Volume - 5, Number - 5 May. 1987 Publisher: C.R Chandarana ELEKTOR ELECTRONICS PVT. LTD. 52, C Proctor Road 8ombay - 400 007 INDIA Telex: (Oil) 76661 ELEK IN Standfast House Bath Place j High Street. Barnet | Herts EN5 5XE U.K. Editor: Lan Seymour Publitron Publicacoes Tecnicas Ltda Av Ipiranga 1100. 9° andar CEP 01040 Sao Paulo - Brazil Editor: Juliano Barsali Elektor sari Route Nationale: Le Seau; B.P. 53 59270 Bailleul - France Editors: D R S Meyer; G C P Raedersdorf Elektor Verlag GmbH Susterfeld-Stralle 25 100 Aachen — West Germany Editor: E J A Krempelsauer Elektor EPE Karaiskaki 14 16673 Voula — Athens — Greece Editor: E Xanthoulis Elektuur B.V. Peter Treckpoelstraat 2-4 6191 VK Beek - the Netherlands Editor: PEL Kersemakers Ferreira ft Bento Lda. R.D. Estelania. 32-1° 1000 Lisboa - Portugal Editor: Jorge Goncalves Ingelek S.A. Av. Alfonso XIII, 141 Madrid 16 — Spain Editor: A M Ferrer Kedhorn Holdings PTY Ltd Cnr fox Valley Road Et Kiogle Street Wahroonga NSW 2076 - Australia Editor: Roger Harrison Electronic Press AB Box 63 182 11 Dartderyd - Sweden Editor: Bill Cedrum Copyright© 1987 Elektuur B.V. The Netherlands Electronics Technology Electronic potentiometers 5.22 Secondary breakdown in power transistors 5.25 Linkwitz filters 5.30 RIMS : counting atoms 5.33 Where electronic messages have the edge 5.37 Local area networking 5.44 Schools equipment for tomorrow's scientists 5.45 Projects Feedback in loudspeakers 5.19 Preset extension 5.28 MSX extensions - 5 5.40 Text display on junior computer 5.47 Information News • News • News • 5.17 New products 5.60 Info/Data sheets 5.73 Guide lines Switch board 5.67 Classified ads 5.72 Index of advertisers 5.72 Corrections 5.72 Selex-23 Sound of the sea 5.50 Touch-Keys 5.53 5.05 FEEDBACK IN LOUDSPEAKERS by R Conell Electrical feedback is the backbone of many an electronic circuit. Acoustic feedback is not nearly so common, but R Conell suggests some ways of experimenting with it in a low-frequency loudspeaker. Ever since Thiele and Small published their works on loud- speaker theory, it has been j possible to calculate fairly ac- I curately what the ideal enclos- ure is for a certain type of loudspeaker, or conversely how a loudspeaker will behave in a certain enclosure. Accord- ! ing to Small, a closed box will I behave as a second-order high- pass filter, while Thiele shows ! that bass reflex and trans- J mission line boxes act as fourth- i or sixth-order filters. From this it J is clear that a closed box will j give better bass reproduction than an open system. | The performance of a filter is I determined by its quality factor | 0 and its resonance frequency j f. This is also true of a complete I loudspeaker system, including ] the enclosure, when the total 0 is designated Qk and the res- I onant frequency fc. In an ideal bass system, these quantities I should have values as follows: 0=0.5 to 0.7, and fc<30 Hz. Moreover, the volume of the enclosure should preferably not exceed 100 litres; the frequency range should be | greater than 300 Hz; and the dis- I tortion should not exceed 1%. J It is virtually impossible to meet I these requirements with a ! passive speaker system, par- | ticularly as regards Qk and fc. ! In an active system, it is far easier to approach the ideal. Frequency response equaliz- j ation is one way to tackle the j problem. Basically, it is better. however, to make use of a con- | trolled system. Unfortunately, j such a system is prone to spurious oscillations, which can, however, be obviated by negative feedback. Basic controlled system I Control is possible by convert- ing some of the acoustic output of the loudspeaker into an elec- trical signal and returning this j to the input of the power ampli- fier. To this end. a low-mass ac- celeration pick-up has to be ! fitted to the cone of the drive The block schematic of a poss- ible arrangement is shown in Fig. 1. The left-hand box con- tains the control electronics, fol- lowed by the power amplifier, which has a gain of about 30 dB, and the loudspeaker system. The control electronics consist of an adder that combines the left- and right-hand signals, a low-pass filter with a cut-off fre- quency of 100 Hz, and a differ- ence amplifier where the fil- Impedance converter : The impedance converter— see [ Fig. 2— consists of a Type TL071 operational amplifier. Its pin- out is shown in Fig. 3. This stage should be fitted as close as possible to the acceleration pick-up, preferable direct onto the chassis of the drive unit as shown in Fig. 7. Control circuits Adder ICr in Fig. 3 combines the two stereo signals into a monaural signal. Potentiometer Pi sets the input level for low- pass filter ICs-ICi This Bessel filter has a cut-off frequency of 100 Hz and a roll-off of 24 dB/oc- tave. A similar filter was de- i Fig. 2. Circuit diagram of the impedance converter. The pin-out ’ of the TL071 is shown in Fig. 3. tered input signal is reduced by ! scribed in the January 1986 the correction signal from the ' issue of Elektor India. feedback loop. j The control amplifier proper is , The power amplifier can be of | formed by ICs: the values of Rs, ; any type, but its gain should R", and C* determine the tran- preferably be about 30 dB A sient response of the overall : smaller gain would require ! system. These values will be I i some adjustment of the control j reverted to under Setting up. j I loop, while a higher gain in- j The control signal is deducted creases the tendency to oscil- j from the filtered audio signal in | lations in the loudspeaker sys- , subtractor ICs. The output of | tern. 1 this stage is fed to buffer IC? via The loudspeaker system con- [ two low-pass sections, Rie-Cn j tains the drive unit, fitted with and R.t-Cu. These sections the acceleration pick-up. M, ) further suppress any tendency and an impedance converter, to oscillation and are absolutely ' fCi. ' necessary. It is possible to omit impedance converter ICi and buffer IC?, but the values of the low-pass sections between IC6 and IC; should then be recalculated with due account of the input impedance of the power ampli- fier. Modifying the drive unit The acceleration pick-up is made from a piezo tweeter from which the chassis has been re- moved as shown in Fig. 4. The connexion wires have been cut at the terminals, not at the crystal end. The remaining cone is then cut to the same size as the piezo disc. The resulting acceleration pick- up may be fitted over or under the dust cap of the woofer. The latter method is preferable, but only possible if the dust cap has been fastened with a ther- moplastic glue. The cap may then be removed quite easily with a heated knife as shown in Fig. 5. The removal of the cap should, of course, be carried out with the greatest care to avoid damage to the cone of the drive unit or its speech coil. Once the dust cap has been re- moved, it should be stiffened with a thin layer of epoxy resin and a piece of glass fibre cloth at its inside— see Fig. 6. The epoxy resin may be used at the same time to fix the pick-up in place. In the mean time, the woofer should be kept upside down to prevent dust entering the air gap. After the epoxy resin has hardened, a thin flexible wire should be soldered to each of the two short connexions of the pick-up. These wires should also be glued to the dust cap to prevent them vibrating in unison with the cone later. Next, the dust cap can be fastened onto the cone again, preferably with thermoplastic glue to enable removal at a later stage if necessary. Before glu- ing it in place, however, pierce a small hole in the cone through which the flexible wires are fed. These wires should be glued to the cone in the same way as those to the speech coil. Finally, they should be connec- ted to the impedance converter board as shown in Fig. 2 and 5.20 eleMor India may 1987 3 0 40 70 100 k 4.5% 1.7% 0.65% 0 8b | 1.5% 1 0.6% | 0.5% | 0-65K Maximum sound p different enclosure iressure at 40 Hz with ! volumes | Volume (litre) 50 70 | 100 Without feedback | With feedback [ System parameter 70 1 enclosure 98 dB 100 dB 1 102 dB 101 dB 103 dB 1 105 dB! s measured in a On I L- I /3 d8 Without feedback With feedback 1.9 1 48 Hz 1 29 Hz I 0.6 | 17 Hz 1 20 Hz | Fig. 7. They should preferably be of about the same length as | those to the speech coil, j The drive unit is then ready for j operational use— see Fig. 7. Setting up All the constituent parts of the system should now be intercon- ! nected as shown in Fig. 1. Short ; out Ri! and Cs with the aid j of a switch to disable the con- trol circuit When the switch I is opened momentarily, one of three things will happen: • the loudspeaker remains • the system oscillates at a low frequency (<100 Hz); • the system oscillates at a high frequency (>1 kHz). In the first