40 W; AW>,»s0.25 Q.

07 08 09 08 0B 0C 00 0E 0F

17 18 19 1ft IB 1C ID IE IF

06 07 08 09 08 0B 0C 0D 0E

16 17 18 19 18 IB 1C ID IE

08 09 08 0B 0C 0D 0E 0F 10

18 19 18 IB 1C ID IE IF IF

0E 0E 0E 0E 0E 0E 0E 0E 0E

0E 0E 0E 0E 0E 0E 0E 0E 0E

07 08 09 08 0B 0C 0D 0E 0F

17 18 19 18 IB 1C ID IE IF

05 06 07 08 09 08 0B 0C 0D

15 16 17 18 19 18 IB 1C ID

09 08 0B 0C 0D 0E 0F 1011

19 18 IB 1C ID IE IF IF IF

0E 0E 0E 0E 0E 0E 0E 0E 0E

0E 0E 0E 0E 0E 0E 0E 0E 0E

87

2

Source: Siemens Components XX |
(1985) No. 6. 1

elektoi mdia Aug Sep. 1986 97

LED revolution counter

A close look at the dashboards of a
number of cars may reveal the use of
three basic types of rev counter: first,
the still most commonly found
needle and round scale, analogue
combination; second, a set of digital
displays (often LCDs); and third, a
pseudo-analogue meter in the form
of multi-coloured LED bar. looking
much the same as a LED-based VU
meter on modern recording equip-

The circuit presented here belongs
to the third category. However, con-
trary to the straight LED bar indi-
cation, this design features a round
scale with a coloured LED needle
imitation, just as the good old
mechanical rev counter.

The circuit is based on the
Telefunken Type U1096B analogue
input LED driver which can light one
of 30 LEDs on the rpm scale, whose
lower and upper indication limits
may be set to individual re-
quirements; e.g. the 30 LEDs may
merely indicate a limited rpm range

1 *

to attain a higher resolution.

The circuit diagram shows ICi to
receive the contact breaker pulses
and to reshape them for conversion
to an analogue voltage in an R-C
filter, which passes the signal to the
input of the LED driver.

The detailed operation of the circuit
is as follows. Zener diode Dei and
parallel capacitor Ci safeguard the
base of inverter transistor Ti against
receiving high voltage pulses in-
duced in the ignition coil secondary
winding. The Type NE555 timer has
been configured to function as a
monostable with an output pulse
period time of 3 ms, during which
time R3 causes Ti to conduct so as to
prevent erroneous triggering of the
monostable. The analogue voltage,
proportional to the engine rpm rate,
is established by means of smoothing
network Rs-Cs, R9-C0 and R10-C". The
indication range for the LEDs may be
set with Pi and P2, the presets for the
lower and upper limit, correspond-
ing to LEDs Di-D 2 and D59-Deo re-
spectively. Note the relative
simplicity of the LED array connec-
tion to IC3; only nine IC output lines
suffice to drive any one of 30 pairs of
LEDs, whose colour may be chosen
to individual taste, while it is also
possible to use series-connected
LEDs to achieve a brightly as well as
functionally lit rpm scale.

The circuit diagram shows two rows
of LEDs; the upper one is the normal
rpm indifcation scale, for which the
following coloured subdivision may
be used; 0 to 5000 rmp are green
LEDs; from 5000 to 6000 rpm yellow

y

*■

2, y

or orange types; 6000 rpm and up ar
bright red types. This range and sub-
division may, of course, be adapted
for the specific type of engine.

The lower row of LEDs may be used
to indicate a number of fixed rpm
rates on the scale, for instance at
1000 rpm intervals.

The PCB track layout and component
overlay with this design should
enable anyone to readily construct
the LED scale revolution counter, but
note that the LEDs are mounted at the
PCB track side to get the correct in-
dication in clockwise direction with
increasing the rpm rate. Also note the
use of the low voltage-drop regulator
IC2 which supplies IC. and IC3 with
a stable, noisefree 10 V rail

IC3 = U1096B
(Telefunkenl
Di.. 0eo= LEO

AS

audio-controlled
mains switch

It is often useful for audio or video drives transistor T: . If the output of sufficient to reset the 555. The
equipment to be switched off auto- either ICi or IC2 is logic 1. Ti monostable may also be reset by
matically after there has been no in- conducts. closing Si, which connects pin 6 of

put signal for a while. The 555 operates as a retriggerable the 555 to + 12 V.

The function of the on-off switch in monostable, whose period is deter- When IC3 is reset, Ci is discharged
such equipment is then taken over by mined by R; and C The device is via its pin 7. Resistor R11 serves as
switch S2 in the accompanying triggered when its pin 2 is earthed by protection, because without it Ti
diagram. It remains, however, poss- the closing of S2. Its output, pin 3. could short-circuit the supply lines,
ible to switch off manually by means then remains high for 1 to 2 minutes, When the output of IC3 goes high, T2
of Si. Automatic switch-off occurs depending on the leakage current of conducts, the relay is energized, and
after there has been no input signal the 555. The monostable resets itself the relay contacts switch on the
for about 2 minutes: this delay makes as soon as the potential across Ci ex- mains voltage as appropriate. To
it possible for a new record or ceeds a certain value. As long as counter the induced potential when
cassette to be placed in the relevant there is an input signal to the circuit, the relay contacts close, which could
machine. Ti conducts and C: remains un- damage T2, diode Di has been con-

The audio input to the proposed cir- charged. As soon as the audio signal nected in parallel with the relay coil,
cuit may be taken from the output of ceases. Ti switches off, and Ci (Sv)

the relevant TV set, amplifier, or charges until the potential across it is
whatever. The input earth is held at
+ 6 V with respect to the circuit earth
by potential divider R1-R2-R3-R4. The
two 741s function as comparators: the
output of ICi goes high when the in-
put signal is greater than +50 mV,
whereas the output of IC2 goes high
when the input signal becomes more
negative than —50 mV. Resistors R6,

Rv, and R« form an OR gate that

190!

