to-date electronics for lab and leisure

a modern grid
dip meter

touch tjuning
for F.MS. tuners

digifarad:

capacitance meter
with digital readout.

, page 1 0-06

Construction of the
touch tuning circuit re-
1 quires a bit of handiwork
with a fret-saw. The
printed circuit board
consists of three sections;
once separated, the two
smaller sections are
mounted perpendicularly
to the main board to
form a compact module.

The gate dipper is a
useful little device that
can be used to determine
the resonant frequency
of tuned circuits.
Basically, it is the modern
equivalent of the old
grid-dip meter.

The 'heart' of the strain
gauge is the 'stress
absorber', which consists
of a sheet of suitable
metal onto which a
| bridge configuration of
* four electric-resistance
strain sensors are bonded.

'I played TV games . . .'
i is a description of the
TV games computer,
written by a novice and
for novices! One thing
has become quite clear:

; with a little practice,
even fairly sophisticated
programscanbe designed.

HHffl

October 1979
Jvolume 5
number 10

nu4Uiiiui|'(d'/ n r

-jhf

touch tuning

An important selling point of modern stereo tuners is the
number of preset stations which can be selected. However for
the home constructor, this is often a feature which must
regrettably be foregone, being regarded in many designs as
something of a luxury. The circuit described is intended to
remedy that situation, by providing for up to 9 touch con-
trolled preset stations. The only restriction is thet the receiver

battery saver ....
impedance bridge

It is often very useful

It is often very useful to be able to match the values of

do this is by using an impedance bridge. The circuit described
is quite adequate for this purpose and it is also capable of
measuring resistances between 100 n and 1 M and capaci-
tances between 100 pF and 1 pF.

new programs for the SC/MP

Good news for SC/MP fans: two new records have been
added to the ESS range. One contains the complete NIBL-E
program: the other includes some games, a 'running script 1
program, 'tracer', ‘disassembler' and 'biorhythm', Some

digital rev counter (a. ohdei

digifarad (J. Gutheri

Given the fact that many types of capacitor - especially
electrolytics - have a wide tolerance (20% is fairly common),
it is often desirable to be able to measure capacitances both
quickly and with a reasonable degree of accuracy. Of course
a capacitance meter also enables one to measure the value of
those piles of unmarked capacitors which end up at the
bottom of one's junk box, or to test 'suspect' capacitors for
potential faults - in short it represents a useful addition to

short-interval light switch

p.c.b. for variable fuzz box

gate-dipper

Tuning resonant circuits in high frequency equipment nor-
mally requires fairly expensive test gear which not every
hobbyist can afford. However there is a reasonably cheap
alternative available, namely a gate dipper, which allows the
resonant frequency of the tuned circuit to be ascertained
simply and quickly.

strain gauge (w. v. oreumei)

There are few projects which have not formed the subject of

gauge falls into that category. This in itself is perhaps slightly
surprising, since there are a number of possible applications
for such a device - a training aid for 'strength sports',
measuring loads on cables, etc. or simple weighing purposes.

I played TV games .

programmable sequencer (c. Vos:
missing link

Elektor Publishers Ltd.. Elektor House.

10 Longport, Canterbury CT1 1PE, Kent. U.K.

Tel.: Canterbury (0227) 54430. Telex: 965504.

Office hours: 8.30 - 12.45 and 13.30 - 16.45.

Bank: 1. Midland Bank Ltd.. Canterbury. A/C no. 11014587
Sorting code 40-16-11, Giro no. 315.42.54

2. U.S.A. only: Bank of America, c/o World Way
Postal Center, P.O. Box 80689. Los Angeles.

CA 90080. A/C no. 12350-04207.

3. Canada only: The Royal Bank of Canada,

c/o Lockbox 1969, Postal Station A, Toronto,

Ontario, M5W 1W9. A/C no. 160-269-7.

Please make all cheques payable to Elektor Publishers Ltd. at the

Elektor is published monthly.

Number 51/52 (July/August) is a double issue.
SUBSCRIPTIONS: Mrs. S. Barber

January to December ii

Semiconductor types
Very often, a large number of

with different type numbers. I
this reason, 'abbreviated' type
numbers are used in Elektor
wherever possible:

• '741 ' stand for pA741 ,
LM741, MC641. MIC741,
RM741.SN72741.etc.
e TUP- or 'TUN' (Transistor.
Universal. PNPor NPN respe
ively) stand for any low fre-
quency silicon transistor tha

of zeros are avoided wherever
possible. The decimal point is
usually replaced by one of the
following abbreviations:

Resistance value 2k7: 2700 «.
Resistance value 470: 470 n.
Capacitance value 4p7: 4.7 pF, t
0.000000000004 7 F . .
Capacitance value 10n: this isth

10.000 pF™r > 0' V ° c - i.

New home-video standard from
Philips

8 hours of TV from one cassette

The latest Philips and Grundig top-hit in
the home-video field, the Video-2000
system, has received world-wide atten-
tion, One cassette, containing about
1 000 feet of Vi-inch tape, can be used to
record eight hours of colour TV. One
cassette costs about £ 20, so that one
hour of TV program costs just over £ 2
to record. By way of comparison, the
first colour video recorder used well
over £ 20 of tape per hour. The picture
quality hasn't suffered by this drastic
cost reduction. All in all, it's not just
another step forward - it's a giant leap!
Obviously, Philips and Grundig hope
that the new video cassette will be
accepted as an international standard,
and that it will prove the same long-
term success as its predecessor: the
compact audio cassette.

The Philips recorder, designed around
this new cassette, is a beautiful piece of
modern technology.

The first obvious difference between the
Video-2000 system and older Philips
systems (and JVC's VHS and Sony's
highly-praised Betamaxl is the narrow
tape-track used. Until now, Vi-inch
tracks were fairly common; Philips and
Grundig use less than half - the Vi-inch
tape in their cassette is used in both
directions, since the cassette can be
'turned over'. It is used in the same way
as the audio compact cassette -the
only difference is that it is used to
record television programs!

The new cassette can not be used on
older video recorders, like the N1700.
That particular machine will be going
out of production in the second half of
1980; suitable cassettes will, however,
remain available for some time to come.
A new video recorder, the VR2020, is
designed to use the new cassette. The
price is expected to be about 30% more
i than that of the N1700— around the
£ 600 mark. Several new features are
! included in the new machine, and these
deserve some further explanation.

Use of the tape

Before discussing the recorder itself, it is
a good idea to take a look at the way in
which it writes the program materia! on
tape. Figure 1 shows where the infor-
mation is written on the tape.

Br 1979- 10-01

Since the tape can be 'turned over', the
upper and lower half of this 'tape map'
are mirror images. Starting from the
outside, the first 650 pm are used for a
(mono) audio track. At a later date, if
stereo sound ever gets off the ground
for television broadcasting, this area can
be divided into two 250 pm tracks with
a 1 50 pm gap.

The next 4.85 mm wide section is
reserved for the video signal. This is
recorded in a single track, like the audio
signal; as in most video recorders, narrow
(22.6 pm wide) tracks are recorded at a
slight angle in this section. In this
particular case, the video tracks are
angled at 3° with respect to the tape
'axis' - for clarity, an angle closer to
30° is used in figure 1 . . . The final
section of tape before reaching the centre
(where the mirror image begins) is
unused at present. This 300 pm wide
strip of tape can be used, at a later date,
for various control signals.

All these sections are repeated on the
other side of the centre line, for what is
called (in analogy with gramophone
records) 'the other side of the tape'.

Video recording

Which is what the whole exercise is
about . . . The video signals to be
recorded run up to fairly high fre-
quencies (approximately 4.8 MHz). In
any type of recording, the 'detail' that
can be written depends on how coarse
or fine the writing implement is. In a
tape recorder, the recording implement
is the tape head; its 'gap width'
determines the detail that it can write

However, the 'space' required to record
one period of a 4.8 MHz signal on tape
depends on the speed with which the
tape runs past the heads. The higher the
speed, the longer a single period will be

stretched on the tape. If the tape runs
relatively slowly, a larger number of
periods could theoretically be recorded
on a small section of tape; however, the
tape head is too 'blunt' to make this a
feasible proposition. The result would
be poor picture quality. Somehow, the
speed of the tape relative to the head
must be increased until the picture
quality leaves little to be desired. If the
tape is run at high speed past a stationary
head, the picture quality can be quite
good - but the recorder will 'eat' tape,
both in feet of tape required per minute
and in life expectancy of the tape . . .
For this reason, it has become standard
practice in video recorders to use a
rotating head drum, incorporating two
or more heads. This drum revolves at
high speed, so that the heads move at
high speed past the tape, even if the
latter is transported relatively slowly.
The head drum is mounted at a slight
angle with respect to the tape and very
narrow video tracks are used, so that a
fairly slow tape transport suffices to
move the tape up sufficiently to write
the next diagonal track adjacent to its
predecessor.

All this may seem rather complicated,
but it is basically similar to typing. Even
if you type a lot of letters, it takes a
while to fill the page - certainly if you
use the minimum line spacing, so that
each new line practically touches the
one above. Something similar occurs in
a video recorder; the main difference
being that the tape is moved slowly and
constantly, instead of jumping up 'one
line at a time' like the paper in the
typewriter. If you can visualise the
paper moving up at a constant speed, so
that it has just moved up one line by the
time you start to type a new line (so
that the lines slope down slightly), you
have the principle of the video recorder.

Figure 2 illustrates how this system
operates in practice. Two heads are
mounted on the drum, and the latter is
mounted at a slight angle with respect
to the tape. As the tape is transported at
a speed of 2.44 cm/s (just less than
1 inch/second, or about half the speed
of an audio cassette recorder!), diagonal
tracks are written on it by the heads on
the upper half of the drum. The lower
half of the drum runs at a much lower
speed, and takes care of the tape
transport — it operates as a large
diameter 'capstan'. The diameter of the
drum is 65 mm and the upper half
rotates at 25 revs per second, so that the
two heads (K1 and K2) move at 5.08 m/s
— or just under 17 ft. per second! The
tape is 'wrapped around' half the
circumference of the drum, so that as
one head leaves the tape the other just
starts to write on it. It will now be
apparent how the tracks are recorded.
The tape is almost stationary with
respect to the upper half of the drum.
One head can therefore record a track
length equal to approximately half the
diameter of the drum -very roughly,
100 mm or 3/8".

We said that thetape is almost stationary
with respect to the head. To be more
precise, the tape is transported over
22.6 Aim as one full track is recorded.
When the second head 'hits the track',
the tape has moved up just far enough
to enable this head to record its track
parallel to the previous one, without
overlapping. The final result is a series
of diagonal tracks, as illustrated in
figures 1 and 2.

A separate, stationary head is used to
record the audio signal. This 'sound
track' is located at the outer edge of the
tape, as illustrated in figure 1. The erase
heads are also stationary.

By varying the voltage applied to it, the
height of the heads can be adjusted.

Dynamic Track Following
It's all very well being able to vary the
position of the heads, but first a control
voltage must be derived in some way.
Figure 3 again shows four video tracks
with an exaggerated angle (the true
angle being only 3°). Track f2 is written
first, then f4 is recorded, and so on.
And this is where it gets complicated.
Tracks f2 and f3 are written by video
head two; tracks f4 and fl are recorded
by head one. Simultaneously with the
video signal, a 'pilot tone' is recorded on

each track. Each head records two
different (relatively low frequency)
pilot tones alternately. Track f2 contains
a 117 kHz pilot tone, recorded by K2
(video head 2); track f4, written by K1,
includes a 164 kHz pilot tone; on track
f3, the pilot tone frequency is 149 kHz;
finally, the pilot tone on track fl is

Vertical positioning of the video heads
With the extremely narrow video tracks
written diagonally on the tape, pos-
itioning of the video heads during
playback is obviously highly critical.
Some way must be found to move the
heads slightly until they are centred
exactly on the corresponding tracks. A
most intriguing solution has been found.
The video heads are both mounted on a
little piece of piezo-ceramic material.
This is the material used in no-battery
electric lighters: when it is compressed,
a voltage appears across the ends,
sufficient to draw a spark. However, it
also works the other way: if a voltage is
applied across the ends of the material,
its shape will vary! The so-called PXE
is used in this way in the video recorder.

I

3

102 kHz. This cycle is repeated for the Azimuth

next set of four tracks, and so on. The video heads are mounted in the

During playback, the pilot tones are head drum at a relative angle of 30°.
retrieved together with the video signal, with respect to the tape, head one is

If the corresponding head (K1) is mounted at 90° - 15° -3° = 72°, head

correctly positioned, a clean 164 kHz two is mounted at 90° + 15° -3° = 102°.

tone will be retrieved from track f4. T h j s j s illustrated in figure 4, where
However, if the head is slightly high, both heads are shown simultaneously

some of the 149 kHz signal on track f3 w j t h respect to the tape. There is good

will be mixed with this 164 kHz signal, reason for the relative angle between the
producing a 15 kHz beat signal; if the two heads,
head is low, a 47 kHz beat signal will
appear (164 kHz (f4) - 117 kHz <f2> =

47 kHz). For the other head (K2), the
opposite is true: if it is too high, a
47 kHz signal is produced; a 15 kHz
beat signal corresponds to 'too low'.

The amplitude of the beat signals is used
as a basis for the 'head height' control
signal.

As audio recorder enthusiasts will know,
if a playback head is tilted slightly with
respect to the recorded tape track, the
high frequency response is drastically
reduced. Correct 'azimuth' setting is
essential for high quality playback. In
this video recorder, the result is that a
track originally recorded by K1 will
only be 'read' by K2 with a severely
reduced high-frequency response. In
practical terms, this means that K2 will
only 'see' frequencies up to a few
hundred kilohertz on Kl's tracks — K2
will reproduce Kl's pilot tones, but it
will not reproduce the video signal!
Cunning . . .

Head positioning during recording
During recording, accurate positioning
of the heads is also required, to keep the
tracks up against each other without
overlap. To this end, one of the heads is
fixed in an 'average' position; the height
of the other is adjusted so that its tracks
are correctly positioned.

A complete TV picture consists of
625 lines, written in two 312.5 line
'frames'. Between each set of 312.5 lines,
a short 'vertical blanking' interval is
required. No video signal is recorded in
this interval. One complete picture
(two frames) is recorded 25 times per
second . . . exactly the rotation speed of
the head drum! Coincidence? Don't you
believe it. Each head records one frame,
and two frames make one picture. Each
track includes a vertical blanking
interval, which can be used to record a
control signal. In the VR2020, a
223 kHz signal is recorded at this point,
for 96 ps. Immediately after this, the
head is switched to playback for a
further 96 ps. The result of these
manipulations is sketched in figure 5.
Bearing in mind that the right-hand
track is recorded first (the tracks
themselves are recorded from lower left
to upper right, but the tape movement
is also from left to right) it will be

apparent that when head K1 is switched
to playback ('read') it will detect the
223 kHz tone recorded on the previous
track by K2 if the latter is high. Simi-
larly, if K2 is low it will detect the
223 kHz signal from the previous track
during its own read cycle. Using this
information, the height of one of the
heads is adjusted until the tracks just
mesh correctly.

As illustrated in figure 5, 960 ps
(15 lines) of video are recorded on each
track before the 223 kHz test signal.
Each track therefore contains the
following sequence: first 15 lines at the
end of a frame, then 96 /is worth of the
test signal, then the head is switched to
playback for a further 96 /is, then the
first 294.5 lines of the next frame. The
other head now takes over on the next
track, recording the last 15 lines of the
frame, and so on.

It will be obvious that some fairly fast,
complicated and accurate switching is
required for the VR 2020 to work . . .
For this reason, the whole system works
under microprocessor control.

'Automatic Tracking'. The control
voltage from a Dynamic T rack Following
'discriminator' is used to control a tape
servo. During recording, a constant tape
speed is maintained by referring the
output of a tacho generator to that of a
crystal oscillator; during playback, the
DTF control voltage is used as a
reference, so that both the tape speed
and the tape position with respect to
the heads is accurately maintained.

The rapidly rotating half of the head
drum must also run at exactly the right
speed. This is achieved by measuring the
speed of the drum with a phototransistor
that gives one pulse for each rotation of
the drum; the frequency of these pulses
is compared with that from the tacho
generator that measures the tape speed.
The VR 2020 is quite easy to operate.
Most of the functions are automated,
and it is possible to 'program' it up to
16 days in advance. The cassettes for
the Video 2000 system include mechan-
ical 'record locks' that can be used to
protect recorded tapes from inadvertent
re-use.

Other gimmicks

It may seem surprising, but the head
drum is heated in the VR 2020. Among
other advantages, this ensures a constant
diameter, reduces the tendency of the
tape to 'stick' to the heads and reduces
the wear on the heads.

If, during playback, both heads are
found to be slightly high or if both are
slightly low, the tape is shifted up or
down slightly instead of adjusting the
position of the heads. This is called

The future

As mentioned in the introduction,
Philips and Grundig hope that this
system will become an international
standard. Apparently, ITT have already
decided to use the new cassette, and the
German manufacturers Loewe Opta
and Metz are taking a long, hard look at
it. It looks as if this system has a good
chance of making the grade!

