1

■ i

n

mm

contents

tup-tun tester

'--i Tester gives an instant check of the 'general health' of a transistor, as well as" its compliance" with' the
-.-.mum TUP or TUN specification, by the very simple procedure of plugging it into test sockets and inter-
prwjr^he messages from two light-emitting diodes. It is also possible to check diodes for excessive capacity

tup-tun-dug-dus

interference suppression in cars

Electrical interference generated by cars can be a source of annoyance, not only to tl
to users of other electronic equipment external to the vehicle. This article discusses sc
causes of interference and their cure.

thief suppression in cars

Thefts of cars, or accessories and/or other articles in them, are becoming more and me
token, anti-theft alarms for cars are becoming more and more of a necessity ....

h-l logic tester

This is a TTL logic probe which, instead of the usual LED to indicate" the logic "states, use’s a "seven-segme'm
Mimtron or LED display to indicate 'H for a high or '1 ' state and 'L' for a low or '0' state.

tap preamp (2)

The first part of this article discussed an audio preamplifier and control unit operated entirely by TAP's and
dealt with the design of the TAP and the electronic switching controlled by the TAP. This month's article deals
scribed 16 appl,catl0n of ,hese circuits to a complete touch-controlled preamp with the facilities already de-

supplies for cars

In order to function effectively, electronic equipment used in cars must have an appropriate power "supply"
which must also suppress interference appearing on the battery voltage from the car electrical systems’

what is cybernetics? — H. Ritz

beetle - Arbeitsgemeinschaft der Hauptschule Rossbach

Beetles, tortoises and the like have often served as models for cybernetic machines which must also have £
reasonable appearance. The beetle described in this article can 'see, hear and feel’ and reacts to information ir
the form of sounds and movements. The animal has a memory and can get tired.

the moth — M. Keul, H. Lohr

quadro in practice

In response to an earlier article on quadrophony ('Quadro 1 -2-3-4 ...', December 19741, we received many
requests for the complete circuit of a quadrophony decoder. Nippon Columbia has now put such a design at our
disposal, so that we can fulfil the wishes of many of our readers.

elektor shorthand

sniff race control — E. Waschkowski .

This tester gives an instant check of the 'general health' of a transistor, as
well as its compliance with the minimum TUP or TUN specification, by
the very simple procedure of plugging it into test sockets and interpreting
the messages from two light-emitting diodes. It is also possible to check
diodes for excessive capacity or leakage.

The principle of operation is simple and
no preliminary calibration is needed
only the use of transistors and diodes
known to be ‘good’ and resistors within
the specified tolerance.

An astable multivibrator generates a
square wave at a frequency of about
2 kHz, and this oscillation is turned on
and off by another multivibrator at
about 2 Hz. The collector-emitter path
of the transistor under test (or the
anode-cathode path of the diode) is
connected in series with another transis-
tor across the supply rails, and the inter-
mittent 2 kHz square wave is fed in anti-
phase to the bases of each of the two
transistors. Figure 1 shows a block dia-
gram of the arrangement, from which a
lot of information about the semi-
conductor under test can be deduced
from the ‘behaviour’, voltage-wise, of
the junction between the two semi-
conductors. This information can be
displayed with the aid of only two light-
emitting diodes (LEDs).

Circuit Description

Figure 2 shows the complete circuit,
which has been divided into three sec-
tions to avoid confusion. Transistors TS
and T6 in figure 2a form an astable
multivibrator which runs at about 2 kHz.
T2 and T3 form another multivibrator
which runs at a much lower speed, about
2 Hz, and turns the ‘fast’ (2 kHz) oscil-
lator on and off through transistor T4,
which also supplies a 2 Hz switching
waveform, via connection ‘Q’, to the dis-
play section T7 ... T9 and LEDs ‘A’ and
‘B’ (figure 2c). A similar 2 Hz switching
waveform, in antiphase to the one which
appears at ‘Q’, is supplied to the display
section by T1 via ‘P\ As will be seen
later, these switching waveforms are

if this is functioning.

2 kHz square waves of equal amplitude j
and opposite polarity are produced inter- I
mittently at the collectors of T5 and T6. I
These two points, which drive the whole I
of the test circuitry, are marked ‘X’ and I
‘Y’ respectively. When the fast oscillator I
is turned off, T5 is cut off and its I
collector (‘X’) is at its higher potential. I
The left-hand half of figure 2b is the I
section in which PNP-transistors are I
tested. It has been shown that 2 kHz I
square waves of equal amplitude and I
opposite polarity are being injected I
intermittently at ‘X’ and ‘Y\

Display

Assume that a (good) PNP transistor is 1
plugged in at the test point Ta in fig- I
ure 2b. When the fast oscillator is off, I
‘X’ is positive and ‘Y’ is negative. (The I
terms ‘positive’ and ‘negative’ are used I
to denote the higher and lower potentials I
taken up by various points in the circuit). I
Both the transistor T 1 0 and the transis- I
tor Ta under test are therefore cut off, I
and the connection joining the collectors I
of T 1 0 and Ta is floating. The diode D 1 0 I
does not pass any current and the I
Darlington pair T 1 1 and T 1 2 is cut off. I
Figures 2b and 2c show that the collector I
of T12 is one of the points connected I
to the base of T9 (point A). When T12 I
is cut off, ‘A’ is positive and T9 is there- (
fore also cut off. LED ‘B’, which is in [
the collector lead of T9, is therefore off, I
and the collector of T9 is negative.

To find what LED ‘A’ is doing, the I
other switching waveforms, derived from I
the slow oscillator via ‘P’ and ‘Q’, must I
now be examined. To switch the fast I
oscillator off, ‘Q’ must be negative;!
therefore ‘P’ is positive. T7 is connected I

ceives a positive drive from the collector!

series with the 680 £2 resistor R9 be- I T8 is ah

also returned to the negative rail!

elektor june 1975 — 609

tTir---c - aad when the fast oscillator
Btanei off, both LEDs are off.
ft kss bees seen that the three points
determine the LED display are
*A\ T astd ’O'. The basic relationship is
S foacrWS

I . Vbes F is positive (i.e. the fast oscil-
kator is turned off), LED ‘A’ will
ijs: up if the base of NPN transis-
tor T7 is driven positively from the
collector of T9.

2_ When Q’ is positive (Le. the fast
oscillator is turned on), LED ‘A’ will
light up if the base of PNP transis-
tor T8 is driven negatively from the
collector of T9.

3. LED ‘B’ lights up when the collector
of T9 is positive, irrespective of
whether ‘P’ or ‘Q’ is positive.

4. When ‘A’ is negative, the collector of
T9 is positive.

These relationships can be combined
in a kind of truth table which will help
in predicting the display for transis-
tors or diodes in different states of
health. They are also summarised, in a
slightly different form, in figure 3a + b.

What happens during the bursts when
the fast oscillator is turned on?

'X and ‘Y’ are being swung alternately
positive and negative with opposite po-
larities at 2 kHz. When ‘X’ swings positive
and ‘Y’ swings negative, the same reason-
ing which was applied to the situation
when the fast oscillator is turned off will
indicate that ‘A’ swings positive and
LED ‘B’ is off. In this case, however, the
fast oscillator is turned on (‘Q’ is there-
fore positive) and LED ‘A’ lights up.
When ‘X’ swings negative and ‘Y’ swings
positive, it will be seen from figure 2b
that both T10 and the transistor under
test in Ta are turned on. The emitter of
Ta is directly connected to supply
positive, while the emitter of T10 is con-
nected to supply negative through the
470 S2 resistor R28. If the current
gain of Ta is high enough, the potential
at the collector of Ta will move positive-
ly , D 1 0 will conduct and the base of T 1 1
will also move positively. (This will be
discussed in more detail later.) The
emitter of T12, the other transistor in
the Darlington pair, is held by R30 and
R3 1 at half the supply rail potential, so
T12 is turned on; its collector potential
(point ‘A’) swings negative and, as can
be seen from the table, LED'B’ lights
up and LED ‘A’ is off.

So the LED display while the fast oscil-
lator is turned on and the transistor is a
•good’ one is that ‘A’ and ‘B’ each come
on during alternate half-cycles of the
2 kHz oscillation. Both LEDs therefore
appear to be on during each 2 kHz burst,
and it has already been seen that both
are off while the fast oscillation is turned
off.

Figure 1. Block diagram of the arrangement
for testing a PNP transistor. For clarity, the
breakdown voltage test and the complemen-
tary test for an NPN transistor have been
omitted.

Figure 2. Complete circuit of the TUP/TUN
tester. Block A is the collector section, B con -
tains the test bridges for NPN and PNP transis-
tors and C shows the breakdown voltage testing
and display sections.

The full display cycle for a ‘good’ tran-
sistor is that both LEDs blink on and off
together (figure 3c). It will be seen later
that this display occurs only with a
transistor which is good according to all
the criteria that are tested in socket T A -

Transistor with low current gain
(a')

When the fast oscillator is turned off,

610 — elektor june 1975

tup-tun

‘X’ swings positive and *Y’ swings nega-
tive, so both T10 and the transistor under
test in Ta are cut off. Their commoned
collectors are floating, and by the same
sequence of events as described for a
good transistor, the voltage at the col-
lector of T9 is low and LED ‘B’ is off. It
can be deduced from the table that this
combination of switching voltages leads
to LED ‘A’ also being off.

When the fast oscillator comes on and
swings ‘X’ and ‘Y’ negative and positive
respectively, T1 0 and Ta are both turned
on. The potential at the base of T10 is
therefore determined by the potentio-
meter R1 5 (figure la), R26 and R27, Le.

20

33

4.7+ 120 + 33

= 4.2 V.

The base-emitter voltage drop in T 1 0 will
be about 0.7 V, so the voltage at the
emitter of T10 cannot rise above
4.2 V - 0.7 V = 3.5 V. T10 is therefore
acting as a current source, its collector
current being stabilised at the value
determined by this latter voltage and the
emitter resistor R28, i.e.

3.5 x 1000
470

mA as 7.4 mA

As the emitter of Ta is directly con-
nected to the positive supply rail, its
base current is determined by the voltage
(about 19 V) between ‘X’ and the posi-
tive rail, and by R25, i.e.

18 x 10 6

270 x 10=" A " s70M
(the base-emitter resistance can be dis-
regarded in this context).

It has been mentioned that T10 acts as a
current source attempting to stabilise the
collector current through both transis-
tors at 7.4 mA, which corresponds to a
current gain of something over 100 for
the transistor under test. If Ta cannot
produce this current, T6 bottoms and
the voltage at the connected collectors
of Ta and T10 becomes too low for T1 1
and T12 to be turned on (figure 3d). So
the potential at ‘A’ remains positive and
LED ‘B' stays off. The table will show
that LED ‘A’ comes on.

When the fast oscillator swings to its
other polarity (i.e. ‘X’ swings positive
and ‘Y’ swings negative) the linked col-
lectors of Ta and T10 revert to the float-
ing condition, so that the Darlington
pair Til and T12 remains non-conduc-
tive and ‘A’ positive. LED ‘B’ therefore
stays off and LED ‘A’ stays on.
Summarising: the LED display with a
transistor of low current gain is that
LED ‘A’ blinks and LED ‘B’ stays on.

Transistor with high capacitances

When the fast oscillator is turned off,
the situation is the same as in both the
cases already examined: T10 and the
transistor under test are both cut off,
and this leads to LED ‘A’ and LED ‘B’

both being off. When the fast oscillator
comes on and swings '.V negatively and
‘Y’ positively, both transistors are turned
on, but if Ta has high collector-to-base
(Cgb) and/or collec:or-to-emitter (Cce)
capacitance, its response is delayed. The
voltage rise at its collector is slowed
down as these capacitances discharge, |
but the voltage will probably level off at |
its ‘final’ value before the end of the
period in which T A is turned on, and
when this happens LED B' comes on I
while LED ‘A’ stays off i figure 3e).

When, however. T A and T10 are once I
more turned off by the swings at ‘X’ and I
‘Y’, the capacitances can recharge only I
through the Darlington pair T 1 1 and T12 I
(which has, by definition, a high input I
impedance) and through the 10 Mf2 re- I
sistor R29. The drop in potential at the I
collector of Ta- as the capacitances
recharge, is slower than it would be with
a normal transistor, and if the capaci-
tances are too large the potential will not
fall far enough to turn Til off (and
therefore LED 'A' on and LED’B’ off)
before the time when Ta and T10 are
turned on once again. So LED ‘B’ will I
stay on, and LED ‘A’ off, throughout |
each period when the fast oscillator is I
turned on.

With slightly smaller capacitances, I
LED ‘A’ may come on dimly if the slow I
recharge of excess capacitance only I
allows this LED to turn on for a small I

Figure 3. Summary of LED displays, based
on the waveforms at the collector of the
transistor under test

Figure 4. Transistor testing chart, showing what
the various displays signify. This chart is
derived from figure 3.

Figure 5. Transistor lor cfio del breakdown test
chart.

Figure 6. The condu c tion test for diodes, in
the ■RNP’ test socket is shoeei in figure 6A.
The leakage test for dodes is shown in
figure 6B.

Figure 7. Tests to ■

A 1 LED B

MEANING

-3

Good transistor

&

a. PNP/NPN reversed b. Leak > 1 «A
c. C-B short d. C-E short

►

•

a. o' <100

b. BE short

►

-:<3

C-B or C-E capacitance > 20p

Y

Leak > 10«A + very low o' + large Ccb

DISPLAY |

LED A

LED B

-:<3

;<3

•

Kev to display as for figure'

Impossible. If this happens something
is wrong with the tester. Check the
power supply!

- = On continuously
Q = Blinks dimly

7

TEST APPLIED

DISPLAY |

1 iillMJ

1. Nothing plugged into any test socket

9D

*arBdfaTA<»T £* B ’ in " J "° C0 “

3. PNP or NPN transistor known to be good, correctly
plugged into T* or Tees applicable

4 As previous test, but with a 22 p capacitor connected

nm

5. Emitter and base leads ol a PNP or NPN tr'amlitS;

UIU

—BMW

known to be good connected to the collector and

Hi

part of each fast-oscillator cycle.
Summarising again: the display for high
capacitance is that LED ‘B’ blinks on
and off while LED ‘A’ remains off or
blinks dimly.

Transistor with high leakage

A transistor with high leakage current
tends to behave, from the tester’s point
of view, as though it were turned on all
the time. In all the cases examined so far,
no collector current flows in the transis-
tor under test while the fast oscillator
is turned off. If, however, there is a
leakage current between collector and

emitter, this will flow through DIO and
R29 to the negative rail even when ‘X’ is
positive and both Ta and T10 are sup-
posed to be cut off. This leakage current
develops a voltage across the 1 0 MS2 re-
sistor R29, and therefore raises the
potential at the base of T1 1-

It will be recalled that the emitter of T1 2
is held at half the supply voltage (i.e. at
about 1 0 volts) by the potentio-
meter R30 and R31. So if the leakage
current is a little more than 1 iiA, it will
build up a voltage sufficient to turn on
T1 1 and T12 and thus light up LED ‘A’
and LED ‘B’ while the fast oscillator is

off. When the fast oscillator is turned
on and the display transistors are
switched through ‘P’ and ‘Q’, LED ‘B’
stays on but LED ‘A’ goes out (figure 3f).
So with a transistor having a leakage
current of 1 /iA or more, LED ‘B’ stays
on and LED ‘A’ flashes.

Transistor with base and collector
or emitter and collector short-
circuited

A transistor with one of these faults
'looks like’ one with high leakage (only
more so). A current can flow from the
positive rail through the emitter-base

612 - elektor june 1975

tup-tun

junction and the base-collector short in
Ta (or directly through the emitter-
collector short), through DIO, and
through the 1 0 Mfi-resistor R29. It has
been shown that a leakage current as low
as 1 /r A can turn on Til and T12 and
therefore make LED ‘B’ light up and
LED ‘A’ go out while the fast oscillator
is on. When the fast oscillator is off,
LED ‘A’ lights up and ‘B’ stays on. So the
display with base and collector or emitter
and collector short-circuited is that
LED ‘B’ stays on all the time, and
LED ‘A’ blinks on and off (figure 3h).

Transistor with base and emitter
short-circuited

When the base and emitter are short-
circuited, no ‘normal’ base current can
flow, and therefore there is no collector
current. So the transistor ‘looks like’ one
with zero a', and the LED display is the
same: i.e. LED ‘A’ blinks and LED ‘B’
stays off (figure 3g).

Combined leak and low current
gain or combined leak and base-
emitter short

While the fast oscillator is off, the display
is the same as for a leaky transistor:
both LED ‘A’ and LED ‘B’ are on.
When the fast oscillator is on and
is turning Ta and T 1 0 off, the leakage cur-
rent holds the collectors of Ta and T10
high enough in potential to turn on T1 1,
resulting in LED ‘A’ being off and
LED ‘B’ being on. When the fast oscil-
lator turns Ta and T10 on, the low
current gain of Ta allows T10 to ‘over-
come’ both the leakage current and the

Figure 8. Alternative power supply arrange-
ments, depending on the components one can

Figure 9. The p.c.b. and component layout
for the TUP/TUN tester. Three alternative
layouts are given, corresponding to the three
power supply arrangements.

collector current (if any) in Ta and pull
down the potential of the commoned
collectors, whereupon LED ‘A’ comes
on and LED ‘B’ goes off. This alternate
lighting up of LED ‘A’ and LED ‘B’ is at
the speed of the fast oscillator, and both
LEDs stay on while the fast oscillator is
turned off, so we have a display cycle
in which both LEDs appear to be on
continuously (figure 3i).