Quite arguably, Nickel Cadmium
(NiCd) batteries are frequently used
as replacements for disposable types
of battery; this is possible because
they can be inserted readily in the
existing battery compartment and
supply the same voltage as dispos-

NiCd battery

able batteries. The fact that one need
not go out to purchase (relatively ex-
pensive) batteries puts the recharge-
able cells in an advantageous pos-

However, one drawback of the use of
rechargeable batteries is the need

chargers

to remove them from the equipment
any time their charge is exhausted. It
would, therefore, be convenient to
leave them where they are, i.e. in the
battery compartment, as they receive
the charge current.

Two circuits are suggested for the in-

1 00 elak.o. in

corporation in existing battery-
operated equipment. Figure 1 shows
the absolute minimum in the form of
a simple current source. The refer-
ence voltage is obtained from the for-
ward drop across LED D. (about

1.5 V for a red LED). R 2 fixes the cur-
rent through the LED, and the voltage
at the base of Ti is therefore about

1.5 V lower than the positive supply
rail. The voltage across R: is about
0.85 V, and this value may be used to
determine the charge current for the
battery, since Ib=0.85Ri, indepen-
dent of the circuit supply voltage.
The value of Ri is thus readily
calculated if it is known that most
NiCd batteries are preferably
charged with a current of one tenth
their capacity in amperes per hour
(Ah). A number of the more popular
battery types and associated values
for Ri have been listed in Table 1.

A noteworthy aspect of the circuit is
the fact that LED Di will go out in the
absence of a battery, since the
voltage across Ri inevitably drops;
the LED current which used to flow
through R2 will now pass through R.
and the base-emitter junction of Ti.
The elaborated version of the NiCd
charger, shown in Fig. 2, includes a
diode to protect the circuit from be-
ing damaged by input voltages

1

T

HP"

_ 2 _

T

having the wrong polarity. R., Rj and
T2 have been incorporated to disable
the charger in the absence of a suffi-
ciently high input voltage; Table 2
lists the relevant values for R2 and R3,
given the number of 1.2 V cells con-
tained in the NiCd battery.

Almost any type of silicon PNP tran-
sistor in the BC series should work
satisfactorily in the T: position if the
charge current does not exceed
about 100 mA. Higher input voltages
and/or charge currents are, however,
better handled by a medium-power
transistor from the BD series.

The input voltage to the charger

need not be elaborately regulated or
smoothed; in fact, any type of inex-
pensive adapter providing the
necessary direct output voltage and
current may be used. Depending on
the number of cells contained in the
NiCd battery, the charge current may
also be obtained from the 12 V car
battery.

The circuits as shown are readily fit-
ted on a small piece of veroboard to
suit incorporation in the relevant
equipment; the input voltage to the
charger is conveniently connected to
a small plug or socket fitted onto the
cabinet. (H)

p,.

charge

Ri

ImA! ^

number of j

V, n I

R2 1

” R3 I

9 V block

no

n

g2

cells !

(min.) |

IQI

IQI

lady RI 1

N

180

18 1

47

2

_ 5

270

22

micro R03

AAA j

180 1

18

47

3 1

6

330 1

27 |

penlight R6

AA

500 |

50

15 1

4

7.5

470 I

39

5 I

9

560 |

47

baby R14 1

C

1200 '

120

6.8 ,

6 1

10

56

180

7 j

12 J

820 |

mono R20 |

D 1

4000

400

2.2 |

voltage inverter

91

Here is a circuit that produces a
negative voltage from a positive one,
for instance, from +5 V to —10 V. The
output voltage. Uo, is determined

Uo = -1.2(Ri/R 2 + 1) [V]

As in other, similar circuits, the maxi-
mum output current depends on the
ratio between input and output
voltage and is calculated from

Ioimax) = 500/(R 1 /Ri + 1) [m A)

The choke is readily made with a

17.5 mm pot core on which 85 turns
of 0.2 mm 2 enamelled copper wire
are close-wound.

The maximum input voltage to the 1C
is 15 V. Efficiency is of the order of
60 per cent. (St)

101

The basic operation of the circuit is
as follows. Activating start switch Si
sets the bistable composed of N 2
and N3, causing N3 to provide clock
pulses to decade counter IC3. The
driver transistors at the outputs of this
1C will light the LEDs one after
another, indicating the countdown
function of the circuit. Rei is ener-
gized the moment the last LED in the
row lights; IC3 is disabled via its CE

input, and the N2-N3 bistable is reset, ranged to sound by means of N<; this
all at the same time. The equipment, function may be disabled by means
powered over the relay contacts, is of S3.

turned on, and the user may take the Rei should not consume more than
desired reading. 100 mA of coil current, while its con-

The power-on interval may be tact(s) should be rated to suit the load
restarted or interrupted during to be switched. fgj

countdown by depressing S 2 , which
resets the bistable and counter.

During the final three stages of the
countdown, a warning buzzer is ar-

pocket frequency meter

941

This easy to construct circuit meets
the demand for a simple, yet versatile
battery operated frequency meter
which can interpret signals with a
minimum rms voltage of 10 mV and a
maximum frequency of 100 kHz. The
quiescent current consumption of
the meter circuit is only 4 mA, which
ensures a long life for the 9 V battery.
Also of interest is the fact that the cir-
I cuit continues to work normally with

battery voltages down to about 5 V.
The meter input is protected up to
250 V AC.