Philips Gloeilampenfabrieken,
P.O.B. 523.

Eindhoven,

The Netherlands.

Grundig AG,
Kurtgartenstrasse 37,

8510 Fiirth /Bayern,

West Germany.

Switching transistors approach
1000 V Barrier

The switch mode power supply has a
number of advantages over other
conventional supplies. However, their
use in high power circuits has been
limited by the absence of transistors
with sufficiently good characteristics to
meet the heavy demands placed on
them under high voltage switching
conditions. An ideal switching transistor
should have characteristics that include:

• very low Vce sat

• very low leakage current

• very good switching characteristics

• very good ruggedness

• good reliability

Characteristics that are difficult to obtain
in high voltage high power transistors.
The SGS-ATES MULTIEPITAXIAL
MESA technology, which was created to
overcome these problems, gives an
excellent compromise between these
characteristics. In addition to this it
gives the possibility of making comp-
lementary NPN-PNP high voltage, high
power transistors, a feature impossible
to find in other high voltage, very high
power technologies.

In the Multiepitaxial Mesa technology a
heavily doped N* substrate is used as a
foundation onto which is epitaxially
grown a normally doped N type layer.
On the N type layer is grown a second
epitaxial layer of lightly doped N~ type
material. These two epitaxial layers
form the collector of the transistor. This
type of collector construction gives
extremely good ruggedness in Es/b
conditions. On top of the collector is
grown a third epitaxial layer which is to
form the base of the transistor (into
which an N** type emitter diffusion is
made). This epitaxial layer, in order to
maintain the high voltage characteristics
of the device whilst giving good switching
times, must be of a P~ type material.
However, if the emitter diffusion was
made into the base as it stands, problems
would arise in the stability of the
transistor due to the very high electric
field between the P~/N~ layer seen at
the edge surface. Therefore an ad-
ditional P* diffusion is made into the
base epitaxial layer that has the effect
of widening the distance between
equipotential lines at the surface thus
reducing the surface electric field. The
P' diffusion does not of course reduce
the intrinsic high voltage characteristic
of the transistor.

Whilst the multiepitaxial layer construc-
tion gives a breakdown value in the
order of 1 000 V it is known that when
transistors are separated by mechanical

1979- 10-05

VcBOIminl

v CEO(min]

VCE(sat)m<

’on(typ)

«s(.yp)

•fltyp)

BUW34 BUW35 BUW36 BUW44

: Vq£( s j is specified ai Iq = 5 A, Ib = 1 A for BUW34, 35,36 and Iq = 10 A,
I B - 2 A for BUW 44, 45, 46. Sw. on characteristics typified at Vcc ■ 250 V,
l c = 5 A, I B1 = I B2 * 1 A ,for 's-'f 1 1 B1 ■ 1 A (for t on l for BUW 34, 35, 36
and V C c “ 250 v - >C * 10 A ' 'B1 3 -'B2* 2 A ,,or t s and *f* *B1 * 1 A ,or
t on ,or BUW 44, 45. 46.

means, after diffusion, irregularities are
caused on the edge of the transistor
which dramatically reduce the collector/
base breakdown voltage. Obviously the
breakdown voltage of the device as a
whole is the breakdown voltage of the
weakest point, in this case the edge
surface between collector and base.
Therefore a method had to be found
which would allow separation of devices
without causing surface edge irregu-
larities.

The method found, whilst extremely
simple in concept, has had dramatic
effects in improving the collector/base
breakdown characteristics. In essence
the method used is to isolate each
transistor on the wafer by a deep
chemical edge. In this way it is possible
to achieve an extremely smooth edge on
the active part of the transistor in the
area of the base collector junction. It is
the cross section of the transistor after
the deep chemical etch that gives rise to
the name Mesa (after the mesas found in
S.W. United States and Mexico). The
deep chemical etch is carried out after
the emitter diffusion and then, to
further enhance stability and preserve
surface cleanliness, glass passivation is
carried out on the channel formed. The
cut made to separate the transistors is
then made in the innactive area between

the dice.

Using this method it has been possible
to produce transistors with a Vmax as
high as 900 V and with extremely good
reliability and ruggedness in high voltage,
high temperature conditions.

Typical electrical characteristics for
transistors constructed using the
multiepitaxial mesa technology are
shown in table 1 .

Using these transistors it has been
possible to build switch mode power
supplies with a performance never
before possible. A typical example of
these transistors in a switch mode power
supply is shown in fig. 1. This circuit,
which uses two BUW 34's in the power
output stage, is capable of delivering up
to 400 W at 24 V or by using two
BUW 45's at 24 V.

10-06 - elektor October 1979

Many FM tuners employ varicap (variable
capacitance) diodes. These are diodes
which are especially designed so that
their capacitance can be varied by means
of a control voltage. If the varicaps are
included in an LC circuit, the resonant
frequency of the latter can thus be
varied by altering the control voltage. In
most tuner designs the control or tuning
voltage is derived from a stabilised supply
and is varied by means of a poten-
tiometer. The main requirements of the
tuning voltage are that it must be stable
and affected as little as possible by
fluctuations in temperature.

touch tuning

touch-controlled preset station tuning

An important selling point of modern stereo tuners is the number of
preset stations which can be selected. However for the home
constructor, this is often a feature which must regrettably be foregone,
being regarded in many designs as something of a luxury. The circuit
described here is intended to remedy that situation, by providing for up
to 9 touch controlled preset stations. The only restriction is that the
receiver be varicap tuned.

Preset tuning can be realised by using
not one potentiometer, but a number of
potentiometers connected in parallel,
these being selected by switches (see
figure 1). Only one switch may be closed
at any given time; for example, when
switch Sb is closed, switch SI automati-
cally opens. By adjusting the preset
potentiometers so that each switch
brings in a different station, a simple
and effective preset tuning facility is
obtained.

In the circuit described here, the basic
design has been further refined, so that
using only two switches a total of
10 preset stations can be selected. By
employing touch switches, the need for
interlocking switch assemblies is avoided,
whilst the physical construction and
appearance of the switches can be
tailored to suit individual requirements.

Circuit

For a range of 87 to 1 04 MH z, the tuning
voltage of most receivers must be capable
of being varied from roughly 2 or 3 volts
to approximately 30 volts. Thus it is
clear that conventional CMOS switches
cannot be used, since they are only
capable of switching voltages of up to
15 V. However, as can be seen from the
circuit diagram of figure 2, CMOS
buffers N1 . . . N4 are used to form a
pair of suitable touch switches.

Under normal conditions the inputs of
N1 and N3 are held high via R1 and R2.
When one of the sets of touch contacts
is bridged, the input of the corresponding
gate is pulled down to ground (logic 0).
The output of the gate is thus taken
high, with the result that Cl or C2
rapidly charges up and the output of the
succeeding buffer (N2/N4) goes low.
Removing one's finger from the touch
contacts takes the output of the first
gate low again, causing the corresponding
capacitor to discharge slowly via the
parallel resistor. Thus each time one of

elektor October 1979 - 10-07

the touch switches is operated a logic 0
is applied to the up or down input of
IC1 (synchronous decade up/down
counter). This 1C counts the pulses
applied to its inputs when the LOAD
input is high, and transfers the result in
BCD form to its outputs.

Upon switch-on the LOAD input of the
counter is held low via capacitor C3, so
that the counter outputs are reset (i.e.
also taken low). When the 'up' touch
switch is operated, the counter in-
crements by one, i.e. the number 1
appears in BCD at the counter outputs.
If the up switch is touched a second
time, the number 2 appears at the
counter outputs, and so on. Touching
the 'down' switch decrements the

number on the counter outputs by 1 .
The outputs of the counter are connected
to a BCD-decimal decoder/driver (IC2).
Depending upon the BCD input data,
one of the outputs of this 1C will go low.
The counter outputs are also connected
to a BCD-7-segment decoder/driver,
which in turn is connected to a
7-segment display. In this way the state
of the counter (and the output of IC2
which is active) is clearly indicated.
When one of the outputs of IC2 goes
low, the corresponding transistor is
turned on. The emitter voltage of the
transistor is determined by the position
of the associated potentiometer wiper.
Only a small saturation voltage is
dropped across the transistor. The
output voltage of the circuit (i.e. the
tuning voltage for the varicap diodes)
can thus be set by adjusting each poten-
tiometer to give the appropriate voltage
when the corresponding output of IC2
goes low.

Altogether 9 preset potentiometers are
used, which means 9 preset stations. If
no preset station is required (the counter
output is zero) tuning through the FM
band is accomplished by means of a
conventional (ten-turn) potentiometer.

Construction

Construction of the touch tuning circuit
requires a bit of handiwork with a fret-
saw. The printed circuit board, which is
obtainable via the EPS service, consists
of three sections, which before the
components are soldered in place, must
first be separated from one another. On
one section of the board are four copper

planes, which form the two pairs of
touch contacts. A second section of the
board is intended to accomodate the
7-segment display. A section is sawn out
of the main board at the point where
the display is to be mounted. The display
board and the touch contacts are
mounted perpendicularly to the edge of
the main board, as shown in the ac-
companying photograph. Of course the
individual is free to choose an alternative
design for the touch switches if desired.
The potentiometers used are 20-turn
presets from Piher. The existing tuning
potentiometer in the receiver can be
used for the 10-turn potentiometer.

In conclusion

Since transistors are used as voltage
switches, the circuit is slightly tem-
perature dependent. However most
tuners have fairly good automatic
frequency control (AFC), which should
ensure that this is not a problem.

The supply voltage is 5 V, whilst the
input tuning voltage should not exceed
30 V. When power is applied, the circuit
automatically selects channel 0, i.e. the
receiver can be tuned by hand. If one
wishes a preset station to be selected
immediately after switch-on, then the
inputs of IC1 can be programmed to
select another channel. For example, if
pin 15 of the 1C is connected to plus
supply, channel 1 will automatically be
selected.

Finally, it is perhaps worth remarking
that if the display is not required, then
R5 . . . . R1 1, IC3, and the display itself
can of course be omitted. M

1

battery saver elektor October 1979 - 10-09

battery

sawr

W. Jitschin

With many electronic games, such as
heads-or-tails, roulette, or any of the
versions of electronic dice, a consider-
able saving in battery life can be
obtained by ensuring that the circuit, or
at least the current-guzzling displays, are
switched off after each throw or turn.
Naturally enough.it would be somewhat
tiresome to have to do this by hand, so
the following circuit is intended to take
care of this chore automatically.

Basically the circuit is a simple timer.
Pushbutton switch SI is the start button
for the die, roulette wheel, etc. When
depressed, it causes capacitor Cl to
charge up rapidly via D1. Transistor T1
is turned on, so that, via T2, the relay is
pulled in, thereby providing the circuit
of the game with supply voltage.

When the switch is released, initially
nothing will happen. Cl discharges via
R1, R2 and the base-emitter of T1,
however it takes several secondes until it
has discharged sufficiently to turn of T1 .
When it does so, however, the relay
drops out, cutting out the power supply
to the die, etc.

With the component values shown in
the circuit diagram, a delay of roughly
3 seconds is provided in which to read
off the display. If that interval is too
short (or too long), it can be modified
as desired by choosing different values
for Cl and/or R1/R2. N

It is often very useful to be able
to match the values of capacitors
and resistors and the only quick,
effective way to do this is by
using an impedance bridge. The
following circuit is quite adequate
for this purpose and it is also
capable of measuring resistances
between 100 n and 1 M and
capacitances between 100 pF and
IpF.

Measuring resistance

Most readers will be familiar with the
basic Wheatstone bridge circuit shown
in figure 1, which represents the
simplest way of measuring an unknown
resistance. The bridge is formed by two
pairs of resistors (voltage dividers)
which are connected in parallel. As
every reader will know (we hope), when
two resistors are connected in series,
the voltage dropped across each resistor
is proportional to the value of that
resistor. Thus if the resistors are con-
nected as shown in figure 1 and we
ensure that the ratio of R a and Rb to
R x and R c is the same, the voltages at
points A and B must also be the same.
To put it another way, for the bridge
to be 'balanced' and the meter to read
zero voltage between points A and B,
Ra x Rc must be the same as R x x Rb-
If now we make Rb variable and provide
it with a calibrated scale, then by
adjusting Rb until the meter shows zero
deflection we can determine the value
of the unknown resistance, R x .

Measuring capacitance

Measuring capacitance is slightly more
complicated than measuring resistance,
however the basic principle involved
is the same. A capacitor also possesses
resistance to current flow, which is
called its reactance, and like resistance
is measured in SI. Unlike a resistor,
however, it is only meaningful to talk
of a capacitor's reactance to alternating
current, since capacitors do not pass
steady current at all. Furthermore, the
reactance of a given capacitor is fre-
quency-dependent, i.e. the greater the
frequency of the voltage across it, the
lower its reactance, and vice-versa.
For this reason, we have to ensure that
the supply voltage to our Wheatstone
bridge is alternating and of constant
frequency (it of course makes no
difference to a resistor whether the
voltage is AC or DC). Once that is the
case, the reactance of the capacitor is
determined solely by its capacitance.
Thus if we replace the unknown
resistance, R x , by the unknown capa-
citance, C x , and one of the fixed
resistors in the bridge by a fixed capaci-
tor, we can determine the value of C x
from the setting of the calibrated
variable resistor, Rb-
Since the capacitors are connected in
series with a resistor, strictly speaking
the meter is measuring impedance,
hence the name, impedance bridge.
When the variable resistor is adjusted
for zero deflection on the meter,
Wheatstone's formula once again applies,
i.e.: Z x • Rb = Ra ' z c- where Z is the
symbol for impedance (in SI).

Circuit

The complete circuit diagram of the
impedance bridge is shown in figure 2.
As already explained, a resistance
remains the same, regardless of whether
the voltage source is steady or alternat-
ing. Thus we can choose an alternating
supply voltage for the bridge. In order

to be able to measure fairly small
capacitance values, a reasonably high
frequency (significantly higher than the
mains frequency) is required, and to
this end a Wien bridge oscillator, formed
by the circuit round op-amp A1, is used.
When the gain of the op-amp is x 3, the
oscillator produces an alternating vol-
tage with a frequency of roughly
1 kHz. The gain of the op-amp can be
varied by means of PI, thus ensuring
that the oscillator can always be started.
Ideally PI should be adjusted such that
the circuit just oscillates and no more. If
desired the oscillator output can be
examined on an oscilloscope and PI
adjusted for as sinusoidal a waveform as
possible, although this step is not
strictly necessary. A2 functions as a
buffer stage, delivering sufficient power
to drive the bridge.

The Wheatstone bridge is clearly recog-
nisable in the circuit diagram. If we
compare it with the circuit of figure 1,
it is apparent that resistor R a is replaced
by four different value resistors, each
of which can be selected by the range
switch, SI. Potentiometer P2 assumes
the function of variable resistor Rb in
figure 1. When the wiper of this poten-
tiometer is turned hard up against the
end stop such that no greater resistance
can be measured, one simply has to
select a larger value for R a .

The fixed value capacitor in the bridge
is formed by C8. This capacitor is
connected in series with another poten-
tiometer, P3. During the measurement
procedure, when P2 is being adjusted
for zero deflection on the meter, P3 is
set for zero resistance. Once the
measurement has been completed, the
quality of the unknown capacitor (C x )
can be determined with the aid of P3.
How this is done is discussed in the
section on using the impedance bridge.
The voltage between points A and B in
the circuit is measured by the differen-
tial amplifier A3. C6/R17 and C7/R15
ensure that only the 1 kHz alternating

impedance bridge

jber 1979- 10-11

voltage appears across the inputs of A3.
The output of A3 is fed via C9 to A4,
which in conjunction with D5 provides
a half-wave rectified voltage, suitable for
driving the meter (which in fact displays
the average value of the rectified signal).

Construction

It should not be difficult to construct
the circuit using Vero-board or similar.
If the circuit is mounted in the same
box as the power supply, then care
should be taken to place diodes D3 and
D4. which stabilise the amplitude of
the oscillator signal, at a reasonable
distance from components which are
liable to run warm. This point should
not prove a serious problem, however,
since the circuit only consumes some
20 mA.

Any readily available meter will prove
suitable since it is not required to pro-
vide u reading which is accurate in
absolute terms, rather it is a question
of determining which setting of P2 gives
the smallest deflection. The meter is
being used to give a 'dip-reading'.

Using the impedance bridge

The general operation of the impedance
bridge should be fairly clear from the
foregoing decription of the circuit. First
of all however, the circuit must be
calibrated. This is done by adjusting
PI until the oscillator starts. The
oscillator can be checked by setting P4
to roughly the mid-position and connec-
ting a wire link between the test ter-
minals (Z x ). When the oscillator starts
the bridge will cease to be in a state of

Figure 1. The basic Wheatstone bridge. In
order to measure capacitance, R c is replaced
by a capacitor and the unknown capacitance

equilibrium (which is another way of
saying that a potential difference
exists between points A and B in the
circuit). It may occur that the oscillator
will stop after a short period; this
simply means that PI was not set to
the optimal position and should be
readjusted.