case, everything is in order and the system can be taken into use. In the second case, the con- I nexions from the pick-up to the impedance converter board 1 must be reversed. In the third case, the oscil- lations must be damped by changing the values of a few | components. First, increase C12 to ln8 and, if this does not help, C11 to 1 pF. If that still does not cure the problem, reduce the value of R. . and increase that of C& Resistor R11 affects the lower cross-over frequency, while Cs alters the Qk of the system. The author has built ( several of these systems and has never encountered oscil- lation problems. Do not forget to remove the switch from across Rn and Rs. 8 IL liliiiiiiililii PRHBiniliBiiu BipepB ilill iiiiiiii =====1 lilii lilii lilii == 11 11 ll ei&iiillilni mniiiiiHi HIHIiSl IP!!! Ill mill liiilii iiiliiii iiiiiiii iiiiiiii iiiiiiii iiiii=== The frequency characteristics in Fig. 8 show the results of the modification: it is quite evident that the lump between 30 and 100 Hz in the response of the system disappears when the feedback is introduced. The response between 20 and 30 Hz is also much improved. A number of pertinent measurements are tabulated in Table 1. The system with feedback was also compared with a number of top quality loudspeaker systems: in all cases, it per- formed equally well over the bass range, in spite of its cost being only a fraction of that of the competition. Fig. 8. The frequ without feedback. 5.21 ELECTRONIC POTENTIOMETERS by T Scherer An exploratory look at all-electronic replacements for potentiometers in high quality AF applications. Potentiometers are, arguably, number of low-cost alternatives on this thin carbon film readily on synchronicity of the set re- not the best way of controlling to potentiometers used in gives rise to scratching noises sistance is often no less than the volume and tone settings in various circuit sections of AF made audible in the loud- 20%, even with linear law types, an AF amplifier. We all know | equipment. speakers. Furthermore, dust The voltages at the wipers of a that they can cause scratching and foreign particles can easily logarithmic stereo poten- noises when operated, collect , , , enter the potentiometer enclos- tiometer can also differ by sonfe dust, and sometimes develop * ” e car bon track ure, and block certain sections 20%, causing a volume differ- contact problems giving rise to potentiometer of the carbon track, so that the ence between the channels of a troublesome discontinuities in j This most commonly used amplifier falls still at particular maximum of 2 dB, which may the operative range. High qual- j voltage divider is generally volume settings, making the ad- be noticeable in listening, ity potentiometers for AF appli- composed of a carbon film justment very difficult. : Potentiometers are generally cations are not only difficult to deposit on a ceramic base ar- Stereo potentiometers of the mounted on equipment front obtain, but also notoriously ex- ranged in a three-quarter cir- carbon film type are a further panels, and are connected to pensive. In the following sec- ; cular form (270°). The poor source of trouble. With most in- , the electronic circuit with the tions we will briefly examine a contact definition of the wiper expensive types, the tolerance aid of shielded wires that often Fig. 1. Experimental stereo volume control circuit based on the use of a LED-LDR optocoupler. 5.22 eleklor india may 1987 carry very low signal level at relatively high impedance. This makes the amplifier suscep- tible to noise, hum and strong RF fields, which can still be picked up by the carbon track in the potentiometer (plastic en- closures!), and even in the cable shield. In conclusion, it is reasonable to say that the standard carbon track potentiometer is un- suitable for a great many critical applications. Stepping switches Rotary (wafer) switches with fixed resistors at the contacts are, in principle, a good way to effect volume and tone setting in an amplifier. The tracking is adequate, and scratching noises due to spindle move- ment are effectively ruled out. However many rotary switches of suspect quality do develop contact problems after pro- loged use. A major difficulty in the designing with stepping switches is the finding of types having the number of positions required to ensure a sufficiently smooth adjustment range. Wire-wound potentiometers Long ago in the history of elec- tronics, all potentiometers and resistors were made from resist- ance wire. For a number of specific applications, the wire- wound potentiometer is still in use. Ganged types with motor drive units can be found in some of the most expensive types of amplifier. This appli- cation, however, requires soph- isticated mechanical engineer- ing on the one hand, and a fairly complex electronic control cir- cuit on the other, making the whole set-up rather cumber- some and expensive at the An LDR-based potentiometer The first attempts at making a fully electronic potentiometer were carried out with combi- nations of LDRs (light depen- dent resistor) and a small bulb. Although the results were quite satisfactory for AF equipment on the market in the early 1960s, we would nowadays reject the LDR-and-bulb control for incor- poration in Hi-Fi equipment in view of the noise production, rumble sensitivity, and poor tracking characteristic of the stereo versions. We all know that each and every electronic component re- mains subject to continuous en- hancement by the joint force of manufacturers and their re- search laboratories. The German firm Heimann. for in- stance, took up the long forgot- ten LDR for further research, and used two of these devices together with a LED to make an optocoupler that has adequate features for Hi-Fi applications. The LDRs in their Types LTlOxx and LT20xx optocouplers are of excellent quality, and es- pecially the LT20xx should do very well as a stereo po- tentiometer with adequate tracking properties— see Fig. la for the pinning and R-Id curves, and Fig. lb for a suggested ap- plication circuit. An OTA-based potentiometer A fairly simple potentiometer replacement can be realized with the aid of an OTA (oper- ational transconductance ampli- fier). which is essentially an amplifier with current-con- trolled gain. The gain range of about 80 dB, the extensive usable frequency range and linearity of the current-gain cor- relation. make an OTA such as the Type LM13600 eminently suitable for the applications we are concerned with here. Those who want to experiment with these devices will find the suggested circuit in Fig. 2 of use for further experiments. The only drawback associated with OTAs is their limited dynamic range, which results in a, maximum attainable signal-to- noise ratio of about 80 dB. Analogue multiplexers The circuit shown in Fig. 3 is a high-quality, all-electronic volume control featuring 16 dB and 2dB steps as controlled from a 6 -bit digital input. The ICs in this circuit are the well- known Type 4051 eight-channel analogue multiplexer/demulti- plexer, which is in essence an electronic version of an 8 -way, single pole rotary switch. The contacts are inputs 0-7, the pole is output Z, and the switch pos- ition is set with the 3 bits at the A-B-C inputs. Example: apply- ing binary code 010 to the A-B-C inputs of the left-hand multi- plexer connects input 2 (pin 15) to output Z. The input signal for opamp A 2 is therefore taken from the —32 dB contact on the resistor ladder. The resistors at the inputs of the second multi- plexer driving A 3 are dimen- .I.Mo< 1 nd,.m.y ,987 5.23 sioned lo give 2 dB attenuation steps, so that the overall range of this electronic potentiometer is from 0 to —96 dB as set with 6 bits. A balance control can be made with two of these circuits, operated on the basis of soft- The tone control section shown in Fig. 4 uses the same prin- ciple as the above volume ad- justment. The resistors as part of the R-C filters in the feed- back loop of A« are selected with 3-bit codes for bass and treble. Use high-stability resistors and capacitors when constructing these circuits, and provide ample decoupling of the supply lines. The opamps should be low-noise types such as the TL074 indicated in the circuit diagram. The digital ad- justment of the volume and tone control circuits is a matter we leave in your hands. You may want to use an up/down counter, a microprocessor port, or a special switch to arrange for the correct bit combinations at the multiplexer control inputs (consult Table 1). SECONDARY BREAKDOWN IN POWER TRANSISTORS by Sue Cain & Ray Ashmore * This article examines the different types of secondary breakdown that occur in power transistors, and investigates the phenomena that cause them. It concludes that