up-to-date electronics for lab and leisure 11-02 -elektori tber 1976 elektor decoder What is a TUN? What is 10 n? What is the EPS service? What is the TQ service? What is a missing link? ir types Semii Very equivalent semiconductors exist with different type numbers. For this reason, 'abbreviated' type numbers are used in Elektor wherever possible- - ’741' stands for pA741, LM741, MC741, MIC741, RM741, SN72741, etc. - 'TUP' or 'TUN' (Transistor, Universal, PNP or NPN respectively) stands for any low frequency silicon transistor that meets the specifications listed in Table 1 . Some examples are listed below. - 'DUS' or DUG' (Diode, Universal, Silicon or Germanium respectively) stands for any diode that meets the specifications listed in Table 2. - BC107B', BC237B'. 'BC547B' all refer to the same "family" of almost general, any other member of the same family can be used instead. (See below.) For further information, see ’TUP, TUN, DUG, DUS', Elektor 17. p.948. Some TUN's are: BC107, BC108 and BC109 families. 2N3856A, 2N3859, 2N3860, 2N3904, 2N3947, 2N4124. Some TUP'S are: BC177 and BC1 78 families; BC1 79 family with the possible exception of BC159 and BC179; 2N2412, 2N3251, 2N3906, 2N4126, 2N4291 . Table 2. Minimum specifications for DUS (silii and DUG (germanium). DUS DUG VR,max 'F,max iR.max P,ot.max c D,max 25V 100mA IpA 250mW 5pF 20V 35mA lOOpA 250mW lOpF Some DUS s are: BA127, BA217, BA218, BA221, BA222, BA317, BA318, BAX 1 3, BAY61 , 1N914, 1N4148. Some DUG s are: OA85, OA91,OA95, AA116. BC107 (-8, -9) families: BC107 (-8, -9), BC147 (-8, -9), BC207 (-8, -9), BC237 (-8. -9), BC31 7 (-8, -9). BC347 (-8. -9), BC547 (-8, -9), BC171 (-2. 3). BC182 (-3. -4). BC382 (-3, -4). BC437 (-8.-9), BC414 BC177 (-8. -9) families: BC177 (-8,-9), BC157 (-8. -9). BC204 (-5. -6). BC307 (-8. -9) BC320 (-1, -2), BC350 (-1.-2). BC557 (-8. -91. BC251 (-2. -3). BC212I-3. -4). BC512I-3. -4). BC261 (-2. -3). BC416. Resistor and capacitor values When giving component values, decimal points and large numbers of zeros are avoided wherever possible. The decimal point is usually replaced by one of the following international abbreviations: 10 ' (nano ) = 10 9 H (micro-)- 10“ m (mill,-) = 10 ’ k (kilo) = 10 J M (mega-) = 10 6 G (giga-) = 10’ A few examples: Resistance value 2k7: this 2.7 kn, or 2700 n. Resistance value 470: this 470 n. Capacitance value 4p7: the 4.7 pF, or 0.000000000004 7 F . . itance value 10 n: th riting 10.000 pF or .01 pF. since 1 n is 10‘ 9 farads or 1000 pF. Mains voltages No mams (power line) voltages are listed in Elektor circuits. It is assumed that our readers know what voltage is standard in their part of the world! Readers in countries that use 60 Hz should note that Elektor circuits are designed for 50 Hz operation. This will not normally be a problem; however, in cases where the lains frequency is used for iynchro ification may be required. Technical services to readers - EPS service. Many Elektor icles include a lay-out for a printed circuit board. Some - but not all - of these boards are available ready-etched and predrilled. The 'EPS print issue always gives a complete list of available boards. - Technical queries. Members of the technical staff are available to answer technical queries (relating to articles published in Elektor) by telephone on Mondays from 14.00 to 16.30. Letters with technical queries should be addressed to: Dept. TQ. Please enclose a stamped, self addressed envelope; readers outside U.K. please enclose an IRC instead of stamps. - Missing link. Any important modifications to, additions to. improvements on or circuits are generally listed under the heading 'Missing opportunity. I* BLEHTOr Volume 2 Number 11 Editor Deputy editor Technical editors Art editor Subscriptions W. van der Horst P. Holmes J. Barendrecht G.H.K. Dam E. Krempelsauer G.H. Nachbar Fr. Scheel K. S.M. Walraven C. Sinke Mrs. A. van Meyel UK editorial offices, administration and advertising: 6 Stour Street, Canterbury CT1 2XZ. Tel. Canterbury (0227) - 54430. Telex: 965504. Bank: Midland Bank Ltd Canterbury A/C no. 11014587 Sorting code 40-16-11. giro. no. 3154254. Assistant Manager and Advertising : R.G. Knapp Elektor is published monthly on the third Friday of each month, price 40 pence. Please note that number 15/16 (July/August) is a double issue, 'Summer Circuits', price 80 pence. Single copies (including back issues) are available by post from our Canterbury office to UK addresses and to all countries by surface mail at £ 0.55. Single copies by air mail to all countries are £ 0.90. Subscriptions for 1976 (January to December inclusive): to UK addresses and to all countries by surface mail: £ 6.25. to all countries by air mail £ 11. — . All prices include p & p. Change of address. Please allow at least six weeks for change of address. Include your old address, enclosing, if possible, an address label from a recent issue. Letters should be addressed to the department concerned TQ = Technical Queries; ADV = Advertisements; SUB = Subscriptions; ADM = Administration; ED = Editorial (articles submitted for publication etc.), EPS = Elektor printed circuit board service. For technical queries, please enclose a stamped, addressed envelope. The circuits published are for domestic use only. The submission of designs or articles to Elektor implies permission to the publishers to alter and translate the text and design, and to use the contents in other Elektor publications and activities. The publishers cannot guarantee to return any material submitted to them. All drawings, photographs, printed circuit boards and articles published in Elektor are copyright and may not be reproduced or imitated in whole or part without prior written permission of the publishers. Patent protection may exist in respect of ci components etc. described in this magazine. The publishers do not accept responsibility for failing tc Natio inal advertising rates for the English edition of Elektor and/or international advertising rates for advertising at the same time in the English, Dutch and German issues are available on request. London N7 7AX. Copyright © 1976 Elektor f Printed in the Netherlands. elektor november 1976 — 11-03 contents selektor albar Albar is an ultrasonic movement detector, as both transmitter and receiver. The circu worth the cost. It can also be used as a elaborate alarm equipment. sensitive metal detector An unusual design approach was chosen fo structor should be able to build a highly s Doppler effect. It uses only one ultrasonic transducer as the basis of a burglar alarm, which could prove well of skill. The article also contains proposals for more little care, the average I be suitable for a wide egg-timer-with-a-difference - J. Schmitz Relatively few components are used in this eggtimer. The 'boiling time' is automatically determined according to the size of the egg and the degree of hardness required. DEW line The Domestic Early Warning System (DEW line) will monitor various vital points throughout the house and give timely warning if anything goes wrong. pocketronics calendar — R. ter Mijtelen •••••. A circuit for producing a calendar display using the pulses from a digital clock has already t Elektor 8. This design provides an alternative circuit ,n display facility to draw attention to dates of particular intercom A domestic intercom doesn't lequire a complicated components, even though an automatic gam contiol hi signal horn Although designed as a warning hom for this issue, the Signal horn may of course lay is small but the sound produced is loui voltage-to time converter . . . minivolt - A.M. Bosschaert . . In tins age of digital things (>) it is nice i outs This handsome little digital voltmete cricket - J. Schmitz Summer is long gone, and all of nature is proceeding through us autumnal eventide towards w Birds are departing, and the casual wanderer will notice the lack of animal and insect sounds A sad state ol affairs However. electronics is coming to the rescue in this article, we proudly present a cricket witn film slave - R.L.A. Trost The film slave is a sound activated switch originally intended or no modification it may also be used for a variety of other recorder in response to sound pulse generator E. van Dee and B. Alberts A pulse generator is an extremely useful aid when testing and here is uncomplicated and is based on TTL IC's. but the facilitie missing link coming soon refrigerator alarm This unit will sound the alarm if the door of a refrigerator is required for this effective food and-energy saver it can be usei •Domestic alarm' described elsewhere in this issue capacicoupling In this article an electrostatic system is described which can be used as part o reading in bed limiter At a certain age. children are often packed off to bed with the final admonition for a quarter of an hour, but then you must turn off the light and go to Sleep know, the children tend to suddenly loose all sense of time in this situation When a member of the Elektor design team was faced with th.s problem, he s solution. The final circuit, as published here, has pioved extremely effective a LOGICal replacement for the carbon-track potentiometer market handful described elsewhere in use with the 'Albar' ultrasonic alarm system ue used with any system requiring an audible v I and extremely penetrating. of CMOS! I a cine camera. However, such as automatic startini left open too long. Very j either as a selfcontained components are electronic tjCawTiir 'lecieso », ,la >*st /] elektor BLBHTB f This is a selection from the contents of the various issues: number 9 — feedback pll for fm — function generator — racing car control — simple mw receiver — digital master oscillator — pll-ic stereo decoder number 10 — call sign generator — morse decoder — morse typewriter — digital wrist watch number 1: - tup-tun-dug-dus - equa amplifier - mos clock - distortion meter - tap sensor - electronic loudspeaker - steam whistle number 2: — minidrum number 1 1 modulation systems how to gyrate number 3: — tap preamp — pll systems — fido — compressor — disc preamp — a/d converter — led displays number 12 rhythm ger number 13 — equin (2) — digisplay — versatile logic number 14: thief suppression ir supplies for cars cybernetic beetle the moth quadro in practice led light show digibell number 5: ‘Summer Circuits' issue, with over 100 circuits: amplifiers, generators, dividers, universal frequency reference, im- proved 7-segment display, receivers, power supplies, rhythm generators, measuring equipment, etc. number 6 — edwin amplifier — versitale digital clock — phasing number 15/16: number 17: - m.p.g. indicator - car service meter - windscreen wiper - rev counter number 18: numbers 7 and 8 sold out 1-10 — elektor nove ir 1976 Albar is an ultrasonic movement detector, based on the Doppler effect. It uses only one ultra- sonic transducer as both transmitter and receiver. The circuit can be used as the basis of a burglar alarm, which could prove well worth the cost. It can also be used as a modern game of skill. The article also contains proposals for more elaborate alarm equipment. Alarm equipment based on ultrasonics has several advantages over other types. Typically, it is quick and simple to install, and its sensitivity can be limited as required. However, it is not the purpose of this article to conduct a survey of available alarms, merely to provide a design for a particular one. When discussing ultrasonic movement detectors, it is necessary to distinguish between two basic principles of oper- ation, vis: 1. circuits which detect the interruption of an ultrasonic beam, and 2. circuits which detect the movement of an object within a given area. Method ( I ) is shown in block diagram form in figure la. The area that can be protected by this sort of device is very limited, so this method is usually employed in situ- ations where the securing of one door or window protects a larger area. A block diagram of a movement detec- tor is shown in figure lb. In this case, it is not necessary for the transmitter and the receiver to be opposite each other (i.e. on the same "acoustic’ axis), since the operation is based on the Doppler effect. This effect occurs when the sound source and the receiver are not stationary, but move either towards or away from each other. The most fam- iliar example of this is to stand beside a railway line and listen to a train as it passes. The sound of the approaching | train appears to the listener to be at a higher frequency than the sound of it receding. The effect can be defined as ^ a frequency shift in the signal, if the albar I source and receiver are moving relative | to one another. The effect can be useful I in the field of ultrasonics. The receiver (transducer) is the target | tor a fixed frequency signal which j originates at the transmitter and is | reflected by stationary objects. The sig- I nal reflected from a moving object will I have a different frequency, due to the I Doppler effect, and this signal will also be picked up by the receiver. | As a result of interference between the two signals, an amplitude fluctuation of the transmitted signal (i.e. ‘beating’) occurs. This fluctuation is detected and is used, eventually, to operate the alarm. The effective range of such a movement detector can be considerable, although this is not always desirable. If the sensi- tivity is too high, then flying and even crawling insects. Hexing doors and j windows (in windy weather), and even air circulation can set off the alarm. If the movement detector is equipped with | a sensitivity adjustment, then the ‘oper- ational range' can be matched not only i to the space conditions but also to the object to be detected. The circuit is then universally applicable. J To achieve a useful range, the two trans- ducers (transmitter and receiver) must be placed at a considerable distance I Irom one another. Consequently, ampli- fication of the signal by the receiver | must be very high, before it is useful to i the delector stage. A suitable receiver . circuit is shown in figure 2. The trans- mitter is shown in figure 3. Both parts of the circuit can be supplied by