With P2 set for minimum resistance
and SI in position 4, P4 is then adjusted
until maximum deflection is obtained
on the meter. Diodes D6 and D7 are
included to limit the current through
the meter to an acceptable value; how-
ever if full-scale deflection cannot be
obtained on the meter, an additional
diode can be connected in series with
D6/D7. Alternatively, should it prove
impossible to limit the current through
the meter sufficiently by means of P4,
then D6 can be replaced by a wire link.
Once the bridge has been set up, the
next question is, how do we provide
P2 with an accurately calibrated scale?

The simplest solution would be to
print a suitable scale in this article.
Unfortunately this is not really feasible,
since P2 must be a linear potentiometer,
and different types have a different
effective electrical rotation. Furthermore
the first and last sections of the poten-
tiometer tracks are not completely
linear, and the extent of the non-
linearity varies from potentiometer to
potentiometer. For these reasons it is
better to experimentally determine a
suitable scale oneself.

First of all, S2 is set to position R
(measurement of resistance). SI is then
set to position 1 and a series of close
tolerance resistors with values ranging
from 100 f2 to 1 kfi are mounted
between the test terminals. For each
resistor, P2 is adjusted until the bridge
is balanced (i.e. minimum deflection on
the meter). At the corresponding
position of P2 a mark is drawn on the
scale, accompanied by the first two
figures of the resistor value separated
by a full-stop. For example, if R x equals
470 n, one writes 4.7. For the different
positions of the range switch, SI, the
following multipliers give the correct
magnitude of the values;
position 1 x 100 S2

position 2 x 1 kS2

position 3 x 1 0 kfi

position 4 xIOOkfi

The calibration procedure need only be
carried out for one range; thereafter the
scale will also be correct for the other
ranges.

To calibrate the scale for capacitors, S2
is set to position 2 and P3 adjusted for
zero resistance. Close tolerance capaci-
tors between In and 10 n are then
connected between the test terminals in
turn, and P2 adjusted for minimum
deflection on the meter. Once again the
scale is marked at the corresponding
positions of P2. For the value 1 n,
switch SI should be set to position 4;
for larger values up to and including
10 n, position 3 is required. The scale
will 'run' in the opposite direction to
that for resistors, i.e. the scale will
decrease from 10 down to 1 from left
to right, whereas with resistors it
increases from 1 to 10.

The multipliers for each position of the
range switch are;
position 1 xIOOn
position 2 x 10 n

position 3 x In

position 4 x 0.1 n

Capacitance is not the only quantity
which can be measured, however. Once
the value of the capacitor has been estab-
lished, it is possible to obtain an idea as
to the quality of the capacitor. This is
done by adjusting P3 (which during the
measurement of capacitance is of course
set for zero resistance). If by so doing
the deflection on the meter can be made

even smaller, then the further the

deflection can be reduced, the poorer
the quality of the capacitor. H

The maximum thrust available from the
engine is limited, so that leaving reverse
thrust too late will result in a crash
landing. The maximum permissible
descent rate at the moment of touch-
down is 01; if this is achieved, the display
will alternate the final results with the
message 'landed'.

It is not an easy matter to control a
LEM, and the result is that landings may
well be rougher than intended. In that
case, the message on the display will

Program 1: Luna (R. Bayer)

This program simulates the landing of the
LEM (Lunar Module) on the moon. The
display gives information on the height
above the surface, the rate of descent
and the amount of fuel left in the tank.

new programs 6r the
SOMPi

Good news for SC/MP fans: two
new records have been added to
the ESS range. One contains the
complete NIBL-E program; the
other includes some games, a
'running script' program, 'tracer',
'disassembler' and 'biorhythm'.
Some further details on the latter
programs are given here.

SC/MP

>ber 1979- 10-13

after which a new attempt can be
initiated by operating any one of the

Program 2: Battleships
(F. Schuldt)

'Battleships' is normally a game for two
players. In this program, the computer
takes the role of one of the players.

The game is played on a 64-square
'ocean', as shown in figure 1. In all, six
ships take part in the engagement: two
of three squares each, two of two
squares and two of one square each.
The ships may only be entered in
horizontal or vertical direction, and
they are not allowed to touch.

When the program is started (at address
0C40), the word 'Ships' appears in the
display. As soon as any key is operated,
the computer draws in its own set of
ships in its memory. It then invites its
opponent to take the initiative: 'Fire'.
The coordinates of the first square to
come under fire can now be entered:
first the line number and then

the column number (or
letter). The computer can

reply in three ways:

1 . If the shot landed on one of its ships,
it will display 'Hit'. After a brief
delay, it will invite a further try:

2. If a ship is sunk, that is to say if all
the corresponding squares have
already been hit, this is indicated by
the word 'Lost'; after a brief delay,
this is again followed by 'Fire'.

3. A miss is indicated by the word 'Fail'.
The computer will follow this by a
shot of its own, indicated as

line and column numbers, respect-
ively. The player can now answer in
three ways:

1. A hit is recognised by operating the
'Down' key. The computer will reply
immediately with 'shot XY'.

2. Operating the 'Up' key indicates that
a ship is sunk. This, too, will be
acknowledged with another shot.

3. A miss is indicated by operating any
other key. The computer will tell
you to get on with it, in that case:
'Fire'.

As soon as all ships of one of the sides
are sunk, the word 'end' will appear on
the display. After a brief delay, the
program will reset and the word 'Ships'
will appear.

Program 3: Keyplay (F. de Bruijn)

This game is known under a variety of
names, 'NIM' being one of the most
popular. It can be played with
matchsticks, coins, or . . . numbers. The
rules are simple: each player in turn
subtracts a number from the original;
the one to get 0 as result, wins.

When the program is started, at
address 0C00, the program will
ask for a four-digit decimal
number ('GE' = Give
Entry); this is the
number from which
the players will sub-
tract in turn. Next,
the program will want
to know the Limit
('LI'): this is the maxi-
number that may be
subtracted at one time.

The human player is allowed to start.
This is indicated by 'U' in the first
display digit. A four-digit number can
now be entered. If it is either 0 or more
than the limit, the computer will refuse
to accept it: it will display the word
'reject', followed by a repeated request
'U'. If a valid number is entered, the
computer will perform the subtraction
and display the result: 'SAxxxx', where
xxxx is the remainder. It then calculates
the number that it wants to subtract, and
displays this with the prefix T; finally,
it performs this subtraction and again
displays the result as 'SAxxxx'. It is
now the human player's turn, and the
game continues until the remainder
becomes equal to 0. Depending on who
reached this point, the display will
indicate either 'I LOSE' or 'U LOSE'.
The program can be re-started by
operating the Halt/Reset key.

Program 4: Runtext
(R. Brinkmann)

This program can display up to
16 different lines of text, each consisting
of up to 256 characters, as a 'running
script' on the 7-segment displays.

The start address for the program is
0C00. Initially, 'runtext' appears on the
display. One of the keys O . . . F is now
used to select the desired one out of the

sixteen texts. Even when a text is
running, it is possible to switch over
immediately to any other text, by
operating the corresponding key.

The program consists of three parts:

1. A selection routine, that uses the
Elbug LDKB1 routine to determine
which of the texts is required. It
places the start address of the text in
pointer 2, and the length of the text
in a memory location reserved for
this purpose (as can be seen from the
listing).

2. A display routine, that transfers the
text (pointer 2) to the display
(pointer 1). This routine also checks
to see if a different text is required
(key entry); as long as this is not the
case, the text originally selected is
repeated. The speed at which the
text runs across the displays can be
varied within wide limits by
modifying the contents of addresses
0D48 and 0D57.

3. The text section, containing the texts
in 7-segment format. Each character
is stored in one memory location
(8 bits). The texts all start with seven
spaces (00), so that a new text always
starts on a blank display.

When this program is loaded from the
ESS record, not only sections 1 and 2
(as given in the listing) are entered, but
also several texts. For this reason, the
memory is used up to and including
location 0E33.

Program 5: Biorhythm (H. Prante)

A few years ago (in October 1977),
Elektor published a program for
calculating biorhythms on an HP 65
calculator. Now, a similar program is
available for the SC/MP system.

As usual, the program is started at
address 0C00. Initially, the word 'today'
appears; the date for which the
biorhythm data are required should now
be entered. The date should be entered
in the following order: day, month, year
(without '19'). This entry is immediately
followed by the display 'birthday'; this
date is entered in the same way.

The computer performs the necessary
calculations and displays the results:
three numbers, corresponding to the
physical, emotional and intellectual
rhythms. A new calculation can be
performed after operating the Halt/Reset
key.

The biorhythm theory was explained in
the earlier article referred to above, but
a brief reminder may be in order. The
physical rhythm has a cycle of 23 days;
the emotional cycle is 28 days the
intellectual cycle lasts for 33 days.
The 'zero crossings' are critical days,
and these include the half-way marks:
11th . 12th d a y for the physical cycle,
14th f or the emotional and 16*h - 17th
for the intellectual. The first half of
each cycle is taken to have a positive
influence; the second half is negative.

Program 6: Tracer (J. Fischer)

This program is a powerful extension of

10-14 -elektor

iSC/MP

the monitor software already available
in the SC/MP system. The CPU routine
in Elbug can only handle one breakpoint,
and it must be reset every time it is used.
'Tracer' constitutes a much more
powerful aid when de-bugging.programs.

It can be used to execute any other
program in 'single-step' mode. The
program uncjgr test is thus executed
instruction-by-instruction; between in-
structions, the contents of all registers
can be examined (PI, P2, P3, Accu,
Extension Register, Status Registerl.
The display gives information on the
position of the program counter and the
following instruction, before actually
executing it. If errors are noticed at this
point, it is possible to correct them
before continuing the single-step scan.
The single-step mode can be executed in
three ways:

1 . H igh Speed : The program to be tested
is executed at a rate of approximately
one instruction per millisecond, until
a specified address is reached. At this
point, Tracer' automatically switches
over to the 'Low Speed' mode. The
display is blanked during the High
Speed mode.

2. Low Speed: The address and the
corresponding instruction are
displayed for approximately one
second. The instruction is then
executed, and the display is blanked
for one second. This sequence is
repeated until the point is reached
where the change-over to 'Manual
Step' is required. This will occur
automatically at a specified address;
however, it is possible to effect an
earlier switch to the Manual Step
mode by operating any of the keys
during the Low Speed mode.

3. Manual Step: The next address and
corresponding instruction remain
visible in the display until one of the
keys (any key except the CPU-routine
key) is operated. The address and
instruction remain visible for about
one second after the key is operated;
the instruction is then executed and,
after a brief delay, the next address
and instruction appear on the display.

In all three modes, the keyboard and
display remain available for in- or
output of data.

When 'Tracer' is started (at address
0C00), the message 'SS . . . ' appears on
the display. Three addresses should now
be entered, in the following order:

1. The 'start address' of the program to
be tested;

2. The address at which the change-over
to Low Speed is required;

3. The address at which the Manual Step
mode must be initiated.

After the third address has been entered,
operating any one of the keys starts the
T racer routine. The first section of the
program will be run through in the
High Speed mode, unless the second
address is equal to the start address (1).
In the High Speed mode, the keyboard
and display seem to function normally.
In the Low Speed and Manual Step

modes, this becomes rather more
complicated.

In the Low Speed mode, the keyboard
must be operated in the time that the
instruction (and address) are visible on
the display. In the Manual Step mode,
the keyboard becomes operational when
the command is given to execute the
instruction; it remains available for
approximately one second, until the
display is blanked and the instruction is
executed.

The time during which the display is
blanked by Tracer (for one second after
the instruction is executed) is used to
show the display that the program under
test would provide after that instruction
is executed. However, it should be
noted that the display is again used by
Tracer before coming to the next
instruction, so that all previous display
data is lost and the 'program display'
can therefore consist only of single
digits.

Both the Low Speed and Manual Step
modes can be interrupted to check the
contents of all registers in the CPU. In
the Low Speed mode, it is first necessary
to switch over to Manual Step, by
operating one of the keys. The display
will then show the address and instruc-
tion that is about to be carried out. If
the CPU-routine key is now operated,
the display 'CP' will apppear. The
keyboard can be used at this point to
select one of the registers; the codes are
the same as those used in the Elbug CPU
routine: '1 ' ■ Pointer 1 , '2' = Pointer 2,
'3' = Pointer 3, '5' = Status Register,
'A' = Accu, 'E' = Extension Register.
There are various ways to leave the CPU
routine:

S(ubtract)-key: Tracer can be re-started.
R(un)-key: Return to High speed. Tracer
now expects the entry of two addresses:
one to indicate the point at which it
must switch over to Low Speed and one
which specifies the first address of the
Manual Step mode.

Having re-started Tracer in either of
these ways, all the facilities described
above are available again.

Program 7: Disassembler
(F. de Bruijn)

A disassembler is a program that can be
used to obtain listings (without
comments, obviously) of programs in
machine language. It is the opposite of
an assembler program.

The listing can be obtained on a printer
or an (Elek-)terminal. In the latter case,
of course, no 'hard copy' of the print-out
will be obtained.

The serial output signal for the printer
or video display is available at flag 0.
The transmission rate is 300 baud. This
speed can be modified, if required,
according to the following table:

110

300

600

1200

address baud baud baud baud

1 59 B 97 64 25 86

159 D 17 06 03 01

15A7 89 F0 50 81

15A9 08 02 01 00

The ( 1 Zz K) program offers the following
facilities:

a) enter the 'begin address' of the
program that is to be 'disassembled';

b) specify the begin and end addresses
of a table;

c) mark a byte used by the program, by
entering '20' at that point;

d) enter the number of consecutive lines
to be disassembled.

The program is started at address 1000.
When 'D1 . . . ' appears on the display,
the begin address can be entered. This
can be followed, if necessary, by
specifying one table; in that case, the
Block T ransfer key must first be operated
- if any other key is operated, the
program assumes that there is no table.
After the Block Transfer key, the begin
address of the table is entered, followed
by the end address plus one.

The next step is to specify the number
of lines to be printed: note that this
number must be entered in hexadecimal.

A suitable value, when using the
Elekterminal for the display, is 0010.
The maximum value is 00FF; this
already makes for quite a lengthy
print-out.

The program will start the print-out
immediately after receiving this final
entry; it will stop when the specified
number of lines have been disassembled.

A further group of lines will be disas-
sembled if the Halt/Reset key is j
operated.

If the program finds an instruction that
it doesn't recognise, it will print '?'.
Jump instructions by means of the
program counter are shown with the
address to which the jump would be
executed. The same applies to other
instructions that use the program
counter. H

The circuit shown here provides a
two-digit display calibrated in hundreds-
of-revs per minute, i.e. 6000 r.p.m. will
produce a readout of 60. There are two
principal reasons for restricting the
display to two digits. The first is quite
simply that accuracy greater than this
is not necessary, and secondly, a much
longer gate time would be required
otherwise, with the result that the
| counter would not be able to follow
sudden changes in the engine speed.
The circuit is a modernised version of
a rev counter published in an earlier
issue of Elektor (see Elektor 1,
December 1974). The input signal is
derived from the contact breaker; the
amplitude of the resulting pulse train
being limited by zener diode D1 and
then 'shaped' by T1 and the monostable

N1/N2.

The pulses are counted by IC3 (dual
decade counter), whose outputs are
connected to two BCD-to-7-segment
latches/decoder-drivers. The reset pulse
for the counters (i.e. the timebase
signal) and the latch enable pulse are
provided by a 555 timer (IC5).

The circuit has three adjustment points.
Preset potentiometer PI sets the width
of the reset pulse. In the majority of
cases it will be sufficient to set this
potentiometer to the mid-position.
However it may happen that the re-
liability of the circuit can be improved
by choosing an alternative position. The
latch period, and hence the rate at
which successive measurements are
displayed, is set by means of P2. Finally,
P3 is used to calibrate the counter. This

can be done using either a tone generator
with a calibrated tuning scale, or else by
using a mains frequency signal. In the
former case the frequency of the input
signal will depend upon the type of
engine with which the rev counter is to
be used. The counter is calibrated for
a nominal r.p.m. of 6000, and cepen-
ding upon the number of contact
breaker pulses produced for each
revolution of the engine in question, a
signal of suitable frequency (see table 1 )
is fed to the input of the circuit and P3
adjusted until a readout of 60 is ob-
tained. If a tone generator is not avail-
able, a low voltage signal of mains
frequency (e.g. from a doorbell trans-
former) can be used. P3 is then adjusted
until the appropriate readout is obtained
(see table 1, 'revs at 50 Hz'). M

10-16 — elektor i

>ber 1979

digifarad

dlgiJhrad

digital capacitance meter

Given the fact that many types of capacitor - especially electrolytics -
have a wide tolerance (20% is fairly common), it is often desirable to be
able to measure capacitances both quickly and with a reasonable degree
of accuracy (e.g. when constructing precision timer circuits, matching
the time constants of several RC networks, etc). Of course a capacitance
meter also enables one to measure the value of those piles of unmarked
capacitors which end up at the bottom of one's junk box, or to test
'suspect' capacitors for potential faults — in short it represents a useful
addition to the test gear of any constructor.