secondary breakdown is a function of transistor technology, and cannot always be improved without some trade-off in other parameters. One of the basic failure I mechanisms in power tran- j sistors is second breakdown. This term includes various physical phenomena which are completely different. They de- | pend on the different use of transistors in the circuits and have in common the electrical j and thermal instability inherent | in transistors themselves, j The conduction behaviour of an | emitter base junction and the current gain of a transistor de- | pend significantly on the tem- perature and increase as a function of the temperature. Electrical and thermal in- stabilities may act simul- taneously within the device, thereby giving rise to destruc- tive second breakdown mech- j An understanding of this mech- i anism is of great importance for | a safer optimum application of a | power transistor. A distinction should be made j between direct second break- I down (/s/b) which is dis- j tinguished by a normal j direction of base current Ib j (entering into an NPN transistor) 1 and inverse second breakdown <£>,/b), when /» is in the op- ] posite direction (extracted from an NPN transistor). The limits to which a transistor may be used without entering into Es/b are defined by the reverse bias safe operating area (RBSOA). Direct second breakdown It is important for the power cir- cuit designer to know the locus of the Ic—Vce points defining the boundary between stable and unstable operation of for- ward biased transistors. This locus defines the SOA safe operating area, that is, the area of the logic — logger plane which may be used without any j risk in DC current conditions or | with different width pulses at a | known temperature. A typical i SOA is shown in Fig. 1. The J limits of this area are as follows: ! l)The A-B section represents the upper limit of the collec- tor current that may normally be j used, generally limited by wire bonds. Operation at higher cur- rents may cause damage to the wires of their bonding. 2) The B-C section is the —I | slope curve section (ie. the I section with constant dissi- I pation) defined by: VceIc = P wv, = (T,ma,-To)/Rs [ 1 ] This section therefore indicates the maximum dissipable power of the device. Tmtx is the maxi- I mum temperature which the collector-base junction may reach, over which the device reliability may be compro- mised. In power transistors, Tn„, varies between 125 and 200 degrees Celsius and gener- ally depends on the metallurgy and the type of package. Rq is the thermal resistance between the collector-base junction and the case, including all the sili- con and package system. The increase in the maximum dissipable power when the pulse width decreases (Fig. 1) corresponds to the decrease of Zg with respect to % 3) Section C-D corresponds to the second breakdown phenomenon (or /s/b) and limits the maximum power that the transistor can dissipate. This may occur even at relatively low Vce voltages. 4) Section D-F is the limit due to the transistor's BVceo. Second breakdown is gener- ated by the electrical and ther- mal instability of the transistor. The main causes of this insta- bility are: 1) The Kbe of a directly biased j base-emitter junction, at con- | stant current, decreases lin- early with temperature, with a I $ = 2 to 2.5 mV/°C slope. The base current of the transistor may therefore be expressed by: /s=/o|exp(eV'Br ' T /kT\ [2j i and. when Vbe is kept constant. | it increases with temperature. | 2) The hFE at the relevant voltage values increases as a function of temperature accord- ing to the law: ItFE—hFEo I expfAfu/AT’) j [3] Where A Eg is an activation energy which is a feature of the transistor. 3) The thermal conductivity of silicon decreases when tem- perature rises, causing a worsening of the thermal resist- ance of the transistor. When these three phenomena are taken into consideration, it may be observed that a pulse of power P= VceI generates: a) an increase of the junction temperature, giving rise to an increase of h and hFE, and therefore to an increase of Ic with a following increase of P and, therefore, a further tem- perature increase. b) a dissipation to the exter- I nal environment, controlled by the thermal resistance Rg - dT/dP which tends to stabilize the device. The situation evolves towards stability when: “2 , yeti A/ci 3 T is smaller than 1, or instability if > 1 . In this way, a stability factor, S, may be defined that will be a function of Vce and Ic: S.R,V [1] 5.2b When S>1, so called "thermal runaway" occurs and the junc- tion temperature increases without any limit, thereby de- grading and possibly damaging the transistor. The failure gener- ally occurs when the surface temperature becomes greater than the eutectic temperature . between silicon and the contact metal (front aluminium) with a consequent melting of the alloy. A localized temperature in- crease may also damage the crystal, or the inner tempera- ture of the device may reach values high enough to melt the silicon. To understand /»/ b phenomena which give rise to a reduction of the maximum power that the transistor can dissipate as Vce increases (zone D-E), it is necessary to take into account that device operation is not homogeneous on all the dice area. There are disuniformities in the emitter base current den- sity that may be due to junction disuniformities, crystal defects and, most of all, to the emitter edge concentration phenom- The voltage drop due to the base current flowing through the cross resistance rbb' gives rise to a disuniformity of Vbe at the junction, and therefore to the disuniformity of the current density Je (see Fig. 2). A side drop of 26 mV reduces the injected emitter current by a factor 1/e, where e is the base of the natural system of logar- ithms (= 2.71828...). A concentration of the current at the emitter periphery is therefore generated, so the ac- tive silicon area is reduced and hot spots occur, leading to an effective increase of the ther- mal resistance. As a result, the maximum dissipable power is decreased. When Vce is increased, the ef- fect of the base-collector elec- tric field is to increase the base current concentration. Different techniques may be adopted to limit the /»/ b phenomenon. Fundamentally, these consist of minimizing the mechanisms that trigger elec- trical and thermal instabilities in the transistor. The basic tech- niques are: 1) minimization of crystal dam- ages, metal impurities, and doping disuniformities; 2) optimization of package and die attach techniques to minimize the thermal resistance 5.26 eleklor india may 1987 on which the stability factor S depends. Disuniformities of sili- con die bondings to the case may give rise to adverse vari- ations of Rg as a macroscopic parameter for the dice as a whole, but also to significant variations between different points, giving rise to premature second breakdown: 3) increase of the base thick- ness to reduce the high cur- rent densities (due to emitter crowding) flowing through the collector base junction (where the electric field is localized), so that the density of the dissi- pated power is decreased High base thicknesses, how- ever, will result in lower cut off frequencies and slower switching times; 4) optimization of the horizontal geometry; 5) introduction of distributed ballast resistances connec- ted in series with the base, the emitter or both, which tend to give a negative feedback to thermal runaway, therefore stabilizing the device. The introduction of a ballast re- sistance in series with the base of the emitter may reduce from Ji to h the current density in the hot spot. The emitter ballast re- Figure 3. The common emitter sistance is generally obtained by opening emitter contacts thinner than the emitter strip. In this way it is possible to limit the current density at the bound- aries of the emitter. These resistances show the drawback of increasing the saturation voltage of the transistor by the amount Vceismi = Rdctsan. On the other hand, the base ballast resistance is obtained through a "N + pocket" (in the case of NPN). around the emit- ter area. This N + diffusion, be- ing unbiased, cannot be tra- versed by the base current, so this is forced to flow below the N + through a small section and. in the case of a diffused base, encounters a higher re- sistance on the way to the edge of the emitter. In this way it is possible to improve h, b signifi- cantly. It should be noted that the SOA limits are temperature depen- dent and suitable derating must be applied. Reverse second breakdown The reverse breakdown phenomenon (£Vb) is also due to thermal and electrical in- stability of the transistor. As already mentioned, it is dis- tinguished from 7s/ b by the