From the circuit diagram it is seen
that the input signal is applied to the
gate of Ti via Ri and C; . C: is an ad-
ditional speed-up capacitor to im-
prove the response at higher fre-
quencies, while anti-parallel con-
nected diodes Di and Dz protect the
FET gate from high voltage surges.

T: functions as a buffer, preceding
Schmitt trigger A:, which has been
dimensioned for a relatively low hys-
teresis of about 18 mV to prevent the
overall sensitivity being too strongly
degraded. The output of Ai is fed
direct to divide-by-two counter ICia,
which is followed by three cascaded
divide-by-ten IC sections. S 2 selects
the divisor and hence the relevant
frequency range. Whatever range is

103

selected, a frequency of 50 Hz at the
pole of Sz corresponds to a full scale
reading on moving coil meter M.
The signal at the pole of Sz is used to
trigger the monostable built around
the Type 7555 low-power precision
timer. The correct operation of this
circuit section can only be achieved
if the time period of the monostable
is less than half that of the full scale
frequency, i.e. ’/z(l/50>s = 10 ms.
Therefore, a monostable time of 8 ms
is used in the proposed configur-
ation.

The output signal from IC3 has a

duty factor which is proportional to
the input signal frequency. The
pulses from ICs are levelled at
2.5 Vpp by 1C;, before being in-
tegrated by R12 and Ce to produce a
direct voltage which is proportional
to the input frequency. The circuit
around A ; and Tz is a simple voltage-
to-current convener with the 100 m A
moving coil meter connected in the
collector supply line to T2. Cs may
be added to stabilize the read-out at
the lower end of the scale.

Though a Type LM393 opamp has
been used, the less expensive Type

LM339 also works all right, provided
the inputs of the unused opamps
contained in this chip are tied to the
positive supply rail to minimize their
power consumption.

The frequency meter is so sensitive
that merely touching the input ter-
minal with a finger causes the meter
to read the mains frequency. This is,
incidentally, a convenient method of
calibration, since Pi may be set to
give a reading in accordance with
the local mains frequency, which is
usually stable within 1%.

CD

951

Stepper motors have either unipolar
or bipolar stators. In unipolar models,
each stator winding has a centre tap.
which enables the magnetic field to
be inverted by switching from one to
the other half of the winding. Bipolar
types have a single stator winding, so
that the direction of the current
through it must be changed to attain
inversion of the magnetic field. From
this, it is clear that, given that the two
motors are of similar size, the bipolar
type will provide a larger couple
than the unipolar model. There is,
however, a price to be paid for this
larger couple: the drive of a bipolar
motor is more complex than that of a
unipolar type.

The drive for bipolar motors may, in

104 .,ekto.iodi.A U 9.Sep> 1986

current drive for
stepper motors

principle, be obtained by means of a

• full bridge circuit, i.e. four tran-
sistors per stator winding;

• half bridge circuit and dual power
supply, i.e. two transistors per
stator winding;

• half bridge circuit with large out-
put capacitor.

The last method is totally unsuitable
for low stepping frequencies or
stand-still. Of the other two, the half
bridge is to be preferred in most
cases, in spite of the requirement for
a dual power supply. In this context,
it should be noted that the supply
need not be regulated, since con-
stancy of current is guaranteed by a
zener diode and emitter resistor,
even with variable input voltage. The

value of the smoothing capacitors in
the power supply is determined by
the total stator current, and is a
minimum of 2000 fi F/A.

Values of R. and Rz are given for
various values of stator current in the
table below.

R , & Rz 1 Is
33 Q;5 W | 100 mA
18 S;1 W ■ 200 mA
6Q8;2 W ! 500 mA
3Q3;4 W ! 1 A

1

Current drive ensures a higher pull- 2
in rate, i.e. permissible starting fre-
quency, because commutation is
quicker with an inductive stator
winding.

The higher the supply voltage, the
more effective the drive, but also, un-
fortunately, the dissipation in Ti and
T2. In practice, a 2x12 V or 2x18V
mains transformer has proved very
satisfactory. Note that freewheeling
diodes have been included in the
darlington circuit to give a good
measure of protection against high
induced voltages caused by
switching.

The prototype was used in the first in-
stance for the control of four-phase
stepper motors via an eight-bit output
port of a microprocessor system. The
interface used to obtain TTL levels
was a Type 7407 which has 30 V
open-collector outputs. The control
instructions may be generated as
shown in the table.

If the stepper motor is required to be

used on its own, this may be done
with the aid of commercially
available control ICs such as the
SAA1027 or the TEA1012. The latter is
dealt with in Circuit No. 97 in this
issue and may be connected as
shown in Fig. 2. (TW)

Phase

Bit

Output byte
Auxiliary byte
New O/P byte
Rotate aux. byte
twice*

12 3 4

7 6 5 4 3 2 1 0
10 10 10 10 initial position
0 0 0 0 0 0 1 1 XOR with output byte
10 10 10 0 1 made one step
0 0 0 0 1 1 0 0 preset for next step

*Direction of turning determines rotational direction
of motor.

105

1

QC tuneable active
Vv aerial for SW

Many of the modern, synthesizer-
tuned, general coverage SW re-
ceivers incorporate the latest types
of high dynamic range RF prestages
and mixer devices, while the good
old tuneable preselector stage
seems to have been eradicated in all
but the most expensive and sophisti-
cated types of multi-mode receiver.
It would seem as if manufacturers
associate a simple tuning control
with an attack on user friendliness of
the receiver, while a well-designed,
tracked or individually controllable
input attenuator would have been a
better solution to the problems
caused by the worldwide escalation
of SW transmitter output levels.