Other combinations of Faults

It would not be a very profitable
exercise to list the LED displays with all
possible combinations of faults, but it
can be said that only a transistor which
is ‘sound in wind and limb’ according to
all the test criteria will give the ‘good
transistor’ display in both test sockets.

PNP and NPN transistors

The foregoing descriptions apply to
PNP transistors. They also hold good
mutatis mutandis, for NPN transistor:
plugged into test socket Tb, which ap
pears on the right-hand side of figure 2b
The functions performed by T10, Til
and T12 and associated components foi
PNP transistors are performed by T13,
T14 and T1 5 and associated components
for NPN transistors. In this case, how
ever, the transistors T13 and T14 which
pass on a voltage drop at the anode of
Dll are not a Darlington pair but al
complementary PNP-NPN-pair.

If a transistor is plugged into the wrong
test socket (PNP into an NPN socket or
vice versa), the base-to-collector path be-
comes equivalent to a forward-connected
diode, and the display is the same as for
a transistor with a base-collector short.
The transistor will not be damaged, and ‘
it is clearly a good thing, when one I
shows up unexpectedly as ‘faulty’, to I
check whether it has been plugged into!
the wrong holes!

Breakdown Voltage Test

The sockets for this test are Tq and Tq, I
shown in figure 2c. The effective break- I
down test voltage is about 20 V, and if a I
breakdown current flows the voltagr
at ‘A’ is pulled down continuously, I
resulting in LED ‘A’ blinking anc
LED ‘B’ staying on throughout the
cycle. For a transistor which passes this
test, LED ‘A’ blinks and LED ‘B’ stays
off all the time (figure 5).

Diode Tests

By plugging the anode and cathode leads
of a diode into the emitter and collectos
sockets of the PNP test points (or the
other way round with the NPN test
points) it can be tested for forward con-
duction, leakage and breakdown voltage.
When the fast oscillator is off, the
junction of the diode cathode and the
collector of T10 will be held positive I
by the conduction of the diode, and if!
the conduction is good enough, thin
junction will remain positive when TlOB
is turned on (through ‘Y’) by the fast!
oscillator. When T10 is turned off the!
junction will still be positive. This leads!
to a display cycle in which LED ‘A’ ’
blinks and LED ‘B’ stays on continuous* '
ly (figure 6).

When the diode is non-conducting, open-1
circuited or connected the wrong wa>l
>und, the junction of T10 collectofl
and the cathode (or anode) will remain
negative throughout the oscillator cycle*
giving a LED display in which ‘A’ blink*
and ‘B’ remains off. When a diode ifl
deliberately connected the wrong wan
round, this display gives an indication
(if the diode is a good one) that it
blocking properly in the reverse direfl
tion. If a diode is short-circuited or
leaking severely, it will give the sanfl
display, when plugged in the wrong wan
round, as a good diode connected thn
correct way round. It is just possiblfl
however, that it is a good diode pluggefl .

Stabilised supply with discrete i
ponents

Tr = 20 V/100 mA
R41 = 1 k
C5 = 100 /i/35 V

Parts list
Resistors:

R1,R7.R24,R30,R31,R34= 10 k
R2,R3, R6.R8 = 22 k
R4,R5 = 220 k

R9 (if used),R16.R21 = 680 £2

R10.R1 1,R14,R1 5 = 4k7

R1 2,R1 3.R32.R33 = 100 k

R17 = 2k7

R18.R19 = 47 k

R22 = 1 k

R25. R40 = 270 k

R26 = 1 20 k

R27,R38= 33 k

R28,R37,R20 = 470 SI

R29.R36 =10 M

R23.R35 = 1 M

R39 = 1 20 k

R40 = 270 k

Capacitors:
Cl ,C2 = 4/l7
C3.C4 = 5n6

Semiconductors:

T1 ,T4,T8.T9.T1 5 = BC307B or equ.
T2,T3,T5,T6,T7,T1 0,T1 2.T1 3,T1 6 =
BC237B or equ.

T1 1 = BC239C or equ.

T14 = BC179C or equ.

D1 ... D1 7 = BAX13, BY126. BY127,
1 N4002, or other general-purpose
silicon diodes
2 x LEDs

Unstabilised supply
Tr = 18 V/100 mA
C6= 1000 /i/25 V

Stabilised supply wi
Tr = 20 V/100 mA
C5= 100 /i/35 V
C6 = 10/1/35 V
IC = /iA78M18HCc

614 - elektor june 1975

tup-tun taster

10

TESTER I

NPN

PNP

e

T

ir

GOOD TRANSISTOR

O

‘

O

•

:

•

9

L 1

o

9

[ICo>10iiA.o -C100
|Ccb >20pF

bI_j

o

O

J ICo>10«A

l . <(0

•

•

o

•

•

!

©(£>-**-

3 4 3

9

•

»._>»»

R5"

o

V~™,<MV

o

•

,9

o

O

o

NPN

PNP

Q

EPS 9Q7B-E

Figure 10. Front panel design for the tester.
This panel is available via the Elektor print
service, with black lettering on an aluminium
background.

in the correct way round, i.e. that one
has mistakenly connected it this way
instead of reversing the connections. So
it is important to make tests in both
polarities if ambiguities are to be
avoided.

If a diode which leaks moderately is
plugged in the wrong way round, the
leakage current may be enough to hold
the anode of DIO positive while T10
is turned off, but not while T10 is
turned on. This will give a display in
which both LEDs appear to be on con-
tinuously - the same as for a transistor
which is both leaky and low in current
gain.

General

Figure 3 summarises the LED indications
when the fast oscillator is on or off, and
for different potentials at the anode of
DIO in the main PNP testing section.
The difference between this figure and
table 1 is that the table is based on the

fast oscillator and point ‘A’, which is
common to the main testing circuits and
to the breakdown voltage testing circuit.
The relationship between the potential
at the anode of DIO and the potential
at ‘A’ when the main PNP testing section
is in use is that a high potential at the
anode of DIO produces a low potential
at ‘A’ and vice versa. Figures 4, 5 and 6
summarise the meanings of the possible
LED indications for various tests. Fig-
ure 4 is, of course, derived from figure 3.
The front panel design (figure 10) sum-
marises all three figures.

Construction and testing

Unless one has access to an independent
means of testing the transistors and
diodes, these should be ones which carry
the manufacturers’ warranty. Resistors
should have 5% tolerance.

Figure 9 shows the p.c. board and the
component layout for the complete
tester (with the exception of the mains
transformer). It should be noted that
the emitter and collector connections to
the test sockets appear to be inter-
changed, but when the sockets are
mounted on the copper side of the board

the connections will be correct. This
facilitates the mounting of the board
flush under the top panel, without the
other components getting in the way.
As it is not possible to give meaningful I
voltage or current test values for individ- I
ual transistors, which might help to local-
ise mistakes or faults, the construction
should be checked very carefully. Even
if one does not intend to use the optional
LED ‘C’, one of the LEDs which will
ultimately serve as ‘A’ or ‘B’ can be
‘borrowed’ at the constructional stage
to serve temporarily as ‘C’ and thus give I
a check whether the slow oscillator (T2 I
and T3) and also T4 is working. The!
actual value of the nominally 680 £2 re- 1
sistor in series with ‘C’ is not critical. If I
the temporary ‘C’ is seen to blink at 1
about the right rate, test number 1 of I
figure 7 will show whether T8 is also 1
working. Test number 2 will show
whether T7, T9, T1 1 and T12 are work- I
ing if the short is put into test socket Ta,
and T7, T9, T13 and T14 with the short ■
in test socket Tg.

To check the complete PNP and NPN I
testing sections, including the breakdown I
voltage test, one will need spare PNP I
and NPN transistors known to be sound, I
in addition to those used in building the I
tester. Only a transistor with normal I
current gain, or with a particular com- 1
bination of faults, can produce a display I
in which LED ‘B’ blinks.

Once the two good transistors have been
seen to give the display for test number 3
of figure 7, the more refined test num-
ber 4, which simulates the effect of
excessive capacitance, can be applied. A
22 pF capacitor should be enough to
make LED ‘A’ black out altogether, but
it may be of interest to experiment with
different lower capacitor values to find
what value of capacitor is just low enough
to allow the waveform at the junction
of Ta and T10 to extend below the
critical level (about 1 0 V) and cause
LED ‘A’ to blink dimly, as shown in I
figure 3e 2 .

The emitter-base junction of a transistor I
forms a diode with a reverse breakdown I
voltage of about 5 V, so one of the tran- |
sistors used in the foregoing checks can I
also be used to check the breakdown ‘
voltage testing section (figure 7; num-
ber 5). One must make sure that it is
connected the right (or is it the wrong?)
way round.

Power Supplies

Three different possibilities are offered I
fora mains power supply unit (figure 8): I
a simple unstabilised unit, a stabilised I
unit with the stabilisation circuit built I
up from three discrete components, or a J
stabilised unit using an IC. The unstabil- I
ised unit uses a transformer with an 1 8 V I
secondary — which may sometimes be 1
difficult to obtain — and a 1 000 fi capaci- I
tor. If either of the stabilisation circuits I
is used, the capacitor can be much J
smaller. The 20 V transformer used in .
these circuits should be more readily I
obtainable.

The transistor or IC should have a heat- I
sink. H

elektor june 1975 •

- 615

TUP

TUP

run

Tun

DUG UUG

uus

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

For further information, see the article 'TUP-
TUN-DUG-DUS' in Elektor 1, p. 9.

type

GBP

Ptot

Hn

NPN

20 V

100 mA

100

100 mW

100 MHz

PNP

20 V

100 mA

100

100 mW

100 MHz

*. Minimum specifications for TUP and

Table 1b. Minimum specifications for DUS and
DUG.

type

UR |

IF

IR

p tot

CD

DUS

Si

25 V

100 mA

1 AfA

250 mW

5 pF

DUG

Ge

20 V

35 mA

100 UA

250 mW

10 pF

Table 2. Various transistor types that meet the Table 4. Various diodes that meet the DUS or
TUN specifications. DUG specification'

TUP

Tun

DUG

DUS

DUS

DUG

BA 127
BA 21 7
BA 218
BA 221
BA 222
BA 31 7

BA 318
BAX 13
BAY61
1N914
1N4148

OA 85
OA 91
OA 95
AA 116

Table 5. Minimum specifications for the
BC107, -108, -109 and BC177, -178, -179
families (according to the Pro-Electron
standard). Note that the BC179 does not
necessarily meet the TUP specification
Oc.max = 50 mA).

NPN

PNP

BC 107

BC 177

BC 108

BC 178

BC 109

BC 179

V ce 0

45 V

45 V

20 V

25 V

20 V

20 V

V eb 0

6 V

5 V

5 V

5 V

5 V

5 V

*c

100 m A

100 mA

max

100mA

100 mA

100 mA

50 mA

p tot.

300 mW

300 mW

man

300 mW

300 mW

300 mW

300 mW

f T

150 MHz

130 MHz

min

150 MHz

130 MHz

150 MHz

130 MHz

F

10 dB

10 dB

10 dB

10 dB

4 dB

4 dB

The letters after the type number
denote the current gain:

A: a' (0. hf e ) = 125-260
B: a' = 240-500

C: a' = 450-900.

Table 6. Various equivalents for the BC107,
-108. . . . families. The data are those given by
the Pro-Electron standard; individual manu-
facturers will sometimes give better specifi-
cations for their own products.

NPN

PNP

Case

Remarks

BC 107
BC 108
BC 109

BC 177
BC 178
BC 179

■o

BC 147
BC 148
BC 149

BC 157
BC 158
BC 159

4

Pmax =

250 mW

BC 207
BC 208
BC 209

BC 204
BC 205
BC 206

■©

BC 237
BC 238
BC 239

BC 307
BC 308
BC 309

•3

BC 317
BC 318
BC 319

BC 320
BC 321
BC 322

<3

150 mA

BC 347
BC 348
BC 349

BC 350
BC 351
BC 352

3:

BC 407
BC 408
BC 409

BC 417
BC 418
BC 419

•<SL

250 mW

BC 547
BC 548
BC 549

BC 557
BC 558
BC 559

31

500 mW

BC 167
BC 168
BC 169

BC 257
BC 258
BC 259

m

169/259
’cmax =

50 mA

BC 171
BC 172
BC 173

BC 251
BC 252
BC 253

•3

251 . . .253
low noise

BC 182
BC 183
BC 184

BC 212
BC 213
BC 214

■3

*cmax =

200 mA

BC 582
BC 583
BC 584

BC 512
BC 513
BC 514

•3:

'cmax =

200 mA

BC 414
BC 414
BC 414

BC 416
BC 416
BC 416

•3

low noise

BC 413
BC 413

BC 415
BC 415

<3

low noise

BC 382
BC 383
BC 384

•3:

BC 437
BC 438
BC 439

B

Pmax =

220 mW

BC 467
BC 468
BC 469

Bi!

Pmax =

220 mW

BC 261
BC 262
BC 263

•0

low noise

616 — elektor june 1975

interference suppression in c

interference
suppression in

Electrical interference generated by cars can be a source of annoyance,
not only to the car occupants, but also to users of other electronic
equipment external to the vehicle. This article discusses some of the
more common causes of interference and their cure.

Anyone who has installed a car radio
and then discovered that the programme
is drowned out by odd buzzes, pops and
crackles will know how difficult it is to
trace and suppress interference. Com-
mercially available ‘do-it-yourself sup-
pression kits often do not effect a cure,
because the interference is not dealt with
at source.

Interference in cars on radio and T.V.
bands is caused by high-frequency
energy, usually produced by arcing
contacts, but also by discharge of static
electricity. The interference may be
continual, for example that originating
from the ignition circuitry, or it may
be transient, such as interference occur-
ring when light switches or brake li gh ts
are operated. Windscreen wiper motors
can also generate substantial interference
when they are running.

Interference may be divided into two
groups:

1 . External Interference.

2. Internal Interference.

External interference affects not only
electronic equipment in the car but also
radios and televisions in the vicinity.
Internal interference is usually restricted
to bad reception on the car radio.

The principal sources of interference in

a car are as follows:

ignition system

dynamo and regulator

screenwiper motor

electric fan (if fitted)

heater fan motor

petrol gauge

brake light switch

light switches

starter motor

starter relay and switch

trafficator flasher unit

relays

ignition switch

Some of these, such as the ignition
switch and light switches, cause inter-
ference of a very transient nature and are
probably not worth bothering with.

The ignition system is the most powerful
source of interference and will be dealt
with first. Interference can be suppressed
at various points in the ignition circuit.
Figure 1 shows a typical arrangement
of the high-tension side of an ignition
system. Screened plug caps provide a
good degree of suppression. The screen-
ing makes contact with the plug base
and covers the porcelain insulator. A
typical screened plug cap is shown in

figure 2. These are available from auto
electricians. Plug caps with a built-
suppression resistor are also effective, i
alternatively ‘carbon-string’ H.T. leai
may be used, though the mechanic-
reliability of these is questionable. Thl
distributor may already have suppressic
incorporated, but if not, suppressor ca|
may be fitted to the distributor cove
These should have a resistance of aboii
1 k. Alternatively, in-line suppressor
may be fitted in the plug leads ne J
distributor.

If a radiotelephone is installed in a cai
an extreme degree of ignition suppressioi
may be necessary in the form of screened
cables for all H.T. leads and connectio J
to the coil. Under normal conditions thq
coil itself may be suppressed by conne«
ting a capacitor of about 3/i 400 i
between the SW (ignition Switch]
terminal of the coil and earth. Figure
shows how this is done.

The second important source of inter]
ference is the dynamo/regulator systen
of the car. Dynamos generally have tw<|
connections, one to the armature (vij
the brushes and commutator), which a
the output terminal of the dynamo, and
one to the field winding, the currenl

t

y

6

7

8

elektor june 1975 - 617

Figure 1. Diagram of the high tension side of a
typical car ignition system, showing places to
incorporate suppression.

Figure 2. Cut-away drawing of a typical
screened plug cap.

Figure 3. Ignition coil showing where to
connect suppressor capacitor. Numbers on
various terminals are DIN standard codes
for these terminals.

Figure 4. Connections to a dynamo. Usually
only the armature and field connections are
brought out to terminals and the common
terminal is earthed to the frame of the dynamo.
Again DIN standard codings are used for the
terminals.

Figure 5. A low-power dynamo with an
integral regulator showing connection of sup-
pressor capacitor.

Figure 6. Circuit showing suppressor con-
nections for dynamo with integral regulator.
An 0.5 /i capacitor may be connected from the
armature terminal if necessary, or if this is not
directly accessible, from the ignition warning
light terminal.

Figure 7. General appearance of a non-
coaxial feedthrough capacitor.

Figure 8. Internal construction of a non-
coaxial feedthrough capacitor.

Figure 9. Suppression of dynamo with
remotely-mounted voltage regulator.

through which controls the output volt-
age. The commonest type of electro-
mechanical voltage/current regulator has
three relays that respond to dynamo
output voltage and charging current and
control these parameters. The contacts
of these relays can generate interference,
as can the dynamo commutator/brush
contact.