presence of a reverse h (ie. with a direction opposite to the normal direction of a transistor operating in the active zone) and by high Vce values of the transistor. The device may be in these working conditions dur- ing turn-off with an inductive load. In Figure 3 the common emitter characteristic curves for a tran- sistor are shown. It is easy to understand the behaviour of these curves when the common emitter gain ex- pression is considered: hrc = Ar/(\-Ar) |5| for high Vce values, Ar is re- placed by MAr. For low Vce values, M is an in- significant factor, being very close to 1: it increases when Vce is increased according to the following expression: M = \I[\-VceIBVcboT\ (61 From expressions (5) and (6) it is clear that hrE depends on Vce, becoming infinite when MAy- 1/BVceo. The negative slope section, which is a feature of the curves with 7e<0, is due to the fact that A decreases at low values of the emitter current. During tum-off with an induc- tive load, the transistor has to operate with negative base cur- rent and a high value of Ic. It often has to reach a working area above Vceo, remaining there all the time required for the inductance to be dis- charged. Fig. 4 shows the be- haviour of Ic, Vce, Ib and the power dissipated by the transis- tor during tum-off. The area of the dissipated power corresponds to the energy stored by the induct- ance. '/zi/*, which is dis- charged into the transistor and this is called second breakdown energy (£s/b). Like 7s/b, the voltage drop due to the reverse Ib flowing through the side resistance rbb’ makes the centre of the emitter strip more biased than its periphery. In this way, a current concentration occurs at the emitter centre. Let us analyse the case of an NPN transistor with diffused base and epitaxial collector, i.e. with constant concentration ND of donors doping particles. Poisson's equation is: 2 m j I5j * \ - ■ i m i Figure 2. The voltage drop resulting from the base current flowing through the cross resistance rbb’ gives rise to a disuniformity of 1/be at the junction, and so to the disuniformity of the current density Je dE/dX=dV/dX=p(x)/f 17] The X axis is normal to the sili- con dice surface, (x) is the charge per unit volume, E is the dielectric constant of silicon. When the collector current is limited to low values, expres- sion ]7] becomes (q being the electron charge): tlE/dX=qNo/i 18] and the electric field behaviour is similar to that shown in fig- ure S for Jc = J',. The voltage Vcb ( = Kcr) is given by the area of the EX graph and is smaller than primary breakdown voltage due to the reaching of critical field Eci. In the presence of significant values of current density Jc. the expression [8] is modified due to the n concen- tration of electrons flowing at the speed V through the deple- tion layer. dEI6X=[q[ND-n)\h 19] where n=Jc/qV At constant Vcb] the area limited by E has to remain con- stant. When Jc increases, the E-X slope varies (J'2) until its sign is changed J'3) and Ea is reached J'c,). At this point ava- lanche local multiplication of electrons occurs with an uncon- trolled current increase— and so a strip is formed with a very high temperature that gives rise to either crystal damage or sili- con melting. Possible crystal defects, metal ions, and junc- tion disuniformities further ex- aggerate this phenomenon. The avalanche multiplication is very fast and very localized so the device remains externally cold. The £»/b behaviour is not in- fluenced by the die bonding quality. High £«/b values can be obtained with a proper geo- metric design to limit the cur- rent crowding and, most of all, by inserting a second epitaxial layer N of intermediate doping between the collector and the substrate. The intermediate layer creates the condition shown in Fig. 6. When the current density in- creases (J'2 ) the electric field at the interface N7N is increased. Before the critical field £cr is reached at the interface, the contribution of layer N becomes significant in sustain- ing the voltage. A further den- sity increase (J'3) reduces the electric field at the interface N /N and the breakdown is not I triggered until the critical field | is reached at interface N/N + . For a good power transistor j with Kcro(St/s)=450V. the cur- rent density J'a corresponding j to £« is of the order of 20A/mm 2 . i.e., greater by a fac- tor 10 when compared to the average current density given | by the ratio between maximum saturation current and emitter The £s/b behaviour is also in- fluenced by the conditions out- side the transistor, Rue. Vbe, L. The base conditions are es : pecially important, as they regulate the crowding phenom- enon. The system most commonly used by power designers to re- duce the £s/b effect during turn off with inductive load is a clamping or snubber' circuit, that limits the voltage peak be tween collector and emitter. The presence of the clamping circuit allows only a minimal amount of the energy stored in the inductance to be absorbed by the transistor, and £s/b j becomes independent of the value of L and practical RBSOA limits may be defined. The presence of high Vce and negative base current, Ib. may give rise at high current to the previously described £s/b phenomenon, even in the presence of the clamping cir- cuit. The multi-epitaxial tran- , sistors show a better behaviour even in the presence of a The reverse bias safe operating area establishes the maximum switchable current with induc- tive load versus clamping voltage in very harsh base con- j ditions that simulate the real | base driving conditions in the I circuits. The temperature is not a major [ factor in the £s/b and so the RBSOA rating can be con- | sidered to be independent of I temperature. Conclusion Second breakdown perform- ance is a function of transistor I technology and cannot always be improved without some trade-off in other parameters. The application conditions have a considerable effect on both 7s/b and £s/b capability. * Sue Cain is with BA Electronics and Ray Ashmore is with SGS. m*.™, 1987 5.27 PRESET EXTENSION FOR FUNCTION GENERATOR by M Kistinger A simple to build, ten-frequency preset unit for the Elektor Function Generator that features an adjustable sweep function, a LED indication, and much more at a very small outlay. The AF Function Generator de- scribed in Elektor India , January 1985, has generated a lot of interest, which is mainly due to the instrument being versatile, relatively simple to construct, and sufficiently ac- curate for a great many appli- cations. The preset extension proposed here is a separately housed, 10-way programmable ancillary intended to drive the generator's VCO input. Fre- quencies that are often used for test and measurement pur- poses can be called up at the flick of a switch, and there is also a facility to successively select all ten of them at variable speed, providing a 10-fre- quency sweep function. Fur- thermore, the extension pro- vides an output signal to trigger an oscilloscope with any one of the ten available frequencies. Ease of control is the key word in this design. Once you have set the ten generator output frequencies with the aid of multiturn presets, you can sel- ect manual operation on the ex- tension and press the single step key until the relevant fre- quency is enabled, as indicated by the associated LED. If the man/auto switch is in the auto position, the VCO voltages are successively output at a rate defined with the speed poten- tiometer and the fast/slow push-button selector. A BCD (thumbwheel) switch is used to select the period of one of the 10 available VCO voltages for triggering an oscilloscope. Standard components are used throughout this extension, which will quickly prove an in- dispensable add-on unit that can save you quite some time in setting the generator’s output frequency. 5.28 elektor india may 1987 Circuit description The circuit diagram of the pro- posed extension is shown in Fig. 1. At the lower left is the power supply, which delivers +5 V for the logic circuits, and + 10 V for the sweep oscillator. ICi, and the VCO output drivers, T.