A likewise argued plea for reestab-
lishing the tuning control could be
entered for the active aerial which,
while not able to offer the perform-
ance of a long wire or multi-band
beam aerial, is none the less gener-
ally recognized as a satisfactory
means for receiving broadcast pro-
grammes in the SW bands up to
about 15 MHz.

As generally known, an active aerial
is composed of an aerial proper and
associated amplifier. As to the latter,
the ciruit diagram shows that the
design has a varactor-tuned, sym-
metrical input using two FETs Type
BF256C which are fed over the coax
cable to the receiver. Opamp ICi
functions as a fast symmetrical to
asymmetrical converter capable of

Table 1

kHz

»H

T

3

150

2200

32

51

71

0.5 !

350

390

!L

1000

47

6_

n

0.5

2000

12

2

3

7

4000

3.9

L_

05

operation up to about 30 MHz. Note
that the varicap diode set is tuned
over a separate cable; twin-lead 75 Q
coax cable is, of course, ideal for
the present purpose. The indicated
varicap set ensures a tuning ratio of
about 1:2 to 1:3.

When constructing the aerial to this
design, it should be noted that
neither the circumference of the
loop aerial nor the total length of the
dipole must be in excess of one tenth
of the relevant wavelength in order to
ensure the correct directivity
characteristics, especially in the case
of the loop aerial; the dipole will
typically fail to match the amplifier
input impedance and thus cause

2

problems in getting the device tuned
properly.

Table 1 summarizes the aerial con-
struction data, given a number of
possible operating frequencies. The
aerial should be mounted in such a
position as to receive a minimum
amount of man-made, short range
interference; the amplifier's sym-
metrical input should ensure suf-
ficient aerial directivity to find a dip
for the interfering source.

The loop aerial is uncritical as to the
height above ground, but not so the
dipole, which is bound to act as a ver-
tical rather than horizontal aerial
when mounted at less than a quarter
wavelength above ground. (B)

106 eleklor india Aug Sep! 1986

stepper motor control

The control of stepper motors is
not simple, particularly when no
specially designed control circuit is
used. The Type TEA1012 is an inte-
grated stepper motor controller that
can cope with most if not all situ-
ations. In addition to controlling the
phases for whole and half steps, it
also sets the current with the aid of
these phases.

The TEA1012 was specially designed
for the control of unipolar stepper
motors, in which the current passes
through the stator windings in one
direction. Because the windings

971

behave inductively, the current
through them will become too large
when the stepping speed is low. The
reason for this is that in that situation
only the ohmic resistance, which is
fairly small, determines the value of
the current. To limit the current, a
limiting circuit is connected in series
with the windings. In the diagram,
the current through Li and Ls is
restricted to 0.3/Ri, and that through
La and La to 0.3/Rn. This enables the
current through the stator windings
to be adapted to any type of motor.
The table shows in what sequence

the various phases are driven with
full and half step control, as well as
for clockwise and anticlockwise con-
trol. The stepper motor is arrested in
the position it occupies with the
STOP input. CL is the clock input: for
each pulse, the motor turns one step
forwards or one step backwar ds.
Because inputs^ CL, STOP,
CCW/CW, and F/H all are TTL com-
patible, it is not difficult to connect
these controls to a computer. Re-
sistors Rii to R« incl. and the associ-
ated switches, enable the circuit to
be manually provided with control

data.

The maximum stepping speed de-
pends on the type of motor and on
switch-off time-constants Toaro and

Letters CW and CCW signify clock-
wise and anticlockwise respectively,
while input F/H enables choosing
whole (F) or half (H) steps. A double
resolution is, therefore, possible.

The supply voltage of the IC may be
between 4.5 V and 15 V. The outputs
of the TEA1012 are open-collector, so
that the operating voltage of the step-
per motor may be made indepen-
dent of the supply voltage to the IC.

(St)

98

The name jumbo dimmer points to its
association with the Jumbo Display
(see August September 1985), but it
can, of course, also be used with
other appliances such as lamps,
pumps, ventilators: in short for all ap-
plications where a direct voltage is to
be controlled by pulse duration
modulation.

With reference to the diagram, Ai is
a rectangular-wave generator: a
useful by-product of this stage is the
(quasi) triangular voltage at its invert-
ing input. This signal is applied to the
non-inverting input of comparator
A>. The reference voltage for this
stage is derived from preset Pi. The
output of the comparator is a rec-
tangular voltage with a frequency of
around 200 Hz and a pulse duration
that is variable between nought and
100 per cent. The onset point of the
pulses is determined by the setting
of Pi. The actual control function is
provided by transistor Ti, which
switches the relatively large display

jumbo dimmer

staircase light controller

This circuit has been designed to
function as an automatic switch-off fa-
cility on the lines of the well-known
hotel switch circuit, i.e. the combi-
nation of two switches and a single
light. While not exactly a replace-
ment of any of the two changeover
switches at the top and the bottom of
the stairs, the proposed controller
may be fitted into one of the relevant
junction boxes in which a live mains
line is available.