the j common power supply given in fig- j ure 4. The purpose of the nickel- cadmium accumulators and the calcu- lation for the resistor R x are given later. The transducer converts the ultrasonic signal into an electrical signal, which is then fed to a two-stage amplifier with Tl and T2 (see figure 2). The detector stage, based round T3, is followed by anothei two-stage amplifier (T4andT5). The signal at the collector of T5 is used to drive the multivibrator consisting of T6 and T7. This in turn drives the final amplifier stage to which the alarm is connected. The output of T9 is rectified by the IN4I48 diodes and is fed back to the base of T5 to bias it on and thus latch on the alarm. The alarm may be reset by switching off the power supply [ and then switching it on again. The circuits requires a fair number of I components and, unfortunately, it is l not possible to simplify it very much by using integrated circuits. This is due to i the fact that the signal level as far as T4 is so low that conventional operational amplifiers are unsuitable. The worst problem is the threshold detector. A simplification is. however, possible if the principles of utilising the Doppler effect are reconsidered. The ultrasonic signal from the transmitter and the fre- quency-shifted signal give rise to inter- ference. This results in an amplitude fluctuation, the frequency of which is the difference between the frequencies of the two signals: the ‘beat frequency’. Interference will also occur if the ultra- Figure 1. Two ways of using ultrasonic equip- ment for alarm systems, la) shows the beam interruption principle and lb) shows the movement detector. The latter has much wider applications. Figure 2. Circuit diagram of the receiver for an ultrasonic movement detector. Figure 3. Circuit diagram of the ultrasonic sonic transducer which transmits, is con- nected so that it can be the receiver as well. Such an arrangement is shown in block diagram form in figure 5 and in detail in figure 6. The performance of this circuit is satisfactory as long as the difference is small between the transmit- ted frequency and that reflected from moving objects. The advantages of this arrangement are that only one transducer is required, and the preamplifier can be omitted, so the component cost is reduced. Block diagram The transducer US produces an ultra- 1-12 - elektor november 1976 Figure 4. Power supply for the receiver and transmitter of figures 2 and 3. Figure 5. Block diagram of Albar. Figure 6. Complete circuit diagram for the movement detector, Albar. Note the saving of components over the circuits of figures 2 and Figure 7. To adjust the oscillator amplitude, connect a universal meter as shown here. Figure 8. A complete alarm system with four detectors, as an illustration. This hybrid diagram (both block and wiring) is presented to indicate clearly the interconnections be- tween the various circuits, including the Signal Horn. The mains unit can supply about ten Albar circuits and in most cases the Signal Horn can be powered as well as the alarm sonic signal which hits a moving object (figure 5). Some of the reflected signal, which has a different frequency from the original, arrives back at the trans- ducer. Due to interference between the two signals, a signal which varies in amplitude occurs at the output of the receiver. This is then fed to the envelope detector; a lowpass filter removes the higher frequency components. The resulting LF signal is amplified, with the amount of amplification adjustable to match the required application. The amplified signal is then rectified and fed to a flip-flop which sets the alarm. The circuit The first stage (Tl and T2) in figure 6 acts as both transmitter and receiver, with the interference signal appearing at the emitter of T2. This stage is basically an emitter-coupled multivibrator. The ultrasonic transducer is situated be- tween the emitters of the two transis- tors where it serves to set the frequency of the multivibrator. R5 and C3 are selected to force the transducer to its resonant frequency, so that harmonics, which are often present with this type of transducer, are eliminated. This filter should prove adequate if the Murala transducers (specified in the parts list) arc used. However, with some other transducers it has proved necessary to use a 2.2 mH coil in place of R5 (e.g. Toko type EL 08 10-222k). The transistor T1 is an amplifier stage I for the reflected signal and T2 is an I emitter follower. Although it might theoretically be possible to equip this stage with only one transistor, practical I difficulties occur particularly in relation I to the higher frequency resonance I points of the ceramic transducer. T3 is a detector stage in which the I transistor has no bias voltage or current preset at its base, i.e. it is class C. It is only the positive half wave which is I amplified, provided the signal amplitude I is sufficient to raise the transistor above its threshold. The voltage which appears | across Cl 1 is dependent on the control j signal voltage, so that modulation of the ! control signal (moving object) appears elektor november 1976 — 11-13 as a voltage change across C 1 1 . T he LF signal is amplified by IC1. The amplification is adjustable with P2. To eliminate any traces of the carrier still present after the first low pass filter (Cl I ). the feedback to the opamp is fre- quency dependent (CIO). The next stage is a voltage doubler consisting of two diodes, D2 and D7, and two capaci- tors C8 and C9. The final stage is a discrete component flip-flop, T4 and T5. It requires a con- trol signal amplitude of about 0.7 V. An alarm can be directly connected be- tween points A and B; alternatively, a relay which controls the alarm can be connected instead. The current con- sumption of the load should not exceed 20 mA. Putting the circuit into operation Connect the supply voltage and then measure the oscillator voltage with a suitable AC voltmeter, as shown in SIGNAL HORN POWER SUPPLY 11-14 -elektori iber 1976 figure 7. Adjustments are carried out as follows: 1. adjust PI to the maximum reading; 2. re-adjust PI so that the reading drops to two-thirds of this maximum ; 3. adjust the overall sensitivity of the circuit with P2. The voltage values marked on the circuit apply only to the quiescent state, i.e. no alarm. Construction and application The circuit, together with the trans- ducer, should be housed in a metal casing. Three external connections are required: 1. supply voltage ; 2. reset to the circuit; 3. connection to the switch contact relay. The relay is not required if the ‘Signal Horn' circuit described elsewhere in this issue is used. The relay can then be replaced by a 4k7 resistor, and the diodes D3 and D4 are not required. Point B is then connected to the Signal Horn circuit (screened leads are not necessary). As an example, figure 8 shows an instal- lation consisting of four detector cir- cuits, a Signal Horn and an alarm relay. A system of multiple movement detec- tors can be arranged in two ways: 1 . every entrance to an area is protected by a separate detector circuit; 2. each room which should not be entered by an intruder is protected by a circuit. The second method is usually the cheaper, especially where none of the rooms in question are larger than the average-sized living room. When the supply voltage is switched on, the alarm operates immediately, which at least enables you to know that it is working. If the noise is not required at switch-on, then the reset should be operated at the same time. The light- emitting diodes (D1 to D4) give a visual indication of the state of the alarm circuits. It is usual to hide such switches as the on/off and the resets, but this is a matter of choice and may not always be convenient. For instance, it may be useful where various rooms in a house are protected, for the controls to be situated in the bedroom. In the case of larger installations, it may be necessary to be able to put some of the units out of operation temporarily. In the case of Albar, this is done by breaking the supply lead at the points shown by crosses in figure 8. The mains supply unit shown in figure 8 is given in detail in figure 1 1. It supplies 33 V for the Signal Horn. To enable the alarm to function even during a power cut (either accidental or deliberate) accumulators are included as stand-by power supply. Whenever the supply voltage from the mains is on, the accu- mulator draws current equal to its leakage current. The internal discharge is approximately one hundredth of the value of the accumulator capacity (in ampere hours). elektc 1976 - 1M5 For the calculation of the value R x , the 1 following formula therefore applies: ^ 1.5 x Vjr- VNiCad R * * 0.01 x Uh where Vxr = secondary voltage of the mains transformer; VNiCad = nominal voltage of the accu- mulator; 1 Ah = accumulator capacity in amp- hours. In the event of a mains failure, the cir- cuit receives it supply voltage via D2. | Under these conditions, the noise out- , put of the signal horn is perceptibly, but not drastically, reduced. If the Signal Horn is not switched on, then the current, consumption is less than 15 mA. | A small accumulator with a capacity of 100 mAh should be sufficient for most applications. Note that it is highly recommended to inform the local police how to put the alarm out of action. Use of the circuit as a game of skill The circuit detects movements. It can therefore be used as a game of skill by attempting to ‘beat’ the detector by moving very slowly. The winner must exercise good muscle control in order to move sufficiently slowly. An element of luck is involved, however, since involun- tary movements and reflexes will activate the alarm if the sensitivity is high enough. Figure 9. Printed circuit board for Albar (EPS 9428). Figure 10. Printed circuit board for the mains unit of figure 11. It is capable of supplying the Signal Horn in addition to the detector circuits (EPS 9437). { Figure 1 1 . Mains unit for the Albar(s) and the Signal Horn (which requires the 33 V supply). Photo A. Some ultrasonic transducers. The Murata types (group 2 on the photo) are highly recommended. Other types (like the ’ first one on the photo) may not be suitable in this circuit. Several readers have asked us to design a good metal detector. When our 'research and development' group took a closer look at the problem, they soon discovered why: most conventional designs are not really suitable for the home constructor! In most designs, extreme care must be taken to screen the various sections of the circuit from one another and the power supply must be carefully filtered. A different design approach was chosen for the circuit described here. With a little care, the average home constructor should be able to build a highly sensitive metal detector, which will be suitable for a wide variety of uses. The normal way to construct a metal detector is to have two oscillators work- ing on the same frequency. One of these is crystal controlled, and the other is connected to the pick-up coil (figure 1 ). The output of the two oscillators is fed to a mixer stage. When both oscillators are working at exactly the same fre- quency, the difference frequency is zero. However, when a metal object is brought into the vicinity of the pick-up coil the frequency of the VFO will change. This causes a ‘beat’ frequency to appear at the output of the mixer. This is filtered and amplified, and fed to a headphone. So far so good, or so one would think. The difficulty with this system is that the slightest amount of coupling be- tween the two oscillators will ‘pull’ the VFO into lock with the crystal oscillator. When this is the case, quite a hefty lump of metal is required to get an audible indication. A further problem is that the coil determines the frequency of the VFO. It has to be wound exactly according to specification, or else the tuning range of the VFO will not include the frequency of the crystal oscillator. Changing coils can be a major problem. Both of these problems can be resolved to a very large extent by resorting to a ; different design approach. This is shown in the block diagram (figure 2) and the practical circuit (figure 3). In the Elektor design the crystal oscil- lator has been replaced by a less expens- | ive ceramic filter oscillator. The oscillation frequency is determined by ! a 455 kHz filter in the feedback path of Tl. Granted, this oscillator will not be as stable as a crystal controlled unit ; but it is sufficiently stable for the purpose. The output from this oscil- lator is passed to a divide-by-10 stage (IC1 ), which will now deliver a 45.5 kHz square wave. This output is passed through two pulse | shapers in cascade (C4, R6, N1 and C5, R7, N2). This produces very short pulses with a repetition frequency of 45.5 kHz. This signal has a spectrum which extends well into the megahertz ! range, with ‘spikes’ every 45.5 kHz. Each spike can be used to produce a beat frequency with the VFO. This signal is passed to the mixer stage (T2). | The main components of the VFO are T4, Cl 7, Cl 8 and the frequency deter- mining components Cj and Ls, the search coil. It oscillates at a very high frequency. As an example, if the search coil is made of three turns on a 3 inches ' diameter form, the frequency can be adjusted to approximately 3.5 MHz. This signal is fed to the mixer, through a buffer stage (T3). The output from the mixer stage is passed through a low-pass filter and an . amplifier to the headphones. The active devices used in these stages are gates N3 | and N4, biassed to work as ‘linear’ amplifiers. Since the ‘fixed frequency reference’ I produces a broad spectrum, it is not difficult to find a ‘zero beat’ setting of I the VFO. This has the added advantage | that changing coils becomes a simple | matter: there are so many spikes in the reference spectrum that almost any coil will produce a beat frequency some- k where within the tuning range. In prac- ; tice, there will be ‘loud’ and ‘soft’ zero I beats; obviously, it is advisable to select a loud and obvious one. The coil diameter will depend on the I