-Tio. The latter voltage is provided by a pre- cision regulator Type LM317 (IC12) to ensure the stability of the ten VCO drive levels. The + 5 V supply is conventionally based on a Type 7805 regulator which can easily handle the 150 mA current demand of the (LS)TTL circuits. With Si set to man . depression of single step push-button S2 causes N- and delay network R10-C1 to provide a trigger pulse to the B input of monostable multivibrator IC2. whose output period is defined with Rn-C2 As Si is open, the pulse at out- put 0 of ICi is passed through gates Nn and N>- and fed to the clock input of counter IC: If Si is in the auto position. Nn blocks the single step pulses from IC2 , and ICi is arranged to be clocked from oscillator IC< via level translator T12 Poten- tiometer Pi and fast slo-.v push- button Si allow precise setting of the VCO sweep speed. Note that Si is actually part of the speed potentiometer, so that turning this fully counter-clock- wise automatically enables manual selection of the direct voltage level from the preset ex- tension, and hence of the func- tion generator's output fre- quency. Counter ICi is advanced by pulses from No, and the BCD code at its Qa-Qd outputs is ap- plied to the XOR gates in ICi . as well as to BCD-to-decimai de- coder ICs. The Type 74LS90 counter is set up to count from 0 to 9, and is reset to state 0 at power-on with the aid of Cs-Ru. The trigger signal for the os- cilloscope is obtained from N13-N14 and Nn-Nn, which function as a 4-bit comparator in conjunction with IC« and a BCD switch for selection of the rel- evant trigger pulse. The output of N19 goes high if the logic state of outputs Qa-Qd on IC! matches that of the A-D lines on the BCD switch. Any one of the 10 outputs of decoder ICs can enable an as- sociated driver stage, whose direct output voltage is defined with a multiturn preset. If, for in- stance. output 9 of IO, goes low, the output of open-collector in- verter N12 goes high. Transistor T10 is turned on, LED D20 lights, and a portion of the emitter voltage is fed to the VCO input of the function generator, via the wiper of P12 and summing diode Dio. The circuit around T11 serves to raise the ground potential of the extension so as to increase the active range of the presets in the analogue out- put stages. It should be noted that this arrangement makes it impossible to feed the preset extension from the generator’s Resistofs ( i 5%l: Ri;Ri = 4K22: 1% R.- - 100K R«:R« 5K6 Ri.Ri.vRjb . flu incl. - 10K Hi. IKS RuRis R» ind.;R«-1K0 Rs;Rs;Rn;Rii = 1K2 Rio;R.. = 220R R.s. Ri. ind. - 270R R«i - 100R Pi = 500K linear potentiometer with SPST switch ISil Pi 5K0 preset Pi . Piiincl.=4K7multitum preset Capacitors:: C.;C* * lOOn Ci = 2|i2; 25 V Ci=10m. '0 V 0;Cr - V; 35 V Ci - 4ji7, 63 V Ci = 1000m. 40 V Ci=47p: 63 V Cm = 10m: 35 V C.i =2*i2; 25 V decoupling capacitors UOOn) as required Semiconductors: Bi = B80C1500 Di . Dio ind.tDu- 1N4148 Dn... On ind.-- LED Du - 1N4007 ICi = 741 ICa = 74121 ICi = 74LS90 ICa = 74LS86 ICs = 74LS42 IC.;IO - 7405 ICs = 74LS02 ICe.lCio - 74LS00 IC» = 7805 plus heat-sink ICii- LM317T Ti. ..Tie incl. -BC549B Tii • BC237B Tii = BCY69 Miscellaneous: P. 100 mA: delayed action. Fuseholder for Pi. Tn = 15 V; 200 mA. Si;Si - push to make button. Si- part of Pi. S« - SPST mains switch. Suggested enclosure: Verobo* Type 75 3007C 1180 > 120 > 40 mml. Prototyping board IVeroboard) BCD Thumbwheel switch supply. Also, observe that the j Veroboard and housed in an After building the circuit, it is | Finally, connect the extension pulse level at the sync out ter- ■ ABS enclosure that can be suggested to adjust the output : to the VCO input on the func- minal is 5V PP with respect to , placed on top of the function voltage of ICu Use a DMM and tion generator, and adjust the 10 the extension ground potential, j generator or the associated set P; for a reading of 10.00 V. j presets for the test frequencies not that of the function gener- sweep generator. Turn Pi to man and check you require, ator. LED Dai serves the double Although not shown in the cir- whether operation of Si causes J 77 , purpose of raising the base cuit diagram, the supply lines to the LEDs to light in succession. 1 potential of T 11 and functioning the logic circuits should be Turn Pi to auto and check as the on/off indicator of the decoupled with lOOn capaci- whether the sweep speed can preset extension. tors. Keep the wires to the be adjusted with Pi and Si. If switches and the speed poten- necessary, adapt C3 or C« to tiometer as short as possible, define the sweep speed. Turn Pnnctri irtinn and The frec I ue ncy indication LEDs Pi back to man and use a DC- J can be fitted in a neat row on coupled oscilloscope to see Function generator Elektor | setting up the front panel, complete with whether the VCO voltages are India, January 1985. The proposed extension circuit numbers 1-10 for easy refer- all stable and free from digital Sweep generator: Elektor 1 is readily built on a piece of ' ence. noise and ripple. I India. December 1985. 5.29 LINKWITZ FILTERS A brief look at the theory and practice of passive and active Linkwitz cross-over networks. An analysis by Siegfried Linkwitz in the January 1976 issue of the Journal of the Audio Engineering Society shows that conventional cross-over filters have a negative effect on the radiation pattern of a multi-way loudspeaker system as regards both directivity and amplitude. 1 On the basis of his research. Linkwitz proposed a new type | of network that gives a uniform j | radiation pattern and constant amplitude. This filter, which is essentially a Butterworth-de- rived type, was first described by Riley and is, therefore, some- times referred to as a Linkwitz- Riley network. For simplicity's sake, the follow- ' ing discussion is based on a I two-way loudspeaker system, j For optimum results, Linkwitz j ! suggested that the filter must j meet three requirements: • there must be no phase shift ] between the outputs of the loudspeakers at the relevant cross-over frequency to pre- vent an upward or downward | displacement of the radi- I 1 ation pattern; • the signal attenuation at each | filter output must be 6 dB in- ! stead of the usual 3 dB to pre- j vent peaks in the sums of the : signals; • the phase shift between the 1 j output signals must be con- I stant at all frequencies to re- tain the symmetry of the radiation pattern above and | below the cross-over fre- quency: this condition is conveniently met by the use ' of symmetrical filters in both | the low-pass and the high- pass sections. Linkwitz found that these re- | quirements can be met by cas- cading two identical second- order Butterworth filters. | Higher-order types may, of course, be used, but in practi- | cal applications these are less j interesting. It should be noted j that in any case the filter must be an even-order type, since each order causes a phase shift I of 45° at the cross-over fre- ' quency. 5.30 .l.kl« indie may 1987 Fig. 1 shows the amplitude and phase shift behaviour of a But terworth filter, and Fig. 2 those of a Linkwitz-Riley network. Note the 3 dB peak of the But terworth filter. This can not be obviated by increasing the separation of the cross-over fre- quencies of the low- and high- pass sections, because this would violate the first require- ment of zero phase shift be- tween the outputs. For clarity’s sake, the two characteristics are combined in Fig. 3 to highlight the difference between them. The Linkwitz curve is rather more rounded in the vicinity of the cross-over frequency, and starts falling off somewhat earlier. The slightly different phase shift of the two filters should also be noted. The foregoing discussion is true only if the signals are [ sinusoidal. The pulse (or step) \ response of the Linkwitz filter { causes the same problems as i that of a Butterworth filter, j assuming that both filters have separate low- and high-pass sections. Even a Linkwitz filter is therefore not perfect. A practical filter A Linkwitz filter may be de- [ ' signed as a passive or as an ac- tive type. The circuit diagram of J an active design is shown in I ‘ Fig. 4: this may be constructed j J on the printed-circuit board j shown in Fig. 5. Note that this ! board is identical to that used | for the electronic cross-over j network published in the ! September 1984 issue oiElektor J Electronics. i The circuit of Fig. 4 is for a J ( three-way loudspeaker system. I The network has cross-over fre- quencies of 500 Hz and 5,000 Hz I | and roll-offs of 24 dB per octave. ! i Stage A. serves as a buffer for j the input signal before this is | split three-way. The low-pass section is formed by As and As; j the middle-frequency section by Plj and As (high) and As and | Aio (low); and the high-pass sec- tion by An and A12. Each sec- tion is provided with a poten- tiometer for setting the level of the output signal (Pi, P2, and Pj respectively), and a stage to -buffer the output (A 1. A3, and As respectively). The power supply lines are stabilized by voltage regulators IC7 and ICs. The cross-over frequencies may be altered with the aid of Table 1 (any frequency) or Table 2 (the 17 most likely fre- quencies). The values in Table 2 have deliberately not been rounded off to the nearest Fig. 1. Butterworth network: amplitude and phase characteristics over the audio frequency range. The fat line represents the sum of the outputs of the filters. of the