Small cars with dynamos of less then
300 W output often have the regulator
integral with the dynamo. In that case
the terminals of the dynamo are not
directly accessible and there are only
two connections to the voltage regulator.
One is the output terminal to the battery

and the other is the connection to the
ignition warning light. The battery
terminal may be decoupled with ah /j
capacitor and the indicator terminal with
a 0.5 n capacitor (no larger!) as shown
in figures 5 and 6.

For larger dynamos with remotely-
mounted regulator, suppression may be
connected as shown in figure 9. For
improved suppression, capacitors of the
non-coaxial feed-through type may be
used. These have a lower series self-
inductance than the normal type of
suppressor capacitor as the current flows
in through one terminal virtually right
up to the capacitor plates, and out of
the other terminal. Any lead inductance
is in series with the current-carrying input
and output cables. With the normal type
of capacitor a length of (inductive) cable
connects it to the terminal it is supposed
to suppress, which increases its im-
pedance to high frequencies, when it
should have a low impedance to shunt
them to earth. A non-coaxial feed-
through capacitor is shown in figures 7
and 8. If capacitors on the three regu-
lator terminals and on the dynamo
armature terminal do not provide suf-
ficient suppression a suppressor filter can
be connected in the field cable as shown
in figure 9. This can be home-made from
a 1 0 n low-inductance capacitor and a
0.05 ohm resistor wound from 2'A m of
19 SWG copper wire.

Electric Motors

All electric motors in a car (windscreen
wiper, heater, electric cooling fan) are
potential sources of interference. They
can usually be suppressed by a 3 n ca-
pacitor on the supply terminal. If this
is not sufficient then non-coaxial feed-
through capacitors of 0.5 to 2.5 n must
be used.

Electrostatic Charges

If annoying crackles and pops are heard
from the car radio when driving on dry
roads, this may be due to electrostatic
charges building up on the tyres because
of friction between the tyres and the
road. Since the grease film in the wheel
bearings forms an efficient insulator
such charges cannot leak away to
chassis except when they achieve a high

618 — elektor juris 1975

the wheel hub and the axle by means of
a spring contact mounted in the axle cap.

Figure 11. Construction of spring contact.

enough voltage to flash over, hence the
crackles and pops. A solution to this
problem is to make a good electrical
contact between the axle hub cap
mounted in the centre of the hub (not
the decorative hub cap) and the axle.
See figures 1 0 and 1 1 .

Fuel Gauge

The fuel gauge potentiometer may cause
interference, as may other contacts such
as the trafficator flasher unit. Such
interference can usually be cured by a
capacitor of about 3 n between the
offending contact and earth.

Conclusions

The ignition system is the principal
source of interference and should be
tackled first. To avoid possible unnecess-
ary expense the results of suppressing
the ignition system should be assessed
before proceeding with further measures.
With a good car radio suppression of the
ignition system will usually be sufficient,
but if a cheap car radio with poor selec-
tivity and sensitivity is employed, ad-
ditional suppression may be necessary.
For car radios the adage ‘you gets what
you pays for’ is certainly true, and a
cheap radio may prove to be more
expensive if a large degree of sup-
pression has to be used. This is particu-
larly true if the radio has an FM band
as well as long and medium wave bands,
as cheap radios often need a high
degree of suppression if FM reception
is to be satisfactory. A more expensive
car radio is usually a sound investment,
especially as a car radio cannot be tested
in the shop under the actual conditions
in which it will be used.

Cars which are fitted with radiotelephone
equipment generally need very effective
suppression and such installations are
best left to specialists in the field. M

thief suppression in cars

thief

suppression
in cars

Thefts of cars, or accessories and/or other articles in them, are becoming
more and more common. By the same token, anti-theft alarms for cars
are becoming more and more of a necessity ....

Each of the four different designs described in this article allows for
variations on the basic concept; this makes it possible to choose circuits
that will suit a large variety of requirements and budgets.

Each year an increasing number of cars
are stolen. The majority of them are
quickly recovered, but at best they have
been abandoned on running out of
petrol, and often they are a total write-
off, with everything of value removed.
Thefts of articles from cars are still more
common, especially now that expensive
in-car entertainment systems are so
popular. When compared with the poss-
ible loss of property and/or no-claims
bonus, the cost of installing a burglar
alarm is negligible, and a number of
designs are presented here, which vary
in cost from about 10 p to £ 10, and
give varying degrees of protection. Even
with an alarm installed, however, do not
ignore the simple precautions advised by
the police. Lock all valuables in the boot,
and make sure that all doors and windows
are securely fastened.

Circuit No. 1

The simplest of the burglar alarms can be
constructed from components that most
enthusiasts will have in their junk box.
It makes use of the door courtesy light
switches and the horn relay, and the
only additional components required are
three diodes and a hidden switch to
activate and de-activate the alarm. How-
ever, one door of the car is left un-
protected.

The circuit operates as follows:

S] and S 2 are the courtesy light switches.
When the hidden switch S 3 is closed,
opening the door protected by Sj causes
the horn relay to operate via S 3 , D2 and
S 2 . When the horn relay contacts close
the horn sounds and the cathode of D3
is grounded via the relay contacts. This
latches the horn relay via S 3 and its own
contacts. The horn will continue to
sound even if the door is closed, unless
S 3 is opened. The door containing Si is,
of course, not protected, as when Sj is
closed D1 is reverse biased and only the
interior light operates.

When the alarm is not armed (S 3 open),
D2 prevents the interior light from light-
ing when the horn button is pressed and
D3 prevents the horn from sounding

when the interior light is switched on.
Additional courtesy light switches (in the
case of a four-door car) may be connected
in parallel with S 2 , and switches to pro-
tect glove compartment and other ancil-
lary equipment may be connected be-
tween point ‘A’ and ground.

This simple alarm will provide a fair de-
gree of protection at little cost, but it
should be noted that alarms which sound
indefinitely after being triggered are
illegal in some European countries.

Theck
L a th

delay

Circuit No. 2 (J. van Kessel)

Figure 2 shows the circuit of a more
sophisticated alarm, which protects all thT'al
the doors and arms itself after the driver • ani>{
glove t

has got out of the car. The alarm is
activated by a concealed switch Si . After
this switch has been closed the occupants
have about 1 5 seconds to get out of the
car and shut the doors. During this time
Cl is being charged via Rl, R2 and the
base-emitter junction of Tl. T1 is thus
turned on, shorting out C2. When Cl is
charged to almost the supply voltage Tl
turns off. If now one of the doors is
•opened the courtesy light switch S 2
closes and C2 charges rapidly via D1 and
R3, turning on T2. C3 begins to charge
slowly via R5, which gives time for the
owner of the car to de-activate the alarm
by opening Si . After about 1 5 seconds
C3 has charged to about 8 volts (with
1 2 V supply) and T3 turns on. T4 and
T5 also turn on and the horn sounds. R8
is effectively connected in parallel with
RIO when T5 turns on, reducing the
emitter potential of T3 and causing it to
turn on even harder. C3 begins to dis-
charge via D4, R7 and T5 until it reaches
a potential determined by the new
emitter voltage of T3, when T3 turns off.
T4 and T5 are turned off, the horn relay
drops out and the emitter voltage of T3
rises causing it to turn hard off. (It is
apparent, therefore, that T3, T4, and
T5 function as a trigger circuit.) C3
begins to charge again and the cycle is
repeated. The horn therefore gives
peated short blasts, which (hopefully)

[ 3 ~

L

1975

' Figure 1. Circuit of the simplest burglar alarm
consisting of three diodes end a switch. The
door containing Si is not protected. Ad-
ditional door switches should be wired in
parallel with S2. other switches such as boot
and glove compartment between point A’ and

Figure 2. A more sophisticated alarm circuit.
The door switches are wired in parallel with S2-
L is the interior light.

Figure 3. Block diagram of an alarm that can
be triggered by a trembler switch as well as
door switches. This alarm has no 're-entry'
delay and must be de-activated by an external
concealed switch.

Figure 4. The number of points protected by
the alarm may be extended using a diode
AND-gate for switches on boot lid, bonnet,
glove compartment etc.

will attract more attention than a con-
tinuous note.

Even if the car door is subsequently
closed it will take a long time for C2 to
discharge via R4 into the base of T2. T5
should have a sufficiently high collector
current rating for the relay used. If the
original car horn relay is used then a
power transistor may be needed, and
R1 1 must be reduced to provide suf-
ficient base current.

To use the alarm with a positive earth car
it is necessary to substitute all the transis-
tors by their complements (i.e. TUP for
TUN) and to reverse the polarity of all
electrolytic capacitors and diodes.

Circuit No. 3 (H. Bernstein)

A different approach is adopted in the

f i

ft

Is I
Hi

-x*- — -M“i J

r-Xt-W*!

circuit whose block diagram is given in
figure 3. This circuit has a switch-on de-
lay to enable the driver to leave the car,
but the alarm sounds immediately one
of the doors is opened, thus reducing the
possibility of a thief breaking into the
car and stealing some article before the
alarm goes off. This does mean, of
course, that the alarm must be de-
activated before the driver enters the car.
This may be done by a concealed switch
outside the car, or by a reed switch
mounted in the windscreen and operated
by a magnet carried on the key-ring. This
alarm also has a circuit operated by a
trembler switch which will operate if the
car is rocked or shaken. This will help
prevent the theft of such articles as mir-
rors, fog and spot lamps, and will also
operate if an attempt is made to force
a door or window. Such switches can be
bought, or can easily be home made.
Figure 4 shows the method of connecting
switches. S represents the door courtesy
light switches. Additional switches may
be wired as shown in dotted lines. The
diodes form an AND gate and isolate the
switches from one another so that they
do not interact. A diode is only required
in series with each switch that performs
some additional function as well as being
connected to the alarm system. Thus,
when there is a switch in the boot that
operates the boot interior light, then a
diode is required in series with this to
prevent the light being switched on by
the other switches. Where a switch is
installed solely for the purpose of the
alarm (e.g. in the glove compartment)
then no diode is required.

The circuit of the complete alarm is
given in figure 5. It operates as follows:
Before leaving the car Si is opened. This
initiates the switch-on delay as Cl
charges. T1 turns off, which turns off T2
and T3. The alarm is now set. If one of
the switches connected to points a-d is
closed, C2 is rapidly discharged via R6
and T4 turns on, sounding the horn. The
alarm will continue for up to two minutes
even if the switch is opened, as C2
charges again.

A separate circuit is included for the

620 — elektor june 1975

ignition off alarm set

(IV* min)

trembler alarm. T2 grounds the base of
T6 via D7 until after the switch-on delay.
If the trembler switch S 4 subsequently
closes momentarily, C3 rapidly dis-
charges via R7, turning on T5. This turns
on T6 and the alarm sounds until C3 has
recharged, when T5 turns off. Re-
entering the car is accomplished by
momentarily closing the concealed
switch S 2 , which discharges Cl and
initiates the switch-on delay so that the
driver can enter the car and close S ( ,
which de-activates the alarm.

Circuit No. 4

A block diagram of the most sophisti-
cated alarm in this series is given in
figure 6. In many modern cars with a
steering lock there is no need to have a
concealed reset switch as a position is

often available on the ignition switch
that only opens when the switch is in
the ‘locked’ position. It is thus possible,
on entering the car, to de-activate the
alarm by inserting the ignition key and
turning it to the first position (without
actually switching on the ignition). On
cars without a steering lock, however,
it will be necessary to use a separate
concealed switch - or a separate lock-
switch.

A timing diagram of the alarm is given in
figure 7. On removing the ignition key
from the lock the driver has about one-
and-a-half minutes to leave the car and
lock the doors. If a door is subsequently
opened there is a delay (adjustable from
5 to 1 5 seconds) to allow the driver to
reset the alarm with the ignition key.
Failing this, the alarm operates and the

Figure 7. Timing diagram of the alarm of

Figure 8a. Alternative method of wiring the
horn and horn relay, for cars with one side
of the horn connected to chassis.

8 ;

-

horn will sound continuously for 30 se- j
conds. After this there is ten 'seconds *
silence, then short three second blasts
with ten seconds silence in between until
the alarm is reset or the door is closed.
The complete circuit of the alarm is given
in figure 8. When Si is opened Cl begins
to discharge through R1 and R2. When
the voltage on Cl falls below the
threshold voltage of gate 2 then the
flip-flop consisting of gates 1 and 2 is set.
This holds the input (pin 13) of gate 4a
high, which means that when one of the
door switches (represented by S 2 ) closes, I
the flip-flop consisting of gates 3a and 4a I
is set. The alarm is triggered, and even
closing the door (opening S 2 ) will not
reset the flip-flop. C2 and C3 were
previously charged through R6 and D3
from the output of gate la. C2 and C3
now discharge through R5 into the out-
put of gate 4a. C3 may be optionally
switched in by S3 to increase the delay
before the alarm sounds. With S 3 closed
the delay is about 15 seconds; with S 3
open the delay is only 5 seconds. If this
facility is not required S 3 and C3 may
be omitted and C2 replaced by a capaci-
tor chosen to give the required delay.

S 4 may be a trembler switch or switch(es)
on glove compartment, boot etc., which
will discharge C2 and C3 rapidly and
trigger the alarm.

When C2 and C3 have discharged the
output of gate 2b goes high. This takes

pin t

inpu

low,

the 1

R15

gate

charf

const

of tl

30 se

Wher

thres

of th

gate :

of 1

switc

D7, 1

slowl

and

‘off

onds)

thresl

goes :

charg

time 1

After

low, |

R12a

Gates

multi

prodi

vals.

switcl

elektor june 1975 — 621

irm of

>phisti-

h

alarm.

ing the
Its side

10 se-
onds

blasts

until
osed.
given
egins
When
the
the
5 set.
te 4a
f the
oses,
id 4a
even
not
were
d D3
d C3
out-
nally
lelay
osed
h Sj
this
may
paci-
elay.
i(es)
hich
and

the

akes

9 n

1

_j

—

|

"1

j

n — f-r»

n — r„

_

I

j

! i

i —

i ;

H-T-l 1 *

—

—

—

!

1 — 1

| — 1

rn

— ! j — VN2

1 ■ 4 m

i

~i

□

! i w

l~i ' hQrn

u

i

n_j

ri i

j

K j

*

i

j

! i

J ■ tinw

5.15s

30s

H

10s ;3

sJlOs

s; 10 s

3s! 1592-9

pin 6 of gate lb high. Since the other
input is held high by R1 3 the output goes
low, turning on T1 and T2 and sounding
the horn. C8 charges through R14 and
R1 5 in about 3 seconds, taking pin 13 of
gate 3b high. Meanwhile C6 slowly
charges through R1 1 and it is this time
constant that determines the duration
of the initial blast of the horn (about
30 seconds).

When the voltage on C6 exceeds the
threshold voltage of gate 3b the output
of this gate goes low, grounding pin 5 of
gate lb through C7 and R12. The output
of lb thus goes high and the horn
switches off. C8 now discharges through
D7, R14 and the output of gate lb. C7
slowly charges through R13 and R12,
and this time constant determines the
‘off period of the horn (about 10 sec-
onds). When C7 has charged to the
threshold voltage of gate lb the output
goes low and the horn again sounds. C8
charges through R14 and R15, and this
time constant determines the subsequent
‘on’ periods of the horn (about 3 sec-
onds).

After this period the output of 3b goes
low, grounding pin 5 of gate lb through
R1 2 and C7, and the whole cycle repeats.
Gates 1 b and 3b thus form an asymmetric
multivibrator which causes the horn to
produce short blasts at 1 0 second inter-
vals. In addition, each time the horn is
switched off a differentiating network

consisting of R8, R9, C4 and C5 feeds
a reset pulse to pin 9 of gate 3a, so that
if the doors are closed during the horn
‘off period, the horn will not sound
again and the alarm will be re-set.

The disadvantage of this alarm circuit is
that it cannot easily be adapted for
positive earth cars. On the other hand,
it has the advantage that it is insensitive
to spurious pulses due to the high noise
immunity of COSMOS. The circuit will
operate over a wide range of supply
voltages without modification (4 to
14 volts), with almost the same delays.

Installation of the alarm

To ensure reliable operation the alarm
should be mounted where it cannot be
disabled by a thief, but not inside the
engine compartment, which gets rather

hot for COSMOS. Wiring should be con-
cealed or made as inconspicuous as
possible, especially wiring into the engine
compartment if the car does not have
a bonnet lock. In this case it is also wise
to install an alarm switch in the bonnet
lid, as otherwise the alarm could be dis-
abled simply by disconnecting the bat-
tery. Another (somewhat expensive)
possibility would be to power the alarm
from a separate battery locked in the
boot. Wiring from the battery to the
alarm should, of course, be direct, not
via the ignition switch, and the simplest
way is to run cables from the battery
side of the fuse box with in-line fuse I
holders in them.

When constructing any of these alarms it |
should be borne in mind that they may I
need to be adapted to suit particular j

622 — elektor june 1975

Components list for
figures 8 and 1 1

Resistors:

R1,R6 = 10 k
R2, R15 = 470 k
R3,R7,R9,R10,R16 = 22 k
R4,R5.R8 = 100 k
R11 = 180k
R12.R14 = 47 k
R13- 1M5

Capacitors:

C1.C6- 220///16 V
C2 -47/i/I 6 V
C3= 100/i/16 V
C4= 10 n
C5 = 100 n
C7,C8 = 1 0/i/1 6 V

Semiconductors:
ICa.ICb = CD401 1 A
T1 = TUP
T2 = BC140 or equ.