The circuit diagram shows that the
controller is fed direct off the mains.
C3 and R21 create a suitable series
impedance which charges Ci to
6.8 V by means of rectifier De and
zener diode D?. Set-reset bistable
T3-T4 keeps track of the position of
S2, which determines which of the
two triacs is to be driven so as to turn
the light on. Any time S2 is operated,
timer ICi is started by means of
Ci-Ru, C2-R1S, Ni, N2 and N3; the out-

put of the latter goes high in this con-
dition, resetting IC: and causing it to
pull all of its counter outputs low.
Note that the reset condition can also
be forced by depressing S
FET Ts is turned off at reset, and
SO Hz clock pulses are applied to the
<t>i (clock input) terminal of ICi. Any
one of the five timer outpus Ob. . .Qi3
may be wired to the inputs of gate N-i
to select the desired on-time for the
light; longer intervals may be re-
alized by adding a further counter.
When the selected light-on interval
has lapsed. Ts conducts and
disables ICi from receiving clock
pulses; the counter state is thus
frozen until a reset pulse is applied at
terminal 12. Finally, T: and T : provide
DC control of the relevant triac, while
AND gate simulators Di-Ds-Rj and
D2-D4-R4 ensure the correct selection
of Trii or Tri2 to power the bulb.

The circuit is readily constructed on

a piece of veroboard and fitted into
an ABS mains wiring junction box, as
a replacement of one of the switches
in the hotel circuit.

As many points in the circuit are at
mains potential, due precautions
should be taken in the construction
and wiring of the controller. Note that
S; should be rated at 240 V AC, in
view of the necessary isolation with
respect to the mains voltage.

Trii and Tri2 require no heatsinks if
the bulb is rated at 100 W or less,
while the maximum power rating for
the triacs is about 400 W. (Sv)

109

listen-in key
for data recorders

© -r 1

The pros and cons of using data
(cassette) recorders for mass
memory storage in a computer
system are likely to be so well-known
that any further discussion as to the
relative cost efficiency of the
cassette tape would seem to be
superfluous.

There is, however, one distinct disad-
vantage to the data recorder that is
relatively easy to get rid of, viz. the
trouble many users experience in
positioning the tape to the leader
note of the desired program or file to
load into the computer. Many
datarecorders, while offering the
highest possible save and load
speed, fail to produce the sound on
tape when the computer audio cable
is plugged into the earphone socket,
forcing the user to plug and unplug
this cable in a desperate search for
the program.

The solution to this sorry plight con-
sists of a simple combination of
resistor and push to make button,
which are to be built into the cassette
recorder. The circuit diagram shows
the method of connecting these

parts; pressing the button with the
earphone plug inserted in the socket
will enable the user to listen to the
recorded data as the tape is played.
The value of the resistor may have to
be adapted to suit the specific output
power of the data recorder, given the
optimum playback level for the
computer.

Now that you have opened the
recorder for the outlined modifi-

cation, it is just as well to mount a sec-
ond button enabling tapes to be
winded and played while the remote
control plug rests inserted in the
associated socket; this simple
modification may also be of
appreciable interest for the
improved efficiency in locating files
on tape.

AS

101

40-track adaptor

Over the past few years, the cost of
5 Zt inch floppy disk drives has gone
down to the extent that modern,
80-track, double-sided drives now
cost less than a simple, 40-track,
single-sided type some three years
ago. It is, therefore, not surprising to
see many computer owners upgrade
their systems with a set of 80-track,
slim-line drives to boost the mass
storage capacity of their micro.
However, 40-track stored programs
are not readily retrievable in the new
system, because the distance be-
tween tracks in the 40-track drive is
twice that in the 80-track model.

This circuit offers a solution to the
problem, in that it doubles the step
distance for the R/W head in the
80-track disk drive, so as to make it
"look like" a 40-track type to the
computer which should, of course,
be programmed with a 40-track disk
operating system (DOS).

It is seen from the circuit diagram

110 elektor India Aug/Sept 1986

that Gate N. receives the FDC con-
troller STEP pulse, which is used in
the circuit as a timing reference for
the automatic generation of another
STEP pulse to follow the first after
3 ms.

It should be noted that, when incor-

porating the circuit in an 80-track
drive, the track-to-track access time
in the 40-track mode is double that as
given in the drive specifications,
which refer to 80-track use.

HS

CPU gear-box

102

i

2

While many computer enthusiasts
are keen on getting their system to
run at the highest possible clock
speed, there are often quite awkward
constraints posed by relatively slow,
bus-connected support chips, and
the ensuing frustration after failing to
get reliable system operation at, say.
double the ‘old’ clock speed may
readily lead to abandoning the
speed-up project altogether, for lack
of precise information regarding the
necessary clock-based synchroniz-
ation between CPU and peripheral
chip(s).

A noteworthy example of this hap-
pening in practice is the go at incor-
poration of the Type 9367 CRT
controller in a 6502-based computer
system running at 2 MHz; the specific
application concerns the high-
resolution graphics card published
in Elektor Electronics,

This circuit ensures a correctly
timed, synchronized slow-down of
the system clock speed, when ap-
propriate for CPU access to a mem-
ory-mapped (E150-E15F) device.
Following the reception of a high
level on the relevant I/O line, the pro-
posed circuit arranges for the clock
signal frequency to be divided by
two, while a low I/O causes division
by four.