Figure 1. Block diagram of a conventional metal detector. The VFO is set at the same frequency as the crystal oscillator. Figure 2. Block diagram of the metal detector described here. The output from the crystal oscillator is divided down to a lower fre- quency and then fed to a pulse shaper, producing a very broad spectrum of reference elektor november 1976 — 11-17 purpose for which the detector is to be used: a large coil is useful for approxi- mate location of relatively large objects, whereas a small coil can be used to pin- point the exact position of even a small object. The coil former can be plastic tube, up to the maximum tube diameter available (8 to 10 inches). For larger coils it becomes more difficult - but who says a coil must be round? A square section is also quite permissible. The best material to use for the coil is screened cable or coax, with a closely- wound screening braid. The core is used for the coil, and the screen is connected at one end to supply common. Remem- ber, the coil is supposed to respond to a magnetic field change — it is not sup- posed to pick up short-wave trans- missions! The value of the tuning capacitor (Cj) is not critical. Any trimmer that will fit on the board and has a value between 100 p and 350 p should be suitable. The construction of the whole circuit, including the coil, must be mechanically rugged. This is one of the reasons for using a p.c. board. Using the detector Several ‘field tests’ have shown that this metal detector is quite sensitive. For instance, a very small piece of metal was buried at a depth of over 6 inches. When the metal detector was used to search in the general area indicated, a large number of objects were located - including several orange-coloured (ferrous? ) stones ... In another test, the position of an old brass farthing was accurately pin-pointed under a two- inch thick pile of papers. The metal detector has also proved useful for locating electrical wiring con- duits and the like in the walls of a house. Results will depend, of course, on the skill of the operator. In general, the following rules should be observed: 1. Start with a relatively large search coil. Hold it close to the ground (or wall) and tune in to a loud and obvious zero-beat. 2. Search the area in slow parallel sweeps (like mowing the lawn on a small scale). If you have ever seen it done on TV you will know exactly OH ho s ensitive metal detector Figure 3. Circuit diagram of the metal de- Figure 4. Printed circuit board and com- ponent layout for the circuit shown in figure 3. what is meant. 3. A change in tone indicates ‘some- thing’. This ‘something’ can be an object, but it can also be that the distance between the metal detector and the ground is not held constant. If several sweeps in the same area give the same indication, there is probably something there. A rela- tively sudden change in tone means that the object is close to the surface, whereas a gradual change in tone indicates an object at greater depth. 4. In the general area indicated by the last sweep, try a further search pat- tern at right-angles to the previous one. This will narrow the possible position down even further. 5. In some cases (specifically, if the ob- ject is small) it is advisable to switch to a smaller coil at this point. Repeat the search procedure described above within the general area already found. It should be possible to get a quite accurate indication of the position of the object. M Relatively few components are used in this eggtimer. The 'boiling time' is automatically determined according to the size of the egg and the degree of hardness required. Eggtimers range from miniature hour- glasses to sophisticated digital clocks. The more modem versions usually give an audible indication that the alotted time has elapsed. The timer described here differs from all conventional units in that there is no time scale on it. Instead, there are two scales that are used to set the timer; one is calibrated in egg sizes (from ‘canary’ through ‘hen’ to ‘ostrich’) and the other in degrees of hardness (from ‘runny’ through ‘medium’ to ‘bullet’). The timing interval is started by touching a TAP sensor; this is indicated by an LED lighting up. At the end of this period the unit produces a dis- tinctly audible tone, signalling that the egg has been in hot water long enough. After a short time this tone is switched off and the unit puts itself completely out of action: the current consumption drops to zero. There is no need for a switch. The circuit The complete circuit (figure 1) consists of two main parts: the timer proper (T1 . . . T4, with associated components) and the warning tone generator (T5 . . ,T7). The timing sequence is started as soon as the TAP sensor is touched. Sufficient current passes through the skin resist- ance to turn on T3, and this transistor drives T4 into saturation. The timing circuit is now under power, and LED D1 lights up. For a very short time, a charging current flows through Cl into the base of Tl. This drives Tl into saturation, charging C2 rapidly. When Cl is fully charged, T1 turns off and C2 starts to discharge through R3 and PI . Initially, T2 is turned off and T3 is conducting. As C2 discharges, the base voltage of T2 rises. After a certain time T2 will start to conduct, the volt- age across R4 increases and T3 is turned off. This happens as soon as the voltage at the base of T2 has become equal to that at the base of T3 (T2 and T3 form a so-called ‘long-tailed pair’). The latter voltage depends on the setting of P2 and P3. When T3 stops conducting, T4 is also turned off. The supply voltage to the timer proper now falls away (the collector voltage of T4), and this volt- age drop is passed via C4 to the oscil- lator circuit. T5 is turned on, enabling the oscillator proper (T6 and T7). This produces the tone which signals "egg ready’. C4 now starts to charge through the timer circuit on the one side and R9 on the other. After a short time, it will have charged so far that T5 is no longer held in conduction. When this transistor stops conducting, the oscillator is also turned off. The timer can be re-used immediately: Cl . . . C3 are already discharged suf- ficiently, and C4 will be discharged via D2 and T4 as soon as the TAP sensor is activated. Calibration Start with P2 and P3 in the mid- positions (‘B’ and ‘2’, respectively). Now try a ‘dry run’ — i.e. no water and no egg - to see how many minutes the 2 Figure 1. The eggtimer-with-a-difference. There are three timing adjustments: PI is preset for normal boiling of average-sized eggs; P2 is for setting the consistency re- quired ('runny' to 'bullet') and P3 is to compensate for differing egg sizes. Figure 2. Printed circuit board and com- ponent layout. timing interval is. If this interval is shorter than required for boiling a medium-sized egg to a medium degree of hardness, the resistance of PI should be increased (turn the slider clockwise). If the time is too long, PI should be turned the other way. After several dry runs, the correct setting of PI should have been found. It is left in this position, and P2 and P3 can be used to vary the timing according to taste (P2) and/or size of the egg (P3). Position ‘A’ of P2 corre- sponds to ‘hard as a bullet’ whereas position ‘C’ is for ‘runny’, position ‘1’ of P3 is for over-sized eggs, and pos- ition ‘3’ is for the miniaturised kind. * The washing machine starts to leak while you're watching TV .... Or the freezer fails and you don't notice it for a day or two .... Or the children turn on the taps in the bath and leave them running .... In all these cases, and many more, some form of Domestic Early Warning System (DEW line) can prevent more or less serious consequences. What is required is a system that will monitor various vital points throughout the house and give timely warning if anything goes wrong. A very simple alarm line is described here. In spite of its extreme simplicity, this alarm system will prove quite adequate in many cases. The alarm line proper consists of a sufficient length of three-core cable. Either three-core mains cable or two- core screened cable can be used. The alarm sensors and the receiving station (or stations) can be connected to this cable at any point along its length. The electronics required are very simple, because the system works on the prin- ciple of audible recognition. In other words, the sophisticated sound recog- nition system which is part of human hearing is used as part of the alarm system. To be more specific, the alarm sensors each produce a different warning tone if something goes wrong. They are all connected to the three-core cable. The receiving station is nothing more than a two-transistor audio amplifier. As soon as one of the alarm sensors registers the fact that something is wrong (the bath is running over, the deep-freeze is becoming too warm, etc.) it puts its own unique alarm tone on the line. This tone is reproduced by all the receiving stations. It is up to the listener to decide from the frequency and rep- etition rate which alarm sensor it is. In this article, alarm sensors for water leaks and for deep-freezers will be de- scribed. The refrigerator alarm described elsewhere in this issue can also be con- nected into the DEW line. In the 'Summer Circuits’ issue (Elektor 15/16) a ‘Bell extender’ was described, includ- ing a telephone bell pick-up. This, too, can be connected into the alarm system. Alternatively, any or all of these sensors can be connected into the original ‘Wireless bell extender’ circuit (Elektor 15/16, circuit 52), if one is not inclined to run three-core cable all over the house. Deep-Freeze alarm The first ‘domestic calamity sensor’ to be described is a unit that produces an intermittent warning ‘bleep’ if the tem- perature inside the deep-freeze becomes so high that the food is in danger of thawing out. The circuit is shown in figure 1. Only a single COSMOS integrated circuit and a good half-a-dozen other components are required. The four NAND gates in the IC are used in two multivibrators: N3/N4 produce an audible tone, and N1/N2 produce a very low frequency square wave. This low frequency oscil- lation is used to turn the audible tone oscillator (N3/N4) on and off, pro- ducing an intermittent ‘bleeping’. As long as the temperature of the NTC is sufficiently low, the first oscillator is blocked and the output of N2 is ‘low’. This blocks the second oscillator, so that no tone is produced. If, however, the temperature rises above a certain point, the resistance of the NTC drops sufficiently far to enable the first multi- vibrator, sounding the alarm. The tem- perature at which this happens can be set with PI. The repetition frequency of the alarm depends on the resistance of the NTC. As the temperature rises further, the repetition frequency increases. The more urgent the alarm sounds, the higher the temperature! Needless to say, the NTC must be mounted inside the freezer. It can be connected to the rest of the circuit via thin, insulated wires. Flood warning The circuit for the flood Warner (figure 2) is very similar to the previous one. As before, gates N1 and N2 produce a low frequency square wave which switches Figure 1. The deep-freeze alarm. The tem- perature sensor is an NTC, which must be mounted inside the deep-freeze. 11-2 2 — elektor november 1976 2 zFV^-q, ninir — J ” ^ " i 4011 ^ r the audio oscillator (N3/N4) on and off to produce the bleeping alarm signal. In this case, however, the oscillators are switched on by a ‘flood sensor’. This is just a nice-sounding name for two nails in a piece of wood. As soon as this becomes sufficiently wet there will be a conducting path between the nails. This turns on the two oscillators, sounding the ‘wet alert’. The frequency of the alarm tone itself and the repetition rate are both some- ' what higher than those in the previous circuit. It should not be at all difficult to distinguish the two different alarm sounds. If more than one flood sensor is re- quired (for instance, one in the bath- room and one on the floor near the washing machine), they can be ‘tuned’ to give different alarm signals. Cl sets the repetition rate and C2 sets the frequency. I If the wood used for the sensor is too green, it may be sufficiently moist to i trigger the alarm unnecessarily. There | are two solutions to this problem. One is to try a different piece of wood, or use bolts in a piece of plastic instead. The other is to reduce the value of R5 until the alarm signal stops. Telephone bell extender ,| This circuit has already been described in a previous article (‘Wireless bell extender’, Elektor 15/16). However, I since it can also be incorporated in the DEW line, the circuit is repeated here (figure 3). The sensor proper is an inductive pick- M up, which can be attached to the ! telephone by means of a suction cup. |' These units are available for use in | ‘loud-speaking’ telephones and the like, j A three-stage amplifier boosts the signal Figure 2. The flood warning circuit This is very similar to the deep-freeze alarm, but the sensor consists of two nails in a block of Figure 3. The telephone bell extender. Figure 4. The receiver is a very simple power amplifier circuit. L to a sufficient level (adjustable with the 2k2 preset potentiometer). The output (point B) can be connected into the DEW line. The receiver The receiver is nothing more than a two- stage amplifier and a loudspeaker (figure 4). No attempt has been made to achieve ‘Hi-Fi’ sound quality. The power dissipation in the output stage is fairly low, since the input signal is a square-wave. Even so, some form of cooling fin on T2 can’t do any harm . . . The type of transistor to use depends on the supply voltage. As shown in the table, the 2N2219 can only be used for . supply voltages of 6 V or less; the ! BC140 can be used up to slightly higher voltages and the BD135 is all right up to 12 V. Note that the 5 . . . 