The circuit described here offers the advantages of a digital display, has
5 decade ranges, measuring from 1 nF to 9.999 pF, and is accurate to
about 2%.

The range of digital test equipment is
growing ever more extensive. Voltage,
current, frequency, resistance, tempera-
ture — all these quantities are now
commonly measured, and displayed,
digitally. This not only applies to
'professional' applications, even the
'amateur constructor' has gone digital
(see, for example, the 'universal digital
meter', Elektor 45). Now it is time to
add a digital capacitance meter to the
range - the 'digifarad'.

The block diagram of the 'digifarad' is
shown in figure 1 . C x represents the
unknown capacitance to be measured.
Depressing the 'start' button momen-
tarily closes the electronic switch, ES,
so that C x is charged to a given voltage
(Uc). When ES reopens, C x is discharged
by a constant current source (I), with
the result that the voltage on C x falls in
a linear fashion. All other things being
equal, this discharge rate is determined
by the value of C x . The voltage on the
capacitor is monitored by a window
comparator, formed by two op-amps
and a set/reset flip-flop. For the period
that U c remains within the upper and
lower reference voltages (U 1 and U2)
of the 'window', the output of the
comparator is low. This enables a three
digit counter, which counts the number
of pulses from a clock generator. Thus
the greater the capacitance of C x , the
longer U c takes to fall below the thres-
hold voltage of the window comparator,
and the more pulses counted by the
counter. Finally, by varying the size of
the constant current, I, we can arrange
for capacitors of widely differing value
to be measured in the same way.

The complete circuit diagram of the
digifarad is shown in figure 2, and a
pulse diagram is given in figure 3. The
latter is not only useful in the (unlikely)
event that trouble-shooting proves
neccessary; it is also a great help in the
following explanation of the circuit.
The various wave-shapes (A ... I) were
measured at the corresponding points in
the circuit.

It is not too difficult to relate the block
diagram, given in figure 1 , to the actual
circuit shown in figure 2. The constant-
current source, I, is formed by op-amp
A1 and transistor T1. The size of the
current is determined by the position of
the range switch, SI (see table 1). The
op-amp varies the current through T1
and the selected range resistor so as to
ensure that the voltage at the inverting
input is always the same as the fixed
reference voltage at the non-inverting

The electronic switch, ES, consists of
transistor T2, which is turned on via the
start button, S2, and flip-flop N3/N4.
The voltage on C x is buffered by op-
amp A2, and fed to the window com-
parator formed by A3 and A4. N1, N2,
Cl, C2, R18 and R19 form a set/reset
flip-flop which is triggered by changes
in the output state of the window
comparator. When C x is fully charged,
the outputs of A3 and A4 are both high.

J. Guther

However when the voltage on C x reaches
the upper threshold of the 'window'
(i.e. the voltage on the non-inverting
input of A2 falls below that on the
inverting input) the output of A2 goes
low, with the result that the output of
N2 also goes low, enabling the counter.
As the unknown capacitance continues
to discharge, the voltage on C x will
reach the lower threshold of the
window, whereupon the output of A4
will go low, taking the output of N2
high and stopping the count.

In addition to turning on T2, the second
flip-flop formed by N3 and N4 provides
the reset and display enable signals for

the counter. The display is inhibited
during the count cycle, thus ensuring a
stable readout. R20, C3 and the two
diodes (D1 and D2) ensure that the two
flip-flops assume the correct state upon
switch-on.

The clock-signal for the counter is
provided by a 555 timer (IC3) con-
nected as an astable multivibrator. The
counter itself (IC6) is a single 1C,
type 74C928. It performs the 7-segment
decoding, and drives the three LED
displays via transistors T4 . . . T6. The
displays are of the common cathode
type (e.g. HP 5082-7760, DL 704, etc.).
In all, four 'supply' voltages are required

for the circuit: the reference voltage
(Uref) and the 16 V, 12 V and 5 V
supplies. The obvious solution is to use
IC's: one three-pin regulator (IC5) takes
care of the 1 2 V supply, and a 'basic' 723
circuit (IC4, T3) provides all the other
voltages, including the reference voltage.

Construction

Once again, printed circuit boards
(available through the EPS service)
reduce constructional problems to a
minimum. Every single component,
barring the mains transformer, is
mounted on these boards - from supply
circuit to displays. To increase the sense
of achievement, three boards are
required instead of one. A display board
(figure 4c) is mounted behind the front
panel, and the other two boards
(figures 4a and 4b) are bolted together
with spacers in a sandwich construction
and mounted behind the display board.
The display board contains the displays.
Obviously. It also provides space for
IC6, resistors R31 . . . R37, switches SI,
S2 and S3; furthermore the on/off
indicator D8 and 'banana plug' con-
nection sockets for the unknown
capacitor C x . The upper board in the
sandwich (figure 4b) is intended for the
supply circuit (all except IC5) and the
clock generator, IC3. Finally, the lower
'sandwich' board (figure 4a) must
provide space for the remainder of the
circuit. Rest assured: it does. The inter-
connections between the various boards
are clearly marked with diagonal arrows.

digifarad

1979- 10-19

Final notes

The capacitance meter is as easy to use
as a multimeter: switch it on with S3,
select the desired range with SI,
connect the unknown capacitor, press
the start button (S2) and watch the
result appear on the display. The
measuring ranges are listed in the Table;
the current I listed in this table is the
constant-current used to discharge C x -
If the capacitance value is completely
unknown, it is a good idea to start in
the highest range (position 5), and then
switch back step-by-step until a useful
reading is obtained.

The circuit contains only one calibration
point, namely preset potentiometer PI .
Calibration can be carried out with the
aid of a close tolerance capacitor of a
known value. Silvered mica capacitors,
for example, typically have a tolerance
of 1%.

One final remark. If a 12V/1 A trans-
former is felt to be rather heavy, or if a
smaller transformer happens to be
available, resistors R31 . . . R37 can be
modified as required. Provided a slightly
less brilliant display is considered
adequate, the value of these resistors
can be increased to 22 Si; a 12 V/'/a A
transformer is then good enough. m

short-interval
light switch

KC.lt.

fel*

variable

liira

box

Even in today's well-equipped modern
houses there are various 'corners' where
additional lighting is required. For dark
cupboards, meter boxes etc. temporary
lighting is usually sufficient, so that
making a connection to the mains is
hardly worthwhile; a simpler and cheaper
solution is to use a battery-powered
circuit which will light a lamp for a
short period of time. As is apparent
from the accompanying circuit diagram,
such a circuit is by no means compli-
cated. Using only one CMOS 1C, three
resistors and one capacitor, the circuit
will switch on a lamp for a presettable
interval.

The operation of the circuit is perfectly
straightforward: when the button is
pushed Cl charges up to the supply
voltage. The outputs of the four parallel-
connected inverters (N3 . . . N6) are
then low, so that the lamp will be lit.
When the button is released, Cl dis-
charges via R1 until the input of N1
reaches half supply. The Schmitt trigger
formed by N1 and N2 then changes
state, with the result that the lamp is
extinguished. The positive feedback
resistor R3 ensures that the Schmitt
trigger changes state very quickly.

With the resistor values shown in the
circuit diagram, the lamp will remain lit
for roughly 2.5 seconds per pF of Cl.

1

Thus a 10 p capacitor would give an
interval of roughly 25 seconds.

The circuit can be powered by four
1.5 V cells connected in series. If a
larger lamp is required, three 4.5 V cells
connected in series can be employed.
Alternatively, for really 'heavy-duty'
applications, the four parallel-connected
inverters can be replaced by a transistor,
as shown in figure 2. The supply voltage
should be matched to the voltage rating
of the lamp and may lie between 4.5
and 1 5 V. The current through the lamp
should not exceed 500 mA in that case.

2

The design for the 'variable fuzz
box' was first published in the
December 1978 issue of Elektor.
Such is the popularity of this
circuit, that we have decided to
produce a printed circuit board
for it.

The variable fuzz box is a special effects
unit for guitarists, which by allowing the
amplifier signal to be clipped in a variety
of different ways (symmetrically,
asymmetrically, soft, hard, etc.) offers a
greater degree of control over the
resultant sound. The circuit of the fuzz
box was described in detail in the
original article, hence will not be
repeated here. However one correction
to the original description has to be
added: symmetrical clipping of the
output signal produces only uneven (not
even, as was stated) harmonics, whilst
asymmetrical clipping generates both
even and uneven harmonics in theoutput
signal.

The alternative circuit diagrams of the
fuzz box for symmetrical (figure 3 of
original article) and asymmetrical
(figure 4 of original article) power
supplies are here combined into one (see
figure 1). The circuit diagram contains a
number of lettered connection points
(a . . . h, j, k, m . . . w) which are
marked on the printed circuit board
shown in figure 2. The circuit diagram
and accompanying parts list provides
the relevant details on which connections
should be made for either symmetrical
or asymmetrical power supply require-
ments.

The current consumption of the circuit
is less than 20 mA. A 741 can be used
for IC1, however an LF 356 is a better
choice. ^

"© (13,5 V)

Literature:

Variable Fuzz Box, Elektor 44,
December 1978

frequency oscillator. The coil is held
near the parallel-resonant circuit (the
equipment containing the tuned circuit
should be switched off for the purposes
of the measurement). Series-resonant
circuits can also be measured by shorting
their inputs, so that a parallel-resonant
circuit is obtained. The coil of the grid-

the modern equivalent of the grip-dip meter dip meter is eiectromagneticaiiy coupled

to the resonant circuit. As the frequency

provides a quick way of checking the resonant of the oscillator approaches the resonant
x , , « . . . frequency of the LC circuit, so the

trequency OT LU tunea Circuits oscillator becomes increasingly damped.

This is registered by the meter, so that
when the needle deflection is at a
maximum, and the oscillator frequency
coincides with the resonant frequency
of the tuned circuit, the latter can simply
be read off a calibrated scale.

The circuit of the gate dipper described
here is based upon a device known as a
lambda diode. As many readers may
well never have heard of such a 'beast',
it is worth devoting a little time to an
explanation of this slightly unusual
circuit element.

Lambda diode

If the term lambda diode is unfamiliar,
the majority of our readers will have

Tuning resonant circuits in high frequency equipment normally requires heard of tunnel diodes. These are diodes
fairly expensive test gear which not every hobbyist can afford. However which exhibit a negative resistance over
there is a reasonably aheap alternati.e available, namely a „a„ dipper,

which allows the resonant frequency of the tuned circuit to be resistance may seem confusing, but in

ascertained simply and quickly. fact it is quite straightforward. As the

a certain portion of their voltage-current
characteristic. The concept of a negative
resistance may seem confusing, but in
fact it is quite straightforward. As the

A grid-dip meter (the name harks back
to the good old days of valves, which
actually had a wire grid between their
anode and cathode and the meter
measured the grid current of the valve)
is a useful little device which enables the
resonant frequency of tuned circuits to
be determined without having to make
any electrical connection to the circuit
in question. The grid-dip meter contains
a coil, which forms part of a variable-

1979- 10-23

gate-dipper

voltage dropped across a 'normal' or
'positive' resistance increases, there is a
directly proportional increase in the
current flowing through that resistance.
A negative resistance, however, ensures
an inversely proportional relationship
between voltage and current, i.e. the
current increases as the voltage decreases.
The typical voltage-current character-
istic of a tunnel diode is shown in
figure 1. Over the range — r the diode
exhibits a negative resistance. Assume
for example that the diode is forward
| biased to point P; if the voltage is now
I increased by a U, the current will fall by
A I. The resistance of the diode is thus:

_ - aU

The negative resistance is smallest (i.e.
there is the greatest drop in current for
a change in voltage) at the point where
the curve is steepest.

The question now is: how can we utilise

this negative resistance characteristic?
Strictly speaking, a negative resistance
can be regarded as an active circuit
element (being just the opposite of
normal resistance) , and it is as such that
tunnel diodes are normally used. Figure 2
shows a simple example of a tunnel
diode oscillator. The average current
through the tunnel diode automatically
settles at a value where the effect of the
negative resistance is at a maximum (i.e.
at the steepest point of the negative
resistance portion of the voltage-current
characteristic. Several advantages of
tunnel diode oscillators are low power
consumption, good frequency stability.

Fqure 2. Only a few components are required
to build a tunnel diode oscillator. The
-Cerent simplicity is the great advantage of
this type of oscillator circuit.

Figure 1 . A peculiar feature of a tunnel diode
is that it exhibits a negative resistance over a
portion of its voltage v. current characteristic.
When biased to this point, the diode
effectively becomes an active circuit element.

and last but not least, their inherent
simplicity.

More recently however, the advent of
FETs has seen the design of oscillator
circuits which offer even better per-
formance, with the result that tunnel
diodes are rarely used for this application
nowadays. Despite this fact, the sim-
plicity of tunnel diode oscillators has
led to the search for ways of improving
their performance whilst continuing to
use the same basic principle. This
attempt has resulted in the lambda
diode, which consists of an N and a
P-channel FET connected as shown in
figure 3. Between the anode and cathode
of the device there is the same negative
resistance characteristic as in tunnel
diodes. Thus the lambda diode can also
be used as the active element in an
oscillator circuit. This is in fact the type
of oscillator employed in the grid-dip
meter circuit.

Gate dipper

The complete circuit diagram of the gate
dipper is shown in figure 4. By using
a voltage regulator (IC1) the circuit can
be powered by a 9 V battery, thereby
making the meter portable and easy to
use. The lambda diode is formed by
FET T1 and transistor T2. Since
P-channel FETs have a relatively shallow
transfer curve, a bipolar transistor is
used in its place. Although the configur-
ation shown in the circuit diagram may
appear slightly different from that of
figure 3, as far as AC current is concerned
its basic operation is the same.

The oscillator circuit is formed by the
fixed inductor, L x , and the variable
capacitor, C3, by means of which the
oscillator frequency is adjusted. The
lambda diode is biased to the negative
resistance region by means of PI . Diodes
D1 and D2 clamp the adjustment range
to suitable values.

The output of the oscillator is rectified
by D3. A negative DC voltage (L x can
be considered a short circuit for AC
currents) appears across this diode,
which serves as the control voltage for
the lambda diode (via the gate of T1).

This voltage is smoothed by C4/R2 and
fed to T3, which is connected as a
source follower. Potentiometer P2 is
adjusted such that a zero reading is
obtained on the meter. If the coil, L x , is
brought near the passive tuned circuit
which is to be measured, the negative
voltage across D3 will fall as the oscil-
lator is increasingly damped. This causes
the source voltage of T3 to rise, thus
causing a deflection on the meter. When
the deflection is at a maximum, the
value of C3 is an index of the resonant
frequency of the tuned circuit under
test. Due to the effect of the lambda
diode, the behaviour of the meter
needle is the opposite to that of other
types of grid-dip meter, where the
oscillator frequency is adjusted for
minimum deflection (hence the term
dip meter).

The grid-dip meter can also be used to
check the operation of an oscillator.
Once again the coil of the meter is held
near the oscillator circuit, and C3 is
adjusted until audible beat frequencies
are obtained. These low frequency beat
notes are not sufficiently smoothed to
prevent them appearing at the source of
T3, with the result that they are fed
through to the output stage round T4
and T5, where they can be heard via a
pair of headphones. P3 then functions
as a volume control.

When checking the operation of tuned
circuits in radio receivers, if the grid-dip
meter is tuned for zero beat, then it is
possible to modulate the r.f. signal (in
accordance with the direct-conversion
principle). The lambda diode oscillator
then functions as a self-oscillating mixer
stage. This fact allows the meter to be
calibrated with a precise frequency scale
(the procedure is described in detail in
the section on calibration).

Construction

The track pattern and component
overlay of the printed circuit board for
the grid-dip meter is shown in figure 5.
The coil, L x , is not mounted on the
board, but rather is connected to the

3

Figure 3. If a P-channel and N-channel FET
are connected as shown, the result is a so-
called lambda diode. Like a tunnel diode, this
has a negative resistance over a portion of its
voltage v. current characteristic.

>ber 1979- 10-25

circuit via a plastic DIN loudspeaker
plug. This provides the option of using D
several different coils to obtain different
measurement ranges. The accompanying
table lists the winding details for each
coil and the corresponding frequency

The coils are wound on the plugs as far
as possible from the metal terminals (see
figure 6). If the qbil is near any metal,
eddy currents wii) cause energy losses
which increase ;with frequency. The
result is that after adjusting C3, the zero
point of the meter will tend to drift.
Admittedly, that is not such a disaster,
since the meter is not being read, but
merely used as an indicator to obtain
the correct position for C3. However if
the energy losses are severe enough, the
meter will deflect to the point where no
'dip' is obtained.