It is important to point out why the
commonly used method of using <t>2
to enable the address decoder chip
is to no avail when it comes to
synchronous and glitch-free clock
speed switching under software con-
trol; the following paragraphs
therefore aim at offering an insight
into the basic operation of the gear-
box circuit and its incorporation in a
6502-plus-graphics card system.
Figure 1 shows the hardware Jo^the
gear-box. A logic level at the I/O in-
put is passed to the D (data)input of
bistable FFa, as well as to the R
(reset) input _of FFa. FF3 toggles and
activates its Q output; this causes the
4 MHz clock signal, divided by two in
FFa, to be output as 2 MHz towards
the CPU <t>m terminal. Division by
four (1 MHz clock output) should take
place in a synchronous timing ar-
rangement as soon as I/O goes low;
just prior to this pulse transition, ®m
has already gone low, so that the
level change at the FF< reset input is
of no consequence to the CPU oper-
ation at that time, however the
bistable can not change state
anymore. Thus, FF3 will have to

supply the output clock signal; the D
input lollows the I/O signal tran-
sitions. since 0 of FF2 was forced to
go low in consequence of S (set)
being activated. The first leading
edge coming from the FF; Q output
will cause 0 of FF3 to go logic high,
ending the set condition of FF.
Given an input clock frequency of
4 MHz. thejmtlined timing sequence
results in Q of FF? going high after
250 ns, followed by a low level at 0 of
FF3 after another 250 ns. The timing
diagram shown in Fig. 2 clarifies this,
admittedly rather complicated,
timing arrangement in the gear-box
circuit. It is noted that a complete
1 MHz period has lapsed, provided
FF 1 is properly synchronized during
the CPU initialisation cycle.
Theoretical research into this matter,
however, has shown that this is not
always the case; the result is an asym-
metrical output clock period with a
logic low and high level duration of
250 and 500 ns respectively. The
remedy for this undesirable effect is
simple, since it merely involves inter-

changing the clock signal connec-
tions to FF; and FF).

It is seen that ®2-based I/O decoding
is less desirable, since it involves too
long a delay; what remains is to indi-
cate the method of obtaining I/O
from the graphics card system (EE,
Nove mber 1985, p. 71).

XX5X is dis misse d for now obvious
reasons, but P = Q at pin 19 of ICi can
be used for our purpose, while the
possible objections to the resultant,
rather coarse address decoding are
readily rendered devoid of relevance
by the incorporation of a single 3-to-8
decoder Type 74LS138, mounted
piggy-back onto IC2 and connected
direct to pins 1. . .5, 16 and 8. The re-
maining pins of the additional IC are
either cut off or bent to preclude
wrong contacts from being made in
the circuit. However leave pins 6 and
10 in function, since the former
should be tied permanently to + 5V
(small wire to pin 16), while the latter
can now be used to supply the cor-
rect I/O pulse for the CPU gear-box.

(D)

11

103

musical greeting cards

The designer of this circuit will
readily admit that it is literally not
much to make a song or dance about,
since what is shown as the circuit
diagram speaks (sings) for itself.
Available in about 30 different song
versions, the Type UM3166-xx is a

fully autonomous melody generator
chip which operates at extremely
low battery voltages (1.3... 3 V),
while capable of directly driving a
small piezo-buzzer from antiphase
output terminals 2 and 4. If you wish,
you may connect an AF amplifier to

either of these pins in order that
more listeners may be captured by
the melody selected from the accom-
panying table. The melody may be
played continuously by connecting
terminal 3 to 7 rather than 1. (St)

Table

TYPE

MELODY

TYPE

MELODY

UM3166 1

JINGLE BELLS • SANTA CLAUS IS COMING

UM3166 16

TOMORROW

TO TOWN • WE WISH YOU A MERRY

XMAS

UM3166 17

WE WISH YOU A MERRY X MAS *

SILENT NIGHT

UM3166 2

JINGLE BELLS

UM3166 18

WEDDING MARCH (WAGNERI

UM3166 3

SILENT NIGHT

UM3166 19

FOR ELISE

JINGLE BELLS * RUDOLPH, THE RED NOSED

UM3166 20 ' WHEN THE SAINTS GO MARCHING IN

REINDEER - JOY TO THE WORLD

UM3166-21

CONGRATULATION * HAPPY BIRTHDAY

UM3166 5

HOME SWEET HOME

UM3166 22

JINGLE BELLS (NEW VERSIONI

UM3166 6

LET ME CALL YOU SWEET HEART

UM3166 23 ! IF YOU LOVE ME

UM3166 7

CONGRATULATIONS

UM3166 24 ! TWINKLE TWINKLE LITTLE STAR

UM3166 8

HAPPY BIRTHDAY TO YOU

UM3166 25

MARCH OF TOY SOLDIER

WEDDING MARCH (MENDELSSOHN

U M3 166-26

ROCKABYE BABY

UM3166 10

1 WILL FOLLOW HIM

UM3166 27

CHORAL SYMPHONY (BEETHOVEN

LOVE ME TENDER. LOVE ME TRUE

SYMPHONY NO. 9)

SUCH A WONDERFUL DAY

U M3 166-28

HAPPY BIRTHDAY TO YOU

EASTER PARADE

(NEW VERSIONI

GRADUATION MARCH

UM3166 29

BLUE BELLS OF SCOTLAND

UM3166 15

ALOHA OE

UM3166 31

LULLABY (SCHUBERT)

104

DC operated 50 Hz
timebase

Many clocks, both of the digital and
the analogue type, make use of a
SO Hz timebase signal which is
usually derived from the mains. In
order that these clocks may also
work in places were there is no
mains supply available, as in cars, on
boats, or, say, on a camping site, this
one-chip circuit provides an accurate
50 Hz square wave output signal,

® ,1 , 1 . 1.1 . 1 , 1 .13

3.2768MHI

while being fed off any DC supply
voltage between 6 and IS V (battery,
solar cell array, etc.). Current con-
sumption of the circuit is only 3 mA
(max.).

The Type SAF0300 by ITT Semicon-
ductors merely requires a crystal to
perform the above task, while also of-
fering the possibility to adjust the
exact output frequency by means of
seven active low bits as listed in the
pin assignment table.

If a 64 Hz output frequency is desired
rather than SO Hz, the crystal may be
replaced with a 4.194812 MHz type.