15 V specified for the supply is valid forTl and for all the sensor circuits, but not necessarily for T2! If it is the intention to use only one of the alarm sensors with one (or more) receivers, the output of the sensor can be connected direct to the input of the amplifiers). However, if more than one sensor is to be used, they must be inter- connected as shown in figure 5. This is the alarm line proper. elektor iber 1976 - 11-23 The DEW line The interconnection of all alarm sensors described and the receiver is shown in figure 5. To avoid clutlering this dia- gram, only the last gate (N4) is shown of the refrigerator alarm, deep-freeze alarm and flood warning and only Ihe last transistor (T3) of the telephone bell extender. As can be seen in this diagram, one core of the (three-core) alarm cable is used for the positive supply connection and one for supply common. The third core is the signal line. All the sensors have been designed to give a high-level output signal under normal conditions. This means that they can all be connected lo the alarm line through diodes, as shown - the ‘quiesc- ent’ sensors will not prevent an active alarm from sounding. The receiver is connected directly lo the line, as shown. As stated earlier, more than one receiver can be used; each i receiver is connected to the line at the point required. There is one point to » watch, however. If the ‘volume’ control in the receiver (the 100 S2 preset) is turned up to maximum, the peak cur- I rent consumption can be more than 1 Amp. For this reason, it is advisable to give each receiver its own power supply. R x should only be incorporated in one of the receivers. While on the subject of the power supply, it is important to note that electrical safety is even more important than usual in this case. The alarm line will be running all over the house, and some of the sensors will be in a moist environment (the bathroom, for in- stance). For this reason, the mains transformer must be absolutely reliable, and the wiring inside the supply must also be such that there is absolutely no Figure 5. The DEW line. Utmost care must be taken when constructing the power supply, so that there is absolutely no possibility of mains voltage appearing on the alarm cable. pocketronics Anybody who regularly looks inside modern factory-produced equipment, and particularly radio and TV front-ends, will often be as- tounded by the number of components per square inch of printed circuit board. On Elektor p.c. boards, we do try to keep the component layout fairly compact but some fairly large components arc used in most cir- cuits. Furthermore, resistors and capacitors arc all mounted flat on the board, since mounting them ‘on end' requires rather more skill and care. However, we felt that many of our readers might like to try their hand at mounting interesting little circuits in the smallest poss- ible space. The first idea was to start a series ‘electronics in a nutshell’, but this was soon dropped as definitely too ambitious. The next suggestion was ‘matchbox circuits’. Several circuits have indeed been developed that can be squeezed into a large-sized match- box. Examples are the egg-timer and cricket elsewhere in this issue. Even this title was found to hamper our stride, however: some interesting projects had to be housed in a cigarette carton. Fi- nally, we settled for the all-embracing title ‘pocketronics’. To be able to fit the units into such a cramped space, the boards have all been designed for chance of a short or even leakage be- tween the mains and the rest of the circuit. If Ihe sensors described here are to be connected into the ‘Wireless bell ex- tender’ circuit, they are interconnected with diodes as shown in figure 5. In- stead of being connected into the receiver circuit, however, they are con- nected to point ‘B’ in the transmitter. M truly miniature components. This may have the disadvantage that some of them arc not so readily available: - Resistors should be 1/8 watt types. - Depending on the value, capacitors are usually miniature ceramic, Siemens MKM, | tantalum electrolytic or miniature electro- lytic. Transistors and ICs are usually standard types, but extensive use is made of the Darlington transistors BC516 and BC517 (Texas Instruments). Batteries arc usually 1.2 V miniature mercury types, as used in hearing aids, photographic equipment and the like. The ‘loudspeaker’, if required, is actually a dynamic microphone capsule (Sennheiser type HM35). Other types may also be suitable; the requirements are: very small dimensions, relatively high impedance (more than 500 SI), and, preferably, relatively high efficiency. Standard potentiometers are not used. Where these would normally be needed, presets arc mounted on the board. Some mechanical ingenuity, of the bent-wire type, may be required to replace the ‘con- trol knob’. M 11-24 — elektor november 1976 calendar A circuit for producing a calendar display using the pulses from a digital clock has already been described in Elektor 8. This design provides an alternative circuit which gives day, date and month display, plus flashing display facility to draw attention to dates of particular interest. The calendar is advanced by the 24 hr (i.e. midnight) pulses from a digital clock. Each pulse steps the date and the day displays by one. To facilitate the initial setting of the display, a higher frequency signal must be provided. The most suitable and convenient is the seconds pulse train from the clock. Logic is included in the circuit to enable the month display to change on the appropriate date. The only manual intervention necessary is the setting of a switch to indicate ‘leap year’ when necessary. The date indication is by a pair of 7-segment units, while the day and month are displayed on illuminated win- dows in a suitably divided rectangular mask. A suggested lay-out for the front panel is shown in figure S. Day of the week display Figure 1 gives the circuit diagram for the day display. The input is normally from the once-per-day 24 hr clock pulse, but it may be switched to the once-per-second pulse, by SI, when required. In either case the input signal is fed, via a decoupling capacitor, to the > input of the counter IC2 (7493) where the negative-going edge triggers the count. The output of IC2, three bit BCD, is routed to the inputs of IC1 (7445). The count is automatically reset to zero when the value seven (BCD ‘1,1,1’) is reached. The values 0-6 of the count correspond i to the days Monday - Sunday respect- ively. The BCD-decimal decoder chip | IC1 takes the binary input and produces 1 0 outputs which can drive 80 mA loads. Setting the day display to the correct value is achieved by feeding seconds pulses to IC2 (via SI). The seconds pulses arc also used in this circuit to provide a flashing display to draw attention to dates of interest, i These one second pulses are gated by J the flash enable input (connection 6). I When they are passed by N1 to the i fourth (D) input pin of 1C 1 , the output | is periodically inhibited. The selected j lamp will then flash once a second. Date display The date display circuit is shown in I figure 2. It consists of two counters con- J structed from flip-flops. A decimal I counter is formed by FF1, FF2 (7476) I and FF3, FF4 (7473), while FF5, FF6 I (7473) form a divide-by-four counter. I The operation of the decimal counter is I best explained by considering the I events which occur when the tenth I input pulse arrives. This pulse must reset I the counter to zero. The tenth pulse I causes the Q outputs of both FF2 and I FF4 to be high, so the output of NAND j gate N33 goes to 0. This signal resets I FF1 (i.e. the 0 output goes to 0) and is I inverted before being used (via N2) to 1 reset FF2, FF3 and FF4. The falling J edge of the Q output of FF4 increases I the count of the divide-by-four circuit. I The units of the date display (Ml) are I supplied by IC3, a BCD-to-7 segment ] decoder (7447), which uses the output I of the decimal counter as its input. The I tens of the date (M2) are supplied by IC4 (another BCD-to-7 segment de- j coder) using the output of the divide- by-four counter. The monthly resetting of both counters : is achieved by gates N18-N22. After this reset, the date must be ‘one’ not ‘zero’, therefore the decimal counter states A,B,C,D must be at 1,0, 0,0 for the units, and the divide-by-four counter J must be 0,0 for the tens. The operation is as follows: in the normal state, the outputs of N 1 9-N22 are all high, so that the output of N18 is low. The outputs of N25 and N26 are therefore high and have no effect on FF1, FF5 and FF6. Similarly, the reset signal to FF2, FF3 and FF4 is high (provided no clear sig- nal comes from N33 as already de- scribed). If one of the inputs 28, 29, 30 or 31 goes to logic I, then the output of the corresponding NAND gate goes to 0. provided that the other inputs ° — 01 — ©0 . Figure 1. Circuit diagram for the day display. Each of the seven lamps corresponds to a day of the week. Figure 2. Circuit diagram for the date display, which consists of two 7-segment numbers. (taken from the two counters) are also 1. Connection 9 (inputs 28 to 31) is taken from the month display circuit and indicates the number of days in the month at present being displayed. As an example: the date count is at 28 and the input pin 28 is high. The next incoming pulse sets the counters to 29 (i.e. the divide-by-four counter has the value 2. and the decimal counter has 9). All the inputs to gate N 19 arc therefore at logic 1 , so the output is 0. The out- put of N18 goes to 1 causing FF2-FF6 to be reset, and FF1 to be set, so giving the desired output of I for the first of the month. The leading zero is sup- pressed by earthing the Rg input of 1C4. This improves the legibility of single digit numbers. Initial setting of the date is achieved by feeding second pulses to the circuit by means of switch S2. Month display The month display (figure 4) is the same in principle as the day display, except that there are twelve lights instead of seven. As the decoder ICS (7445) only has ten outputs available, two NAND gates (N30 and N31) are used for the other two. The output state of these gates is low for the values 10 and 11 from the counter, corresponding to the months November and December, respectively. These gates drive the lamps L18 and LI9. Resetting to zero, i.e. January, occurs at the thirteenth input pulse, which causes the counter to go to the value twelve, momentarily. This resets the 11-26 — elektor november 1976 counter to zero because of the connec- tions from D to R 0 i and C to . The input to IC9 is normally supplied from FF5 (see figure 2) via connection 2. However, switch S3 is used for setting. Automatic resetting of the displays re- quires a further item of information, i.e. whether the month has 28, 29, 30 or 31 days. This information is coded by the inverter N28 and the NAND gates N23 and N32. If, for example, L9 is lit, which means the month is February, then either lead 28 (for a ‘normal year’) or lead 29 (for a ‘leap year’) is at logic 1 . depending on the position of switch S4. Inputs to the NAND gates are, however, all high, so the leads 30 and 3 1 are both low. Each light, when lit, produces the correct output at connection 9. Only the item ‘leap year’ needs to be fed in manually. Date reminder ! Figure 4 shows the circuit of the date reminder, which enables the day display j for certain selected dates to flash, as | explained earlier. This facility is an i optional extra to the calendar, included j for those who have difficulty remem- bering such things as your nearest and dearest’s birthday. j The memory consists of a diode matrix, j the NOR gates N3A...N13, the decoder 1C 10 (7442) and the NAND gates N14...N17. IC10 decodes the units of the date and the gates N 14 . . , N17 decode the tens. Eleven of the twelve month inputs are always high, with only the month at present j being displayed providing a low input, j Every date gives a unique output. For example, consider the coding for February 1 6th: IC10 : the sixth output is low (pin 7) and so the input to gates N4C, N8A and N 1 1 B are also low. NI4 . . . N17 : the output of N16 is zero, and also therefore, the second input to N8A. Input 5 : ‘February’ is at zero, so the third input to N8A is taken to zero via diode Dl. The output of N8A is therefore at logic 1 which provides the flash enable signal at output 6, and the day display will now flash (see text: day display). As an example, the memory is shown with several diodes (Dl ...Dll) con- nected, each one of which corresponds to one date in the year. K Figure 3. Circuit diagram for the month dis- play, which is the same in principle as the day display circuit, but of course, twelve out- Figure 4. The programmable memory for important dates. As an example, several dates have been programmed by means of the diodes shown. Figure 5. A suitable front panel for the calendar. A red sheet of Perspex over the whole front gives good legibility and a pleas- ing appearance. 