D1 ... D8 = DUS

Figures 10 and 11.
Board and component
layout for figure 8.

S*-— • Z — L

l ~

r ‘^V

pc

*&■

31

types of car. For instance, some horn re-
lays have one end of the coil and the
contact commoned, so they would not
work in the circuit of figure 1 . A separate
heavy-duty relay (6A contacts) would
have to be used. This problem does not
arise with the more sophisticated designs, i
and a normal horn relay may be used. A
printed circuit board layout is given in I

figures 1 0 and 1 1 for the alarm circuit
of figure 8 and this has alternative sets
of connections for the relay, as shown
in figure 8a, to suit different car wiring.
S s in both circuits is the original push-
button for the horn.

Final Remarks

The designs discussed in this article give

string sound

This circuit can be used to give quite a
natural imitation of a vibrating string.
An astable multivibrator consisting of

transistors Ti and T 2 produces the
fundamental tone. As long as switch S
is not operated, the two transistors T 3
and T 4 are conducting, and thus the
multivibrator output is short-circuited.
When S is operated, the short-circuit is
removed, and T s is turned on whilst T 3 is
cut off. At the same time T 6 conducts
momentarily, C s discharges and T 4 is
thus also cut off. The generator signal
now occurring at the output slowly dis- I
appears as capacitor C 5 charges again. If I

varying degrees of protection at varying
cost. It should be remembered that any
protection is better than none - the
majority of thieves are amateurs and will
be deterred by even the simpler circuits
described.

S isclosed before the tone has disappear-
ed, it is abruptly cut off because T 3 is
turned on again.

With this circuit a tone can be produced ,
whose frequency depends on the setting »
of P! and on the values of Ci and C 2 .
The values for these capacitors can be
determined by experiment; they will
generally be chosen in the range 1 ... 1 0 n. -
If several of these stages are intercon-
nected via resistors, a simple synthesizer
can be built. m

elektor june 1975 - 623

seven-segment Minitron or LED display to indicate 'H' for a high or ' 1'
state and 'L' for a low or '0' state. The circuit also detects when the

probe input is open-circuit and the readout is suppressed, thus indicating
that contact with the desired test point has not been made. This avoids
the false readings that may occur with some types of probe when the
input is not connected.

The circuit makes use of a 7447 decoder
driver. The input circuitry to this IC is
designed so that when the input to the
probe is high a ‘1’ is applied to the ‘C’
or 4 input of the IC. When the input to
the probe is low a ‘1’ is applied to the
‘A’, ‘B’ and ‘D’ or 1, 2 and 8 inputs of
the IC. This results in the display of the
number 4 and the symbol □ respectively
in accordance with the truth table for
the 7447. However, the connections
from the outputs of the IC to the seg-
ments of the display are rearranged so
that the display is actually H and L. When
the input to the probe is open-circuit all
four inputs to the 7447 are high ( A = B =

I7447

output

No.

13

12

connected to
display segment
not connected

d 10

e 9

f 15

g 14

d

not connected

:r=)E> —

C = D = 1, i.e. ‘15’) and the display is
completely suppressed.

The input circuitry operates as follows:
Ni and N 2 are exclusive-OR gates. When
a ‘0’ is applied to the probe input both
inputs of N, are ‘0’ so the output is also
‘O’. One input of N 2 is held at ‘0’ via R!
and the other is held at ‘ 1 ’, by R 2 , so the
output is ‘ 1 ’. This output is connected to
the A, B and D inputs of the 7447. When
the probe input is ‘1’ one input of N t is
‘0’ and the other is ‘1’, so the output is
‘1’. This output is connected to the
C input of the 7447. Both inputs of N 2
are ‘1’, so the output is ‘O’. When the
probe input is open-circuit the input of
Nj is not connected to ground floats at
just above the ‘1’ threshold level, so the
output is ‘1’. The forward voltage drop
of D! and D 2 prevents this from holding
the input of N 2 high, so the input is held
low by Ri . The other input is, of course,
held high so the output is ‘1’.

Figure 1. Connections from outputs of 7447
to display segments.

Figure 2. Complete Circuit of the H-L tester,
showing the alternative connections for Mini-
tron and LED display.

624 — elektor june 1975

measurement

results

Input impedance: 60 ... 160 kfl,
depending on

Input sensitivity: 70 ... 170 mV
(adjustable)

Output impedance: Ikn
or up to 4k7, depending on
Maximum output level:

180 mV (or up to 850 mV)

S/N ratio: better than 60 dB
Input selector: suppression of
unwanted inputs: better than
60 dB

Crosstalk: adjustable;
in stereo position: —40 ... —50 dB,
100 Hz... 10kHz
Current consumption:

approx. 200 mA (10 V)
Distortion, as a function of the
output voltage from the input
selector stage: see graph A

graph B

port 2

tap

preamp;

The first part of
this article discus- 1
sed an audio pre-
amplifier and
control unit
operated entirely ]

by TAP's and dealt with the design of the TAP and the electronic
switching controlled by the TAP. This month's article deals with the
application of these circuits to a complete touch-controlled preamp with ]
the facilities already described.

A block diagram of the preamp and
control unit is given in figure 1. The
input selector, with inputs for four signal
sources, is followed by a tone control
that provides bass lift, presence (middle
lift), treble cut, or a flat response. (It
should be noted that the touch control
panel shows a symbol which could be
interpreted as ‘treble lift’ in the fourth
position.) The signal is then fed into a
circuit that controls the image width
from mono to ‘enhanced stereo’ by
introducing crosstalk between the
channels. The signal is fed finally to a
volume control that provides four preset
gains.

The disc input must be preceded by a
suitable RIAA-equalised preamplifier,
which may be mounted in the control
unit, but preferably in the record deck

as this will give better hum figures and! 1
(provided the disc preamp has a lovij’
output impedance) the frequency i
sponse will be unaffected by cable! 1
capacitance.

The TAP, which controls all the funcJ
tions, is shown in figure 2. Its operation
was described in detail in last month’!
article, but basically, touching any o
of the inputs causes the corresponding d
output to become ‘1’ and all the othel
outputs to become ‘O’. Only one Q oua
put can be ‘ 1 ’ at any time. The Q output!
of the TAP are connected to the corrf
sponding Q inputs of the input selectoj
tone, width and volume controls.

The Input Selector

The input selector of figure 3 makes us|
of the electronic ‘break contact’

elektor June 1975 - 6:

and scribed in last month’s article to short
low out the unwanted signals. When one of
re - the inputs 1-4 is selected the correspond-
able mg transistor (T1-T4) is turned on. The
corresponding pair of transistors for left-
inc - and right-hand channels (T5/T6-
tion T11/T12) are turned off so that the
th’s desired signal can reach the base of T1 5
one and T1 6. All the other pairs of transistors
g Q are turned on and short the unwanted
ther signals to ground,
out- The presets on each input allow adjust-
puts ment of input sensitivity and channel
rre- balance to correct channel imbalance in
tor, the signal sources. The channel balance
of the preamplifier itself may be adjusted
by presets in the volume control stage.

use The Tone Control

de- The tone control circuit is shown in

Figure 1. Block diagram of the complete
touch-controlled preamplifier consisting of
input selector, tone, stereo image width and
volume controls. The four units each have a
nominal gain of one, so any unit or units may
be omitted without affecting the sensitivity.

Figure 2. Circuit of the four-position TAP.
Touching one of the inputs causes the appro-
priate output to become T. The Q outputs
are used to control the preamplifier functions.

Figure 3. Circuit of the input selector. T13 and
T14 provide additional amplification for low-
output tuners and can be dispensed with
if not required. Presets R27-R34 are used to
adjust for the same nominal output for all
«gnal sources. For high-level inputs the value
of these presets can be increased to 1 Mil

figure 4. It consists of input and output
buffer amplifiers T13-T16 with three
switched filters for bass lift, presence
and treble cut interposed between them.
The circuit operates as follows:

T5/T6 and T7/T8 act as normally closed
(break) contacts. Til and T12 act as
make contacts. T9 and T10 are omitted
as they correspond to the flat position
(position 3) where no filter is in circuit.
When position 1 (bass boost) is selected
T5/T6and T1 1/T12 are turned off while
T7/T8 are turned on. This means that
C5, C6, R55 and R56 are shorted out
and C7 and C8 are open-circuited. A
filter as shown in the simplified circuit
of figure 5 thus appears between the
two buffer amplifiers. This boosts fre-
quencies below 400 Hz.

When position 2 (presence) is selected
T5 and T6 are turned on, shorting out
Cl and C2, whilst T7/T8 and T11/T12
are turned off. This connects the filter
of figure 6 in circuit. This is a band-pass
filter, which boosts the signal over the
range 200 Hz to 4 kHz, with a maximum
mid-band boost of about 10 dB.
Position 3 is the flat position with T5/T6
and T7/T8 turned on and T11/T12
turned off. The circuit of figure 7 there-
fore results, which is of course simply
an attenuator with no frequency-
dependent characteristics. In position 4
(treble cut) all the transistors T5-T 1 2 are
turned on and the circuit of figure 8
results. At frequencies away from those
at which the filters have their effect the
nominal gain is one in all positions. This
ensures that switching from one filter
position to another does not result in
large changes in volume.

Stereo Image Width Control

This control alters the separation be-
tween channels by introducing crosstalk
from one channel into the other. The
image width may be varied from a mono
signal, where each output channel con-
tains an equal proportion of left and right
inputs, to ‘enhanced stereo’, where cross-
talk is introduced in antiphase to in-
crease the image width beyond that of
normal stereo. The circuit (shown in
figure 9) operates in the following
manner:

T13 and T14 function as a difference
amplifier. That is to say that the signal

elektor june 1975 - 627

1 appearing across their collectors is pro-
e portional to the difference hetween the
e left and right input signals to their bases.
If, for example, the right input is
grounded and the left input is fed with a
i signal then the signal on the collector of
s T 1 4 is proportional to the left input and
i 180° out of phase with the signal on the
1 collector of T13. The same argument
holds true if the left input is grounded,
i When both inputs are driven the outputs

e at the collector of T13 and T14 consist

t of left channel with a proportion of anti-

t phase right channel and right channel

i with a proportion of antiphase left

t channel. That is to say, the crosstalk

s appearing in the signal at the collector

of T13 is 180° out of phase with the
right channel signal appearing at the

1 collector of T 1 4 and vice versa.

t In-phase crosstalk may be introduced

2 ^ into these signals by mixing with the

r opposite channel at the base of T15 and

s T16 respectively. When no in-phase

: crosstalk is introduced the signals ap-

' pearing at the collectors of T1 5 and T16

consist of the original signal plus the
> antiphase crosstalk and the image width
- is increased. When the proportion of
in-phase crosstalk is the same as the
' proportion of antiphase crosstalk the

two cancel and only the original signal
remains. This is normal stereo. When the
' in-phase crosstalk exceeds the antiphase
’ crosstalk the net result is a proportion
: of in-phase crosstalk and the stereo image

‘ width is reduced; finally, when the
’ crosstalk equals the signal a mono output
: results.

When position 1 (mono) is selected tran-
sistors T7-T12 are cut off. This means
that crosstalk from the collector of T1 4
is fed into the base of T1 5 together with
the left channel signal from the collector
of T13 and vice versa. When position 2
' (reduced width stereo) is selected T9 and
• T10 are turned on, grounding R37 and

R38 respectively. R51, R37 and R52,
R38 thus form attenuators that reduce
the amount of in-phase crosstalk. The
same applies to position 3 (normal
? stereo) when T9 and T10 are turned on,
i but R39 and R40 are chosen so that
: the in-phase crosstalk equals the anti-

phase crosstalk and the two cancel. R39
'■ and R40 may be replaced by presets so
I that the circuit can be trimmed to cancel

Figure 4. The tone control unit. The four
potitions give bass boost, middle lift (presence),
flat response and treble cut respectively.

Figures 5-8. Simplified equivalent circuits of
the tone control in the four positions men-
tioned.

Figure 9. Circuit of the stereo image width
control. This has four positions from mono
to enhanced stereo with different amounts of
crosstalk introduced between the channels to
vary the separation.

—

the crosstalk exactly. In position 4 T1 1
and T12 are both turned on and the in-
phase crosstalk signals are shorted to
ground leaving only the original signals
plus the antiphase crosstalk. This results
in an enhanced stereo image width.
Looking at the operation of the circuit
mathematically we can derive the follow-
ing:

ignoring the gain of the difference ampli-
fier, which affects all components of the
signal equally, we can say that
L c = -L + k,R

where L c is the signal at the collector
of T13, L and R are the left and right
inputs and k! is a constant determined
by the parameters of the difference
amplifier. Similarly

R c = -R + k,L

The minus signs are due to the 180°
phase change in T13 and T14 respective- 1
ly. After mixing with in-phase crosstalk
the signals appearing at the collectors of '
T15 and T16 (again ignoring the gain,
which affects all components of the j
signal equally) are

L 0 = — L c -kjR c ,

and

R 0 = — Rc — k 2 L c

where k 2 is a constant whose value is
selected by switching in the different
attenuators that introduce varying pro-
portions of in-phase crosstalk.

tap-preamp

. 'h.

For enhanced stereo there must be only
antiphase crosstalk i.e.

kj =0.

The value of k 2 for a reduced width
stereo signal (position 2) is purely a mat-

Figure 10. The four-level volume control,
which may be preset to the desired listening
levels and may be used to adjust channel
balance.

Figure 11. Pattern of the universal p.c. board
used for each of the four units of the pre-
amplifier.

Therefore

Lo = L-k,R + k 2 R-k,k 2 L
= U1 -k,k 2 ) + (k 2 — kj )R
k] was chosen subjectively and it was
found that a value of 6 dB (X Vi) of anti-
phase crosstalk gave the best results.
This immediately gives some of the
values for k 2 .

For a mono signal the proportions of L
and R in the output must be equal Le.

1 -k,k 2 =k 2 -k,
which means that k 2 = 1 .

For a normal stereo signal the crosstalk
must be zero Le.

%

i

let of personal taste depending on the
image width required and may be ad-
justed by changing R39 and R40 in
Spire 9.

Volume Control

Hus was discussed briefly in last month’s
article and the complete circuit is given
in figure 10. Selecting one of the pos-
itions turns on the corresponding pair of
transistors T5/T6-T1 1/T12, grounding
the potentiometers connected to each
collector. These form attenuators with
R51 and R52 which control the levels
of the signals fed into the bases of T15
and T16 respectively. The degree of
attenuation produced in each position
may be altered by the potentiometers
to suit personal taste and to adjust the
channel balance.

Construction and Adjustment

The four units described are each con-
structed on a universal printed circuit
board, the pattern for which is given
in figure 1 1 . The component layouts for
the different units are detailed in fig-
ures 12-15 and the parts lists are given in
the tables 1 and 2. The components
common to every board are given in
table 1 and those particular to one unit
are given in table 2. The capacitors
marked * in figures 4, 9, and 10 may be
omitted if all 4 boards are used together
but should be included if any board is
used on its own.

Setting up of the units is a simple matter.
The input potentiometers of the input
selector stage are adjusted so that the
output of this stage is about 100 mV
when fed with the nominal signal level
of each source. Thus, if the system is to
be used with a tuner of nominally
1 00 mV output the tuner input should be
adjusted with 1 00 mV input signal from
an oscillator. If no test equipment is avail-
able the circuit may be adjusted using the
actual signal sources (disc, radio, tape
etc.) and listening on headphones each
input potentiometer may be adjusted to
give approximately the same volume
level. Balance between channels should
also be adjusted to compensate for im-

fr Vqj

AUX

Figure 13. Component layout for the
(figure 4).

tone control stage

if fij

T T • ♦

T

:»

H i-if

• H • _• •

• 3 •

-£ h • •

4 4 • a

s •

|KjMn

• t m~

• ® •

'it

.**i •..«

rf-

g ••

$ .

flp! ,
r^FTT

r

Kg

* * «

HI

iiSis

ill

T

■

mmim

b.-.lance in the signal sources. The volume
control settings are next adjusted to give
the desired listening levels. Channel
balance may also be adjusted to com-
pensate for any imbalance in the pre-
amplifier itself or in the power ampli-
fier and loudspeakers. The unit is now
ready for use.

The output level is 200 mV; if this is
insufficient for full drive of the power
amplifier, R61 and R66 (figures 10 and
15) can be increased to 4k7. The output
level then becomes 1000 mV.

Conclusion

All the units in the touch-controlled
preamplifier have a nominal gain of
unity and so may be used in any com-
bination without affecting the perform-
ance, or they may be used in con-
junction with other equipment. It is

hoped in a later article to publish
details of a touch station selector for
radio and other additions to the system.

Design Modification

A small modification has been made to
the portions of the circuit using the

‘break contact’. Referring to figure 6 of
last month’s article, when T1 is turned
off there is still a residual current of
about 6 mA flowing through the LED
via R3, R4 and the base-emitter junc-
tions of T3 and T4. With certain types
of LED, notably those with a clear plastic
encapsulation, this may give rise to a

noticeable glow. This can be eliminated
by connecting a 220 J2 resistor across
the LED. Current will flow through this
resistor but the voltage drop across it will
bf less than the tum-on voltage of the
LED. This modification applies to the
following: figure 3, D13-D16; figure 4,
D13 and D14. M

supplies for

In order to function effectively, electronic equipment used in cars must
have an appropriate power supply, which in addition to providing a
regulated voltage from the battery supply, must also suppress inter-
ference appearing on the battery voltage from the car electrical systems.
To be effective, such power supplies must of course be used in con-
junction with the measures described elsewhere in this issue for the
suppression of radiated interference at source.