1 Output 1 150 Hz)

2 Adjustment pin 122 ppm

3 Adjustment pin 61 ppm

4 Adjustment pin 30.5 ppm

5 Adjustment pin 15 ppm

6 Adjustment pin 7.6 ppm

7 Adjustment pin 3.8 ppm

8 Adjustment pin 1 .9 ppm

9 Test pin M (F x /4)

0 Crystal connection

2 Bridge output

3 Bridge output

4 Ground, 0

5 leave vacant!

6 Supply voltage

Finally, the SO (64) Hz output pulse supply voltage, and a duty factor of
has a voltage swing of nearly the IC 0.5. HS

loudspeaker protection

This is an all-transistor design for
incorporation in AF amplifiers that
produce nasty clicks in the loud-
speakers when turned on or off,
jeopardizing the voice coils by pass-
ing a large current surge.

Assuming that AF amplifier and pro-
tection circuit are off. C> and C2 are
empty of charge and Re is deac-
tivated. At power-on. Di rapidly
charges Ci. Provided both the
negative and the positive supply
voltage are present and at the correct
level, T2 and T3 conduct, while T. is
off, enabling C2 to be slowly charged
via R4. If the voltage across C2 is suf-
ficiently high for T< to conduct, Ts
will draw base current and energize
Re, which connects the loudspeakers
to the amplifier outputs. Zener diode
D4 fixes the voltage across the coil of

Re, so that differently rated relays
may also be used in the circuit,
provided D4 is changed accord-
ingly. However the relay coil current
should not exceed about 50 mA,
while the changeover contacts
should be rated in accordance with
the amplifier output power and im-
pedance; for a 2 x 100 W at 8 Q type,
for instance, the relay contacts
should be rated at least 8 A.

Should either one or both supply
voltages (— Ub; +Ub) disappear for
some reason or other (amplifier mal-
function, short-circuited smoothing
capacitor, etc.), the relevant transistor
T2 or T3 will be disabled, causing T>
to receive base current via R-; C2 will
be discharged forthwith and Re is
deactivated in consequence since T4
and Ts are turned off. The amplifier

channels can now produce clicks
they like; the output is safely applied
to two resistors matching the output
impedance.

The protection circuit is fed off the
voltage across Ci, which is purposely
rated at only 100 jiF to enable Re to be
deactivated almost immediately after
the amplifier has been switched off.
Power-off clicks, if produced, will
therefore end up in the dummy re-
sistors rather than the expensive
loudspeaker voice coils.

The protection unit is most readily fit-
ted on a piece of veroboard, while Re
should be mounted close to the loud-
speaker output terminals to keep
contact losses as low as possible.

AS

113

106

Both safe and remarkably simple to
construct, this circuit detects the
zero crossing moments of the mains
voltage, in order to provide other cir-
cuitry with timing information about
the correct instant for switching
mains-connected loads; in other
words, when the least possible
switching dissipation is involved,
and, therefore, least interference is
induced on the mains lines.

The proposed circuit operates direct
off the mains, while comprising no
more than two opto-couplers and two
resistors. It is seen that photodiodes
Di and D2 are connected in anti-
parallel while being fed with the
mains voltage via a resistor, which
limits the current through the rel-
evant diode to about 2 mA as it con-
ducts (i.e. lights) during the negative
or the positive half wave (D2 or Di
respectively) of the mains sinewave;
in either case, the circuit output
voltage is low, since the associated
phototransistor conducts and draws
current from +Ub via R2.

107

A calibration generator is of par-
ticular use with many older gener-
ation receivers, which have no, or a
poor, frequency read-out. However,
the RF section of these receivers is
invariably far superior to that of most
modern models, and consequently
there are still many of these 'oldies' in

The circuit in the accompanying
diagram provides calibration signals
at multiples of 100 kHz and 1 MHz, all
of which are available simul-
taneously, so that no switching is
necessary.

The output signal of the crystal oscil-
lator, Ti, is divided by 10 in ICi.
Astable Ni operates at a frequency
of around 22 Hz, which is low enough
to allow zero beat tuning even in SSB
operation. The 100 kHz harmonics
sound (on AM) like a sort of wood-
pecker.

Astable N3 operates at about 1.5 kHz
and is gated with the 22 Hz signal.
Consequently, the 1 MHz signal ap-
pears for 22 ms as a carrier wave.

mains zero-crossing
detector

However, at the moment of zero
crossing, neither one of the diodes
conducts, and the voltage at the cir-
cuit output rises to near +Ub level,
whence the 100 Hz pulse train.

The value of Ra may be adapted to
suit the level of + Ub and the manu-
facturer-specified typical collector

current through the phototransistor.
For the Type TIL1U, the current
should not exceed about 50 mA. The
type of optocoupler used in the cir-
cuit should not be very critical, but
the value of Ri had best be left at the
indicated 100 k so as not to run into
excessive diode dissipation. CN

calibration generator

which is modulated with the 1.5 kHz The circuit is usable up to 30 MHz
signal during the next 22 ms. This when CMOS devices are used, and
signal is also easily tuned for zero up to around 300 MHz with HCMOS
beat. ICs. (B)

ELECTRONIC
| INSTRUMENTATIONS.
\ CONTROL

( j M/S STATIC POWER
SYSTEMS PVT. LTD.