outputs of the filter sections. standard E12 or E24 value. The sections may also be given a slope of 12 dB per octave by using Ae, As. A.o, and Ait as buffers. Resistors Rio, Rn, Ris, and Rio, as well as capacitors Cs7, Css, Cjs, and C36, are then replaced by wire links, while Ru, RlS, Rss, R 2 3, CS3, C 2 4, C31, and C 32 are omitted. The circuit may be adapted for use with a two-way system by the omission of the entire middle-frequency section, ex- cept for As which is housed in the same package as As. If the slope is changed to 12dB per octave, the connexions to one of the loudspeakers must be reversed, because the phase shift at the cross-over frequency is 180° here. In a three-way system, this should be done at the middle-frequency speaker; in a two-way system at the tweeter. A passive filter may be con- structed as shown in Fig. 6. The values of the actual compo- nents used should be as close as possible to the calculated ones, otherwise the filter will become a cross between a Fig. 5. The printed-circuit board for constructing the Linkwiu filter of Fig. 4. LinkwitZ and a Butterworth 1987 5.31 Fig. 4. Circuit diagram of an active Linkwiu filter. lype. If the filters are given a I ascertained as detailed in Loud- 12 dB per octave slope, the con- j speaker Impedance Correction nexions to the middle-fre- (Elektor India, June 1986), quency loudspeaker (in a three- way system) or those to the j tweeter (in a two-way system) i should be reversed. The loudspeaker impedance j must be corrected in a manner j that ensures that it is constant j and ohmic at the cross-over fre- quency. The corrected im- j pedance of the loudspeaker, R in Fig. 6a and 6b, should be RIMS: COUNTING ATOMS by Dr Kenneth W. D. Ledingham, Department of Physics and Astronomy, University of Glasgow Resonant Ionisation Mass Spectroscopy (RIMS) is a unique, ultra-sensitive analytic technique which can detect down to the level of a few atoms. It is applicable to any sample, whether solid, liquid or gas and can be used to assay every element in the periodic table apart from helium and neon, as well as any stable or radioactive isotope. It is likely to find important applications in fundamental and applied physics, and to become a valuable tool in the semiconductor industry and The need to develop new analytic ways to measure ultra- trace quantities of elements in various substances is becoming urgent in many branches of science, engineering and medicine. There are already many sensitive analytic tech- niques. including neutron or photon activation analysis, in- ductively coupled plasma spectroscopy, atomic absorp- tion and various kinds of mass spectroscopy, particularly secondary-ion mass spec- troscopy (SIMS). The sensitivity of these techniques for trace analysis is usually limited to the order of parts in 10' or 10®. In the last few years problems have arisen that require ultra- trace analysis at the previously unheard-of sensitivities of parts in 10 8 to 10' 2 or even further. Already three areas which re- quire such analysis have been identified and as techniques are developed many more ap- plications are likely to become apparent. Firstly, it is essential to reduce The minimum detection limit of impurities in silicon if improve- ment, especially in miniaturis- ation, of the semiconductor manufacturing process is to be maintained. Secondly, Pro- fessor M. Baxter of the Scottish Universities Research and Reactor Centre, near Glasgow, has speculated whether there is a health risk from the presence of very low-activity p emitters in the environment. They are very difficult to monitor because they are likely to be below the sensitivity range of conven- tional nuclear counter tech- niques. Finally, the presence of trace in diagnostic amounts of certain elements in human body fluids and tissues is considered to be essential to health. This is a poorly under- stood branch of biochemistry and widely divergent figures for trace metal concentrations in apparently healthy people have been published. But there is growing evidence that many of the studies are flawed by gross analytic inaccuracies and that new. reliable techniques are necessary at sensitivity levels of parts in 10 8 . During the middle and late 1970s the possibility of applying laser techniques of single-atom detection to ultra-trace analysis attracted interest. The tech- nology had been pioneered largely by Professor V. S. Letokhov of the Academy of Sciences in Moscow and Pro- fessor G S. Hurst of Oak Ridge National Laboratory, USA. Res- onant Ionisation Spectroscopy, RIS as it has come to be known, can detect one atom of a specific type in a background of 10'* others in gaseous phase. The implications of this degree of sensitivity for many disparate fields of research are likely to be enormous. Resonant Ionisation Spectroscopy With the development of in- tense, tuneable, pulsed lasers the simultaneous absorption of several photons by a single atom or molecule to produce a free electron and a positive ion became experimentally feas- ible. In the simplest RIS process, a pulsed laser is tuned precisely to the wavelength required to medicine. excite the atom or molecule from its ground state of energy to an excited state that is unique to the element under study. A second photon, of the same wavelength and from the same laser pulse, interacts with the atom in its excited state and causes an electron to be re- leased from it. thereby creating a positive ion. This process can be made more selective by ad- ding further resonant steps in the excitation process, using a second laser tuned to another frequency. Five different laser schemes, represented in the first illustration, can ionise all the elements in the periodic table, except helium and neon. From left to right in the diagram they are: (a) A(u>i,aiie~)A' This reaction means that two photons of the same wave- length (that is, with angular velocity w>) create the ion (b) A(2a.,a»*)A- The laser wavelength is fre- quency doubled into the | ultra-violet and then mixed with the fundamental to create the ion pair. (c) A(&mua,a>i or iu*e‘)A' In this process three photons are absorbed with two colours being involved, indicated by um and w>. (d) A(2u.i. uiuuje )A' One colour is frequency doubled (2un) and another | photon of a second colour is j absorbed as well as one of the original photons. (e) Afoim.., c* )A- In this case usually three photons of the same colour are absorbed to create the The second diagram is the periodic table of elements with one of the five schemes being ascribed to each, after Pro- fessor Hurst. In the early days of the technology, the electrons created in the resonance pro- cess were detected by ionis- ation or proportional counters. Soon, however, it became ob- vious the ultra-trace isotopic selectivity was needed, too, so mass spectrometers were in- troduced to detect the positive ions. Although both magnetic sector and quadrupole mass spectrometers have been used by different research groups, the arrangement preferred now includes a time-of-flight mass spectrometer. Resonant Ionisation Mass Spectroscopy When laser techniques are used to detect ultra-trace amounts of elements or iso- topes in a substance or matrix, three separate steps are in- volved. A typical laser time-of- flight mass spectrometer is shown in the third illustration, indicating the steps. Firstly, a pulsed, charged, argon beam or a neutral argon beam, ablates or creates neutral atoms from the surface of the solid sample to be assayed. Ideally, the atoms created should be accurately representative of the solid under analysis and to date argon ablation has been shown to be largely matrix-free. This technique is now considered to be superior to the laser ablation technique, which is a high- temperature method known to cause matrix problems be- cause it favours the easily etekior India may 1987 5.33 1 II in IV V VI VII VIII 0 1H 4B..I 48 •*<6C *♦ -VtnV-. y *-So* '11 NaN WW N ;:i2Mg> 013AI-0 ////< ’+* 15 P V *.16's’V f 4*7(2.*+ '////. x 20Ca> sSSSS VAV X 21 ScO sSXNX NW\ >22Ti\ c\\\S '24Cr\ \W\\ '25Mn N PWP xxxp K28Ni\ \\x\ N 29 Cu 30 Zn '//// '31 Ga' '//// .32 Ge: ::::: ::: *33 As* * 34Se*+ 4*35 Bri-*- 4*36*K*r*. 'ft**' ////<> ^\\N TaOZit- SS1\\ V41 Nb x ^\\\ -,42 Mo/ X\\\x E§Tc]= x\\\\ '44 Ru\ \\\\ N \XX\X '45Rh\ \X\P : 46Pd i /49ln/ '//// \\V\'- x 50 Sn x \\W N *4531*4*4 -*MXe*4 •47 Ag-i 48 Cd! :51 Sb: : :52Te i 80 Hg ; 72Hf; jlmab; :”/4W> 1- 75 Re; _760s : I ; 78 pt : ; 79 Au 7777 , //// 87 Pb 83 B. 84 Po H 'M/'. yet, team 0 '63> es'/j/GB Jb'Zpy '/>67/ //Ho'- '70 sYb S '71^ T [\92\j r Am T \99\ Q one of the five schemes of the first diagram ascribed to each. I vapourisable materials. | SIMS, already mentioned, also uses an ion beam to ablate the | surface of the sample but I analyses only the charged ions, I which are created in numbers some two or three orders of j magnitude fewer than the neutrals. Because they are charged, these ions are emitted at a rate that is a function of the chemical composition of the surface. RIMS and SIMS are I made quantitative by making j comparisons with well charac- ■ terised standard samples, so if there are any matrix problems, j any quantitative analysis is likely to be inaccurate. 