Two power supplies will be described;
a simple zener stabilizer with built-in
suppression, for low-power circuits such
as instruments (electronic tachometer
etc.) up to about 170mA and a high-
power stabilized supply, for powering
such things as portable cassette re-
corders, up to about 2 A.

Low-power circuit

Figures 1 and 2 show the low-power
circuit configuration for negative- and
positive-earth cars respectively. LI and
D1 provide high-frequency decoupling.
LI may be wound on a wire-ended
cylindrical ferrite core, with a diameter
of about 1 0 mm, using 45 turns of
0.5 mm enamelled copper wire
(25 SWG). Zener diode Z1 stabilizes the
output voltage at 5.6 V, which is suit-
able for TTL circuitry, but 5.1V or
4.7 V types would also do, as they are
within the supply voltage limits for TTL.
Other voltages may, of course, be used
for different equipment.

If a supply voltage is required that is
almost equal to the available battery
voltage, then R1 and the zener diode
may be omitted. Of course, the circuit

will then not provide stabilization of
the supply voltage, but only interference
suppression. This would be quite ad-
equate for COSMOS IC’s, which are
fairly tolerant of supply voltage vari-
ations.

The difference between battery voltage
and output voltage is dropped across R 1 .
In the example given, for a 6 V battery
R1 would be 8.2 ohm 14 W, and for a
12 V battery 47 ohm 2 W.

C2 should be a low-inductance type
such as ceramic and D1 can be any diode
that will carry 200 mA.

High-Power Circuit

This circuit (figure 3) is designed to
provide a stable, interference-free supply
for cassette recorders, audio amplifiers
and other equipment. Car radios gener-
ally have their own inbuilt suppression
circuits.

The circuit is a simple feedback stabil-
izer. T5 and T6 form a differential
amplifier with a constant-current
emitter load (T7). This compares a
portion of the output voltage at the
slider of PI with a reference voltage
provided by D6. T3 provides a constant

current through this zener diode.T4 is a
constant-current load for the collector
of T6, which drives the Darlington-
connected transistors T8 and T9. Current
limiting is provided by T1 and T2. When
the current through R4 is such that the
voltage drop across it exceeds the base-
emitter voltage of Tl, this transistor
turns on, which turns on T2, pulling
down the collector of T6 and reducing
the drive to T8. The maximum output
current required determines the value
of R4:

LI is wound from 45 turns of 1.0 mm
copper wire (19 SWG) on a ferrite core
and thus provides interference sup-
pression due to its high inductance. PI
will adjust the output between about
5.6 V and 12 V, although the higher
figure can only be obtained when the
(12 V) car battery is fully charged and in
good condition, so that the supply
voltage is around 14 V, since some
voltage must be dropped across the series
regulator transistor and R4.

A printed circuit layout for this supply
is given in figure 4. The connections
shown encircled in figure 3 are shown
on the board layout. The circuit of
figure 3 is for use with negative earth
cars, but may be modified for positive
earth simply by reversing all diodes and
electrolytic capacitors and replacing
transistors by their complementary
equivalents.

If higher output currents are required,
an external transistor (e.g. MJ3055) can
be added. The connections to this tran-
sistor (Tl 0) are shown dotted in figure 4;
it should be mounted on an adequate
heatsink, with mica washers for isolation.
If this transistor is not required, the
connections 5 and 6 on the pcb must be
bridged.

To avoid deterioration of the board in
the car due to humidity and dirt it is
best to encapsulate the circuit (once it
has been tested) in epoxy resin or
silicone rubber. M

mos~ clock

The mos-clock described in Elektor 1 has been extended with a crystal
time base and an emergency supply. These extensions do make the clock
a bit more expensive, but they are at the same time elements changing
the clock into a highly-accurate and universal instrument.

Furthermore, the total extra current consumed by these extensions is
practically negligible. Both the emergency supply and the crystal time-
base are mounted on one printed circuit board.

In recent years especially, the mains
voltage has been liable to cut out
momentarily (or even for quite some
time!). Depending on the capacitance
of the electrolytic in the power supply,
the d.c. voltage driving the clock will
have disappeared after about 200 ms.
The clock will then forget what time it
is, so that it must be reset. This is in
itself not such an enormous problem,
but a number of brief failures in one
day could be annoying!

The emergency supply circuit described
here uses a 9-volt battery. This provides
an emergency drive for about 20 hours;
a period within which even the most
serious mains breakdown will have been
repaired. The emergency drive also offers
the possibility of moving the clock from
one room to another without it stop-
ping.

Design

As the clock-IC MM5314 consists of
one monolithic MOS circuit, its current
consumption can be neglected in com-
parison with that of the displays. At
15 V the total current consumption of
the clock is about 250 mA, 240 mA of
which is drawn by the displays. This is
one of the reasons why the clock-IC is
provided with a so-called strobe input
(pin 1 of the IC) which is ‘1’ when the
clock is running normally. If, however,
the strobe input is made ‘O’, the display
is suppressed, and only the divider
circuits and the memory still draw
current.

These circuits still function satisfactorily
at a supply voltage of 7 V.

So for an effective emergency supply
from a small battery it is essential that
as soon as the normal supply cuts out,
the strobe input becomes ‘0’ so that
current consumption drops to about
8 mA. Of course, it must be possible to
switch on the display circuits now and
again on the emergency supply. A circuit
which provides for this is given in
figure 1.

If the mains voltage is available, the
circuit is driven via point BS from
the secondary of the transformer. This

Parts list for figures 1, 2, 5, and 7.

resistors:

R1,R3= 100k
R2.R7 = 27 k
R4 = 10 k
R5 = 120 J2
R6 = 15 k
R8 = 47 k

Cl =0.1 H

C2 = 5 . . . 30 p trimmer

C3 = 1 5 p

C4 = 220 H, 10 V

semiconductors :

T1.T2.T3 = BC 107 B or equ.

D1 = BAY 61, BA 127
D2.D3.D4.D5.D6 = DUS
IC1 = ICM 7038 A (intersil)

miscellaneous

SI = push button with break contact
X-tal = 3.2768 MHz crystal; parallel
resonance with external 12 p
8-pin DIL IC-base for IC1

point is indicated on the clock P.C.
board. The transformer secondary volt-
age on BS is rectified via Dl, so that
capacitor Cl is charged. Then transistor
T1 is driven into saturation via resistor
R 1 . The collector voltage of this transis-
tor is then so low that transistor T2 is
not driven. Via resistor R4 and push-
button SI the collector of T2 is con-
nected to the strobe input of the clock-
IC. This point (SB) is also indicated on
the clock p.c.b. When transistor T2 is
off, the strobe input of the clock-IC is
connected to a relatively high-impedance
load so that it ‘sees’ a ‘1’, causing the
displays to light up.

If, however, the mains voltage cuts out,
point BS in figure 1 no longer carries
a voltage, and Cl discharges so rapidly
that transistor T1 is cut off within
5 ms. Now the base of T2 is driven via
resistors R2 and R3, so that its collector-
-to-emitter resistance drops to about
200 ohms. As a result point SB becomes
‘0’ via resistor R4 and the supply for
the display circuits is cut off inside the
clock-IC.

elektor june 1975 - 635

When button SI in figure 1 is depressed,
the strobe input of the clock-IC becomes
‘1’ again and the displays light up.
Instead of a push button, a single-pole
switch can be used for SI.

Practical version

For the emergency supply, only a bat-
tery with a series resistor need be
connected across the supply electrolytic
capacitor of the clock proper. Figure 2
shows the relevant detail of the clock
supply: the extra battery and the series
resistor R5 are in parallel with the supply
elco. A so-called minipower pack will
do as the battery. The design of the
p.c.b. is based on this type of battery.
Two alternative arrangements are also
possible. The first is to replace R5 by a
diode, with the anode connected to the
battery and the cathode connected to
the supply rail. This has the advantage
that the full battery voltage is available
to drive the displays for short periods
on emergency drive. However, it also
means that no ‘refresher’ current runs
through the battery when the mains is
on.

Perhaps the best arrangement of all is to
replace R5 by a diode, as above, and add
an extra resistor of 1 00 k in parallel to
the diode. This will trickle-charge the
battery.

In all cases it is advisable to check the
battery condition occasionally (by dis-
connecting the mains). This is particu-
larly important when ‘dry’ batteries are
used - they sometimes become very wet
after a period of time, as many owners
of portable radios and torches have dis-
covered to their cost!

Crystal timebase

Although the standby supply will main-
tain the information in the memory of
the clock in the event of a mains failure,
the counting circuits will not operate in
the absence of a 50 Hz signal. This is
where the crystal timebase comes in. It
ensures good timekeeping accuracy
(approx. 1 0 seconds per month) and
makes the drive to the clock independent
of the mains.

Table 1 . Some specifications of the Intersil IC type ICM 7038 A.

supply voltage

1.6 V ... 4 V (max. 5 V)

current consumption

60 IJA (V b = 2.2 V); 130 /JA (V b = 3.6 V)

output resistance

230 12 both for p- and n-output
condition (l 0 = 3 mAI

minimum oscillator frequency

0.2 MHz (V b = 1.6 V)

maximum oscillator frequency

10 MHz

power dissipation

input voltage oscillator

—30^ C . . . +125° C maximum

case temperature

ambient temperature

-20° C . . . +70° C maximum

output voltage

as V b at all outputs

output current at V Q = 0

80 mA maximum

400 Hz output current

18 mA maximum

(V a as V b )

400 Hz output resistance

200 £2 (l a = 3 mA)

4

t

VdoE

•

e0~

VssCZ

lie

7038A

Unc

2|g:

J

D^400Hz

Figure 1. With the mos-clock, the displays
draw most of the current. If the dock is
provided with a circuit that drives the 'strobe'
when the mains cuts out, the emergency
battery will keep the clock itself running for
about one day.

Figure 2. The mos-clock can easily be provided
with an emergency supply. A 9-volt battery
and a series resistance of 120 £2 will do.

Figure 3. The intersil 1C type ICM 7038 A
comprises a.o. a 16-stage divider and a crystal
oscillator. For the oscillator only the crystal
and two capacitors need be connected exter-
nally. The 1C is built up from complemen-
tary MOS-circuits, so that the power dissi-
pation is extremely low.

Figure 4. The external connections of the
ICM 7038 A. Owing to the COS/MOS proper-
ties, it is always advisable to use a base or
socket for this 1C.

For this crystal time base use is made
of a complementary MOS IC which,
besides an oscillator, also comprises the
dividers necessary for obtaining the
50 Hz square-wave voltage with which
the clock is driven.

Figure 3 gives a simplified block diagram
of this IC, the INTERSIL ICM 7038 A.
Transistors T1 and T2 in figure 3 are a
part of the crystal oscillator. Two exter-
nal capacitors and the crystal are con-
nected between points 7 and 8. The
oscillator is followed by an inverter (I)
which is turn is followed by 16 divider
stages. The 16th divider is followed by
two inverse output stages at which a
relatively low-impedance square-wave
voltage is available for further pro-
cessing. The divider stages D1 . . . D16
are all binary so that after the 16th
divider we have a frequency of:

f = f 0 /2 16 = f 0 /65536
Here f Q is the oscillator frequency and
f the output frequency behind the 1 6th
divider.

It is also apparent from figure 3 that all
important points are protected by means
of zener diodes. This does not imply,
however, that the IC can be handled
as if it were TTL: due care is always
recommended. Touching the connecting
pins of the IC must be avoided as much
as possible, whilst the IC can best be
mounted on the p.c.b. by means of a
socket.

636 — elektor june 1975

The most important specifications of
the ICM 7038 A are given in table 1.
The low supply voltage (1.6 . . . 4 V)
and the low current consumption
(average 90 fiA) are worthy of note.
Figure 4 gives the pin connections of
the ICM 7038 A.

The circuit

Figure 5 shows the circuit diagram of
the complete time base.

Since the clock supply lies between 8 V
and 18 V, a special circuit must be
provided which ensures that the time
base IC gets no more than 5 V. The
simplest solution is shown in figure 5 :
the IC is fed via a resistor R6, and an
electrolytic capacitor (C4) is connected
across the IC. The diodes D2 . . . D6
ensure that the supply voltage can never
rise above about 3.5 V.

In figure 5 the capacitors C2, C3 and
the crystal are the external components

X-Tsl ■ 3.2768 MHj ,

for the oscillator. The crystal must be
of a type that has been ground for
parallel resonance with an external
parallel capacitance of 12 pF nominal.
For 50 Hz output reference the crystal
frequency must be 3.2768 MHz. The
oscillator can be adjusted with capacitor
C2.

Transistor T3 has been included in the

circuit to make the output voltage of
IC1 suitable for clock drive. The collec-
tor of this transistor is connected to
point X on the clock p.c.b., after which
resistor R22 (100 k) is removed.

The printed circuit board

Figure 6 shows the lay-out of the
printed circuit board on which the

circuits of figures 1, 2 and 5 can be
mounted. The component arrangement
of these circuits is given in figure 7.
Figure 8 shows a photograph of the
board. It can be mounted on the mains
transformer on the original clock board.

Figure 5. The complete 50-Hz reference source
for the mos-clock. The diodes D2 . . . 06 are
for protection and ensure that the supply
voltage for the IC cannot rise above about
3-5 V. T3 is needed because the output
voltage of the IC is too low to drive the
clock input directly.

Figure 6. The lay-out of the printed circuit
board for the circuits of figures 1, 2, and 5.

Figure 7. The component arrangement on the
printed circuit board of figure 6.

Figure 8. Photograph of the complete printed
circuit board with the circuits of figures 1, 2,
and 5.

is cybernetics?

elektor june 1975 - 637

The 'cybernetic beetle' is the
first of a series of articles in
'Elektor' on designs which are
; on the one hand meant to be a
game, but which on the other
I hand should be regarded as
serious and sometimes scien-
I tific imitations of animal
behaviour.

These cybernetics projects
have an electronic 'nerve
system' enabling them to react
specifically to external
acoustic, optical and tactile
stimuli.

Such models can moreover be
designed so that the reaction is
not only governed by the type
of stimulus, but also by the
condition of the model at the
moment when the stimulus is
applied. The more 'natural' the
relationship between the
stimulus and the relevant
reaction, the more the
behaviour of the model ap-
proaches that of the animal
example and the more will an
ingenious toy become a scien-
tific object.

It was found to be
extremely difficult to imitate
the behaviour of more highly
developed animals in a cyber-
netic model, because an elec-
tronic nerve system can only
perform a limited number of
functions. Experiments on a
scientific basis therefore con-
fine themselves to certain
aspects of behaviour or to
living models of a very simple
nature.

Although the nature of the
nervous system is as yet un-
known, it is interesting to see
that at any rate the models
behave in a typically 'animal-
like' way so that their reaction
to a certain stimulus cannot be
predicted.

These cybernetic models derive
their attraction from the fact
that the spectator uncon-
sciously interprets the pur-
posefulness or the learning
capacity of artificial animals as
a sign of intelligence. Here the
question immediately arises
whether it is possible to build
a 'thinking' machine.

It would carry us a little too
far to use this series of articles
for entering into these un-
doubtedly highly interesting
and topical problems. The
cybernetic aspects of the
electronic animal models will,
however, be dealt with more
extensively. (Ed.)

H. Ritz

what

is

cybernetics?

The word ‘cybernetics’ which has in
recent years taken a permanent place
in our vocabulary, is of Greek origin and
means approximately ‘art of steering’.
The Greeks used it to denote the skill of
a pilot who steered a ship safely into
harbour. At present the term denotes a
new branch of science which still has
something to do with the art of steering.
In 1948 Norbert Wiener founded this
branch when he wrote his: ‘Cybernetics—
the science of control and communi-
cation, in the animal and the machine’.
The title of this book already shows that
the sphere of influence of cybernetics
encompasses all dynamic systems, both
natural and artificial. For a cybernetic
investigation of such systems it is irrel-
evant of what material they are made and
by what force they are driven. The only
thing that matters is the abstract system
detached from all technical details.

It will therefore be clear that cybernetic
investigations may also concern other
dynamic systems, such as economics, but
of course it is technology that profits
most by cybernetics.

It is the different ways in which the
dynamic systems behave that are of
interest to cyberneticians. The behaviour
of living organisms is then frequently
compared with that of a technical imita-
tion, a model.

A toy train driven by a spring motor
running in circles on its rails is a dynamic
system. Yet, this model track is of no
interest to cyberneticians, because no
behavioural patterns can be observed. A
system which has to develop certain
behavioural patterns must also have a few
properties other than dynamic ones only,
viz.:

1. It must be able to receive external
information coded in signals;

2. It must be able to store this infor-
mation and hence have a memory;

3. It must be able to convert stored infor-
mation into reactions. These reactions
must be controlled, which means that
external interference must always be
effectively reacted to. Reactions do

not conform to a previously fixed
programme, but are always compared
first with previous reactions and their
results. All cybernetic control systems
are based on this feedback principle .