D- 1 48. Bonanza Indl. Estate
Ashok Chakravarthy Road
Kandivali IEI
Bombay-400 001

ROTARY SWITCH WITH
ADJUSTABLE STOP TRS-12
Comtech' TRS-12 is a Rotary
models, FirslL 1 pole 1 2
positions and Third: 4 pole 3

PCB/lnstruments and
transformers. These are
available in 2/3/6/8 ways in
ratings of 5,10,15 A ir, 130L.
130LS types. Some of the
important features of these

NOVOELEX CABLE CARE

SYSTEMS

Post Box No. 9159

range of Push Button Switches.
These switches are compact in
size with Silver plated Phospher
Bronze Contacts. Stainless

US’ ii -

For more details, please co
M/S RL COMPONENTS

115

ALL THAT SIANDS
BETWEEN DARKNESS AND HER
ISYOUR SIGNATURE.

IMAGINE A WORLD OF LENGTHENING
SHADOWS. OF VISIONS FADING INTO
OBLIVION. A WORLD STEADILY GROWING
DARKER. CATARACT. A DISEASE THAT CLOUDS THE
LENS OF THE EVE OBSTRUCTS THE PASSAGE OF LIGHT.
MILLIONS IN OUR COUNTRY. ABOV E THE AGE OF SIXTY ARE
ENGULFED IN THIS FOG. WE. ATLOKKALYAN S \M1TI ARE ASKING YOU TO HELP
"nnH a^ahSaS OV | R -iKf? 00 EYES EVERY YE VR »l H XV I . IKCAMZED 71 EYE! AMI'S IN

LCD Thermometer Issue
No. 26 EPS No. 82 156

Simple! Eosy to assemble. Based on
1C 7 1 06 This thermometer is very
accurate with a temperature range
- 50°C to + 1 50°C. Idea! for house
or laboratory.

Capacitance Meter issue
No. 11 EPS No. 81 1012- 1/2

To measure those elusive farads, this
project features

— Read out on a 3 1/2 digit LCD

— A measuring range from 0. 1 pf to
20 mfd.

— high accuracy.

It is a complete instrument at a
hobbyists price Complete kit
(without cabinet) Rs. 750.00

Digilex Digital Trainer

Teach yourself digital electronics in a
simple, unique manner. Broaden
your horizons and improve your
fundamental understanding of the
changing world of electronics
around you. This trainer gives you an
opportunity to learn electronics at a
very reasonable cost.

Complete trainer at Rs. 325.00

Junior Computer Kit

This kit enables you to have a
practical orientation of the concepts
explained in the Junior Computer
Book I. Get actual 'hands — on'
experience. Leam the basics of
hardware and software in an
innovative step - by - step manner.

And that too at an incredibly low
price-!

Complete kit (without cabinet)

Rs. 1500.00

Elektor Binder — reda.sh

binder collects your loose copies of
Elektor into one handy volume. A

Kits currently available

Issue No. EPS No. Title

33 85110.

36 9827

37 80054

39 86013

9967
85447
85431
85476

simple system enables you to add
each copy as it arrives and to
remove a single copy without
disturbing the others.

Price Rs. 25.00 plus Rs. 5 for Packing
& Postage.

Telephone Exchange

Magnetiser

Talk Funny

Single Trace CRT Convertor
VHF/VHF Modulator
Fault finding probe for Ups.
Hi-fi headphone amplifier
Brakelight Monitor
Electronic Dog

Special offer - 1 0% oft

on Video Amplifier. Stereo Balance Indicator, p

After Burner, Elekterminal Bell, Electronic Pocket Pin Ball game, I
7400 siren, hfe tester Offer valid for Aug/Sept. 86 only. I

Buying one of our kits will save you the frustration of tracking down
those elusive components that hold up your projects. General
information.

Ordering Information

1. All payments in Advance by M.O.RO.orD.D. only.

2. Items will be send by R.PP only

3. Price includes Packing & Postage.

4. For orders out of Maharashtra state, Please add 10% as a 'Incidental Charges'

Send payment to #

precious

* ELECTRONICS CORPORATION

Journal Division 1 1 , Kiln Lane, Off. Lamington Road, Bombay - 400 007.

123

TH€ SIMPl€ (1NSUJ€R FOR
€L€CTRONIC COMPONENTS

^odlac

ELECTRONICS

CORPORATION

• DISPLAYS: 3" 5" & 1" IN COMMON CATHOD &
COMMON ANODE INCLUDING 3V* DIGT LCD.
TANTALUM CAPICITOR5 OF ALL

ZENER DIODES SILICON DIODEsJH
TRANSISTORS PHOTO TRANSiST^B

PHOTODIODES LEDS CRYSTa^^H Wg?

INDENTOR FOR AL^
ELECTRONIC COMPO

ll^Bsof

mponPnts,

n new DrtWn tSKINGS THE

POSSIBILITY OF A HEW PROGRESS.

• 1C s : TTL CMOS. MOS, LSI, LINEAR
MICROPROCESSOR MICRO COMPUTER ere.

• 1C SOCKETS: MOLEX & SMK MAKE INCLUDING
MOLEX PINS & TEXTOOL O INSERTION SOCKETS.

• TRIMPOTS MULTI TURN AS WELL AS SINGLE TURh
OF OOURNS, VRN & BAKEMAN MAKE

-SHPDANT NIWAS
CHUNAM LANE

-RD FLOOR. OFF LAMINGTON ROAD
•ANT-ROAD, BOMBAY-400 007

ino I KU/Mtri I S, GREEN PHOSPHER
AS WELL AS RGB
MONITORS ETC.

-ONES: OFFICE: 369192 5137225.5135845

Our Principles:

M/s. ZIEX TRApiNr HJJiJ Trv
~ Ir^TFlSAPORE 'HONGKONG

i YSS1 I II I*VKS S»vWER»SI»»S niiNTAWES ltMHI W.S

cosmic

WE ARE SOUND

R.N No. 39881/83