1 1 -28 — elektor november 1 976 wired in such way that the output is fed to the loudspeaker 'that is not being used as microphone at that moment. Understandably. . . . The automatic gain control is derived from T1 with its associated com- ponents. This transistor is connected as a current source. The maximum current it can supply to the bias input of the OTA (pin 5) is 0.6 V _ 470H = r = 1.2 mA. A domestic intercom doesn't require a complicated circuit. The unit described here uses only a handful of components, even though an automatic gain control has been incorporated in the circuit. This intercom was designed to meet two requirements: low cost and high per- formance. To keep the cost down, the output stage is nothing more than a two-transistor class-A stage that can deliver 100 mW. To maintain good intelligibility in spite of this low power output stage, an automatic gain control is incorporated. This ensures that the output stage will nearly always be fully driven. To keep the cost of this gain control down, an OTA is used as the preamplifier - its gain is a function of a DC bias current. Small 1 50 £2 loudspeakers are used both as microphone and as loudspeaker. The circuit Switch SI is the master ‘press-to-talk’ button. When this switch is depressed, the loudspeaker in the main station (the upper loudspeaker in the diagram) is connected to the input of the ampli- fier. When the button is released, the other loudspeaker is connected. Resistors R2 . . . R4 and capacitor Cl produce a smoothed DC bias voltage for the OTA. R5 and C4 are included in the input circuit to reduce clicks during switch-over from ‘talk’ to ‘listen’; they also reduce the effect of inter- ference pulses picked up by the (long) leads from the substation. The output signal from the OTA is fed to the ‘power’ output stage T2 and T3. This stage only gives a voltage gain of a factor 2; its primary function is current gain, to drive the loudspeaker. SI is This corresponds to a maximum gain of the OTA of 2500 x. When the output voltage rises above a certain level, current starts to flow through D3. This raises the base voltage of T1 towards the supply level, thereby reducing the current through T1 . This in turn reduces the gain of the OTA. This automatic gain control action is ‘tamed’ by including C5. Installation In most cases, the wiring to the sub- Resistors: R1 =47 k R2= 22 k R3.R4.R6 = 10 k R5.R10 = 1 k2 R7 = 10 M R8 = 470 O R9 = 33 n R11 = 1 M R12.R13 = 2M2 R14.R15 = 1 k C1.C3- 100 m/16 V C2.C5 = 10 m/16 V C4= 100 n C6 = 2m2/16V C7= 100 m/25 V Semiconductors: IC1 - CA3080 T1 = BC557B.BC177B.2N1711 T2 = BC517 (darlington) T3 = BC160.BC161 .2N3553 D1 . . . D3 - DUS.1N4148 Sundries: Cooling fin for T3 2 loudspeakers, 150 SI/1 Watt Switch or pushbutton, 2P2T power supply: Trafo 12 . . . 15 volt, 100 mA Fuse 100 mA slow blow (and fuse holder) Bridge rectifier B40C400 1000 m,25 V elco station can be normal two-core cable. However, if the distance is too great or if the leads run close to mains wiring it may be necessary to use single-core screened cable. Since the output stage is running in class-A, the current consumption is too high for battery operation. It would be possible to use batteries if an on/off switch is included in the main station, but this would mean that the substation cannot initiate the conversation. For battery operation, the supply voltage can be reduced to 9 V, although this will reduce the output level. A better solution is to use a mains- driven supply. A transformer with a 12 ... 15 V secondary which can supply 100 mA is sufficient. Connect this to a bridge rectifier followed by a 1000 /i/25 V smoothing electrolytic. The gain of the intercom will vary be- tween a maximum of 5000 x and a minimum of 150 x, depending on the Figure 1. The circuit of the intercom. T1 is the main component in the automatic gain Figure 2. The printed circuit board and com- ponent layout for the intercom. Switch SI and the two loudspeakers are connected to points 1 ... 4. as shown in the circuit diagram. Photo. A finished unit. Note the way the cooling fin is mounted on T3. A common mistake is to mount it 'upside down’, but this makes it much less effective. automatic gain control. The maximum is set by R8; increasing the value of this resistor will reduce the maximum gain. Do not decrease the value, as this can damage the IC. The minimum is set by the value of R7; this should not be altered. SI can be either a switch or a push- button, according to taste. In either case, it should be a break-before-make type. 11-30 — elektor 1976 signal horn smmlcw The supply voltage to T1 and the multi- vibrator should not exceed 12 V, but the supply to the loudspeaker may be up to 60 V (with an 8 loudspeaker) to obtain a louder signal. Alternatively, if the entire circuit is to be run from a single 12 V supply (e.g, a car battery) then several speakers may v be connected in parallel (minimum 1 fl paralleled resistance) to obtain a louder signal. Although designed as a warning horn for use with the 'Albar' ultrasonic alarm system de- scribed elsewhere in this issue, the signal horn may of course be used with any system requiring an audible warning. The parts outlay is small but the sound produced is loud and extremely penetrating. Construction A p.c. board and component layout for the signal horn are given in figure 2. T1 should be provided with a clip-on heat- sink for cooling. T2 may or may not require a heatsink depending on the number of loudspeakers driven and the supply voltage, but plenty of space is reserved on the board for the ‘multi- finger’ type heatsink. Operating Hints To make the maximum amount of noise, the loudspeaker should be mounted in a box made of a suitably resonant material (e.g. a biscuit tin). If the unit is intended for mounting out- side then the enclosure should be water- tight and the loudspeaker cone should be waterproofed with a few coats of !j It was decided at the outset that to B keep the cost down the output trans- ducer should be a cheap loudspeaker such as might be cannibalised from an old TV set, radiogram or other equip- ment. In view of the limited power handling capacity of such speakers it I was decided that the best drive wave- |i form would be a rectangular waveform with a low duty-cycle. In this way a j high peak sound level could be obtained I whilst staying within the average power rating of the speaker. This short duty- cycle also keeps down the current con- sumption of the circuit. The circuit The basis of the circuit is an astable multivibrator using the ubiquitous 555 timer. ! With the control input (R1 ) open-circuit | the output of the 555 is high, so T1 is turned off, as is T2. When the input is I grounded the circuit begins to oscillate I at a frequency of about 400 Hz (deter- mined by R3, PI and C4). The duty- ] cycle of the waveform T is given by _P i + R i R 3 +2(P, +R 2 ) ! The duty-cycle may be varied between about 1% and 10% by means of PI. i Since PI is part of the frequency deter- j mining network this will also cause the ■ frequency to vary slightly. ! The output of the 555 switches T1 on i and off, which in turn switches T2. In j order to ensure fast rise and fall times at ; the output, the base resistors of T1 and ; T2 are kept small. The loudspeaker is connected in series with the collector of j T2, and diodes D1 and D2 protect T2 against the back e.m.f. from the speaker. Figure 1. Circuit diagram of the signal horn. The duty-cycle may be adjusted by means of PI. Figure 2. Printed circuit board and com- ponent layout. Note the cooling clip on T1 and the heatsink on which T2 is mounted. Figure 3. An automatic interrupter for the signal horn, which may be required in some areas to comply with local regulations. model aircraft dope or something similar. Before mounting the signal horn on the outside of a building it is advisable to check the local laws concerning the use of alarms. In some areas continuous operation of an alarm may be pro- hibited and only intermittent operation is permitted. Figure 3 shows the circuit of an auto- matic interrupter consisting of a 1 Hz astable multivibrator and a switching transistor. When point B is grounded the multivibrator will switch the horn on and off at one second intervals. This circuit may satisfy local regulations, and in addition the on-off switching of the alarm is more likely to attract attention. M signal He /ember 1976 — 11-31 voltage-to-time converter The well-known TTL IC type 74121 is a monostable multivibrator that can be set to give an output pulse with a duration of anything between 40 ns and 40 s. This pulse duration is determined by a single external RC combination. If the fixed resistor in the basic circuit is replaced by a variable resistor (i.e. a potentiometer), the length of the out- put pulse can be varied over a wide range. In the circuit shown here, this principle is used to obtain a monostable multi- vibrator with a pulse duration that is determined by an external (DC) voltage. The original fixed resistor, which would have been connected between pin 11 and the positive supply, has been re- placed by transistor Tl. Resistor R1 is included to protect the IC from inad- vertent overdrive. The transistor works as a sort of variable resistor (with a ‘resistance’ that depends on the DC voltage applied to its base) and it determines the charging current flowing into the external capacitor C. The duration of the output pulse (t) is now set by the input voltage Vj n , approximately as follows: t«0.7Cx(5-V tr )x — Vtr-0.7) where t = duration of the output pulse (seconds) C = value of capacitor C (Farads) R 2 = value of resistor R 2 (Ohms) Vt r = trigger voltage (typ. 3 V) Vj n = input voltage (volts) a = DC current gain of the transistor. The monostable is triggered in the nor- mal way, via the A 1 , A2 and/or B inputs. H 11-32 — elektor november 1976 dowrni? A.M. Bosschaert In this age of digital things (?) it is nice to dress up the front panel of home-brew equipment with digital readouts. This handsome little digital voltmeter is easy to build and is inexpensive. In the form presented here, it is a single range unit, intended for use with power supplies similar to the ones described in last month's Elektor ('dual regulators'). Since this DVM was designed to replace conventional moving coil panel meters, on such things as power supplies, the price was one of the main design con- siderations. The unit uses only 6 tran- sistors and three C-MOS integrated cir- cuits: one IC contains four Schmitt triggers, the other ICs each house a decade counter with built-in BCD-to- seven segment decoder/drivers. It fea- tures 2V4 digit readout and has an acceptable accuracy. Basic operation Conversion of the positive input voltage into a quantity that can be digitally displayed is accomplished by converting the input voltage to a current. This cur- rent is then used to control a variable frequency oscillator; the output fre- quency of this oscillator is a linear function of the input voltage. This frequency is counted by the counter unit and displayed. The counter unit is controlled by a free running oscillator that determines the gate time for the counter stages and resets them just before the start of each count. It also blanks (turns off) the readout during the brief count cycle. Circuit operation At first glance, the input circuit may appear to be confusing, but if one breaks it into smaller units it is much easier to understand. The heart of the DVM is a current- dependent oscillator. This oscillator consists of the following parts: gate Nl, D2, Cl in the charge path and Cl, T2 and R2 in the discharge path. The input voltage is converted into a discharge cur- rent by T1 and T2. These transistors keep the voltage across R2 equal to the input voltage at all times. Since R2 is 1 k, the current through it (the dis- charge current!) in milliamps is equal to the input voltage in volts. The dis- charge time of Cl is therefore a linear function of the input voltage. The time required to charge Cl is always the same. The charge current is supplied from the low impedance out- put of N 1 . If for a moment we assume that pin 2 of Nl is held high, the oscillator circuit can be more easily understood. The secret that allows N 1 to operate as an oscillator is the fact that its switching levels are not the same (hysteresis). If Cl is completely discharged pin 1 is low, making pin 3 high. In this state Cl will be charged rapidly. When the voltage on Cl reaches the upper switching threshold of Nl , pin 3 goes low. D2 prevents Cl from discharging into the output of Nl, and since the input (pin 1) is a very high impedance, the only discharge path for Cl is through elektor november 1976 — 11-33 Summer is long gone, and all of nature is proceeding through its autumnal eventide towards winter hibernation. Birds are departing, and the casual wanderer will notice the lack of animal and insect sounds in the hedgerows. A sad state of affairs . . . However, electronics is coming to the rescue: in this article, we proudly present a cricket with a heart of CMOS! T2 and R2. This discharge current (and, therefore, the discharge time) is deter- mined by the input voltage, as de- scribed earlier. Once Cl has discharged to the lower switching threshold of Nl, Cl is re- charged rapidly and the cycle repeats. The higher the input voltage the faster Cl will be discharged, and the higher the output frequency will be. This fre- quency is gated into the counters by the internal time base (shown in the dashed box in the circuit diagram). The time base really has three functions: gating the incoming frequency, blanking (turning off) the display during the count cycle, and resetting the counters to zero just before the start of the count. The duration of the positive portion of the waveform being produced by the time base must be ‘spot on’, otherwise the unit will not be accurate. Therefore PI is provided for full scale calibration | of the unit. The timebase output has a very low mark to space ratio so that the | count cycle is very short compared to | the readout time. During this ‘enable’ | time the output of gate N2 is high. T3 j is switched off, blanking the display i (this is perceptible as a short blink). The reset pulse (the positive edge of the enable pulse) is passed via C3 to the counter section. Resistors R5, R6 set i the voltage level at pin 9 of gate N3 and pin 1 3 of gate N4 between the switching thresholds of these Schmitt triggers. A positive pulse will now switch the gate output to ‘0’ and a negative pulse will switch it to ‘1’. The reset pulse will turn on T4 briefly, so that the outputs of N3 and N4 will both switch to ‘1’. T5 and T6 are turned off and the ‘100 s’ display (D7) is blanked. C4 is discharged. Positive pulses are supplied by the ‘carry’ output (pin 5) of the decade counter IC2 at each transition from 9 to 0, and resulting step is differen- tiated by C5 and RIO. The first time this occurs, N3 switches to ‘O’. This turns on T5, so that a ‘1’ is displayed (segments b and c of the ‘hundreds’ display). C4 is still dis- charged, so the output of N4 is held at T. However, C4 is now charged through R8 and T5. The next zero-to-one tran- sition at the output IC2 (overflow) gives a second positive pulse. This time, the output of N4 does switch to ‘O', causing the display to read ‘H'(for ‘HELP’). At the same time IC1 and IC2 pins number 5 (display enable) change to low, in- hibiting the readout by D5 and D6. As a result the D7 character ‘H’ is the total display. Calibration Apply a known voltage between 1 and 2 Volts to the input and adjust PI for the correct readout. Since the DVM only reads to 1.99 V, an external volt- age divider mutt be added if higher voltages are to be measured. M Miniaturisation of electronic com- ponents and the increasing use of ICs has now reached the stage where quite interesting circuits can be fitted inside a match-box. This particular circuit simu- lates the chirping of a cricket, and in spite of its ultra-low power output it should have a useful range of about 10ft. In the immediate vicinity of the ‘insect’ the chirping sound meets Hi-Fi standards - in the sense that it is true to the original. The circuit (figure 1) uses 6 CMOS inverters and a TUN. The inverters are used in three astable multivibrators (AMVs). The N3/N4 multivib modu- lates a further AMV, consisting of N3 and N6. The interaction of these two AMVs produces the chirping sound. A Figure 1. The circuit of the ‘cricket with a heart of CMOS'. 