4. Because they are capable of self-
correction, cybernetic systems always
display a purposeful behaviour.

5. Cybernetic systems are stable.

By this we mean a dynamic stability
which enables the system to return
to a certain situation after a break-
down. Non-cybemetic machines can-
not cope with external interference.
They only function if possible causes
of interference are known beforehand
and if the appropriate counter-
measures have been incorporated in
the machine.

A cybernetic machine, on the con-
trary, can also deal with interference
not previously taken into consider-
ation.

6. Higher-level cybernetic systems can
‘learn’. They can adapt their behaviour
if experience gained gives them cause
for doing so.

Arbeitsgemeinschaft der
Hauptschule Rossbach (Sieg).

beetle

Beetles, tortoises and the like
have often served as models for
cybernetic machines which must
also have a reasonable appearance. The beetle described in this article can
'see, hear and feel' and reacts to information in the form of sounds and
movements. The animal has a memory and can get tired. The beetle is
the first of a series of articles on cybernetic models.

The behaviour of the beetle: action
and reaction

In daylight (vertical light) the model
moves slowly forward in a circle (coun-
ter-clockwise) with a diameter of approx.
80 cm. If during its journey the beetle
arrives in a place where there is less light
(for instance in a dark comer or below a
chair), it rests in the shade for approx,
half a minute to a full minute and then
resumes its circular journey.

It is also possible that, instead of con-
tinuing the model decides to take a
prolonged rest in the shade: it ‘falls
asleep’. This condition of prolonged rest
can only be altered by external stimuli.
The object casting the shadow is removed
or the model is wakened by a loud sound
(clapping). If the beetle ‘sees’ a horizon-
tal light source on its journey, it makes a
beeline for it. Any deviations from its
course are automatically corrected.

If the model hits an obstacle on its path,
it gives a brief cry of fright and immedi-
ately shrinks back. This reverse move-
ment is, however, preceded by a short
turn, so that when the beetle resumes its
forward movement after 3-4 seconds, it
is positioned obliquely with respect to
its previous direction of travel. Because
the beetle can now no longer see the
light source, it again starts a left turn,
which results in a second collision with
the obstacle, somewhat more to the left
than the first time. Thus, after a few
repetitions, extensive obstacles are also
dodged, after which the beetle heads for
its goal (phototropism).

When the horizontal light source is
reached, the model will collide with it
and the above-mentioned swerving move-
ments will follow. As a result, the model
will find its way behind the light source.
If this horizontal light source does not
emit light backwards and if there is also
a shady area at the rear of the lamp, the
model goes behind the lamp to enjoy a
short or prolonged period of rest.

The model can also learn. If at the
moment of the collision a warning sound
(clapping) is made, the model concludes
that the collision (pain) and clapping

belong together. For this reason, if a
warning sound is made when the model
is moving forwards, it will now first give
a cry of fright and then perform the same
swerving movement which would other-
wise be performed in the case of an
actual collision.

The memory fades, however, during a
rest in the shade. After such a pause the
model hesitates briefly when the warning
sound is made; it listens for a moment
and immediately after the hand-clapping
the forward movement is resumed.

In general this simple cybernetic model
behaves in a typically animal-like way.
It can move, see, hear, feel, it reacts
efficiently to certain external inter-
ferences, acts purposefully and has also
certain reflexes in addition to the senses
of touch and hearing.

Things become even more interesting
if two models are available, each with a
light source on its back. If moreover the
cry of fright of one beetle is tuned to the
warning sound of the other, an exchange
of experience may take place. The
innumerable possibilities of searching,
pursuing and swerving guarantee an
interesting course of experiments.

As is indicated by dashed lines in the
block diagram (figure 1 ) a ‘hunger’ sensa-
tion can be added. This part immediately
interrupts a journey or a period of rest
when the battery voltage drops below a

certain minimum. The model then heads
for a second lamp placed near a charging
unit. If on the way to the charging unit
the beetle starts to use less current for
some reason, it records that sufficient
energy is still left to resume the normal
search for the first lamp. If this effect
is undesirable, a flip-flop can be used
instead of gates Ni S /N 16 . After charging
the reset function must be performed by
hand.

For demonstrations it is advisable to use
a potentiometer as supply control; the
beetle’s ‘hunger’ may then be artificially
generated by increasing the potentio-
meter’s resistance.

The various functions

The drive motor (DM) runs only if the
vertical-light receiver (VLR) receives
light from above and the warning re-
ceiver (WSR) does not receive sound, or,
in darkness, if a short rest (SR 40 sec.)
initiated by the VLR has elapsed and
the flip-flop (LR) for a prolonged rest
is not in ‘prolonged’ position. A warning
sound makes the motor stop briefly and
trips the flip-flop for prolonged rest.
Furthermore the drive motor starts run-
ning if a warning unit (BW) detects a
drop in battery voltage. (The latter
applies only if the relevant control unit
consisting of a sensor, a change-over
switch COSj and a receiver for the light
from the charging unit (CLR) has been
incorporated.)

When an obstacle is touched (OS closes)
CF 1/2 sec. changes over to cry of fright
and RM 3 sec. to reverse movement.
CF 1/2 sec. starts the cry of fright by
means of an oscillator (FO). RM 3 sec.
switches the drive motor to position
‘reverse’ by means of a relay (RRL).
CF 1/2 sec. also ‘enables’ the memory
flip-flop (MFF) for half a second, i.e. it
renders the flip-flop able to receive a
triggering signal. The memory flip-flop
(MFF) is switched on if within this time
a pulse from the warning-sound receiver
(WSR) reaches MFF. Once MFF has been
set, a warning sound causes CF 1/2 sec.
and RM 3 sec. to be activated via a

elektor june 1975 - 639

Figure 1. The block diagram shows the various
'sense organs' of the beetle; three light receivers
(for vertical, horizontal and charging light),
a warning-tone receiver and an impact contact
for tactual stimuli. The information obtained
from these organs affects the behaviour of the
model via the motor, the steering motor and
the reversing relay. A special sensor checks the
operating voltage and ensures that the beetle
makes for the 'food station' in time.

Figure 2. The circuit for the light receiver.
Three of such receivers are required.

Figure 3. The warning-sound receiver.

Figure 4. Fright oscillator with two NAND's.

NAND network. This reaction is switched
off by SR 40 sec., which resets MFF.
The light receiver for horizontal light
(HLR) switches the steering wheel to its
central position via a change-over switch
(COSj) and the steering motor (SM).
The change-over switch furthermore en-
sures that after the collision with the
obstacle the reverse movement starts
with a short turn.

If the automatic ‘hunger sensation’ is
built in, connection A runs through
change-over switch COS 2 . In this way it
is ensured that at a given voltage the
steering motor is no longer operated by
the light receiver HLR, but by the light
receiver for the charging unit CLR (via
COS 2 and COSj).

Light receivers

The circuitry of the three light receivers
is identical (figure 2). Each consists of a
NAND gate with a voltage divider con-
sisting of two resistors and an LDR at the
input. When illuminated, the LDR has a
very low resistance, so that the NAND
input is logically ‘0’ and the output is
logically ‘1’. In the absence of light, the
resistance of the LDR is high. The NAND
input is at positive potential and the
output is at low potential (logically ‘0’).
The LDR for vertical light (VLR) is
mounted on the model’s back. The out-
put of the relevant receiver circuit con-
trols the period of rest (SR 40 sec.). The
LDR of the receiver for horizontal light
(HLR) is accommodated in a black
cylinder, approx. 10 cm long, together
with a converging lens. In the model this
‘viewing tube’ is fitted horizontally and
pointing forwards. If light from the light
source reaches the LDR, the output of
the receiver becomes logic ‘1’.

Via the change-over switch(es) (COS 2
and) COS! the receiver HLR controls
the steering motor of the model.

The LDR of the receiver for the light
from the charging unit must be screened
from other light by means of a cylinder
and, for instance, be fitted to the under-
side of the model so that only light from
the charging unit can be received. Change-

over switch COS 2 is controlled by the
output of the charging light receiverCLR.

The warning receiver WSR

To prevent the beetle from reacting to
any arbitrary external noise, the receiver
must be selective. The ‘clap sensor’ of
figure 3 is eminently suitable. It consists
of a preamplifier (Ti ), a selective ampli-
fier with twin-T network (ICi), a recti-
fier circuit (Di , D 2 ) and a trigger circuit
(Ni , N 2 ). In the digital part of the model
(figure 5) the output of the trigger is
followed by a one-shot (ICi 2 ) set by
clapping once. This one-shot can be
considered as part of the receiver WSR in
figure 1.

The fright oscillator FO

An astable multivibrator consisting of
two NAND gates constitutes the fright
oscillator (figure 4).

Switch S is the on/off switch. In the
model the output of a one-shot replaces
this switch. The LF output drives a low-
power audio output stage.

The complete circuit

The description from the behavioural
pattern shows that the model may move
when sufficient light falls from above.
In the absence of vertical light the model
may only move if the flip-flop LR for
the period of rest is in position Q = ‘1’
and if furthermore the circuit for the
pause in the shade SR 40 sec. is reset
(Q=T).

These two possibilities are checked by
the AND circuit consisting of N s and N 6
(AND!) and the subsequent OR circuit
(N 7 , N 8 and N 3 ).

Furthermore the model may not move
when the sound receiver WSR receives
a signal which makes the Q output a
logic ‘O’. This condition is combined with
the conditions previously mentioned by
means of N 4 (AND 2 ). The output of N 4
is only a logic ‘0’ if the conditions for
moving have been fulfilled and no warn-
ing sound is received, because only in
that case are the two inputs of N 4 logical-
ly ‘1’ and the output logically ‘O’.

The flip-flop for a prolonged period of
rest LR is tripped by output Q of the
warning sound receiver (IC i2 ) at every
clap. For this reason the model will
sometimes rest for the normal period

(SR 40 sec.) and at other times will not
move on after the SR time has elapsed.
Instead, it will fall asleep.

If the model hits an obstacle, impact
contact OS is closed, and the mono-
stable CF is triggered (IC 2 ). For half a
second output Q changes from ‘1’ to ‘O’,
so that the fright oscillator (N| 3 , N l4 )

starts and produces a LF signal which
is made audible via an amplifier stage
(Tj , T 2 ) and the miniature loudspeaker.
The ‘impact pulse’ simultaneously
switches the one-shot RM (1C 2 ). Its
output Q changes from ‘0’ to ‘1’ for
3 seconds. Reversing relay RRL is switch-
ed on via a power amplifier (T s ); the

elektor june 1975 - 641

model moves backwards for approx. RM is at ‘1’ in this case, the steering

3 seconds and thus moves away from the motor is switched on (straight ahead),

obstacle. The Q output of CF and the via N„, T 3 and T„, if the second input

Q output of RM control gate N, in of N 10 is also logically ‘1’ and conversely,

change-over switch COS!- During for- The signal to this second input does not

ward movement the inputs of N 9 are at affect the steering however, during the

‘ 1 ’ and ‘O’, so that the output is logically reverse movement for 3 seconds because

‘1’. As the input of N 10 controlled by the Q output of RM is ‘O’. This means

that the reverse movement is started with
a turn (0,5 sec., determined by CF via
N 9 ), after which the movement is con-
tinued in a straight line backwards for
another 2,5 sec. Then the relay lets go
and the model starts moving forward
again in a circle.

The memory flip-flop (MFF, IC S ) has an
input J connected to the Q output of CF.
Every collision causes J to become ‘l’for
0.5 sec. If within this time a warning
sound is heard the output Q of MFF
changes to ‘1’. Consequently a negative
pulse occurs at the output of N 12 every
time a clap is heard, which causes the cry
of fright and the reverse movement even
without a collision occurring (reflex).
When the beetle rests in the shade MFF
is reset via the ‘clear’ input (Q = ‘O’). The
memory has now been erased, so that a
warning sound can only change over LR
and cause the model to stop briefly (via

n 4 ).

As the time available for the reception
of information by MFF is fairly short
(0.5 sec.), it is often difficult to set the
flip-flop by means of clapping. This is
similar to a normal learning procedure,
where it may also happen that only the
second or even third combination of a
warning sound together with the collision
results in the desired reaction.

The circuit for a rest in the shade SR
also comprises a one-shot (IC 10 ) which in
this case is controlled by the receiver for
vertical light VLR via gate N, . The motor
is stopped for 40 sec. via N s ... N 7 , N 3 ,
N 4 and N 2 , provided that meanwhile the
model does not receive any light from
above. The circuit for ‘hunger’ comprises
among other things a threshold-value
detector consisting of Pj andN 16 . P, can
be set so that the input of N i6 becomes
logically ‘0’ if the operating voltage has
dropped to such a level that the accumu-
lator must be recharged or the battery
must be renewed. There must be suffi-
cient energy left, however, to enable the
beetle to reach the ‘.food station’. Because
the life of the beetle is in danger when
the voltage drops, protective functions,
such as rest and panic stop are of second-
ary concern and are therefore cut out.
The motor runs constantly via N 1S and
N 2 .

The output of Nig (in COS 2 ) is ‘1’,
whether there is horizontal light or not.
As long as no collision occurs, only the
receiver for the charging light (CLR)
affects the steering motor via the input
of gate N 10 controlled by N 20 . A collision
is dealt with as usual; the effect of CLR
is then cut out for 3 seconds.

The mechanical construction

A commercially-available servomotor
(max. 3 V) for model planes is eminently
suitable as a steering mechanism. The
motor must automatically return to its
central or zero position. Building-in
should be done at the front of the model,
on the centre-line of the bottom plate
(see figure 6). The motor must be fitted
with approximately 20° slant to the left,
so that the model describes a left turn

5 V/600 mA

D1 -e.g. ZX3.9I1 W),wi
if drivemotor draws i

i adequate cooling
ore than 250mA.

Figure 5. The complete circuit diagram of the
digital section. As no particular types are speci-
fied for the drive motor, steering motor, relays
and loudspeaker, the indications of voltage
values near the positive connections have been
omitted.

Figure 6. The position of the drive motor with
the driving wheel and the free-running wheel,
steering motor complete with steering wheel
and the impact contact.

Figure 7. An example of a supply unit based
on components that are usually commercially
available, such as a 6-V drive motor and turning
relay.

M. Keul, H. Lohr

the moth

The car has two motors, one to propel
it and one for the steering mechanism.
The principle is quite simple. A light-
sensitive element is mounted obliquely
to the right at the front of the car.

Normally the steering motor keeps the
car on a left turn, but this motor can be
reversed by a relay, so that the car turns
to the right. This happens each time the
light-sensitive element receives light.

This is a design for a simple
cybernetic model, based on an
electric toy car, that will be
attracted towards a light source
like a moth, negotiating obstacles
in its path.

If the car, travelling to the left, passes a
light source, it swerves to the right in the
direction of the light. However, it keeps
turning to the right, so that after some
time the element receives no more light.
Then the car will automatically swerve
to the left again until the element again
receives light from the source, and so on.
Thus the car zigzags towards the light
source. It behaves more or less like a
moth, hence the name of this apparatus.

The circuit

The light-sensitive element used here is
an LDR placed in a cardboard tube. This
tube must be about 3 ... 5 cm long to
screen off the daylight. Via an amplifier
consisting of Ti and T 2 (figure 2), the
LDR drives a relay Re t which reverses
the steering motor M[ .

Pi serves to adjust the sensitivity of the
car. That is to say, it determines how far
the car will swerve to the left or the
right on its way to the light.

In this circuit the moth will hardly ever
reach the light source because in a fur-
nished room it will usually collide against
a wall, chair, table or something else,
and come to a standstill there. So it must
be able to avoid an obstruction, an idea
which is not so difficult to realize. For
that purpose a mercury switch is fitted to
the car chassis in such a manner that the
contact just remains open. Now if the
car collides, the mercury moves about in
its tube with the result that the contact
is closed momentarily (figure 3).

In the circuit diagram S 2 is the mercury
switch. Via this switch Ci is charged
rapidly and then slowly discharges again.
Thus the short pulse from the switch is
artificially lengthened. As long as Ci has
a potential greater than 0.7 V, T 3 is
turned on so that relay Re 2 is energized
for a few seconds. This energizing time
can be varied by means of P 2 .

The second relay has two functions.

the moth

elektor june 1975 - 645

Firstly it reverses the drive motor so that
the car backs away from the obstruction.
That alone would, however, not be suffi-
cient, for if the car were then to start for-
ward again, it would hit the obstruction
a second time, and so on.

It is, therefore, also necessary that when
the car backs away, the steering wheel be
turned in the opposite direction in order
to avoid the obstruction. So the second
function of Re 2 is to reverse the steering
motor. Thus the motor is controlled in
two ways: by the optical circuit (via Re! )
and by the mercury switch (via Re 2 ).
Now the moth will always reach the light
source, provided there is a road.

The mechanical part

Electric toy cars that can be remotely
controlled by a cable make ideal basic
material.

The authors used an old lorry; the drive
motor remained as it was, and a smaller
motor was mounted at the front. On the
spindle to which originally the remote
control wire was connected, a pulley
from a construction kit was mounted;
another pulley was mounted on the
motor shaft. Then these two were con-
nected with a rubber drive belt.

It was found that it is better to drive
steering mechanism via a slipping clutch.
If the motor engaged directly with the
steering mechanism, it would remain
stalled most of the time because the

steering circuit is limited on both left and
right.