The ends of the coil are fed through the
inside of the plug and foldered to the
terminal pins. The coil consisting of
only one turn is mounted directly on
the pins, and the plastic cap is omitted. J®'
The socket for the plug is mounted on sho
the case of the grid-dip meter and the
connected to the printed circuit board m i r

via short lengths of fairly thick wire. In

this way it is a simple matter to inter-
change coils should a different measure- 7
ment range be required. The variable
capacitor, C3, is also mounted off board
and connected to the circuit via short,
thick wiring. If the wires are too long,
measurements above roughly 80 MHz
are no longer possible.

Calibration and use
Before providing the gate dipper with
a calibrated scale one must first know
how to use it properly. PI and P2 are
set so that as positive a 'dip' as possible
is obtained. The meter is here used
essentially as an indicator, rather than a
measuring instrument. Thus at this stage
P2 is adjusted not so much in order to
set the zero point of the meter but
rather to ensure that the needle remains Fj)
within the scale range of the meter. Af
Thus P2 can be adjusted to compensate
for energy losses induced by metals in an
the vicinity of the coil etc. 10

As already mentioned, PI in fact deter- su

mines the biasing of the lambda diode,
and hence the sensitivity of the circuit.
The optimum setting of PI can be deter-
mined as follows:

The wiper of PI is turned fully towards
the cathode of D1 . The oscillator is then
inoperative and the meter defelction at
a maximum. Ensure that the needle is
not hard up against the end stop,
however (if necessary adjust P2 accord-
ingly). Now turn the wiper of PI in the
opposite direction. At a certain point
the needle deflection will decrease (the
oscillator is now running). Continue to
turn PI until the deflection is at a mini-
mum (here again it may be necessary to
adjust P2). The meter range is then set
between these two extremes by ad-
justing P2 (note, P2 will need to be read-
justed when the coil, L x , is changed).

To gain proficiency in using the meter it
is advisable to practice with a tuned
circuit whose resonant frequency is
already known. At the same time one
can experiment with different settings
of PI to obtain optimum sensitivity.

Once accustomed to using the meter,
one can proceed to provide a calibrated
scale for the variable capacitor C3. For
this, the gate dipper is used as an
AM demodulator. A length of wire
(minimum 10 metres) which can be
positioned either horizontally or verti-
cally is used as an aerial. The latter is
coupled to the coil of the grid-dip meter
via a single-turn coil (see figure 7). One
end of the coupling coil should be
earthed (to for example a water pipe
(etc.). Capacitor C3 is then adjusted
until a known AM station can be heard
via the headphones. The oscillator fre-
quency will then be the same as the
carrier wave frequency of the trans-
mitter. The scale for the variable capa-
citor can be calibrated simply by tuning
into a number of different stations. If
desired, higher frequencies can be cali-
brated by employing a number of tuned
circuits of known resonant frequency.
The position of PI at which reception is
the strongest corresponds to the position
which gives maximum sensitivity when
using the circuit as a grid-dip meter. To
facilitate tuning, it is recommended that
a tuning^, capacitor with slow motion
drive be use3. H

1 0-26 - elektor

STRAIN

Although the device described here is
called a strain gauge, it is in fact being
used to measure stress, i.e. the forces
which are applied to it. Strain denotes
the deformation of a material (change in
form or bulk) as a result of the action of
stress. However in all elastic materials
(such as e.g. steel) there is a linear
relationship between stress and strain,
which is expressed by the following
equation: 6 = e. E, where 8 is the stress,
e. is the strain, and E is a coefficient
termed the modulus of elasticity. Every
elastic material has its own modulus of
elasticity which remains constant within
certain limits of stress. Since strain is
proportional to stress, it is thus possible
to measure the one via the other.

The basic design of the strain gauge is
shown in figure 1 . An electrical signal is
derived from a transducer. This signal is
then amplified and used to drive an
LED scale display. If one looks ahead to
figure 3, it can be seen that the elec-
tronics involved are in fact extremely
simple. The heart of the strain gauge is
the stress absorber, the object upon
which the forces to be measured actually
act, and whose strain is measured. This
part of the device cannot be bought,
and must be made oneself.

Stress absorber

As is apparent from figure 2, the object

which bears the brunt of the forces
be measured is formed from a sheet of
suitable metal, with a hole drilled in
each end. The central portion is made
narrower than the top and bottom,
since it is at this point that the defor-
mation of the metal is measured.

The amount of strain is actually
measured by a special type of transducer
called an electric-resistance strain gauge.

In its simplest form it consists of a grid
of resistance wire cemented between
two sheets of paper. The gauge is
bonded to the metal, so that it undergoes
the same deformations. The resultant
changes in the length and cross-sectional
area of the wire causes a proportional
change in its resistance.

As figure 2 makes clear, four resistance
strain gauges are mounted in a bridge
configuration, two on the front of the
stress absorber and two on the back.

The changes in the resistance of the W. van Dreumel

few projects which have
not formed the subject of an
article in Elektor at one time or
another, however a strain gauge
falls into that category. This in
itself is perhaps slightly surprising,
since there are a number of
possible applications for such a
device - a training aid for
'strength sports', measuring loads
on cables, etc. or simple weighing
purposes.

vertically oriented gauges (R2 and R3)
are summed, whilst the horizontally
oriented gauges provide temperature
compensation. A further advantage of
this arrangement is that flexing of the
metal in the lateral plane will have no
effect, since the bridge remains in
equilibrium.

The bridge is provide with a stabilised
supply voltage. A current of roughly
20 mA can flow through the strain
gauges, and since they have a resistance
of approximately 1 20 S2, the voltage
across the bridge is fixed at roughly 5 V.

Circuit

The circuit of the strain gauge is shown
in figure 3, and, as has already been
mentioned, is fairly modest in dimen-

The low level output voltage of the
measuring bridge must be considerably
amplified before it can be displayed.
This is performed by two 747 ICs, each
of which contains two 741 type op-amps
(it is of course also possible to employ
four conventional 741's). A1 and B1 are
connected as unity-gain amplifiers with
high input impedance, so that the bridge
is not loaded by the amplifier circuit.
The latter is formed by A2 and B2,
which are connected as a differential
amplifier with a gain of approximately
1000, adjustable by means of P2. Under
quiescent conditions (no force applied
to the gauge), PI is adjusted for zero
output voltage.

The display takes the form of a column
of LEDs, which are driven by the
well-known UAA 170 LED voltmeter 1C.
Depending upon the input voltage, this
chip lights one of the LEDs D3 . . . D18.
The input is protected against negative
and excessively large positive voltages
by zener diode D2.

The power supply circuit is also quite
straightforward. Two integrated voltage
regulators (7812 and 7912) provide
the + and -12 V rails, whilst the 5 V for
the resistance bridge is obtained by the
inclusion of two resistors (R9, RIO) and
a zener diode (D1).

Construction

The amplifier, display driver and displays
can easily be mounted on a strip of

1-10-27

Veroboard, or similar. The stress ab-
sorber, however, is slightly more compli-
cated, since it involves a certain amount
of mechanical handiwork.

The dimensions of the stress absorber
will depend upon the type of material
used and upon the desired measurement
range. To obtain optimum sensitivity,
the material should undergo as great a
deformation as possible when under
maximum load conditions. As can be
seen from column 3 of table 1, the most
suitable material from this point of view
is hard brass, with duraluminium a good
second. Column 2 of the table is used to
calculate the cross-sectional area of the
stress absorber (X x Y in figure 2). This
is done by dividing the maximum
permissible stress into the required
range of forces to be measured.

The ratio of X to Y can be chosen
individually, however X should not be
smaller than approximately 10 mm
(because of the size of the strain gauges)
and the overall shape of the stress
absorber should remain similar to that
shown in figure 2. The values of R6 and
P2 in the circuit diagram are calculated
on the basis of a stress absorber made of
duraluminium and with a cross-sectional
area of 20 mm 2 .

Although electric-resistance strain gauges
are not widely used by the amateur,
various types are available commercially.
For this particular application their
dimensions should be in the region of
5x10 mm. Suitable types are (among
others) the EA-XX-2506G-120 from
Micro Measurements, the 3/120 LY 11
from HBM, and the PR9833 k/01 from
Philips.

y Calibration

e Under zero load conditions PI is adjusted

h such that the first LEO in the scale

e lights up. A known weight is then

t. suspended from the gauge and P2

configuration of four electric-resistance strain
Figure 1. Basic principle of the strain gauge. sensors are bonded.

E; kg/mm’ 6; kg/mm 1 stress: %

hard brass 9000 42 0.46

duraluminium 7000 26 0.37

semi-hard brass 9000 24 0.27

hard aluminium 7000 14 0.20

sheet steel 21000 18 0.09

Table 1. The information contained in the table allows the suitability of various metals to be
assessed, and the cross-sectional area of the "stress absorber" to be calculated in each case.

adjusted until the corresponding LED However if PI is mounted such that it is
lights up (obviously this will depend accessible externally, this should not
upon the measurement range chosen), present too many problems.

If, for example, 10-turn potentiometers

are used for PI and P2, a fairly accurate

scale can be obtained. For a variety of Literature:

reasons, it is possible that the zero point Linear Applications, National

of the scale may tend to fluctuate. Blektor 12, April 1976. M

10-28-

jctober 1979

It is interesting to note that, by and
large, our readers' comments and queries
— yes, and problems, too — run parallel
to our own. It is even more interesting
that all our problems have been solved,
as will be described.

To make full use of a microprocessor,
one should normally have access to the
instruction manual. For the 2650, this is
a 174-page book ... Fortunately, the
main points can be summarised rather

specifying negative numbers means that
00 . . . 3F are positive; 40 . . . 7F are
negative; greater than 7F don't exist.

All this may or may not seem simple in
theory; in practice it has proved a source
of endless programming errors. . . It is
easier to miscalculate a relative address
than to get it right! For simple programs,
one may as well use 'absolute addressing'
— the additional memory space required
(the corresponding instructions are
longer) is rarely a problem.

I plapd TV games....

Everything you want to know about making software for the TV games
computer, in two easy lessons . . .

In Elektor 48, April 1979, we
described how to build a
'TV games computer'. Included
was a brief explanation of how it
works; the 'instructions for use'
consisted of little more than the
Read Cassette routine, so that the
programs given on ESS records
can be entered.

Apparently, however, the majority
of our readers want more: they
want to do their own programming.
'This will prove relatively easy',
we said - and to prove it, the
(sometimes fairly sophisticated)
programs on the second ESS
record for the TV Games computer
were developed by a novice. The
following article is based on the
experience gained . . .

Addressing modes
When fetching or storing data, or
jumping to and fro in a program, it is
essential to specify the 'address'
concerned. Obviously. In the TV Games
computer, there are several different
ways of doing this.

However, practice makes perfect, and as
programs get more complex it becomes
worthwhile to start using relative
addresses wherever possible. As an aid
to the beginner, one of the programs on
the new ESS record contains a calcu-
lation routine for relative addresses — a
useful check!

Absoluts or relative

An 'absolute' address is simply the
address itself. For instance, in machine
language the instruction for 'Load
Absolute into register zero' starts with
0C (more on this later!); if the data is to
be fetched from address 0F00, the full
instruction will therefore be 0C0F00.

A 'relative' address, on the other hand,
specifies a small jump in the program.
Basically, the processor will calculate an
'absolute' address by adding the specified
number (between -64 and +63) to the
address that follows that particular
instruction. As an example, if the
instruction 'Load Relative into register
zero, 2F' (in machine language: 082F) is
located at the two address bytes 093E
and 093F, the following address is 0940.
The 'absolute' address corresponding to
this instruction is therefore 0940 + 2F =
096F, and the data will be fetched from
there.

The negative number required for a
'backwards' jump is entered as a '7-bit
two's complement number'. In simple
language, this means that you count
down from 80 hex- For instance, if in
the previous example the data was to be
loaded from address 093D, the relative
address would be 70: the 'following
address', 0940 corresponds to 80, so
093F corresponds to 7F, 093E to 7E
and 093D to 7D. The full instruction is
therefore: 087D. Note that this way of

1

Direct or indirect

The two types of addressing explained
above are both referred to as 'direct'
address modes: data is transferred from
or to the specified address. An alternative
possibility is a two-step operation:
specify an address where the desired
address can be found. This is referred to
as 'indirect' addressing.

Although both absolute and relative
indirect addresses are possible, only the
latter are useful in the basic TV games
computer. A relative address is converted
to an indirect relative address by adding
80. In the example given above, the
'load relative' instruction 082 F was
located at addresses 093 E and 093F; the
data was then fetched from address
096F. However, if the instruction is
modified to 08AF (2F + 80 = AF) the

elektor October 1979 - 10-29

08F0

08F4

08F8

08FC

0900

0902

C0 60 50 CE
3D CE 50 60
C0 00 28 FF
63 FE 00 00
7620
05C3
0400

(-► CD5F00
1 — 5978
050E

► 0D48F0
CD7F00
- 5978

CC1 FC0

c ?l

I 5!

0922

0924

0926

0929

c

CC1 FC1

0C1 E88

F420

9879

3F05CD

1F0014

PPSU. II
LODI. R1
LODI. R9
STRA. I-R1
BRNR, R1
LODI. R1
LODA. I-R1
STRA, I/R1
BRNR. R1
LODI. R0
STRA. R0
LOO I. R0
STRA. R0
LODA. R0
TMI. R0
BCFR
BSTA. UN
BCTA, UN

3 instructions ar

already 00 aft

by modifying the data in the LODI
• This slightly extended 'return to monitor'

However, other cc

contents of addresses 096F and 0970
will be used as the absolute address for
this instruction: if the data stored at
these addresses is 0A and 00, say, the
'load indirect relative, 2F' instruction
will be carried out as if it read 'load
absolute from 0A00'.

Once again, for simple programs it is
easier, quicker and more reliable to use
the corresponding 'absolute' instruction,
and forget about the 'relative indirect'
mode. As an aid to courageous novices,
the calculation routine mentioned above
actually gives two results: if the relative
jump in the previous examples is calcu-
lated, the answer will appear as '2F Or
AF'-for direct and indirect, respect-
ively!

Indexed

In contrast to the 'relative' and 'indirect'
addressing modes, 'indexed' addressing
can prove extremely useful in even the
simplest of programs. The basic idea is
that the data stored in one of the
registers is added to a specified 'absolute'
address; the result of this addition is
used as the absolute address for the
instruction. The register containing the
additional data for the address is referred
to as the 'index register', and this
register must be specified in the instruc-
tion. The data are always transferred to
or from register zero when indexed
addressing is used.

To specify the basic indexing mode,
6000 is added to the absolute address.
Thus 0D6900 is not interpreted as
’load register one from absolute address
6900'; if we assume that the data
already in register one is 0A, the instruc-
tion will be read as 'load register zero
from absolute address 090A' — i.e. from

tion make it invaluable: 'indexed with
auto-increment' and 'indexed with auto-
decrement', specified by adding 2000 or
4000 to the absolute address. In both
cases, the final address is calculated in
the same way - by adding the data in
the 'index register' to the specified

absolute address. However, before
calculating the final address, the data in
the index register are increased by one
('auto-increment') or one is subtracted
from the data ('auto-decrement').

The value of this instruction is best
illustrated in an example. Let us assume
that we want to clear all 'background
data' in the PVI. This means storing 00
all addresses from 1F80 . . . 1FAC:
in all! Instead of using 45 individual
ire absolute' instructions, a single
>re absolute, indexed with auto-
decrement' instruction can be used,
with a bit of padding:

052D LODI, R1

0400 LODI, R0

r>CD5F80 STRA, I-R1

I — 597B BRNR, R1

The 'shorthand abbreviations' given
after the actual machine-code instruc-
tions are referred to as 'mnemonics'.
They are simply a quick way to jot
down what the instruction does.

This brief section of program is executed
as follows. First, the 'index register', R1,
is loaded ('LODI, R1'= Load Immediate,
Register 1 - more on this later) and
'00' is loaded into Register 0. This is
followed by the 'Store Absolute, Indexed
to Register 1 with auto-decrement'
instruction - incidentally, the value of
mnemonics is clearly illustrated here: it
is a lot quicker to write 'STRA, l-RT
than the mouthfull given above. At this
point, the value in R1 (2D) is reduced
by one and the result (20 is added to
the basic absol'*** address 1F80
(5F80 = 1 F80 + 4000 for 'auto-
decrement'). The value in R0 (00) is
then stored in the resultant absolute
address: 1 F80 + 2C = 1 FAC. One down,
44 to go! The next instruction, which
will be explained in greater detail later,
is 'Branch if Register 1 is Non-zero,
Relative'. Since R1 is most definitely
non-zero (it is still 2C at this point), the
'relative branch' is executed: the
program 'jumps back' to the beginning
of the previous instruction, as indicated
by the arrow. This whole performance is
repeated, storing 00 in progressively
lower PVI addresses, until the data in

0993
0905
0907
O90A
O90C
090 E
0911
0913
0915
0918

054E

0400

-* CD5F00

- 597B
0469
CC1 FC6
052D
04FF

-»CD5F80

- 597 B

LODI. R1
LODI. R0
STRA, I-R1
BRNR, R1
LODI, R0
STRA. R0
LODI. R1
LODI. R0
STRA, I-R1
BRNR. R1
HALT*

:o end a program, as w

1979

played TV i

R1 becomes zero. At this point, the
BRNR, R1, instruction does not result
in a jump back, since R1 is zero, and the
rest of the program is carried out.