11-34 — elektor i tber 1976 third multivibrator (N1/N2) switches the chirping on and off periodically, thus determining therest periods. The square-wave output from N5 is fed to buffer-stage Tl. This transistor drives the 'loudspeaker', which is actually a high-impedance dynamic microphone. The Sennheiser type HM35 is a good (and cheep sorry, cheap) choice, but many other types will also do quite well. The current consumption is very low, so the miniature mercury batteries should have a very long life. Open-heart surgery to replace the batteries should be i required only very infrequently. A printed circuit board for the circuit is available (figure 2), which should I guarantee rapid reproduction of the 1 creatures. H Figure 2. Printed circuit board and com- ponent layout. (MM R.L.A. Trost 1 } The film slave is a sound activated switch originally intended to control a cine camera. However, with little or no modification it may also be used for a variety of other purposes, such as automatic starting of a tape recorder in response to sound. The film slave is intended for use with cine cameras or tape recorders that have a remote control jack (socket), or which can be modified to take one. The jack is connected in series with the motor supply of the camera, typically 3-9 V. In normal operation the internal con- tacts of the jack are closed and the camera is controlled by the manual button. However, when a plug is inserted into the jack the supply is inter- rupted and the camera may be con- trolled by a shorting switch connected across the plug terminals. With the shorting switch open the Figure 1. Complete circuit of the film slave. When the camera is running the film slave obtains its supply from the camera via the remote control jack connected to points A and B. When the camera is started by a handclap, A and B are shorted out and the camera starts. The film slave then receives its power from an internal battery via T5. elektor november 1976 — 11-35 supply voltage of course appears across the switch terminals and a few milliamps may be taken from it without causing the camera motor to run. This is made use of in the film slave, the circuit being supplied from the camera whilst the camera is stopped, and being supplied from its own internal battery whilst the camera is running. In this way the life of the film slave’s battery is extended. The circuit The circuit (figure 1) comprises three sections, an input amplifier and pulse shaper, a bistable multivibrator (flip- flop) and output switching circuits. Sound is picked up by a crystal micro- phone and the resulting electrical signal is fed via Cl to IC1, which is connected as a non-inverting amplifier whose gain may be varied between about 7 and 320 by PI . Since the circuit is operated from a single supply voltage instead of the normal ± supply used by the IC, the non-inverting input is biased to half- supply voltage by R1 and R2. C2 and C3 cause the gain of the ampli- fier to roll off below about 200 Hz so that pickup of mains hum and other low-frequency interference is not a problem. This arrangement may seem slightly unusual, but is necessary in view of the supply voltage fluctuations that occur when the circuit switches from the camera supply to its own internal battery. If a single capacitor to ground were used then charging currents flowing into it when the supply switched would upset the biasing. The split arrangement of C2 and C3 obviates this. The output of IC1 is rectified by D1 and is used to turn on Tl, which provides the trigger pulse for the flip- flop T2/T3. Any loud transient sound such as a handclap will cause Tl to turn on momentarily. This will take the trigger inputs of the flip-flop (C8 and C9) low, and the flip-flop will change state. 11-36 — elektor november 1976 Figure 2. Printed circuit board ponent layout for the film slave. film slave The flip-flop is symmetrical apart from the trigger capacitors. C8 is much larger than C9 so that when power is initially applied to the circuit the flip-flop will always be reset with T3 turned on and T2 turned off. In this initial state T4 will be turned off, as will T5, T6 and T7. The film slave will receive its supply from the camera via the diode bridge D4 to D7. This ensures that the supply volt- age will always have the correct polarity irrespective of the polarity of the volt- age from the camera. When the flip-flop is triggered (set) by the sound signal T2 will turn on and T3 will turn off. T4 will be turned on, turning on also T5, T6 and T7. T6 and T7 short out points A and B, and again polarity is unimportant. If B is more positive than A current will flow through T7 and D5, but if A is more positive than B current will flow through T6 and D6. The supply voltage to the film slave from the camera is now lost, but the cir- cuit operates on its own internal bat- tery, which is now switched in via T5. R15 and CIO slow down the switching of T4 so that the supply transition is less abrupt to prevent spurious trig- gering of the flip-flop. Reservoir capaci- tor C7 also helps prevent this. Shorting out of the film slave battery by T6 and T7 is prevented by the fact that D5 and D7 are reverse-biased. The LED in the collector circuit of T4 indicates that the camera is running. The camera may be stopped by a second sound signal which will reset the flip- flop, turning off T4 to T7. The circuit then reverts to operation from the camera power supply. A bonus with the automatic supply switching is that the film slave does not require an on-off switch. After the completion of filming all that is necessary is to disconnect the film slave from the camera and (if necessary) reset it to the off state (LED ex- tinguished) by tapping the box. The Figures 3 and 4. Photographs of the com- pleted film slave. Figure 5. Modification to allow switching of any equipment by the film slave. internal supply is now switched off and the external supply is disconnected so the circuit cannot respond to any further sounds until it is reconnected to the camera. Construction A printed circuit board and component layout for the film slave are given in figure 2, whilst figures 3 and 4 show views of the completed unit in its box. 5a It must be stressed that the microphone used must be a crystal type, not a cer- amic type which may have insufficient output for the circuit to operate satis- factorily. Apart from this point there is nothing unusual about the construction except to note that PI should be accessible through a hole in the box so that the sensitivity may easily be adjusted to suit particular conditions of Applications and Modifications The circuit can, in principle, be used with any cine camera or tape recorder where remote control is effected by switching the motor supply voltage, which may be between 3 and 9 volts (DC or AC!) If, however, the remote control operates by switching the con- trol voltage to an electronic motor speed regulator, then it is unlikely that the film slave can be powered from the remote control jack. This may also be the case where the equipment to be con- trolled is other than a cine camera or tape recorder, so that the current and/or voltage to be switched is more than the circuit can handle. In this case the general solution is to use the circuit of figure 5a. T5 and T7 are omitted, as are D4, D6 and D7. The base, emitter and collector connections of T5 on the p.c. board are interlinked so that the circuit runs from its own internal battery all the time, and T6 is used to switch a relay whose contacts control the equipment. In this case the film slave does require an on-off switch. If only very low currents are to be switched the relay may not be required, so that the circuit of figure 5 b can be used. 11-38 — elektor november 1976 pulse generator pflOse ©©(nisffafta? E. van Dee and B. Alberts A pulse generator is an extremely useful aid when testing and servicing logic circuits. The generator described here is uncomplicated and is based on TTL IC's, but the facilities provided are quite comprehensive. A pulse generator is considerably more the pulse repetition frequency f G . The complex than a simple squarewave oscil- duty-cycle of the clock generator is not lator. The facilities offered vary widely, 50%, so the output is divided down by but basically the pulse repetition fre- two cascaded flip-flops to give 50% quency and pulse length are indepen- duty-cycle squarewaves at 2f 0 and f 0 . dently variable. In addition a pulse gen- These outputs are available externally, erator usually has a ‘prepulse' output, The clock generator may be gated on which is a short needle pulse that and off by a signal at the Gate input, so appears before the main pulse. This that bursts of pulses may be obtained may be used to trigger an oscilloscope if required. before the main pulse so that the The trailing edge of the f G output leading edge of the main pulse can be triggers a one-shot, which introduces a observed. delay before the main pulse appears. This delay can be varied by altering the one-shot time constants. When the first Principle of operation one-shot resets, its Q output triggers a The functions of the pulse generator second one-shot that produces the main can be explained by reference to the pulse. The length of the pulse may block diagram of figure 1. be varied by altering the one-shot time The whole circuit is driven by the clock constants. The outputs of the second generator G, which runs at four times one-shot are buffered by an output Brief specification Pulse repetition frequency : 0.1 Hz - 1 MHz in 7 decade ranges with fine control Pulse length : 1 0 ms - 1 00 ns in 5 decade ranges with fine control Pulse delay . 1 0 ms - 1 00 ns in 5 decade ranges with fine control Rise time of outputs :10 ns Clock outputs : f 0 -0.1 Hz - 1 MHz 50% duty cycle 2f 0 - 0.2 Hz -2 MHz 50% duty cycle Width of prepulse signal : 1 00 ns Output impedance : 50 £2 - short-circuit proof Amplitude of all outputs, unloaded : 5 V stage and normal (positive-going) and inverted (negative-going) pulses are available at the outputs. When the first monostable is triggered its Q output is differentiated to give a short spike — the prepulse output. The value of the prepulse output is obvious when one considers that no electronic circuit is free of delay. If the leading edge of the main pulse were used to trigger an oscilloscope timebase then all or part of this leading edge would be invisible due to delays in the oscillo- scope trigger and timebase circuits - it would simply have finished before the timebase started to run. With the pre- pulse this does not happen since the timebase can be triggered before the main pulse begins. In addition to continuous triggering by the clock generator, the pulse generator 11-40 — elektor november 1976 Figure 4. Suggested layout for the pulse generator front panel. may also be triggered either repetitively or in a one-shot mode by an external signal - either positive or negative going. Triggering may also be accomplished by a manual pushbutton. These different triggering modes are selected by switch S. A timing diagram showing operation in the clocked mode is given in figure 2. Complete circuit Clock Figure 3 shows the complete circuit diagram of the pulse generator, including the power supply. The clock generator consists of a Schmitt trigger, 'AIC1, con- nected as an astable multivibrator. Coarse frequency selection is by means of S 1 , which selects the timing capacitor, while fine control is by means of PI. Emitter follower T1 increases the input impedance of the Schmitt trigger so that fairly small value time capacitors may be used without leakage into the IC input becoming a problem. The oscil- lator may be turned off by applying a positive voltage to the Gate input, thus turning on T2. The output of the clock is divided down by flip-flops_ 1C2 go give outputs 2f 0 and f 0 . The Q output of the second flip- flop is connected to one of the positions of S2 and thence to the A inputs of the delay one-shot. The Q output of the delay one-shot is connected to the B input of the main pulse one-shot. Both one-shots are equipped with range selection switches S3 and S4 and with fine controls P2 and P3. The pulse out- puts are buffered by power NAND gates IC4. Prepulse The prepulse output^ is obtained by differentiating the Q output of the delay one-shot with a 10k resistor and 100 p capacitor. This pulse is then squared up and buffered by a NAND gate. T rigger modes The external trigger input may be used with S2 in the second position by feeding positive pulses into the trigger Three channel mixer Summer circuits «(£)a£) itei At a certain age, children are often packed off to bed with the final admonition: 'All right, you can read in bed for a quarter of an hour, but then you must turn off the light and go to sleep'. However, as most parents will know, the children tend to suddenly loose all sense of time in this situation . . . When a member of the Elektor design team was faced with this problem, he started looking for an electronic solution. The final circuit, as published here, has proved extremely effective. In the situation outlined above, what is really required is a unit that will auto- matically turn off the bedside reading lamp after the specified time has elapsed. This time switch must have a few special features: - It should only be possible for the parent(s) to switch on the lamp. This Figure 1. Complete circuit of the reading-in- bed limiter. SI must be a key-switch that can only be operated by the parents. 