As a result, current consumption would
rise substantially. Figure 4 shows a
home-made slipping clutch. Next to the
wheel mounted on the steering spindle,
a larger loose wheel is placed which is
forced against the fixed wheel by aspring.
First pieces of felt are glued to the con-
tact faces of these wheels. The loose
wheel is coupled to the motor and via
friction of the felt between, it drives the
bottom wheel and thus the steering
circuit.

The supply

Owing to the high current consumption
of the motors, it is recommended that
they be driven from an accumulator. A
separate 9 V compact battery can thus
be used for the electronics. The supplies
of the two systems are separated to
prevent interference generated by the
motors from affecting the electronics.
Should you find another motor suitable
for the steering circuit but requiring a
different supply voltage, a second accu-
mulator will be needed.

The on/off switch can be of the three-
pole type, and the two contacts con-
trolled by Re 2 can be supplied separately.

M

M. Ginolas

micro-

squeaker

This circuit is by way of being an
electronic joke. The complete circuit
comprises only one transistor, one ca-
pacitor, a miniature transformer and a
headphone. The transistor can be any
germanium type; the transformer can be
any miniature type with a turns ratio
between 3 : 1 and 1 0: 1 . At supply voltages
as low as 0.2 V the headphone produces
a distinct sound. Current consumption
is then of the order of 10/iA, power
consumption is less than 2|iW. The joke
of this microsqueaker is that it is not fed
from a ‘normal’ current source, but that
the gifts of nature are called upon. The
positive connection is a piece of bare
copper wire, the negative connection is a
bare piece of steel or silver wire. If both
ends are stuck into an apple, a lemon or
a potato, at some distance from each
other, the apparatus produces a tone. A
solar cell could also serve as the voltage
source. The squeaker may also be used
as an indicator for D.C. voltages in the
range of 0.2 V... 10 V.

646 - etektor june 197S

quadro in practice

quadro in practice

In response to an earlier article on
quadrophony ('Quadro 1 -2-3-4
December 1974), we received
many requests for the complete
circuit of a quadrophony decoder
- preferably one which is suitable
for all systems. However, the
problem here is that the patentees
for QS and CD-4 will not consent
to the ICs developed for their
systems being supplied by the
retail trade*. They are only to be
supplied to original equipment
manufacturers.

As a result, a home-made quadro
decoder will have to consist of
discrete components, so that it
becomes far more complex than
would otherwise be necessary.
Nippon Columbia has now put
such a design at our disposal, so
that we can fulfil the wishes of
many of our readers. For com-
pleteness' sake a survey is first
given of the decoder principles of
the four different systems: SQ,

QS (RM), CD-4 and UD-4.

* It is extremely difficult to write a truly
up-to-date article on the subject of quadro-
phony, as new information is pouring in
continuously. Although this article has been
up-dated as far as possible (the first draft was
written in February ...), it is already slightly
out-of-date: according to our latest infor-
mation, the Signetics CD4-392 is now avail-
able via the retail trade. Elektor laboratories
are now studying the new possibilities which
this offers.

For a good understanding of quadro-
phony decoders in general, and of the
decoder described here in particular, it is
necessary to know something about the
theory of the various systems. Therefore
the systems SQ, QS, CD-4 and UD-4 are
first dealt with in this article. For a
description of the results that can be
obtained with the various systems, the
reader is referred to the above-mentioned
article ‘Quadro 1-2-3-4 ...’.

SQ

This system (by CBS) is a 4-2-4 matrix
system, that is to say, that the four orig-
inal channels (Rp, Lp, R B , L B ) are
combined in a matrix (‘encoder’) to two
channels for transmission (Rp and Lp,
where ‘T can be translated as ‘trans-
mission’), and then separated again in a
second (decoder) matrix into four chan-
nels (Rp, Lp, R b , L b ) for reproduction.
With SQ the encoder is described by the
formula:

Rp = Rp + 0.707 (jR B - L B ), and
Lp = Lp + 0.707 (R b - jL B ).

Here the ‘j’ corresponds to a 90° phase
shift. Rp and Lp are signals which are
transmitted via, for example, a gramo-
phone record; for reproduction, four
different signals must be derived from
them via the following formulae:

R£ = Rt.

Lp = Lp,

R b = 0.707 (-jRp + Lp),

L b = -0.707 (jLp + Rp).

These formulae define the SQ decoding
matrix. For completeness’ sake we can
also calculate what the four reproduction
channels have in common with the
original four channels. For front-right
we have:

Rp = Rt = Rf + 0.707(jR B - L B ).

So this signal is composed of the original
front-right signal (Rp), the right-back
signal attenuated 3 dB and shifted
through 90°, and the left-back signal also
attenuated 3 dB and shifted through
180 .In the same manner it can be seen

that the signal reproduced left-front also
consists of a mixture of three signals: the
original left-front signal, and in addition
the original left-back signal attenuated
3 dB and shifted through 270° (-j), and
the original right-back signal attenuated
3 dB and in phase. Expressed in formula:

Lf = Lp = Lp + 0.707 (R b - jL B ).
The signal reproduced at the right-back

R b = 0.707 (-jRp + Lp)

= R B - 0.707 jRp + 0.707 Lp,
whereas left-back is reproduced as

Lr = Lr + 0.707 jLp - 0.707 Rp.

In both cases the crosstalk products are
similar to those in the front channels.
The channel separation can be improved
by using two approaches. The simplest
is the addition of two resistors, by means
of which extra crosstalk, say 10%, is
introduced between the front channels
and 40% between the back channels. In
formula:

Rf - 0.9 R£ + 0.1 Lp,

Lp = 0.9 Lp + 0.1 Rp,

R b = 0.6 R b + 0.4 L b ,

L B = 0.6 L b + 0.4 R b .

Other ‘blend’ ratios are also used. This
decrease of channel separation at the
front and back results in a 6 dB im-
provement of channel separation be-
tween ‘centre-front’ and ‘centre-back’.
The second approach is the so-called
‘gain-control logic’. In its simplest form
this consists of automatic level controls
for the front and back channels. The
operating principle is that if, for example,
the front channels are louder than the
back channels (in other words, the main
sound sources are at front) the difference
is emphasized by extra amplification of
the front channels and simultaneous
attenuation of the back channels. The
more expensive version (‘full logic’ or
‘wave-matching logic’) can in a similar
manner boost one channel at the expense
of the other three. So in that case only
the left-front channel, for example, is

eUktor june 1975 — 647

Figure 1. Block diagram for an SQ decoder.
The basic decoder corresponds to figure 1A.
The phase shifters are so designed that within
the audio band there is an almost constant
mutual phase difference of 90° (corresponding
to±j) between the two pairs of output signals.
Figure IB gives an addition (so-called blend),
with which the channel separation front-back
and along the diagonals can be improved at the
expense of the channel separation left-right.
Alternatively, 'logic' (figure 1C) can be used,
for the same purpose.

Figure 2. The diagram of a complete SQ de-
coder with 'logic'. The MCI 31 2 comprises
the phase shifters and matrix, the MC131S
comprises the control circuits for full logic and
drives the MCI 31 4 which comprises the buffer
amplifiers with gain control.

amplified whilst the other three channels
are attenuated.

With suitable demonstration records this
‘logic’ can lead to an almost discrete
reproduction of single channels. The
block diagram of an SQ decoder is shown
in figure 1 . Motorola is one of the firms
offering a complete series of ICs for an
SQ decoder with full logic: the MCI 3 1 2,
MC1314 and MC1315. The first com-
prises the basic decoder (except for a
number of resistors and capacitors!)
whilst the other two contain variable-
gain amplifiers and the full-logic control.
A practical circuit with these ICs is
given in figure 2. In contrast to the other
ICs, these types can be obtained via the
retail trade.

QS

This system (by Sansui) is also a
4-2-4 matrix system. The matrix itself
is standardised in Japan as RM (regular
matrix), so that on universal decoders
the QS position is sometimes indicated
as ‘RM’.

In this case the encoder is described by
the formulae:

R t - 0.924(Rp - j R b ) +
0.383(Lp - jL B ), and
L t = 0.924(Lp + jL B ) +
0.383(R F +jR B ).

The factors 0.924 and 0.383 are
cos 22%° and sin 22 Vi, respectively. The
decoder matrix is as follows:

R£ = 0.924 R t + 0.383 L T ,

L£ = 0.924 L t + 0.383 R T ,

L b = -0.924 jL T + 0.383 jR T , and
R£ = 0.924 jR T - 0.383 jL T .

After some calculation we find that for
reproduction the output signals are com-
posed as follows:

Rp = R F + 0.707 L F - 0.707 jR B ,

L£ = Lp + 0.707 Rp + 0.707 jL B ,

L B = L B + 0.707 R B — 0.707 jLp, and
Rr = Rb + 0.707 L B + 0.707 jRp.

In contrast to the results obtained with
SQ, crosstalk is almost symmetrical: the

648 - elektor june 1975

quadra in practice

Figure 3. Block diagrams for QS decoders.
Figure 3A shows the phase shifters and basic
matrix, and figure 3B shows the block diagram
of an extended QS decoder with Variomatrix.

Figure 4. Diagram of a complete QS decoder
with ICs. This has been divided into two
sections, for clarity. The HA1327 comprises
the phase discriminator, the HA 1328 contains
the control logic and matrix, and the HD3103
contains a fourfold MOSFET control ampli-
fier. This circuit can easily be extended for
decoding SQ as well; or a simpler version can
be used, from which one HA1327, the eight
transistors and corresponding components are

original right-front signal, for example,
is reproduced at right-front, and also
(attenuated 3 dB) at left-front and right-
back. Owing to this greater symmetry
the results with this system are better
than with SQ.

To improve channel separation, this
system, too, can be equipped with an
automatic level control: the so-called
Variomatrix. The principle is the same as
with the SQ logic: the position of the
main signal source is located (left, right,
front or back) and the gains in the four
reproduction channels are modified to
emphasise this effect. The difference
from SQ is, however, that in this case
the gain control influences the matrix
coefficients (0.924 and 0.383); hence
the name Variomatrix.

Figure 3 gives the block diagrams for a
simple and a complex QS decoder. A set
of ICs for QS — with which SQ can also
be decoded — is manufactured by
Hitachi: the HA1327, HA1328 and
HD3103PA. A complete QS decoder
with these ICs is shown in figure 4;
unfortunately the ICs are not (yet)
obtainable from the retail trade.

CD-4

CD-4 (by JVC/Nivico) is a discrete sys-
tem. This means that, in principle, the
four channels are transmitted indepen-
dently of one another via the gramo-
phone record : a so-called 4-4-4 system.
The basic principle is fairly simple. Where
the left channel for a normal stereo
gramophone record is cut - that is the
left groove wall - the sum signal Lp + Lg
is recorded on a CD-4 record. Moreover,
the difference signal Lp — Lg is recorded
as a frequency modulation of a 30 kHz
carrier (in the band from 20 ... 45 kHz).
For the difference signal a separate fre-
quency correction is applied with break
points at about 800 Hz and 6 kHz. This
is sometimes referred to as ‘FM-PM-FM
equalisation’. In the same manner the
two signals Rp + Rg and Rp — Rg are
recorded as ‘baseband’ and ‘carrier chan-
nel’ in the right hand groove wall.

So it is not entirely true that the four

elektor june 1975 - 651

channels Lp, L B , Rp, R B are transmitted
independently of one another: in fact
two sum and two difference signals are
used. This does not affect the discrete
character of the system, but it does mean
that in the playback equipment a simple
‘sum-difference matrix’ is needed — as is
the case with an FM-stereo decoder. In
spite of the inclusion of this matrix, CD-4
certainly does not belong to the ‘matrix
quadrophony’ systems!

In the practical realisation of this system
the real problems crop up ... Without
going into details it can be said that
transmitting two baseband channels (the
two sum signals) and two wideband
FM-modulated carrier channels (the two
difference signals) in one record groove
can easily give rise to various kinds of
crosstalk and distortion; in this con-
nection phrases like ‘uptalk’, ‘downtalk’,
‘carrier-channel crosstalk’, ‘energy spill’,
et al. are used. The first CD-4 records
were often of inferior quality as a result
of this. In the meantime, however, a
change-over has taken place to a ‘Mark II’
recording system in which various forms
of distortion compensation (Neutrex I
and II) are built in, whilst in the near
future the ‘Mark III’ version is to come
into operation. The fact that this gives
designers (and recording studios?) a
headache is of no importance to the
consumer; what is important is that the
quality of the newer CD-4 records is in-
deed far better than that of the first

To ‘protect’ the carrier channels even
further, the later equipment also con-
tains the possibility of suppressing the
low-frequency components of the differ-

Figure 5. Block diagram of a CO-4 decoder. In
contrast to the previous block diagrams this
diagram also shows the dynamic preamplifier.
This forms an essential part of the decoder
since not only the normal stereo channels are
used, but also two FM-modulated 30 kHz car-
rier channels.

Figure 6. Diagram of a CD-4 decoder with ICs.
The input circuit is more complicated than is
strictly necessary, because it has also been
made suitable for the new semiconductor
pick-up cartridges. These do not require
RIAA correction, but they do require a bias
current; furthermore, the outputs are in anti-
phase.

The ICs are only supplied to original equipment
manufacturers.

Figure 7. Block diagram of a simple UD-4 de-
coder, suitable only for BMX.

ence signals (Lp - Lg and Rp - Rg). As
a result these low-frequency signals are
reproduced as ‘centre-left’ or ‘centre-
right’, but this will generally not be
noticed.

The next problem concerns noise. Par-
ticularly when wideband carrier channels
are used, this could easily reach un-
acceptable levels. A noise suppression
system is desirable, and the so-called
ANRS system was developed for CD-4.
According to JVC/Nivico, the initial
problems have by now been solved, by
adding an automatic carrier level control
at the recording end.

The- final problem especially concerns
the consumer. Since a CD-4 gramophone
record contains frequencies up to
45 kHz, the pick-up element must be
able to reproduce these. At the moment
suitable cartridges are supplied by a.o.
JVC, National Panasonic and Pickering,
but they are not cheap.

Figure 5 shows the block diagram of a
CD-4 demodulator. Note that in contrast
to all other quadrophony systems the
CD-4 system does not require 90° phase
shifters. This means that the system is
suitable for extensive integration, which
can save costs and space.

For CD-4, various types of ICs exist: the
CD-4 392, manufactured and supplied
by Signetics, and the QSI 5022 supplied
by Quadracast Systems Inc. The factory
application for a complete CD-4 decoder
with the latter IC has already been
published several times. For complete-
ness’ sake this diagram is given once again
(figure 6); the necessary low-pass and
band-pass filters are supplied by Matsu-
shita. It should be noted that this

circuit includes the preamplifiers for
magnetic cartridges, as well as changeover
switches and biassing arrangements for
semiconductor cartridges. These circuits
are not included in the SQ and QS de-
coder designs shown previously.

UD-4

This system (by Nippon Columbia) can
be thought of as a combination of a
matrix system such as QS and a discrete
system such as CD-4. As is the case with
CD-4, use is made of a four-channel
gramophone record (the normal two
baseband channels and two sub-channels
which are frequency-modulated on a
30 kHz carrier), but with UD-4 these
four channels are used to transmit a
four-channel matrix.

In the basebands of the gramophone
record — the ones which are reproduced
by normal stereo equipment - the two-
channel basic matrix (BMX) is recorded.
This matrix can first be treated in the
same way as the other matrix systems
(SQ and QS). The encoder for BMX is
described by the formulae:

R T = (1.707 — 0.707 j)Rp +
(1.707 + 0.707 j)R B +
(0.293 - 0.707 j)Lp +
(0.293 + 0.707 j)L B ;

L T = (1.707 + 0.707 j)L F +
(1.707 - 0.707 j)L B +
(0.293 + 0.707 j)Rp +
(0.293 -0.707 j)R B .

The factors 1.707 and 0.293 are 1

and 1 respectively.

V2

7

652 — elektor june 1975

quadro in practice

The decoder matrix for BMX is as
follows:

Lp = (1.707 - 0.707 j)L T +
(0.293 + 0.707 j)R T ,

Rp = (0.293 - 0.707 j)L T +
(1.707 + 0.707 j)R T ,

Lp =(1.707 + 0.707 j)L T +

(0.293 - 0.707 j)R T ,

Rb =(0.293 + 0.707 j)L T +

(1.707 - 0.707 j)R T .

From calculation it is found that these
signals are composed of the following
blends of the original signals:

L*i = 4 Lp + 2(1 — j)L B + 2(1 + j)Rp,

R£,l = 4 R F + 2(1 — j)Lp + 2(1 + j)R B ,

L6,l =4 R B + 2(1 — j)R B + 2(1 + j)Lp,

Rb,1 = 4 Rb + 2(1 - j)Rp + 2(1 + j)L B .

Here, crosstalk is fully symmetric: the
original left-front signal, for example,
is reproduced at the left-front, and also
(attenuated 3 dB and phase shifted ±45°)
at left-back and right-front respectively.
This signal is not reproduced at the right-
back. As a result of this symmetry, the
practical results with this matrix are
quite good.

With the four-channel gramophone rec-
ord the carrier channels (30 kHz, fre-
quency-modulated) are used for trans-
mitting two extra matrix channels, *T’
and ‘Q’, as sum and difference signals:
Tj + Tq and Tj — Tq, respectively. The
encoding matrix for these channels is:
T T = 0- 1 )L F + (j + 1)R F -

(j+DLe-O- DR b .

and

Tq = 1.414 jLp — 1.414 jRp —
1.414 jL B + 1.414 jR B .