For those who feel like trying out this
program, it is more interesting to turn
the background on instead of off. In
that case, the background and screen
colour must also be specified: '69' in
address 1 FC6 gives yellow on blue.
Furthermore, the objects will have to be
cleared, since they are also used by the
monitor program. A complete program
is given in Table 1; the reason for starting
at address 0903 (instead of 0900) will
be given later.

While on the subject of indexed
addressing, one final point should be
noted. In general, this mode is available
as a variation of all absolute addresses,
with the exception of branch instruc-
tions. The only two indexed branch
instructions, BXA and BSXA, will be
discussed further on.

FC1 STRA, R0)

80 LODI. R0

1 F91 STRA. RO

1 F93 STRA, R0

1 F9F STRA, R0

1 FA1 STRA. R0

0C LODI. RO

1 F98 STRA, R0

30 LODI. RO

1 F99 STRA. RO

01 LODI. RO

1 FAB STRA. RO

49 LODI, RO

1 FC6 STRA, RO

E88 LODA, RO

120 TMI. RO

!79 BCFR

: 05CD BSTA, UN

-0014 BCTA, UN

Nearly all instructions involving transfer
or manipulation of data require the use
of a register. Obviously, the register to
be used must be specified in the instruc-

In the examples already given, and
Table 1 in particular, this principle is
clear. The first byte of each instruction
specifies the basic instruction and the
register involved. For instance, the basic
instruction for 'Load Immediate' is
04xx (where 'xx' is the data to be
loaded); adding the number of the

register to this gives the complete
instruction: 04xx for Register 0, 05xx
for R1, 06xx for R2 and 07xx for R3.
In practice, this means that four vari-
ations exist for most instructions: one
for each register. It also means that the
second digit in an instruction specifies
the register involved: 0, 4, 8 and C for
register 0 (0803, for instance); 1, 5, 9
and D for register 1 ; and so on.

Finally, some instructions refer to data
transfer or manipulation involving two
registers, one of which is always register

zero. The instruction 'Load Register 0
from Register 1', for instance, is 01.
Similarly, 'LODZ, R2' (to use the
mnemonic) is 02. It should be noted
that in some cases, but not all (I), both
registers can be specified as Register 0.
This can sometimes be useful, as will
be explained under 'programming tricks',
next month.

Registers

We have already mentioned ‘registers'
several times. It is now time to take a
closer look at them. To put it in a
nutshell, a register can be visualised as a
memory location inside the micro-
processor itself. In the 2650, 8-bit
registers are used; this means that they
can store any data value from 00 to FF.
In all, seven 'general-purpose' registers
are available: register 0 and two 'banks'
of three registers (R1, R2, R3 and R1',
R2' and R3’). Of these seven, register 0
is always immediately available; at any
given moment, however, only one of the
register banks ( R 1 . . . R3 or R 1 ' . . . R3')
is accessible. The other bank, and the

Table C.

data contained in those three registers, — sense: this bit is 'V for the duration
is 'in cold storage'. (The way in which of the vertical reset pulse, at the end
one or other of these banks can be of each 'frame'. It can be used, for
selected will be discussed below: see example, to synchronise the program to
"Program Status Word'). Any instruction the actual display on the screen,
referring to R1, R2 or R3 is performed — flag: can be set, reset and tested at

only on that register in the selected will, as an indication of some con-
bank - it has no effect on the corre- dition relating to the program - for
sponding register in the other bank. instance, to distinguish between the first

and following runs through a particular
Program Status Word section in the program.

The 'Program Status Word' refers to two - Interrupt Inhibit. The PVI generates
special-purpose 8-bit registers: the 'interrupts’ at the end of each frame
"Program Status Upper" (PSU) and and eaclr time an object is completed. If
"Program Status Lower' (PSL). Each bit this bit is set, these interrupt requests
n these registers has a special meaning, are ignored; otherwise, program
as illustrated in figure 1. Briefly, the execution 'jumps' from wherever it
-nost important points as they relate to happens to be to address 0903 and runs
the complete TV games computer are as the program section that it finds there
follows: as a subroutine. Note that this can cause

chaos, if one isn't aware of the
mechanism; for this reason, it is advisable
to start every program with the instruc-
tion '7620' (i.e. set Interrupt Inhibit).
In the simple programming example
given in the original article (and in the
corresponding program on the ESS
record) this was forgotten . . . More on
this later. The Interrupt Inhibit bit is set
automatically by the processor when
the interrupt routine is executed; it is
only reset by an explicit command in
the program.

- Stack pointers. These three bits are
set and reset by the processor, to

keep track of the 'subroutine levels'.
The stack is eight levels deep, which
means that the main program may
branch to a subroutine, that may branch
to a further subroutine, and so on up to
eight times before starting to 'climb
back up' by means of Return instruc-
tions. It is possible to modify the stack
pointers deliberately, as part of a
program, but this is unwise for
beginners . . .

- Condition Code. These two bits are
set by (the results of) several different

instructions, as shown in the Instruction
Set given elsewhere. For instance, if the
data loaded into a register is 00, the
condition code will also be set to 00.
Most of the branch and return instruc-
tions can be made 'conditional', by
specifying a particular condition code
setting: in that case, a 'Branch on
Condition True' instruction, for
instance, will only be executed if the
actual condition code at that point
corresponds to the one specified. If the
two don't correspond, the instruction is
ignored.

- I DC, WC, OVF, COM, C. These five
bits will be dealt with later; see:

Arithmetic and Compare.

- Register bank Select. This bit is used

10-32 — ele

1979

Load registar zero (LODZ)
Load immediate (LODI)
Load relative (LODR)
Load absolute (LODA)
Store register zero (STRZ)
Store relative (STRR)
Store absolute (STRA)

92 from R2 to R9

08yy 'yv' = displacement

OCzzzz 'zzzz' = address

Cl to R1 from R0

C8yy -yy’ - displacement

CCzzzz 'zzzz' = address

to select one or other of the two
'register banks' described above.

Various manipulations are possible on
the two Program Status registers, as will
be described. The Clear, Preset and Test
instructions will prove the most useful;
these can be used to set any bit (or
combination of bits) to 0 or 1, as
required, and to test the setting of any
bit(s). An example was given above:
'7620' is the code for 'Preset Program
Status, Upper, Masked 20'; as can be
seen in figure 1, this sets the Interrupt
Inhibit bit.

Instruction Set

Several instructions have already been
mentioned briefly; having laid the
groundwork, it is now possible to
examine the instruction set in greater
detail.

Load and store

The principle of these instructions is
obvious: data is transferred into (Load)
or from (Store) a specified register.

Load Register zero and Store Register
zero transfer data between R0 and one

of the other three registers. 'Cl', for
example, transfers data from R0 to R1.
Note that the instructions '00' and
'C0‘, for 'LODZ, R0' and 'STRZ, R0\
don't exist.

Load immediate transfers the data given
in the instruction to the specified
register. '07CA' (= LODI, R3) loads the
data 'CA' into register 3.

Load relative and Store relative refer to
the relative addressing mode described
earlier. Relative Indirect addressing can
also be used, as described earlier.

Load absolute and Store absolute are
used when absolute or absolute indexed
addressing is required.

are set according to the sign of the data
transferred: they become 01 if the data
is a positive number, 00 if it is zero and
10 if it is negative (i.e. 80... FF,
corresponding to — 128 . . . -1).

The Load and Store instructions can be
summarised as shown in Table 2.

(Subroutine) Branch

Normally speaking, a program is
executed step by step: in other words,
the instructions are carried out in the
order in which they are stored in the
memory. If a jump to a different section
of the program is required, a so-called
Branch instruction must be used.

There are two basic types of Branch
instruction: those for a (main program)
Branch and those for a Branch to
Subroutine. In the former case, the
main program itself jumps to a different
point in the memory; a Branch to
Subroutine, on the other hand, can be
considered as an interruption in the

main program: the main program is
stopped at the branch-to-subroutine
instruction, the subroutine (elsewhere in
memory) is carried out, after which the
main program continues at the point
where it was interrupted. Several
variations of both types of Branch
instruction are available:

Branch (to Subroutine) on Condition
True, Relative or Absolute. For each of
these four basic instructions, a particular
setting of the Condition Code bits can
be specified; the branch will only be
executed if the actual condition code
corresponds to the one specified. For
example, the basic instruction for Branch
on Condition True, Absolute (BCTA) is
'ICzzzz', where zzzz is the absolute
address to which we want to jump. As it
stands, this branch instruction will only
be carried out if the condition code is
00. Similarly 'IDzzzz' and 'lEzzzz'
specify the condition codes 01 and 10,
respectively. Finally, 'IFzzzz' would
seem to refer to a condition code 1 1 ,
but this code doesn't exist. In fact the
corresponding instruction is used for an
unconditional branch: a branch that is
always carried out, no matter what the
condition code.

Branch I to Subroutine) on Condition
False, Relative or Absolute, These four
instructions are similar to those described
above; the only difference is that the
branch is executed if the actual condition
code does not correspond to the one
specified. The 'BSFA' instruction
BCzzzz, for example, will cause a
branch to subroutine if the condition
code is either 01 or 10, but not if it is
00. Note that no 'unconditional'
variations of these instructions exist:
the corresponding codes 9Byy, 9Fzzzz,
BByy and BFzzzz are used for other
instructions.

Branch (to Subroutine) on Register
Non-Zero, Relative or Absolute. As part
of these instructions, one of the registers
(R0 . . . R3) is specified. If the content
of this register is not zero, the branch
instruction is carried out; otherwise it is
ignored. ‘BRNA, R0' (5Czzzz), for
instance, will cause a jump to address
zzzz provided the data stored in
Register 0 is not zero.

Branch on Incrementing (Decrementing)
Register, Relative or Absolute. These
instructions are an extension of the
previous set. Once again, a register is

1979- 10-33

I played TV games . . .

Branch Itc

example

(BCTR) 18yy
(BCTA) 1 Czzzz
(BCFR) 98yy
(BCFAI 9Czzzz
(BRNRI 58yy
(BRNA) 5Czzzz
(BIRRI D8yy
(BIRA) DCzzzz
(BDRR) F8yy
(BORA) FCzzzz
(ZBRR) 9Byy
(BXA) 9Fzzzz

Branch to Subroutine:

On Condition True, Relative
On Condition True, Absolute
On Condition False, Relative
On Condition False. Absolute
On Register Non-zero, Rel.

On Register Non-zero, Abs.
Zero Relative, Unconditional
Indexed Absolute Unconditionf

(BSTRI 38yy
(BSTA) 3Czzzz
(BSFR) B8yy
IBSFA) BCzzzz
(BSNRI 78yy
(BSNAI 7 Czzzz
IZBSRI BByy
(BSXA) BFzzzz

Return from subroutine:

Conditional (RETC) 14

And Enable Interrupt. Conditional IRETEI 34

IByy- unconditional
1 Fzzzz = unconditional
9Byy: see below
9Fzzzz: see below

R3only!

3Byy - unconditional
3Fzzzz = unconditional
BByy: see below
BFzzzz: see below

R3 1

■sly I

The function of the various 'bits' in the
Program Status Registers was explained
above. At this point, we are only
interested in the available instructions
(as summarised in Table 4).

The Load and Store instructions refer
to data transfer between one of the
Program Status Registers and Register 0
only. 'Load Program Status Upper'
(LPSU: 92) for instance, loads the
contents of R0 into the PSU.

In practice, these instructions will not
be used often, since in most cases
Clear, Masked or Preset, Masked instruc-
tions are more suitable. 'Clear Program
Status Upper, Masked 40' (7440) will

Program Status. Tast, Compare, at

Store

Clear

Status. Upper
Status. Lower
Program Status, Upper
Program Status. Lower

(LPSU) 92
(LPSL) 93
(SPSUI
(SPSLI

ogram Status. Upper, Masked ICPSU) 74 m
Program Status. Lower, Masked ICPSLI 75 m
Program Status, Upper. Masked (PPSUI 76 m
Program Status. Lower. Masked (PPSL) 77 m

Test Pro|

Test Under Mask Imp
Compare to Register
Compare Immediate
Compare Relative
Compare Absolute

No Operation

(TPSL) B5mm

(TMI) F4mm . . . F7mm
ICOMZ) E9 . . . E3
(COMI) E4xx . . . E7xx
(COMR) E8yy . . . EByy
(COMA) E Czzzz . . . EFzzz

specified. In this case, however, 01 is
first added to (increment) or subtracted
from (decrement) the contents of the
register, after which the branch instruc-
tion is only carried out if the new
contents are non-zero. Note that no
'Branch-to-subroutine' version of these
instructions exists.

Zero Branch (to Subroutine! Relative,
Unconditional. These two instructions
are relatively useless in the TV games
computer, since they specify a branch
relative to address 0000: the start of the
monitor program!

Branch (to Subroutine) Indexed,
Absolute, Unconditional. These two
instructions are the only two indexed
branch instructions that exist. The value
in the index register (which must be R3)
is added to the basic absolute address
given, and the branch is executed to the
resultant address.

Return from subroutine, conditional. As
before, a condition code is specified as
part of this instruction; if the actual
condition code matches, the subroutine

is terminated. An unconditional end of
the subroutine is indicated by the
condition code 11, so the instruction
RETC, UN is 17. A variation on this
instruction exists (RETE) that not only
ends the subroutine, but also resets the
Interrupt Inhibit bit. Not a good idea,
until one has gained enough experience
to start using the interrupt facility . . .
The complete set of branch instructions
is summarised in Table 3.

clear the 'flag' bit, without having any
effect on the other bits in the PSU.
Similarly, 'PPSL, RS' (7710) will select
the second register bank.

Finally, any bit (or combination of bits)
in each of the program status registers
can be rested : 'Test Program Status
Upper, Masked 40' (B440) will cause
the Condition Code to be set to 00 if
the 'flag' is set; otherwise the Condition
Code will become 10.

10-34 -i

played TV gar

With the complete program given so far (in tables A ... C), it is possible to get the object into
your sights. Now, what about shooting it down?!

First, modify the instruction in address 0962: instead of '9C0991'. enter '9C099B'

The existing program, from address 0990 is then extended as follows:

(098E

0992
0994
0996
0998
099A
099B
099D
099 F
09 A 1
09A3
09A5
09A7
09A9
09 A 8
09 AO
09 A F
09B1
0983
0986
0989
09BB
09BO
09C0
09C2
09C4
09C6
09C8
09CA
09CC
09CF
09D2
09D4
09D7
09DA

F80O
COCO
COCO
CO CO
CO CO

coco

Once the object is accurately (!) i
the ’F* key.

BORR. RO)
2x NOP
2x NOP
2x NOP
2x NOP
2x NOP
RETC, UN

BCFR
PPSL, COM
COM I. R3
BCFR
COMI, R3
BCFR
COMI, R2
BCFR
COMI. R2
BCFR
LODI, R1
LODA. I-R1
STRA. I/R1
BRNR, R1
LODI. R1
LODA. RO
TMI. RO
BCFR
BDRR. R1
2x NOP
2x NOP
2x NOP
BCTA. UN
LODA, RO
TMI, RO
BCFA
BSTA, UN
BCTA. UN

Test under Mask; Compare

With all the conditional branching
facilities available, is is obviously useful
to have instructions that set the
Condition Code. Basically, all types of
data transfer to or data manipulation in
a register do this; furthermore, the Test
Under Mask Immediate (TMI) and
Compare (COM) instructions set the
condition code bits without altering the
data in any way.

The TMI instruction is the easiest to use:
a register is specified in the first part of
the instruction ('F4' for register zero.

'F5' for R1, and so on) and a 'mask' in
the second part. The mask simply
specifies the bits to be tested: '81', for
instance, is 10(30 0001 in binary and so
the first and last bits will be tested. If,
in the data contained in the specified
register, these two bits are 'Is', the
condition code will be set to 0(3; if not,
CC will become 10. An example. If the
data in R1 is 05, the instruction F501
(TMI, R1, 01) will set the condition
code to 00 —the data, 05, is 0000 0101.
By contrast, F581 will set the CCto 10:
0000 0101.