11-46 - elektor november 1976 can be achieved by using a key switch. - It should be possible for the child to turn off the lamp before the alotted time has elapsed, if it finds that it is getting too sleepy to read. Since the child hasn’t got the key to the main switch, a further reset button is required. - For safety reasons, it is essential to use a low-voltage lamp. The whole circuit, including the lamp, should be run off a reliable mains transformer. Since they are easy to obtain, a logi- cal choice is to use a 1 2 V lamp as used in cars. The circuit The obvious choice for the timer itself is the 555 timer IC, since this can be set to give delay times up to several hours with complete reliability. Furthermore, the obvious transistor type to use for switching the lamp is the well-known ‘work-horse’ the 2N3055. Having chosen these two components, the cir- cuit design is almost finished! The com- plete circuit is shown in figure 1 . The IC is used as a monostable multi- vibrator (MMV). The duration of the output pulse is set by a single RC- network, R1 and C2. In this particular application, the pulse duration is practically equal to the RC time. If R1 is 1 M and C2 is 1000/i, as shown, the RC time is 1 000 seconds, or just over a quarter of an hour. Note that any leakage in C2 will extend this time appreciably; for this reason it is advis- able to use a tantalum electrolytic, and not to increase the value of Rl any further. Initially, C2 is discharged. When the cir- cuit is switched on via the key-switch S 1 , C2 starts to charge through Rl. During this time, the output of the IC (pin 3) is at positive supply level. This turns on transistor Tl, lighting the lamp. R2 limits the base current to the transistor. With the type of lamp shown (12 V, 10 . . . 15 W), the dissipation in Tl should be so low that a heat sink is not required. The supply to the lamp is the raw, full-wave rectified supply voltage. There is nothing to be gained by smoothing this supply. An extra diode (D5) and a relatively small smoothing capacitor (Cl) are used for the supply to the IC. When the RC-time has elapsed, the out- put of the IC switches to 0 V, turning off Tl and the lamp. Pushing the reset button (S2) will switch the lamp off sooner. Since the ‘set’ input (pin 2) is not used, the only way to switch the lamp on again is to first turn the supply off, wait until C2 has discharged, and then switch on again. Officially, this should be done with the key-switch. It is not advisable to demonstrate even once that the same effect can be pro- duced by pulling out the mains plug for A printed circuit board layout for the unit is shown in figure 2. Note that sufficient care should be taken with the mains connection. Use good cable, a rubber grommet where the cable enters the box, and some form of clamp over the cable just inside the box so that there is no ‘pull’ on the con- nection to the transformer. M Figure 2. Printed circuit board and com- ponent layout for the unit. Tl can be mounted on the board, since a heat-sink should not be necessary (EPS 1660). a LOGICal replacement for the carbon-track potentiometer (or TAP out the music!) Ever since the onset of this Electronic Age, the zippy electron has been steadily taking over from the cumbersome mechanical and electro-mechanical equipment in a wide variety of applications. Signalling and other data transmission systems, calculators and business machines, clocks and watches, for instance, are obvious examples. ’ A firm last stand, however, seems to have been made by the familiar carbon-track potentiometer. Despite its inherent deficiencies such as inaccuracy and short useful life it has retained its popularity, particularly in the field of entertainment equipment, because it is cheap, easy to use for all kinds of controls in electronic circuits and is often easily replaceable. The trend now is to limit the use of these devices which have a habit of becoming noisy in their old age, and in advanced circuitry they have been largely replaced by controls such as varicaps, variable-slope transistors, biased diodes and other electronic devices. This article proposes further relegation in that the control voltages for a circuit are not even derived remotely from carbon-track pots, but from logic which produces the voltages entirely electronically, from finger-tip touch sensors. elektor november 1976 — 11-47 A convenient circuit to consider con- verting is the 730-740 stereo control amplifier published in the December 1975 Elektor. In this design, volume, balance and tone are all DC voltage con- trolled from conventional potentio- meters. At first sight, the requirements for generating control voltages elec- tronically would seem to indicate com- plicated circuitry, but this does not turn out to be the case. First consider the properties of the electro-mechanical potential divider. A wiper contact can be moved along a carbon track (by the operator) to produce an output voltage which is proportional to the distance of the wiper along the track. The output volt- age remains the same while the wiper is stationary. An electronic equivalent must, therefore, have two properties: A) an output voltage must be produced which varies continuously in accord- ance with some external command; B) the output voltage must remain con- stant in the absence of a command, i.e. it must be stored. Now consider applying these require- ments to the 730-740 amplifier. The control voltages in question can vary between 1 .0 and 9.0 V and are used for gain, balance, bass and treble adjust- ment. The electronic generation of these voltages does not present a serious problem; one solution would be to charge or discharge a capacitor via a constant current source. Storing the magnitude of this control voltage is, however, another kettle of fish. Long-term electronic analogue memor- ies based on circuits with extremely long time constants do exist, but their cost is prohibitive for this type of application. An approximation to non- volatile analogue storage could be made by loading a capacitor with a high impedance device (such as an MOS- FET), whereby the capacitor could be made to retain its charge for quite some time, up to, say, a couple of hours. In practice, though, we have found that such a circuit is far from reliable, and is difficult to construct. At this point, as with so many analogue problems, a solution is offered by using digital tech- niques. The following actions provide an elegant and simple solution: (1) convert the analogue information into a suitable digital form; (2) store this digital information in a memory; (3) retrieve this digital information and convert it back into an analogue signal. A major advantage of this procedure is immediately evident: digital information is very easy to store and a number of devices that can do the job are readily available. By using TTL ICs, the initial outlay, maintenance and constructional com- plication are all reduced. The problem has now been reduced to that of A/D conversion, step one from above. This is achieved by taking an instantaneous sample of the varying analogue quantity, measuring the sample 11-48 — elektor november 1976 logical replacement against some standard and expressing the result as a digital (usually binary) number. In this desigri, A/D conversion is achieved as follows: an analogue input causes a flip-flop register to count either up or down until the desired control voltage (produced by D/ A conversion) is achieved, at which point the input is removed so the counting stops and the register retains its digital value until a new sample (i.e. further input) is received. The desired control voltage has been reached when the listener detects correct stereo balance, for instance. Since, by definition, digital measure- ments are made in discrete steps, the accuracy of the A/D conversion depends on the number of steps available over a given range of values, i.e. the smaller the size of each step, the better the accu- racy of the A/D conversion and the greater the amount of information retained. The number of available steps may, therefore, be said to determine the ‘resolution’ of the system. This circuit uses four-bit flip-flop registers, so that sixteen steps are available for quanti- fying each analogue signal. Block diagram The block diagram in figure 1 shows that the four functions (gain, balance, bass and treble) are controlled by four almost identical circuits which are clocked by a common slow-rate pulse generator. Each functional unit marked ‘A’ in the diagram contains the following circuits: • a four-bit pulse-controlled (‘syn- chronised’) counter; • a ‘sampling’ input circuit to convert the state of the sensors into count up/down instructions for the counter (A/D conversion); • an enable/inhibit circuit ‘freezing’ the counter in the absence of counting instructions , • a max./min. inhibit circuit to stop the counting when the register is either full or empty; • a D/A conversion output circuit which translates the state of the flip- flop register into a DC control signal. As long as one of the touch sensors is activated, the appropriate output signal varies according to a 16-step staircase (up or down) function, where each step coincides with a clock pulse. Each unit marked ‘C’ in the diagram performs the following functions: • smoothing the discontinuities of the step functions; • amplifying the output signal (after D/A conversion) to the level required by the 730-740; • serving as impedance matching and preventing feedback from the 730- 740 into the logic circuitry. The unit marked ‘B’ also performs all these functions, but in addition it pro- vides a non-linear relationship between the D/A output and the input to the 730-740. 'A' in more Detail Figure 2 shows the functional units of which the blocks ‘A’ in figure 1 are each Table 1 Truth table (74191) counter state Qa Qb Qc □d 0 0 0 0 0 1 0 0 0 2 0 i 0 0 3 1 i 0 0 4 0 0 1 0 5 1 0 1 0 6 0 1 1 0 7 1 1 1 0 8 0 0 0 1 9 1 0 0 1 10 0 1 0 1 11 1 1 0 1 12 0 0 1 1 13 1 0 1 1 14 0 1 1 1 15 1 1 1 1 Figure 1. Block diagram showing the func- tional units necessary to convert the input from the TAP sensors into control voltages for the 730-740 stereo control amplifier. Figure 2. Functions performed by each of the four 'A' blocks in figure 1. Figure 4. Pin configuration of the 74191 comprised. The circuit ‘Al’ is common to all four blocks and generates (slow) clock pulses to synchronise the coun- ters ‘A3’. Input is received from the two touch sensors. Note that the arrows shown next to the sensors refer to the direction of the command, such as ‘increase vol- ume’ for the upward arrow and ‘de- crease volume’ for the downward arrow. This is not the same as the counting direction performed by the flip-flops. The command ‘increase’, i.e. count down, is given by touching S2 (in fig- ure 2) which causes it to become con- ductive. Touching SI causes count upwards. The amount of the change in the control voltages to the 730-740 is proportional to the duration of the con- ductivity. With either sensor operating, an enable signal is produced for the counter, but if neither is being iouched the enable is inhibited. This enable signal and the direction of counting are produced by the input circuit (A 2). The output of the counter is shown by the four flip-flop outputs, Qa, Qb. QC> Op. To convert this information into an analogue signal, the outputs are fed to an additive resistor network, A4, whose logical replacement . . . elektor iber 1976 - 11-49 SN 74191 synchronous up/down counter with down/up mode control Enable 'O': counter is enabled Enable 'V: counter is inhibited Clock: reac's to the positive-going edge Up/down: '0' = upwards '1' = downwards load: ’O' transfers states of Aj„ . . . Dj n to counter output states. Max/min: after 15th pulse = ‘V Ripple clock: after 15th pulse = 'O’ output is a staircase-type signal, as shown. As long as the counter is not enabled, and the power remains on, the state of the counters remains the same. The preset unit A5 provides reset values to the flipflops which are loaded im- mediately after switch-on. These presets are enabled by the ‘load' signal which is a short duration pulse present only after power switch-on. The presets are selected so that reasonable levels of vol- ume, balance, bass and treble are ob- tained when the equipment is first switched on. The Circuit The complete circuit diagram is given in figure 3. The heart of each of the func- tional groups is the counter (IC3, IC4, IC5, IC6), which is the standard TTL type 74191. Apart from the power supply (not shown here) the only com- mon circuit is the clock pulse generator; two Schmitt triggers from a 7413 (IC8) with biasing produce a pulse train with a frequency of approximately 3 Hz. The pin connections for the 74191s are shown in figure 4. The four outputs, 0 a- QB- Qc- Qd. are available on pins 3, 2, 6, 7 respectively. The states of the flip-flops during counting are given in table 1. At switch-on, the A, B and C flip-flops are reset to ‘O’ and the D flip-flop to ‘1’, so that the mid-range output is obtained (see table 1). The preset values are only available to the flip-flops while the ‘load’ signal (pin 1 1 ) remains at 0 V, i.e. until capacitor CIO becomes charged via the resistor R27. Counting takes place when the enable signal (pin 4) is ‘O’, and will be upwards if the level at pin 5 is ‘0’ and downwards if it is ‘1’. Overflow or underflow, i.e. when the counter state reaches either 1 5 or 0, is indicated by a ‘1’ on the max/ min output pin 12. Each of the four output control voltages is controlled by two TAP sensors, of which SI, S3, S5 and S7 cause a voltage decrease (the logic counts upwards), and 52, S4, S6 and S8 cause an increase. While the sensors are not operating, SI, 53, S5 and S7 are connected to the posi- tive supply rail via 10 M resistors and S2, 54, S6 and S8 are similarly connected to the earth rail. The smoothing capacitors (Cl . . . C8) absorb any hum or spurious electrostatic induction. When a sensor is touched, the relatively low resistance of the skin causes the quiescent logic potential to invert. The output state of the sensors is fed to the input circuits (A2 in fig. 2), which operate as follows. Consider the circuit controlled by SI and S2. In the non-operative state SI is at logic ‘1’ and S2 at ‘0’ and the outputs of gates N1 and N2 are ‘O’ and ‘1’ respectively. With the max/min signal from IC3 at ‘0’ (i.e. any condition other than overflow or underflow) transis- tor T1 is non-conducting since both its emitter and base are at zero potential. Transistor T2 is conducting, though, so that a ‘0’ is present at the up/down input (pin 5) of IC3. Counting up does not occur, however, since the ‘ 1’ at the collector of T1 inhibits the counter via pin 4. If SI is now touched, the output of N1 changes and T1 conducts so that the signal on pin 4 becomes ‘O’, thus enabling counting. If the overflow con- dition is reached, the max/min signal Table 2 Logic State of IC1 Control Signals SI S2 8 12/13 9/11 10 enable* up/down -0I>- 1 0 1 0 1 0 -0D- 0 0 1 1 0 0 ■c* 1 ’ 0 1 0 1 * whenever the counter reaches either 15c to 'V and the enable signal becomes "1', r 0, the max/min egardless of tbe s )te of the >gical reple elektor november 1976 - 11-51 changes to ‘1’ so T1 stops conducting. This results in the enable signal being removed so that further counting is inhibited. If, on the other hand, S2 is touched, T2 stops conducting and the up/down signal therefore changes to ‘1’ which resets the max/min signal on pin 12. Also, T1 is now conducting, so the enable signal to pin 4 is present. Counting will be down- wards this time because pin 5 is at 4 1'. Underflow will cause T1 to stop con- ducting and so counting will again be inhibited. The preset values for the counters (A5 in figure 2) have already been described, and are input on Ai n , Bj n , Cj n and Dj n P>ns. The weighted resistors R28 . . . R43 convert the digital output from the counters into analogue 16-step staircase signals. If the resistors are 5% tolerance (or better) the steps will be sufficiently even. At the lower counting limit (underflow) the resulting analogue level is approxi- mately 0.2 V and at the upper limit approximately 3.3 V. The 730-740 requires control voltages in the range 1 V ... 9 V so some conversion is necessary. The active devices linking the D/A con- verter outputs to the 730-740 inputs are Norton’ type DC amplifiers. Contrary to conventional opamps, the Norton types have differential current inputs, not voltage. For this reason they are most suited for a virtual earth input configuration, as shown in figure 5. Returning to figure 3, the opamp quiescent current is obtained from a reference current flowing through R45, R48, R51 or R54 into the non-inverting input. The ‘current mirror’ input circuit of the opamp reflects this reference cur- rent by feedback through the resistors R47, R50, R53 and R56. With the para- meters chosen, the quiescent current supplied by the Norton amplifier causes a potential drop of approximately 10 V to develop across the feedback resistor. Figure 5 shows that the voltage gain of the circuit is determined by the ratio of the feedback resistance to the input re- sistance (on the inverting input). The capacitors Cl 1 . . . C14 smooth the staircase signals so that a more con- tinuous rise or fall of the control voltage is obtained. The three DC inverting amplifiers IC7b , . . IC7d which supply the bal- ance, bass and treble controls, are ident- ical. The relationship between the D/A converted voltage (at the junction of the resistor network) and the output of the current amplifier is linear, with each step of the staircase being approximately 0.2 V. IC7a, which provides gain control, has a slightly different arrangement for quiescent current and gain settings, this is necessary to compensate for the non- linear drop in the TCA 730 control re- sponse. The compensation is achieved by R44 and D5 between the resistor network and earth. For low voltages the combination is non-conductive. but as 11-52 — elektor november 1976 logical replacement . . . the voltage increases to more than about 0.7 V, the bypass D5 .resistance gradu- ally drops, reducing the staircase steps to approximately one third of the orig- inal 0.2 V amplitude. The inversion performed by the lC7a turns this nega- tive departure from linearity into a positively incremental control gradient compensating for the negative bend in the TCA 730 control response. Power Supply The equipment needs 1 5 V and 5 V supplies, but the current demands are not very stringent. The 1 5 V line for the sensors supplies about 6 mA which can easily be derived from the existing supply for the 730-740. The 5 V supply should be stabilised, preferably by means of an integrated stabiliser such as the LM309 or LI 29. For this circuit the transformer secondary should supply about 8 V. Constructional Notes All the integrated circuits in this control equipment are conventional types from well-known manufacturers and should be readily available. Apart from these active components, the equipment uses eight transistors and five diodes, all of which appear in the TUP-TUN-DUG- DUS lists. The resistors marked with an asterisk in the diagram and parts list should have tolerances of 5% or better. Parameter Selection This unit’s operational parameters were determined by the designer’s likes and dislikes. Other users may wish to alter some of the settings such as clock fre- quency, counter presets or the time constants determining the response of the final amplifiers. The clock frequency in the circuit shown is approximately 3 Hz, which means it takes about about 5 seconds to count all the way up or down between 0 and 15. This count speed is deter- mined by the value of C9, an increase in capacitance slowing down the counting process. The presets can be changed simply by strapping the Aj n - - . Di n connections to the power lines to give the desired state of the flip-flops (see tabel 1 ). The time constants for the final ampli- fiers are determined by the capaci- tors Cl 1 . . . Cl 4. The constant chosen here is approximately 0.5 seconds, which gives reasonable smoothing of the stair- Figure 5. The LM 3900 used as a DC ampli- Figure 6. A factory-made touch sensor which is not generally available through retail out- Resistors: R1.R2.R7.R8.R13.R14.R19. R20 = 10 M R3.R4.R9.R10.R15.R16.R21, R22 = 15 k R5,R11.R17.R23,R26,R27 = 1 k R6.R12.R18.R24 = 470 n R25 = 330 n R28.R32.R36.R40 = 22 k" R29.R33.R37.R41 .R49.R52. R55 = 47 k" R30.R34.R38.R42 = 12 k” R31,R35,R39,R43 = 5k6” R44 = 1k8” R45.R48.R51 ,R54 = 180 k” R46 * 18 k” R47 = 100 k** R50.R53.R56 = 120k” R57.R58.R59.R60 = 4k7 C1 P C2.C3.C4,C5.C6.C7.C8= 10 n C9.C10 - 470 ill 6 V C11 ,C12.C13,C14 = 680 n C15 = 22 n Semiconductors: IC1 ,IC2 = CD 401 1 IC3.IC4.IC5.IC6 * 74191 IC7 = LM 3900 IC8 = 7413 T1 .T2.T3.T4.T5.T6.T7.T8 = TUN D1.D2.D3.D4 = DUS D5 = 1N4148 ” * 5% tolerance case steps. The smoothing, however, introduces delay in the response time: after the sensor has been released the control voltage continues to change for ' a while. The settling time can be de- creased by lowering the capacitances, but the steps will then be more distinct. Preliminary Tests and Some Further Constructional Notes After work on the printed circuit is fin- I ished and inspected, testing can begin, I for which a DC (20 k fl/V) multimeter » is the only instrument required. The first check is to make sure the coun- ters come-up in the correct preset con- I ditions. After applying the power to the j unit, the counters should be in the preset condition, therefore the output I voltages should be at a mid-range value, I i.e. between 4 and 5 V. The next test is I to determine that touching each sensor I produces the predicted result. Any I ‘bugs’ found at this point should not be I too difficult to find. Logic levels for the I digital circuitry can be checked with I reference to table 2 (Logic ‘O’ should be less than 0.8 V and M’ more than 2 V.). I The counter outputs can be checked I with the help of table 1 . Testing the clock pulse generator is even I simpler: pulses at pin 6 of IC8 should I cause the multimeter to deflect. Once I the digital circuits have been verified, | the only possible source of remaining I faults is the final amplifiers. Check all I the resistors and capacitors and if the I unit is still faulty, replace the LM 3900. | The only remaining problem is the | mechanical construction of the TAP j sensors. The most obvious design would I be to use copper tracked p.c.b. sheets, I but the snag with this is that conden- I sation in the sensor gap remains conduc- I tive for some time after the sensor has I been released; complete evaporation I may take a number of minutes. A better I design is one in which raised contacts are used, so that the deep gap provides sufficient insulation. Figure 6 shows such a commercially-made sensor, but, of course, such a design is beyond the scope of the average DIY amateur. As long as factory-made sensors or sensor | fields are not stocked by retailers, the I DIY man will have to rely on his own I resources. A familiar sensor array is I made from dome-headed screws, round- head furniture tacks or something simi- I lar. These are very simple to make: for .i each sensor mount two such tacks at 3/16 in spacing, the heads being the business ends, in a sheet of compressed insulating material or epoxy resin. The legs sticking through to the reverse of the sheet can be used as soldering pins. , When using screws, the obvious connec- j tion is one using soldering tags, washers i and nuts. Although these and similar home-made 1 sensor arrays do little to enhance the 1 appearance of a front panel, their ■ electrical performance is quite satisfac- _ i tory. It is therefore left largely to the j individual to arrive at an acceptable : I solution. H advertisement elektor november 1976 — 11-57 WHO and WHERE A directory of electronic component suppliers to Elektor readers IF THERE IS A COMPONENT SHOP IN YOUR AREA NOT LISTED BELOW PLEASE LET US KNOW AREA 1 The Amateur Radio Shop 13, Chapel Hill, Huddersfield, Yorkshire, Tel. 20774 Bell's Television Services 190, Kings Road, Hairogate, Yorkshire, Tel. (0423) 55885 Derwent Radio 28 Hillcrest Avenue, Scarborough, North Yorkshire, Tel. 0723 63982 I Suttons 1 30 Blue Boar Row, Salisbury, Tel. 27171 M & B Components Ltd 36 Bishop Gate Street, -eeds 1 LSI 4BB, Tel. 0532 35649 Greenbank Electronics [ 94 New Chester Road, 1 New Ferry, Wirral, | Merseyside L62 5 AG, J Tel. 051-645 3391 AREA 2 Campbell Electronics Ltd Tweedale Industrial Estate, Telford, Salop TF7 4JR, Tel. Telford (STD 0952) Chas. H. Young Ltd 170-1 72 Corporation Street, J Tel. 021-327 2339 The Radio Shop 16 Cherry Lane, Bristol BS1 3NG, Tel. Bristol 421196, S.T.D. Code 0272 Magtronics Ltd Snarestone, Burton-on-Trent, Staffs. DE12 7DB, Mail order only. AREA 3 Ambit International 25 High Street, Brentwood, Essex, Tel. (0277) 216029, Telex: 995194 Arrow Electronics Coptfold Road, Brentwood, Essex, Tel. 0227 219435 Amtron UK 4-7 Castle Street, Hastings, Sussex TN34 3DY, Tel. Hastings 437875 Basic Electronics Ltd 18 Epsom Road, Cruilford, Surrey B. Bamber Electronics 5 Station Road, Littleport, Cambs. CB6 1QE, Tel. ELY (0353) 860185 Bi-Pre-Pak Ltd 222-224 West Road, Westcliff-on-Sea, Essex SS0 9DE, Tel. Southend-on-Sea 46344 Candis Limited P.O. Box 25 Canterbury, Kent, Tel. Canterbury (0227) 52139 Cavern Electronics 94 Stratford Road, Wolverton, Milton Keynes, Tel. Milt. K. 314925 Charles Town 89 Carrington Street. Nottingham. Tel. Nottm. 868933 and 55489 Chromasonic Electronics 56, Fortis Green Road, London N10 3HN, Tel. 01-883 3705 Cl iff palm Ltd 13 Hazelbury Cres., I Luton, Beds., I Tel. 0582 415832 C.T.S. Ltd 20 Chatham St., Custom Electronic Controls 45, Picardy Road, Belvedere, Kent DAI 7 5QH, Tel, Erith 34476 Direct Electronics 627 Romford Road, Manor Park, London El 2 5AD, Tel. 01-553 1174 Eley Electronics 100/104 Beatrice Road, Tel. Leicester 871522 Electrovalue Ltd 28 St. Judes Road Road, Englefield Green, Egham, Surrey, Tel. Egham 3603 Frank Mozer Ltd 5 Angel Corner Parade, Edmonton, London N18, Tel. 01-807 2784 GB Garland Bros. Ltd Chesham House, Deptford Broadway, London SE8 4QN, Tel. 01-692 4412 Harrington Colorvision 9 Queen Street, Colchester, Essex, Tel. Colchester 47503 HB Electronics 54 Montagu Street, H.G. Rapkin 22 Wellingborough Road, Abington Square. Northampton. Tel. 37515 M.M. Electronics Church Road, Tunstead. Norwich, Tel 0603 614154 MTRS 20 White Hart Street. Mansfield, Notts., Tel. Mansfield 23107 Maydale Electronic Services 2 Wellesley Parade, Godstone Road, Whyteleafe, M. & B. Services Maijs of Church Gate 1 2/14 Church Gate, Leicester LEI 4AJ, Tel, 58662 Orchard Electronics Flint House, High Street, Wallingford, Oxon 0X10 ODE, Tel. Wallingford (0491) 35529 Philcom Electronics Ltd Electronic Centre, 45 Lower Hill Gate, Stockport, Cheshire SK1 1 1JQ, Tel. 061.480,9791 Phoenix Electronics (Solent) Limited R.S.M. Television 38 Carrington Street, Nottingham, Tel. Nottm. 868933 Sintel 53(K) Aston Street, Oxford, Tel. 0865 49791 Smiths of Edgware Road 287-289 Edgware Road, London W2 1BE, Tel. 01-723 5891 Stephens Electronics Akeman St., Tring, Herts., Tel. 044282 3389 Swanley Electronics P.O. Box 68 Swanley, Kent, Tel. Swanley 64851 Technomatic Ltd 54 Sandhurst Road, London NW9 QC Trading 1 St. Michaels Terrace, Wood Green, London N22 4S, Tel. 01-889 7593 AREA 4 WM. B. Peat & Company Ltd 25/26 Parnel Street, Dublin 1, Ireland, I Tel. 749973/4 The easy way to a PCB... ...theSeno33 system! Seno33- The Laboratory in a box From your usual component supplier or direct from: DECON LABORATORIES LTD. Ellen Steet, Portslade, Brighton BN4 1EQ Telephone: (0273)414371 Telex : IDACON BRIGHTON 87443