1 Having created a cloud of 1 vapour above the target, RIS lasers then selectively ionise atoms of the chosen element in | the vapour cloud, which are j subsequently accelerated into 1 the time-of-flight mass spec- trometer. Secondary ions created by the ablation process can be rejected by electrostatic fields or by varying the time be- tween the ion-beam pulse and the RIS laser pulse. The normal laser arrangement to achieve | total elemental coverage is an j Nd:YAG laser powering two dye lasers, one of which has fre- quency doubling facilities, i Typical lasers of this kind operate with pulse lengths of several nanoseconds at rep- etition rates of some tens per ! second. The transverse spatial I dimensions of the beam are typically a few millimetres. One of the strengths of RIS is that the I photo-ionisation process can be j made almost 100 per cent ef- ! ficient, that is, it reaches satur- ation. By saturation of the RIS ( process, we mean that every j atom of a quantum selected j species which was in its ground state before being subjected to the photon field of a pulsed laser is converted to a positive ion and a free electron during j the short duration of the laser pulse. Because saturation oc- j curs when laser fluences, by j which we mean energy per I unit area, are typically about | 100 m] cm ', conventional com- mercial lasers require modest ( focusing of a 3-mm beam. It is | hoped that RIMS will become a [ routine ultra-trace analytic tech- nique, so a short analysis time is desirable, of the order of minutes. For this purpose the low repetition rate of Nd:YAG lasers (30 sec ’) is a limitation. Two of the severe limitations of 5.34 etaktof tndia may 1987 Periodic table of elements, with (After Professor Hurst. I conventional mass spec- troscopy can be eliminated when tuned lasers are used to produce the ions for mass analysis. In a conventional mass i spectrometer the ions to be 1 analysed are normally electron- beam induced, so molecular in- terferences and isobaric effects I cannot be avoided. A mass j spectrometer cannot easily j distinguish between CO and I N2, for example. This is a 1 phenomenon known as mol- ecular interference. Nor can it distinguish between *°Ca, 10 K and 40 Ar because they are isobaric; that is, they have simi- lar masses. These ambiguities are avoided when RIMS is used. The final step of the RIMS tech- tuque is to count and measure the mass of the laser-induced ions, using a time-of-flight (TOR mass spectrometer. A TOF in- strument is a non-magnetic system in which the ions first accelerate through a series of closely spaced electrodes and then pass through a field-free region (D) of Considerable di- mensions, of the order of one metre, to be detected by an ion detector such as a chan- neletron. In its simplest form the transit time (r) of the ion in the field-free region is pro- portional to the length of the field-free region and to the square root of the ion mass (m). For an accelerating voltage of 1000 V, with D equal to 1 m, f is about 20 fis for a singly ionised mass of 100 atomic mass units. There are several advantages to be gained by using a TOF mass spectrometer: firstly, entire mass spectra can be accumu- lated in a very short time and an entire spectrum can be re- corded for each laser pulse; secondly, TOF systems measure isotopic ratios very ac- curately, because they measure them under identical con- ditions; finally, the accuracy of a TOF spectrometer depends on electronic circuitry instead of extremely accurate mechanical alignment, so it is simpler to make. The time-honoured dis- advantage of TOF instruments is low resolution due to the poorly defined spatial and temporal character of conventional ion formation. But that scarcely ap- plies when the ions are formed by lasers, because the laser spot has a tight focus and the laser pulse is so short, between 5 and 10 ns. In the last year a number of groups in the USA, the Soviet Union and Europe have been •set up to exploit the sensitivity of RIMS, Already it is claimed that the technique is capable of detecting impurities at the level of 1 part in 10’° in a routine analysis time of 5 minutes. Future Development The design of the RIMS instru- ment so far described is by no means optimised. A number of promising lines of research have yet to be investigated which may lead to better sensi- tivity. Each of the three steps in the RIMS process will be con- sidered, to see whether im- provements are possible. Block diagram of a TOF resonant in its principle of operation. During the past few years, a great deal of attention has been paid to photon, electron and ion ablation of solids. At present, argon ablation of the sample is the most popular technique, though recent developments in metal-ion beams such as those of gallium and caesium might increase the ion-sputtered yield per unit incident current. What is not in question is that these metal-ion beams can be fo- cused to far smaller spots than an argon beam, down to submi- cron focal dimensions, so they are likely to be of great import ance in future for precision scanning of sample surfaces. Over the next few years, it is im- provement of the RIS step that is likely to contribute most to greater sensitivity. While an Nd:YAG pumped dye laser system has a repetition rate of 30 pulses s ', copper vapour lasers have recently been developed, in particular by Oxford Lasers (UK) which have a repetition rate of 6500 pulses s ', capable of pumping dye lasers to provide saturation in- tensities. This is likely to in- crease the efficiency of RIMS considerably, especially in detecting minute quantities of the actinides, recently demon- strated by Professor Kluge and Professor Trautmann of the Uni- versity of Mainz. At present, however, there are electronic difficulties in handling data at such a large rate. The problems arise from not having enough storage capacity and from the transfer rates of available high- speed transient recorders. One possible improvement in sensitivity may be understood by considering the stepped photo-ionisation process in the final diagram. In (a) an electron in its ground state absorbs a photon and is promoted to an excited state with a cross-sec- tion that is typically about 10 " to 10''* cm 2 . Another photon is absorbed and the excited atom is ionised. The photo-ionisation step is characterised by a small cross-section oion of 10 " to 10 " cm 2 and therefore by a large laser fiuence being needed to achieve saturation. The fiuence is achieved by focusing, so that the volume of interaction with the ablation cloud is small. If, however, pro- cedures (b) and (c) are adopted the probability of ionisation is greater by two or three orders of magnitude. In process (b) the atom is excited to close to the continuum (a Rydberg state) and then finally ionised with high efficiency using a pulsed electric field. Another possibility of improve- ment is shown in (c) where the final ionisation step is to a so- called auto-ionisation state, above the ionisation level but having a large cross-section. Considerable research is necessary to identify the auto- ionisation states in a number of elements before this powerful procedure can be adopted. If processes (b) and (c) can be used then the saturation fluences of the laser are greatly reduced, so that the beam need not be focused. The volume of interaction is then bigger. Future Applications One exciting aspect of this technology is that there are likely to be important appli- cations in both fundamental and applied physics. In connection with fundamental physics, ap- plications of RIMS to solar neutrino experiments, double I aiappeu pnviv-n»iw«iv» _ photon ; is excited to a Rydberg state and finally ionised by a pulsed electric field, (cl The final step is to an auto ionisation state, to give a I large cross-section. beta decay, baryon conser- vation and magnetic monopole searches as well as detection of quark atoms and superheavy atoms are being actively pur- sued. In particular, a detector based on the ”Br (v, e) *'Kr to measure the 'Be neutrino source in the Sun has been shown to be feasible because the long-lived (2 x 10 s year) "Kr can now be counted with RIMS In applied and commercial science, the applications of RIMS are likely to be very far reaching. In the semiconductor and electronic industries