The corresponding decoder matrix for
these additional channels is:

Lp,2 = -(l+j)T T - 1.414 JTq,

Rf, 2 = (1 — j)T T + 1.414 jTq,

Lb, 2 = -(1 -j)T T + 1.414 jt Q ,
Rfc-d +j)Tr - 1.414 jTq.

After calculation it is found, for example,
that the result of these additional chan-
nels for reproduction of the left-front
signals is:

Lf, 2 =4 Lp - 2(1 + j)Rp - 2(1 - j)Ls.
Of course, this signal is combined with
the signal of basic matrix already derived
above:

L£,l = 4 Lp + 2(1 — j)L B + 2(1 + j)Rp,
so that the final result of the complete
QMX matrix (after addition of the two
component signals) is:

Lp = 8 Lp.

As the final result for the other three
original signals (Rp, L B and R B ) is the
same, this can be considered as a fully
discrete system: the crosstalk between
the four quadrophony channels is theor-
etically zero!

For completeness’ sake we can also
investigate the final result of the three-
channel matrix TMX. In this case the
two basic channels and the T channel
are used; there is no Q channel. This
system is particularly interesting as a
possibility for ‘FM-quadro’ and, further-
more, the latest theories show that an

Figure 8. Block diagram of an extensive
UD-4 decoder, suitable for the four-channel
matrix (QMX). The first section can be com-
pared with the CD-4 decoder, and also con-
tains the dynamic preamplifiers; the second
section comprises the phase shifters and
matrix. To economise on phase shifters, the
signals L, R, T and Q are not routed through
separate phase shifters (in that case eight of
these would be needed); instead the blend
signals L+R, R-L-T, R-L + TandQ
are formed first, so that four phase shifters are
sufficient.

Figure 9. No ICs are available for UD-4 as yet.
However, starting from imaginary ICs (derived
from existing types, designed for other sys-
tems) this diagram has been set up. It gives an
impression of what a UD-4 decoder with ICs
could look like. Even more thorough inte-
gration is conceivable: by integrating both
input preamplifiers in IC1, sufficient room
would be made available in IC2 to accommo-
date IC3.

optimal three-channel system is in fact
inherently better than a four-channel
system (used non-redundantly) for giving
four-loudspeaker quadrophonic repro-
duction! The output signal of the left-
front channel is then composed as fol-
lows:

Lp = 6 Lp + 2 Lg + 2 Rp — 2 Rp.

So channel separation is far from com-
plete (about 12 dB), but owing to the

high degree of symmetry and the lack of
phase shift in the final output, the
practical results are particularly good.
As regards the extra carrier channels,
the practical problems encountered with
UD-4 run parallel with those of CD-4.
Here, however, a different solution has
been found for the noise (and crosstalk)
problem: the bandwidth of the Tp and
Tq channels is limited to about 6 kHz,
so that above this frequency the system

elektor june 1975 - 653

is actually reduced to BMX. So localis-
ation at high frequencies becomes less
accurate, but this is hardly noticeable be-
cause the human ear itself is less direc-
tional at these frequencies.

To reduce low-frequency interference
between the carrier channels (or to make
matters even more complex?) the Tq sig-
nal is attenuated with respect to Tp for
lower frequencies. This means that the
difference between the two carriers

654 - elektor june 1975

quadro in practice

(Tj + Tq and Tj — Tq, respectively) be-
comes less at lower frequencies, so that
there is less chance of mixture products
occurring. During playback the Tj and
Tq signals must therefore undergo differ-
ent frequency corrections to restore
them to equal levels. A block diagram for
a simple UD-4 decoder (BMX only) is
shown in figure 7, whilst figure 8 is the
block diagram of a complete four-
channel decoder for QMX. Figure 9
shows what a circuit diagram with ICs
could look like. For UD-4 no ICs have
been developed yet, so that this diagram
is based on ICs similar to those which are
currently available for other systems.

An SQ/QS/CD-4/UD-4 decoder

A universal quadrophony decoder for all
systems is, of course, a highly complex
apparatus, as it must comprise all com-
ponents for all systems. Moreover, since
at the moment hardly any suitable ICs
can be obtained via the retail trade,
currently-available discrete components
must be used ...

The block diagram for such a decoder is
shown in figure 10. Figure 10A com-
prises the dynamic pre-amplifier and
30 kHz FM detector for the left channel,
and figure 10B the same part for the
right channel. These two parts corre-
spond with the section in figures 5 and 8
up to and including the FM demodulator

in the carrier channels. In this case, how-
ever, the RIAA frequency correction
is divided over two blocks, a and g (a'
and g', respectively). Furthermore, a
number of suppression switches for the
carrier channels have been added (e, i, e'
and i') which become operative as soon
as the level of the left carrier channel
becomes too low. They are driven via
block h; this carrier signal detector also
drives a pilot lamp. However, it should
be noted that the switches e and e' can
be left out: they are in fact redundant
and have been found in practice to give
rise to problems! This will be discussed
in greater detail later on.

The outputs 1 , 2, 3 and 4 of the blocks A
and B are connected to the blocks in
figures 10C and 10D. Figure 10C con-
tains the centre section of the UD-4 de-
coder of figure 8, namely the part up to
and including the second sum/difference
matrix. The blocks 1, 1', m, n and n' are
successively the low-pass filters
(4.5 kHz), the first sum/difference
matrix and the frequency correction
for the T-p and Tq signals. They are
followed again by two more signal-
suppression switches (o and o') which
become operative as soon as the level
of the Tj signal from block o' becomes
too low. The blocks q and r are the sum/
difference matrix and following buffer
amplifiers.

Figure 10D comprises the second part
of the CD-4 decoder, (figure 5). The
blocks s and s' contain the 1 5 kHz low-
pass filters, delay circuits and FM/PM /
FM frequency correction networks for
the carrier channels.

Finally figure 10E contains the system
selection switches, phase shifters and
matrices for the four systems or stereo.
So it is connected to the baseband chan-
nel outputs of blocks A and B, and also
to the four outputs of block C and the
four of block D. Depending on the
position of the switches in blocks v and
y, this section operates as a simple 1
SQ decoder (figure 1 A), as a simple
QS decoder (figure 3A), as the last part
of the UD-4 decoder (last section of
figure 8, from the phase shifters), or as
the output buffer amplifier for stereo
or CD-4. However, it does not include
‘logic’ for SQ nor a ‘ Variomatrix’ for QS.

The complete circuit

Figure 1 1 shows the complete circuit
of the universal decoder described above.
The parts A up to and including E in this
diagram correspond to the same parts
in the block diagram (figure 10).

The total circuit has purposely been
left completely in accordance with the
original design by Nippon Columbia (the
‘Denon’ UDA-100). For this reason, vari-

Figure 10. Block diagram of the universal
decoder described in this article.

The blocks A and B comprise the dynamic
preamplifiers and 30 kHz demodulators: a, a',
g and g’ are low frequency preamplifiers with
RIAA correction; c, c', d, d', f and f' are the
30 kHz demodulation chains, and e, e', h, i
and i are carrier channel suppressors.

Block C is the central part of the UD-4 de-
coder; it contains a.o. the first (m, n and n’l
and second (q) sum-difference matrix.

Block D is the second part of the CD-4 de-
coder: the ANRS (t and t') and the sum-
difference matrix (u).

Finally, block E contains the system selector
switches (v and y), the phase shifters (w) and
the matrices for SO, QS and UD-4 (x).

quadro in practice

ous resistors from the E-24 series occur
in it. For the resistors and capacitors it
is recommended that 5% types should be
used. The choice of alternative semi-
conductors is generally not critical. The
transistors 2SC1000BL can be replaced
by TUNs, the 2SC1000BL/CR,
2SC1000GR and 2SC1213 by BC 107B,
the 2SC732GR by BC 109C and the
2SA493GR by BC 1 79B or C. The diodes
can be replaced by DUSs, whilst for the
FETs almost any type of N-channel FET
is suitable (E 300, BF 245, etc.).

The ICs TA7122AP are low-noise
differential amplifiers; when replacing
these types by alternatives, the connec-
tions and some of the resistor values will,
of course, have to be changed. For IC3
and IC6 two possible types are indicated:
the NE565 (Signetics) and the HA 1 173,
with slightly deviating resistance values.
Various adjustment points are also indi-
cated in the circuit (figure 1 1 A/B). The
trimmers near IC3 and IC6 (VR2 and
VR5) serve for adjusting the centre fre-
quency of the PLL (30 kHz). VR3 and

VR6 (near FET2 and FET4) set the
level of the demodulated carrier chan-
nel separation for CD-4 and UD-4.
The potentiometers (‘B’) in the output
circuit of the first low-pass filters serve
to adjust the level of the baseband chan-
nels; in principle, their function is the
same as that of the adjustment resistors
VR3 and VR6. Trimmer VR7 (near
TR16/17 adjusts the trigger level of the
suppression switches (blocks i and i').
In figure 10C only one trimmer occurs:
VR5 (near TR26/27). This serves to ad-

656 — elektor june 1975

practice

just the trigger level at which the
suppression switches for the T and Q sig-
nals (blocks o and o') become operative.
Finally, in block D there are four trim-
mers for adjusting the ANRS. VR1 and
VR2 serve for zero adjustment: they set
the control signal level at which the

ANRS begins to function. VR3 and
VR4 adjust the ANRS control range;
they also influence the zero setting
mentioned above. Furthermore, it should
be noted that the trimmers in the carrier
channels ( VR3 and VR6 in blocks A and
B) also influence this zero setting.

Final notes

As will be clear from what has been said
above, the complete apparatus is not
required for every system. For SQ and/or
QS block E is sufficient; at the most an
SQ logic with MCI 3 14 and MC1315
(figure 2) or extra ‘blend’ resistors could

quadro in practice

Figure 11. Complete diagram of the universal
decoder. The blocks A up to and including E
and the blocks a up to and including z corre-
spond to the indications in the block diagram
(figure 10). The blocks e and e' (in A and B)
may be left out (see text). Replacement types
for the semiconductors are indicated in the

be added and also a Variomatrix for QS.
For the ‘blend’ resistors (see figure IB)
one can experiment with a resistor of
about 68 k between the outputs Lp and
Rp and a resistor of about 1 0 k between
the outputs Lg and Rg.

CD-4 requires the blocks A, B and D, in

combination with a simple CD-4/stereo
switch and output amplifiers as in
block E. For UD-4 only block D is re-
dundant; if only UD-4 is required, the
system selection switches can be sim-
plified and various matrix resistors can
be omitted from block E.

658 - elektor june 1975

quadro in practic*

Finally it should be noted that two 113, respectively,

of the four signal suppression switches replaced by one 4k 1

in blocks A and B can be left out. This

concerns the part from Cl 2 up to and

including Cl 8 (block e) and the part Literature:

from R86 up to C52 (block e'). The Quadro 1-2-3-4 ...:

resistor combinations R35/36andR101/ 1974, p. 33-3 7.

can then each be Proposed universal encoding standards
resistor. for compatible four-channel mixing:

R. Itoh, Journal of the Audio Engin-
eering Society (JAES), April 1972,
p. 167.

Elektor, December Discrete-matrix multichannel stereo:

D.H. Cooper and T. Shiga, JAES, June

1972, p. 346 and July 1972, p. 493.

The Japanese Regular Matrix:
P.B. Fellgett, Hi-Fi-News, Dec. 1972,
p. 2393.

SQ Matrix Quadrophonic Discs:
B.B. Bauer, G.A. Budelman and

D.W. Gravereaux, JAES, Jan. 1973,

p. 19.

The Principles and Performance of the
Sansui QS Variomatrix: S. Takahashi
and K. Hirano, JAES abstract,
Feb. 1973.

Discrete 4-channel records: W.R. Isom,
JAES abstract, Feb. 1973.

The CD-4 Mark-II Modulation System:
Y. Kokubun, S. Muramoto and J. Eargle,
JAES, July /August 1974, p. 416.

j

F" i™ : F^

iTTiiwiHF*

IIP

sbiiESB

5?

KlHj

!

LJ ^ p—

il >^-4

3

!

_p.

i

660 — elektor june 1975

elektor shorthand

sniff race control

elektor shorthand

From various enquiries it has become
clear that some of our readers feel that
they have been plunged in at the deep
end. Elektor’s ‘shorthand’ style of sym-
bols and conventions seems to have led
to some confusion, in spite of our efforts
to the contrary, so some further ex-
planation seems to be called for.

Resistor and capacitor codes

When giving the values of resistors and
capacitors, decimal points and large num-
bers of zeros are avoided as far as possible.
To this end, extensive use is made of the
international abbreviations:
p (pico-) = 1 0 -12 = one millionth of
one millionth;
n (nano-) = 10 -9 = one thousandth
of one millionth;
fx (micro-) = 10 -6 = one millionth;

m (milli-) = 10“ 3 = one thousandth;

E = 10° = unity;

k (kilo-) = 10 3 = one thousand

times;

M (mega-) = 1 0 6 = one million
times;

G (giga-) = 10 9 = one thousand

million times;
T (tera-) = 10 12 = one million
million times.
Furthermore, the symbols £2 (ohm) and
F (farad) are usually omitted, since it is
normal practice to state resistance values
in ohms and capacitance values in farads.
Finally, the decimal point is usually re-
placed by one of the abbreviations (p, n,
H ...) listed above (This has also been
accepted practice for some years).

A few examples may serve to clarify all
this:

Resistance value 2k7: this is 2.7 k£2, or
2700 £2.

Resistance value 470: this is 470 £2.
Resistance value 3M9: this is 3.9 M£2, or
3,900,000 £2.

Capacitance value 4p7 : this is 4.7 pF, or
0.000 000 000 004 7 F ...
Capacitance value 100 n'. this is 100 /xF.
Capacitance value 4700 /x: this is
4700 n F, and could have been written
as 4m7 — but never is.

Capacitance value 10 n: this is lOnF,
and is also sometimes written (but
not in elektor!) as 10,000 pF or
0.01 n F; or even as lOkpF (lOkilo-
pico-Farad), which is a horrible con-
fusion of symbols. In the same way
one sometimes finds fx/xF (micro-
micro-Farad) instead of pF.

Semiconductor type numbers

Very often, a large number of equivalent
types for one integrated circuit exist with
different type numbers. On closer exam-

ination, a group of digits are often found
to be identical, but they are pre- or
suffixed with letters and digits which
denote the manufacturer. As an example,
a popular op-amp is variously denoted
as /1A741, LM741, L741, MC1741,
MIC741 , RM741 , SN72741 orZLD741.
to name a few. To cut through this
confusion, this IC is referred to in elektor
as a ‘741’ which means that we
couldn’t care less who makes it, provided
it meets the specifications...

In the same way, ‘7400’ (or sometimes
even ‘00’) stands for SN7400, SN74H00,
DM7400, MC7400, etc., and the last two
figures are used in the same way for
other ICs in the 7400 series.

Finally, transistors are sometimes listed
‘TUP’ or ‘TUN’. This is explained else-
where. Transistors can also be listed as
BC107, for instance; a long list of
equivalent types for the BC107 series is
also given in the TUP/TUN list.

TV sound

The layout for the printed circuit board
for ‘tv sound’ (elektor no. 2, p. 236)
shows two capacitors marked ‘C x ’. These
are 1 00 n decoupling capacitors.

It should be noted that the circuit in its
present form will not work with some
of the latest models of tv receiver, as
these use ceramic filters throughout
(instead of IF coils) so that there is little
or no stray field for the coil to pick up.
Elektor laboratories are presently
working on a front-end that will convert
this design into an entirely independent
receiver for tv sound.

Modifications to
Additions to
I mprovements on
Corrections in

Circuits published in Elektor

E. Waschkowski

sniff race
control

With many model car race-tracks the
circuit consists of two concentric tracks,
so the inside track is always shorter than
the outside track. A widely employed
method to have both race-cars still cover
the same distance is to constantly switch
them from the inside to the outside
track and back again by means of cross
points.

An entirely different approach is the
so-called sniff race. During this race the
two cars follow the same track. The
cars start with half a track distance be-
tween them. The car that catches up with
its opponent until it bumps into it
(sniffs at it) is the winner of the race.
With ordinary race-tracks it is not poss-
ible to have two cars ride the same track
and be controlled independently, but
by means of a small modification to the
d.c. supply and the cars this becomes
possible.

Figure 1 shows what the supply of the I
race-track usually looks like. Via a bridge I
rectifier each of the two tracks is fed I
with a pulsating direct voltage. In the I
modified version the positive and nega- 1
tive half-cycles of the alternating voltage I
are separated with diodes and made in- 1
dependently adjustable. Furthermore, I
the motors of the car are fitted with |
a series diode so that two cars can be I
controlled simultaneously and indepen- 1
dently of each other (see figure 2). One I
car then runs on a positive voltage (in I
this case car 1) and the other on a nega- 1
tive voltage. For the latter car, how- 1
ever, the polarity of the motor must be I
changed, or the car will run backwards. I

advertisement

elektor june 1975 - 661

advertisements
are eye-catching
don’t you agree?

If you supply electronic components or
equipment and would like to advertise in
Elektor, please contact us at Canterbury,
tel. (0227) 54439 or telex elektor cantbry
965504.

you haven’t
yet missed
the boat,

you can still get copies of Elektor nos. 1, 2 and 3,
or a complete subscription including these back
numbers, if you fill in the reply card on page 664.