The compare instruction is basically
similar, but it is both more precise and
more versatile — and also more compli-
cated to use. In this case, a data value is
specified instead of a mask, and the
condition code can be set in three ways:
01 for 'greater than', 00 for 'equals' and
10 for 'less than'. There are two main
points to watch, when using this instruc-
tion: what is meant by 'greater than'
(data in register greater than data
specified, or vice versa; see the footnotes
in the instruction Set) and what type of
comparison is required. With the 'COM
bit' in the PSL set to 0, an 'arithmetical'
comparison will be performed: all values
from 80 to FF are treated as negative
numbers (two's complement)! If the
COM bit is set to 1 (by means of the
instruction 7702 = PPSL, COM) a
'logical' comparison will result: the data
is treated as a positive 8-bit binary
number.

No Operation

A surprisingly useful instruction, thisl
When the processor finds the code 'C0'
it simply carries on to the next instruc-
tion. There are two cases where this can
be particularly useful: to 'delete'
instructions that prove unnecessary,
without having to re-enter the rest of
the program, and to 'leave a gap' into
which further instructions are to be
added at a later date.

This stops the processor, quite drasti-
cally. The only way to start it up again
is either to operate the 'reset' key or
provide an interrupt — provided the
interrupt inhibit bit is not set. in general,
this is not a good idea; in the TV games
computer, a 'return to Monitor' instruc-
tion (1F0000 = BCTA, UN, for instance)
will usually be more suitable.

I played TV gan

aber 1979- 10-35

A few tips

The instructions explained so far are
sufficient for simple programs. The
remaining facilities will be dealt with
next month. Meanwhile, however, a few
practical tips on how to program should
prove useful.

First and foremost: remember to block
the Interrupt facility if this is not
required in the program! For the time
being, it is advisable to start every
program with the instruction ‘7620’
(PPSU, II).

There are several ways to end a program.
Usually, one of the keys ('PC', for
instance) is used to initiate a jump back
to Monitor. A few variations are given at
the end of Tables A . . . D. The jump to
Monitor itself can be done in several
ways. The shortest is to use a ZBRR
instruction: '9B00' should do the trick,
but we've never actually tried it. A

similar solution is '1 F0000', as
mentioned earlier; this we have tried,
and it usually works. Sometimes,
however, for no apparent reason
problems occur: in particular, a row of
black squares or lines down the left-hand
edge of the screen when the program is
restarted. Without knowing why this
happens, yet (maybe we will know more
next monthl), we can offer three
solutions:

- return to monitor by means of the
two instructions

0400 LODI. R0

1F0011 BCTA. UN

Note that, in this case, the original value
in R0 is lost; for that matter, returning
via address 0000 always causes the data
in R0 to become 09, as several readers
have noticed!

- similarly, but untried :

20 EORZ, R0

9B11 ZBRR

This has the advantage that if fits in the
same memory space as '1F0000', if the
latter causes trouble.

- finally, if the value in R0 is to be
stored:

3F05CD BSTA, UN

1F0014 BCTA, UN

Don't ask us to explain this one — that
would require an extensive discussion of
the monitor software!

When it comes to the program itself, the
first thing is to work out what you want
to do. Obviously. For simple programs,
this can usually be put into words quite

[T

060A

0F1FCB

F440

9879

FA77

LODI, R2
LODA. R3
TMI, R3
BCFR. *
BDRR, R2

0900

0902

0904

0906

0909

0900

0910

0912

0914

0917

0919
091 C
091 E

0920
0922
0924

7620

05AD

0400

C CD5F0O
597 B

[C

CC1 FC6

052D

04FF

CD5F80

060A

0F1FCB

F440

9879

FA77

5970

PPSU, 1 1
LODI, R1
LODI, R0 L

STRA, I-R1 |

BRNR, R1 J

LODI, R0 "1

STRA. R0 J

LODI. R1
LODI. R0
STRA, I-R1
LODI, R2 "V
LODA, R3
TMI. R3 >

BCFR. * (

BDRR, R2 )
BRNR. R1
HALT

background

easily. The program given in Table 1 , for
example, was originally specified as
'clear objects; define background colour;
load FF in all background bytes'. For
more complicated programs, some more
extensive advance plotting may be
required - using a flow chart, say - but
usually a complicated program can be
'broken up' into several simple routines.
These can each be tried and tested
individually, before 'tacking them
together' to obtain the complete final
program.

As each semi-complete routine is
entered, it is highly advisable to store it
on tape before the first test run. This
lesson was learned the hard way, when
one misplaced relative address caused
'garbage' to be stored at all sorts of
awkward places throughout the program.
The only solution was to laboriously
re-enter the whole program . . .

When it comes to 'de-bugging' a
program — it always does, no program
works perfectly first time! — the
'Breakpoint' routine can be very useful.
There are two points to watch, however.
In the first place, as mentioned in the
original article, the Breakpoint address
given must be the first address of an
instruction. For instance, in the
following section of program:

0900 7620 PPSU, II

0902 0400 LODI, R0

0904 0605 LODI, R2

breakpoints can be specified at addresses
0900, 0902 and 0904, but not at 0901 ,
0903 or 0905! The second point to
watch is that breakpoints modify the
program at that point. If the breakpoint
is found in the normal way, the original
data will be restored automatically.
However, if things really go wrong so
that the reset key must be used to
return to Monitor, it may be necessary
to restore the data by hand!

The PVI and keyboard

The main points regarding the PVI were

10-36-1

• October 1979

I played TV gar

Table E.

Finally, what about adding a time limit? As follows:

- first, fill in the space in the program starting at address 0990:

(098E

0990

0994

(09C4

09C6

09C8

09CA

F800

7710

E700

F977

7710

0700

7510

BDRR, R0)
PPSL, RS
COM I, R3
BCFR
LODI. R3
CPSL, RS
ting at 09C6:
BDRR, R1I
PPSL. RS
LODI. R3
CPSL. RS

i data at address 0981 : instead of '1

— at address 09D4. the instruction is modified i

set the 'dock' IR3 in the upper
register bank! going as soon as the
object is first mowed

| the object is hit

1 F0945'. the instruction becomes T F09DD'
io '9C0976' (instead of 9C0945I.
on. as follows:

if 'clock' stopped (R3 - ■ 00).
preset R2’ for one-second count

} decrement R2'

1 reset R2' and
/ decrement R3‘
branch if R3' - 00
| repeat key check routine

store 00 in score, make screen
white ('you lose'!) and repeat

all discussed in the original articles and
the data supplied with the p.c. board.
One point, however, did not receive all
the attention it deserves - at the time,
we didn't realise how useful it was!

The 'VRLE' bit, at address 1FCB, goes
high at the end of each frame; it is reset
at the end of the VRST pulse or when
read. This means that it can only be
read as '1' once for each frame. As an
example, a simple 'delay' routine is
given in Table 5. Basically, what happens
is that the processor waits until it finds
VRLE = 1; it then decrements the value
in R2 and repeats the VRLE scan if R2
is not yet zero. The result is a delay,
approximately equal to the value in R2
times the frame period (20 ms). By way

of demonstration, this routine can be
included inside the 'load background'
loop in Table 1. The result is given in
Table 6.

Finally, the keyboard. Each column
corresponds to one address: 1E88 for
the column above the '-' key, 1E89 and
1E8A for the next two columns, 1E8B
for the column that includes the 'reset'
key (note that this key itself is not

scanned in the keyboard layout
suggested!) and 1E8C . . . 1E8E for the
last three columns. When reading the
keyboard in this way, the four left-hand
bits retrieved as data correspond to the
four keys in the column — and the other
four bits are all ones! 'IF' at address
1E88, for instance, means that the

'-' key is operated; '4F' at 1E8A
corresponds to key '8'. Note that
contact bounce can sometimes be a
problem with this type of rapid
key-scan. A more sophisticated routine,
using part of the Monitor software, will
be described next month.

RCAS, WCAS and ESS

The Cassette routines were discussed in
the earlier articles. Apparently, some
readers have had problems loading the
first ESS record, so a few words of advice
may be appreciated . . ,

Assuming that a cassette recorder is
used, the first point to check is that
programs can be stored on tape from
the computer and retrieved without any
problems. This can be done without
even loading a program: there is always
some kind of data in the memory! The —
test sequence is as follows: T ,

- operate the 'reset' key;

- operate the 'start' key ('1111' should
appear) ;

- press the "WCAS' key CbEG =');

- enter 0900, followed by '+' I

('bEG = 0900, End =');

- enter 0FFF, followed by '+' I

('End = 0FFF, SAd =');

- enter 0900, followed by '+'1

('SAd = 0900, FIL =');

- enter but not '+'('FIL = V);

- start the tape in the Record mode,
and set the level to about half way;

- operate the '+' key.

Hopefully, the recording level meter
should indicate approximately nominal
full modulation during the first second
or so after the '+' key is operated; it will
then drop back slightly (to a few dB
below full modulation). If this is not the
case, the level setting can be corrected,
after which the whole sequence described
above will have to be repeated. Having
found the correct level, it is wise to
make a note of it, for future reference.

z* ' ' n * o i i;

• . played TV games . . .

elektor October 1 979 - 1 0-37

Having made a complete recording at
correct level, the various addresses
entered above will reappear on the
screen. The test can now be concluded:

- operate the 'RCAS' key ('FIL =');

- enter the file number, '1' ('FIL = V);

- press the key ( not '+'!).

The text 'FIL— 1' will jump to the top
of the screen. The tape can now be
played back, and the data recorded on it
Mill be compared with the original data
in the memory. During this time
'approximately 35 seconds) two dots
will flash below the sign on the
screen. At the end of this time, all the
t original data will reappear on the screen
a with the added line 'PC = 0900'. If this
d nappens, all is well and the cassette
i, nterface is working.

II In the unhoped-for event that the check

routine breaks off before the end of the
recording, with the message 'Ad = 090A',
for instance, then something is wrong . . .
In our experience, moving the recorder
further away from the TV set invariably
cures the problem.

Next step. The ESS record. You would
expect that recording it on tape and
then playing it into the computer
should work. In practice it does, most
of the time, but sometimes the computer
rejects the program for no apparent
reason. (Message: 'Ad = ...'). Since the
programs are on the record (with the
exception of the missing 'Interrupt
Inhibit' instruction in file 6, as
mentioned earlier) it must be possible to
load them. In one particularly stubborn
case, the following solution was found.
The output from the preamplifier, after

H LITTLE
PrHCTICE
dOE5

H LOT DF
CUDd

Table 7.

d

0900

0902

0905

0907

0909

090C

*■ 0913

0916
0918

ir 091 A

ll 091 D

d 091 F

|| 0921

3 0924

0929
093 B

>0508

- 0E492D
CD4890

- 5978
7710
3F020E
7510

-5A0A

► 0C1E89
F410

-1879
1 F0038
>-►7710
3F02CF
7510
1B5A

PPSU, II
BSTA, UN
LODI. R2
LODI. R1
LODA, I-R2
STRA, I-R1
BRNR, R1
PPSL, RS
BSTA, UN
CPSL. RS
BRNR. R2
LODA, R0
TMI, R0
BCTR
BCTA, UN
PPSL. RS
BSTA. UN
CPSL. RS
BCTR. UN

(clear/initiate PVI)

(data)

(load MLINEI

'+' key

the tone and volume controls, is fed to
the TV games computer. A 'high' file
number is entered (8 or 9) in the RCAS
mode, and the record is started. After
some manipulations with the volume
control, two dots will start to flash
rapidly under the '=' sign, and the actual
file number should also appear on the
screen. The trick is now to manipulate
the volume control (and, if necessary,
the treble control) until the dots flash
regularly and the second file number
remains constant for the duration of
each program on the tape. Once this is
achieved, the volume and treble controls
are left strictly alone and each program
in turn is loaded into the computer (this
should work, now) and from there to
the tape. From now on, the programs
can be retrieved reliably from this tape.
Rest assured, we are doing our utmost
to make the second ESS record for the
TV games computer easier to load . . .

ESS 006

. . . the second record with software for
the TV games computer, which is what
this article was to have been about.
However, it's long enough as it is.

Some idea of the programs can be
obtained from the photo's distributed
liberally throughout these pages. One
program converts the computer into a
fairly comprehensive colour TV test
pattern generator; the other can be
considered as a programming aid. It
contains routines for composing object
shapes and background on the screen
- so that you can see what you're
doing-, the 'relative address calculation'
mentioned earlier, and a routine for
scanning a character set available in the
monitor program as will be explained
next month.

Full details of how to use these programs
will be included with the record, which
will be made available next month.

In conclusion

With the information given in this
article, it is possible to write simple
programs. Some examples are included
on these pages. Now is the time to start
practicing — next month we'll discuss
the rest of the instruction set, and
give some rather more complicated
routines . . . After that, you will know
as much as we do!

10-38 - elektor October 1979

programmable sequencer

There are a variety of different ways in circuit and the 'subsidiary' address
which the control voltages can be counter, however, even longer (or)
programmed and stored: e.g. via poten- indeed shorter) sequences are also
tiometers, switches, sample-and-hold possible. The note length is controlled
circuits or digital memories. The by a D/A converter and VCO, the
method adopted here is to encode output of which varies the clock fre-
the voltage digitally and store it in a quency of the main address counter.
RAM. When the contents of the mem- The analogue voltages from output A
ory are read out, they are fed to a D/A are fed to the synthesiser VCOs; at
converter, which provides an analogue output B a gate pulse is generated to
signal suitable for feeding to the syn- accompany each note. The gate pulse,
thesiser VCOs. In addition to the pitch whose width can be varied, is used to
of the notes (i.e. their frequency), their determine the start and duration of the
relative length can also be programmed, envelope control voltage generated by
The duration of each note can be the ADSR module of the synthesiser,
selected in the ratio of 1 : 2 : 4 : 8. The complete circuit diagram of the
The block diagram of the programmable programmable sequencer is shown in

programmable sequencer

Sequencers are extremely popular
add-on units for music
synthesisers. They are used to
store pre-programmed sequences
of control voltages for the
synthesiser VCOs/VCFs; the
control voltages can be 'played
back' into the synthesiser, thereby
generating note sequences which
can be used for example to
provide the backing to a
manually played melody.

C. Voss

sequencer is shown in figure 1. The
pitch (i.e. its position on the musical
scale and its octave) and length of
the note are set up in binary code on
switches which are connected to the
data inputs of the RAM (see figure 3).
The address into which the data is
stored is determined by an address
counter. In actual fact, two address
counters are employed, one of which
(the 'subsidiary' counter) is clocked
by the other ('main' counter). When
the stored melody is to be played back,
the address counter steps through
each of the memory locations in turn.
The data is read out and fed to the
digital-analogue converters, which pro-
vide the actual control voltages for the
VCOs. During normal operation the
circuit can store 16 sequences of 16
notes apiece, i.e. a combined sequence
of 256 notes; with the aid of the reset

figures 2a and 2b. Figure 2a contains:
the digital section of the sequencer,
comprising the memory, address counter
and reset circuit, whilst figure 2b shows,
the D/A converters and output stages.
Two 2101 's, 256 x 4-bit R AMs, connec-
ted in parallel from the memory in
which the digitally encoded control,
voltages are stored. The higher order:
addresses of the input data are set up
on switches S2 . . . S5. The flip-flop
(IC11) interposed between the switches’
and the RAMs ensure that the new
address set up on S2 . . . S5 is only
presented to the address inputs of the
RAMs after the previous note sequence!
has ended.

The main address counter is formed by
IC10. The counter is clocked, via IC6,
by the analogue section of the circuit!
shown in figure 2b. This counter
generates the 'low order' addresses,!

iiiy

i.e. it clocks from '0000' to Mill',
whereupon the high order address is
incremented by one (via S2...S5),
before the counter resets and starts to
cycle through another sequence of 16
addresses.

The reset circuit is formed by N12 and
N13. When the data outputs a ... f of
the RAM all go high, N12 and N13
ensure that the binary counter is reset
to zero. Thus the address containing
the data word Mil 111' represents the
reset address. Inverters II ... 14 form a
NOR gate (the inverters all have open-
collector outputs), so that only when
the address counter resets (i.e. its out-
puts all go low), is IC11 clocked. This
ensures that a new (high order) address
cannot be presented to the address
inputs of the RAMs before the previous
note sequence has ended.

When S2 . . . S5 are set to position c,
the ’subsidiary' address counter (IC12)
is connected to the address inputs of
the RAMs. This counter is clocked by
IC10, via II . . . 15, so that it receives
a clock pulse every time IC10 resets
(i.e. every 16 addresses). Thus if all
the outputs of IC12 are connected to

the RAMs, the entire contents of the
memory can be read out in sequence.
Switch SI determines the operating
mode of the sequencer. In position a
the switch blocks gate N10, with the
result that the address counter is im-
mediately inhibited. In position b the
sequencer operates normaily, whilst
in position c the counter will stop once
it reaches '0000'.

To actually program a sequence of notes
into memory, pushbutton switch S9 is
first pressed, resetting the address
counter via N12 and enabling the
RAMs. The information relating to the
pitch and length of the note to be
stored is then written into the RAMs
by pressing S10. Each of the mono-
stable multivibrators MMV1 . . . MMV3
are now triggered in turn. The output
pulse from MMV1 clocks the address
counter (IC10) via N9. The pulse from
MMV2 temporarily puts the RAMs
into the write mode, so that the infor-
mation present on the data mputs is
in fact stored in memory. The Q output
of MMV3 takes the output of N8 high,
so that N13 is capable of recognising
the reset code ('111111') on the data

outputs of the RAMs.

The next note is written into memory
in the same way; the input data is set
up on the corresponding switches
whereupon S10 is pressed and the data
is written into memory. Once the
desired sequence of notes is stored,
pressing S11 writes the reset code into
the memory by taking the inputs of
N1 . . . N6 low and hence the data
inputs of the RAMs high. When N13
recognises the reset code, the address
counter (IC10) is reset, so that via
II... 15, flip-flop FF2 is triggered
and the RAMs are returned to the read
mode. Schmitt trigger N21 ensures that
FF2 assumes a definite state upon
switch-on and that the RAMs are
inhibited for a brief initial period.
The digital-analogue converters and
output stages of the circuit are shown
in figure 2b. IC1 ... IC3 produce the
analogue control voltages which deter-
mine the frequency of the notes, whilst
the D/A converter round IC4 is used
to control the length of the notes.
Unlike the other two D/A converters
(IC1/IC2), the output voltage increases
in an exponential, not linear fashion.

That is to say, when the digital input (potentiometers PI . . . P6), the circuit Fine tuning is performed by means of

signal increments by '01', the output can be set up fairly simply, without the PI. A change in state of input 'a'

voltage doubles. need for any special measuring equip- should correspond to a change of

The output of IC4 is fed to a sawtooth ment. The circuit is adjusted correctly 1/12 V in the output signal. P5 should

generator formed by IC4 and IC5; this when a change in the 'e' input from '0' be adjusted such that the output fre-

both clocks the main address counter to 'V causes the voltage at output A to quency doubles when input 'g' goes

(IC10 in fig. 2a) and, via the Schmitt increase by 1 V. This voltage can be high. P4 is used to compensate the

trigger IC7, provides a variable-width adjusted by means of P3. P2 should be offset voltages of IC4 and IC5. Finally,

gate pulse. set such that the output voltage changes P6 determines the width of the gate

In spite of the six adjustment points by 0.5 V when inputs 'b' and 'c' go high, pulse. M

July/August 1979

- Grcuit 25: linear them

basic computer

Elektor 49, May '79, p. 5-34. The
circuit diagram (figure 3) contains

of IC6 should be as
D-C-B-A along the top s
pin numbers 5, 4. 3 and 1
ively; along the bottom

N1 . . . N4 are ANDs, not N ANDs;
the output of N1 is point 27c.
that of N2 is 31a and that of N3

contains Circuit 27: moisture sensor
1 pinning The sensor should be connected
follows: j n parallel with R1. not in series
nould be with R2 .

! respect- Circuit 30: automatic heated rear
the pins windscreen

3. Gates To avoid co n f us j on; p j n 3 Q f |C6
NANDs; is the c i ock it is i ndeed

JJ ’ operation, are

types designed

Long range photoelectric * appi

switch 100% tested at

The E3N-30 Photoelectric Switch for 100°C par
from IMO Precision Controls has transistors, types
an operating range of 30 metres. 2N6689-93, featL
The E3N-30 incorporates the capability, with 1
latest innovations in design and voltage ratings of
circuitry including the unique switching speeds,
BI-COLOUR indicator to speed voltages and high
setting up time and indicate the area ratings -

2N6674-78 and
ire high voltage

This can be achieved quite easily
by removing the wire link con
necting these pins to pin 24; :

measuring only (including inductive turn-off tii

32a and 32c can then be ac
Finally, it was perhaps no
sufficiently clear that tl»

Circuit 80: programmable digital
function generator
IC1 and IC2 should be type 2101,
not 2102.

Circuit 90: uP-programmable
speed controller for model rail-

elekdoorbell

Elektor 50, June '79, p. 6-1 2. The
numbering of the switches as used

. ' ' bottom, the pin numbers should V

, read: 7 6 ' 5. 4. 9. 10. 11 T

t ere ore («Q7). | n figure 3, the supply

voltage should be 5 V.

Grcuit 100: 256-note sequencer
In figure la, pins 4 of IC1 . . . IC3
i. 6-1 2. The should, of course, be connected
:hes as used to -15 V. In figure 1c. a In
rrespond to capacitor is shown between pin 3
and on the of N38 and supply common. This 90.5 m

pin 14 can be left as it is), but
‘or some other semi-equivalents
'notably the Motorola 14034BCP)
■twill prove essential,
aquarium thermostat
Sektor 50, June 79, p. 6-29. In
sw circuit diagram, D8 is shown
=roected between pins 1 1 and
*2 of IC3. This diode should in
■act be connected between pin 1 1
^ 1 C3 and the +1 2 V line; pin 1 2
O' IC3 is connected only to D7.

Pin 12 of N1 6 is shown connected
via a diode to pin 7 of IC3b; it
should, however, connect to pin 6.
Furthermore, R28 is shown con-
nected to the right-hand side of
the Xtal, whereas it should be
connected to the top of Cl.

The author has suggested two im-
provements to his circuit: the
resistor values for R38 . . . R45
can be decreased by a factor of
10 (e.g. 47 k instead of 470 k),

2N6674-78 are
JEDEC TO-20
metic packages

supplied in the jedec
MA (TO-61) package
teel hermetic shell, a solid
header/stud for low ther-

RCA Solid State - Europe,
Sunbury-on- Thames,
Middlesex. TW16 7HW.

Switching transistors
designed for harsh
environments

10-42 -ele

October 1979

Miniature LCD panel
clocks

all usual timekeeping functions
in both UK and US formats for
time, day, date.

The unit is quartz controlled

fine adjustment), and includes
an incandescent backlight feature.

to drive a bleeper or some other

external means of indication.

With a running consumption

of only 6pA, the PCIM161A is
suited to a variety of applications
- ranging from all types of

to instrumentation, telephones.

communications equipment etc.

Ambit International.

2 Gresham Road,

Brentwood, Essex,

Telephone: 102771227050.

11280 M)

New product range
catalogue from Marshall's

Marshall ‘s announce the publishing

of their new 1979/1980 product

range catalogue on October 12th
1979.

This edition contains many new

including an increased range of
IC's, micro-power LCD clock
modules, data and educational

Marshall's retail branches. Mini-
mum monthly repayment is
£ 5.00 and goods 20 times this

amount may be purchased.
Further details from Marshall's

branches after" October 1 2th 1979.

Marshall's have reduced prices on
their top line products, resulting
in very competitive prices.

Twin reply paid order forms are
supplied in each catalogue to

The catalogue costs 50 p from
any Marshall's branch or 65 p
post paid from their head office.
Marshall's I Head Office),

Kingsgate House,

Kingsgate Place,

London NW6 4TA.

Tel.: 01-6240805.

(1294 M)

Four-colour plotter

An easy-to-use 11 by 17 in. (A3
size) microprocessor-controlled
plotter that produces low-cost,

high-quality multicolour graphic

plots with data sent from virtually

any computer or controller has
been introduced by Hewlett-
Packard. Interface is via RS-
232/V24 asynchronous serial

ASCII at any of eight switch

2400. Special design features
also enable the plotter to be
coupled into an existing com-
puter terminal RS-232/V24 inter-
face.

The new 7220A four-colour
plotter generates character sets,
dashed lines, and implements
scaling and other high-level func-
tions internally. Plotting colours
are selected and changed under
program control to produce high
resolution graphic plots and over-
head transparencies. There are
seven colours available for clear
film plotting.

Character plotting speed of over
two characters per second allows
fully annotated graphs to be
produced in minutes. A buffer
with over 1100 bytes has been
incorporated to store incoming
graphic plot data so that 1/0

cations between computer and

plotter are minimised. As an

option, an additional 2048 bytes
of buffer storage can be made

The plotter's built-in language
contains two categories of in-
structions: device control and
graphic instructions. The graphic

45 two-letter mnemonic instruc-
tions from the Hewlett-Packard
Graphics Language (HP-GL) which
equips the plotter with such

capabilities as relative and absol-
ute plotting, point digitising,
labeling, character sizing, integer

scaling and window plotting. No
specialised programming experi-

ence is needed to use the new

For applications requiring unat-

of the new plotter features
automatic page advance, an inter-
nal paper supply and paper cutter,
and a detached paper tray to
collect full- or half-page plots.
Hewlett-Packard Ltd,

King Street Lane,

Winnersh,

Wokingham,

Berkshire RGI1 5AR

(1286 M)

Single-knob measuring

bridge from Siemens

Siemens Ltd is marketing a single-

operated by one hand - a feature

that greatly facilitates simul-

taneous note-taking. The bridge
is available in two types: model

M273-A1 comprising a Kelvin

double bridge for measuring low

resistance values (200 p-ohms to

2200 milli-ohms) and model

M273-A2 with a Wheatstone
bridge for medium resistance
values. The low resistance model

has built-in measuring lead com-

The bridge is balanced by an
easy-to-read galvanometer read
against a deviation scale. Both
types are accurate to +/— 1% of
the measured value (+/— 1.5%
of the lowest range of model
M273-A1 ) and can withstand

a 2 kV voltage surge. The power

source is two standard IEC R14

1.5 V cells. Alternatively, 2 or

can be used which, generally,
increase the reading accuracy.
Both models are in a rugged
moulded plastic housing. A leather
case is available (extra) for
heavy-duty use. Typical dimen-
sions are 112 mrn (4.4in) wide
X 84 mm (3.3in) deep x 192 mm
(7.5in) long. The weight is 1 .1 kg
(2.41b).

Siemens Limited,

Siemens House, Windmill Road,
SUN BUR Y-on- THAMES.
Middlesex. TW16 7HS.

Tel: 109327185691.

High quality transient
signal processors

Bryans Southern Instruments

just announced the introduction
of a new range of high-sampling-
frequency, high-resolution tran-
sient signal processors. Priced to
suit modest laboratory budgets.

Series 523A.

They offer a 10-bit resolution
(i.e. 1 part in 1024), at the high

sampling rate of 10 MHz, so that

the shortest sampling interval is

100 nanoseconds.

When switched to dual timebase
mode, the first part of the re-
cording is recorded at timebase
rate A, and the second part at
rate B. The point at which the
change is made, is set-up by the

tiples of sweep time, with a resol-
ution of 0,05 from 0 to 0.95 of '
the sweep time.

These dual timebase facilities can
be used even when pretriggering
is in operation -a unique feature
for this type of instrument. There- |
fore, one can reference a short
event of just a few microseconds
during a trigger condition, many
seconds in advance of a fast signal
and still record the pre-trigger ■
information. At the same time,
the 523A maintains an exception- i
nally high resolution of the time
and amplitude of the signal.

These transient signal processors
are available in single- or dual-

channel versions, each with 4096-
word memories. They can be

interfaced to many peripherals
and computers by RS232C,
IEEE488/1975 (G.P.I.B.), or the
general purpose interface board.
If required, up to seven instru-

digital output purposes.

Bryans Southern Instruments Ltd.,
Willow Lane,

Mitcham. Surrey, CR4 4UL.

Tel.: 016403490

mlnp

New, miniature, low-cost
temperature recording
spots

The temperature responsive tri-
angle turns irreversible black from
original white after having been
exposed to its rated temperature
for fractions of a second. Such
single temperature spots, or mul-
tiple temperature sequenced strips

and inductive loads.

The amplifier front-end incor-
porates a *iA 741 operational
amplifier with additional voltage
and current gain stages so enabling

required for servo systems. The
output is protected from voltage
transients caused by inductive
surges. Output current limiting is
selected by placing appropriate
resistors between the supply pins
and the respective current limit
pins. The case is electrically

Absolute maximum ratings in-
clude an internal DC power dissi-
pation of 70 W with a case tem-
perature of 25°C, input voltage
differential 30 V end DC output
current of 10 A.

Camera & Instrument (UK I Ltd..
230 High Street, Potters Bar,
Herts EN6 5BU.

Telephone: 10707) 51111

(1231 M)

Switching regulator
power supplies

Now available from Amplicon
Electronics Limited are four ad-
ditional models to their existing
range of switching power supplies.
These new models, designated
RT153, RT154, RT303 and

RT304. are designed to meet
increasing micro processor appli-

They employ isolated auxilliary
outputs of 5 V and 12 V or 5 V
and 15 V in addition to the main
5 V output.

RT153- 5 V @ 30 amps
12 V @ 5 amps
5 V @ 2 amps
RT154 — 5 V @ 30 amps
1 5 V @ 4 amps
5 V @ 2 amps
RT303 - 5 V @60 amps
12 V@ 5 amps

cutting down the amount
movement necessary on boar
person to person exchanges.

, The system is designed for con-
> tinuous operation and can be left
, on in the standby position for
monitoring from lookout posi-
tions in bad visibility.

The Sea Com control, or master
unit consists of a heavy duty
watertight aluminium case coated
in Rilsen nylon and the system is
fitted with high output loud-
speakers, hand microphone, vol-
ume control, speaker selector
switch and tone alert button.

ing 32.5cm wide x 18cm high x
46.5 cm deep, and weighs 10 kg,
and the carrying handle also
functions as a fully adjustable

The instrument has two input
channels, Y1 and Y2, which
provide maximum sensitivities of
2 mV/cm over the full 60 MHz
bandwidth, and a special control
circuit is incorporated to nullify

The wide range of operating modes
available on the OS 3500 includes
comprehensive delayed timebase
and triggering facilities. Vernier
control of sweep delay timeallows
accurate timing measurements to
be made, and the delayed timebase

base sweep or triggered after a
preset sweep delay. The main and

completely separate.

For the study of complex wave-
forms, an alternate timebase

which allows the main timebase

can display and be triggered from

input channels remain free to
study other important infor-

Another feature which simplifies
the triggering of complex wave-
forms is variable trigger hold-off.
which can be continuously ad-
justed up to approximately one
sweep length of the main timebase

Gould Instruments Division.
Roebuck Roed, Heineult.

Essex 1G63UE.

Telephone: 01-500 1000

(1221 M)

market

liiiL'iiJol pAirij

a separate floating input, as well
as increased voltage and time
accuracies when switched to oper-
ate with the oscilloscope.

‘bright-up’ section of the oscillo-
scope’s main timebase sweep is
introduced and controlled from
the DM3010. The period between
the first and second bright-up sec-
tions is accurately displayed on
the instrument's light-emitting-
diode display. For amplitude

sweep of the channel 2 signal is
introduced, and the bottom of

with the top of the basic display
to provide an accurate digital

Operated as an independent
digital voltmeter with the separate
floating input, the DM3010
measures voltages from 200 mV
to 1000 V DC, with a resolution
of 100 mV and an accuracy of
t 0.1 5% of reading : one digit.

measured. The combined accu-
racy of the DM3010 and OS3500
is : 1% of reading j two digits for
time and i 2% of reading * two
digits for amplitude measure-
ments up to 5 MHz. Above 5 MHz,
the accuracy is conditioned by
the vertical amplifier roll-off to

-3 dB at 60 MHz.

The additional accuracy offered
by the DM3010 is of particular
use in applications such as the
measurement of digital-circuit
time relationships, including
memory timing and propagation

Gould Instruments Division.
Roebuck Road. Heineult,

Essex 1G63UE.

Telephone: 01-500 1000 ^222 I

Model 771 ASCII Keyboard is 1
especially suited for use with the
latest inexpensive video terminal I

The combination of the 771 key- j
board subsystem, and a video |
terminal board mounted in the |
user’s mainframe, provides an j
attractive and versatile cost-saving
alternative to conventional one- 1
piece CRT terminals. Compact. .
reliable, and rugged, the 771 Key-
board is ideal for use in small

software development applications
or educational microprocessor I

Standard features include full |
ASCII alphanumeric section; con- |
venient cursor control and nu- 1
meric pad; two-key rollover for
low error rate; Upper & lower case
plus control codes; TTY mode for I
upper-case only operation; Timed
autorepeat on all keys; all modes
standard parallel interface; detach-

non-glare keycaps; robust steel
desktop enclosure.

The 771 Keyboard is supplied
fully assembled and tested, with
complete documentation, requir-
ing only power and data connec- I
tions to the user's system for I
operation. Supplied mounted in I
an all-steel desktop enclosure, I
finished in textured IBM blueanjl
black, this Keyboard is a perfect'
complement to modern micro- 1
processor hardware.

The KB771 is priced at a modest i
£ 95.00, with discounts for quan- 1
tity.

Electronic Brokers Limited,

49/53 Pancras Road.

London NWI 2QB.

Telephone No: 01-837-7781.

(1216 M)

UK 21 - elektor October 1979

advertisement

Science of
Cambridge Ltd

Now, the complete MK 14 micro-computer
system from Science of Cambridge