RIMS can identify impurities that restrict performance of high- speed. high-density integrated circuits. The technique can ex- tend downwards the present minimum detection limits for contaminants by perhaps three orders of magnitude or greater. In the medical field, early diag- nosis of certain diseases by using trace-element concen- trations in body tissues and fluids is a very attractive possi- bility but must be carried out in a non-invasive way by using as small quantities of material as possible. Finally, RIMS can assist in selecting sites for stor- ing hazardous nuclear wastes by using ground-water dating techniques as well as allaying public concern by ensuring that environmental monitoring be made as sensitive as poss- ible. 0314/6 NEWS • NEWS • NEWS • NEWS • NE1 Advanced universal digital filter A real-time universal digital filter, developed by Fern Devel- opments for use in speech pro- cessing, audiology, psycho- acoustics, electrophysiology, and geophysics, offers linear filtering capabilities which are said to be superior to those that can be achieved with conven- tional analogue techniques. The benchtop EF8, based on a design conceived by the Medical Research Council, is a 512-coefficient finite-impulse- response, non-recursive filter that offers an unlimited number of totally different filtering ac- tions. the anti-alias (pre-pro- cess) and post-process sections use high-precision programm- able low-pass filters. The filter unit has an operating bandwidth of 0-30 kHz, attenu- ation rates of typically 4000 dB per octave, and up to 512 I weighting coefficients for sym- 1 metrical responses. I Fern Developments Ltd i 7 Springburn Place j College Milton North | Glasgow G74 SNU Grundig do it with robots Helping Grundig on the road to success is a new robotic VCR production line which, the company says, is in advance of any other in the world today. Making VS400 machines, which will retail in the UK at around the £400 mark, the production i line cost over £5 million to in- i stall. It was designed and built i entirely by Grundig engineers j and took a mere nine months to j complete from putting pen to paper in the drawing office to the first complete machine coming off the 130-metre long production line. When on full production, the automated plant is expected to produce at least one million VCRs per year: each one taking just 35 minutes to make plus another two hours in soak testing. Each machine goes through 87 work stations and through automatic quality tests on its way to completion. The new VCR line is just the first of a planned series of developments which will con- tinue to radically change Grundigs approach to video production. Grundig International Ltd 42 Newlands Park London SE26 5NQ SMA assembly of PCBs The WS1500 combination work- station from Surface Mounted Production Systems Ltd is in- tended for the surface-mounted assembly (SMA) of printed- circuit boards (PCBs). It incor- porates a precision dispenser, vacuum pick-up, infra-red soldering unit, and a soldering iron. PCBs up to 7 x 4 inches can be accommodated. The WS1500 enables prototype design and development, single or small batch pro- duction and repair work on surface-mounted circuit to be carried out at one workstation, j It is priced at less than £2000. Surface Mounted Production Systems Ltd Unit 5 Sandbank Industrial Estate Dunoon PA23 8PB 5.36 WHERE ELECTRONIC MESSAGES HAVE THE EDGE by Barry Foxj The new age of infor- mation technology is founded on one simple truth. It is quicker, easier, and cheaper to send pulses of electricity down a telephone wire or over a satellite link than it is to transport people or pack- ages by road, sea or air. The telex service has until now been the standard means ot sending text. Telex is a reliable war horse but has its own snags. The equipment is bulky and expensive, trained operators are needed to send mess- ages, and the service relies on dedicated lines — that is to say special circuits designed to carry telex pulses rather than speech. It is still not widely recognized that almost every personal computer, either desk top or port- able, can be used tor electronic mail through one of the available ser- vices. It is the modern alternative to sending cor- respondence by telex. Text is sent from one computer to another along a con- ventional telephone line via a central message- handling computer. Already script writers, translators, bankers, ■journalists, and lawyers are using electronic mail to send text from home to office. Sales teams use it to keep in touch with their headquarters while mov- ing round the country. Mu- sicians use it while on tour. Electronic mail terminals can work equally well from an office desk or a hotel room far away. Digital pulses In its simplest form, a home computer sends messages either to the screen or to a printer. If it is programmed with ad- ditional communications software, it can send similar messages from its output - usually an RS-232 — socket. This output is in the form ot a stream ot digital pulses, similar to telex, but much taster. They can be sent down a short wire cable to a matching computer sys- tem. This is how several computers are networked in an office. The pulses will not travel reliably down a conventional telephone line so they must first be converted into audible tones which the telephone network handles like speech. | A special device called a modem — short for modu- lator, demodulator — is needed to convert the computer pulses into . sound tones. It is connec- ted between the computer output and the telephone line socket, while a com- | puter at the other end of j the telephone line has a matching modem. I This converts incoming • tones back into digital pulses which are then displayed on the com- | puter screen or printed on to paper. Four services Electronic mail provides a mail-box system into I which messages can be I dropped by one user to j be picked up later by another. A host computer I handles the messages [ with a system of passwords j to ensure that messages can only be picked up by | the people to whom they are addressed. In Britain there are four electronic mail services. The most successful so far is Telecom Gold which is run by British Telecom and has around 30 000 sub- scribers. Rival services are offered by Easylink— a sub- sidiary ot Cable and Wireless; Comet from Istel — a subsidiary ot British Leyland; and One-to- One— a private company now owned by United States Telecom's company Telesis. Each of these ser- vices otters a message drop facility. When someone working from home or a hotel room wants to contact an office, he or she calls the relevant electronic mail telephone number and sends a message which is held in a message ser- vices computer. Later, the person at the office calls the same electronic mail number and reads the message off the computer The text can be viewed on screen, stored on mag- netic disk for subsequent word processing, or printed direct on to paper like a telex. Any office wanting to use electronic mail should first find out what services are on offer. The Telecom Gold service in Britain is derived from the ITT Dialcom sys- tem developed in the United States of America. It is now used in over a dozen countries around the world, and is proving increasingly popular. How to buy Most businesses that decide to install an elec- tronic mail system will find it cheaper in the long run to buy the hardware and software through a dealer whose purchase price in- cludes the cost of instal- ling the equipment, get- ting it up ahd running, and teaching the staff how to use it. Once a system has been installed, staff may very soon won- der how they ever lived without it. 5.37 MSX EXTENSIONS - 5: EPROM PROGRAMMER (2) The supporting software for the programmer is an EPROM-resident block of Z80 machine code that provides a deluxe menu, help pages, a built-in test routine, and, of course, EPROM status information plus error reports. After last month's discussion of the programmer hardware, we will now study the way it is actu- ally controlled from the MSX computer. To begin with, how- ever. we will briefly detail the workings of intelligent pro- gramming, already hinted at last month. The intelligent pro- gramming algorithm As the holding capacity of their EPROMs increases, it is logical for manufacturers to devise pro- gramming methods that enable loading these devices within an acceptable time. Should the "old" 50 ms per address pro- gramming method apply to, say, a Type 27256 EPROM (32 K x 8), roughly half an hour would be needed for the device to be completely loaded. Intel, Fujit- su, National Semiconductor, and other leading EPROM manufacturers have, therefore, come up with various versions of an intelligent programming algorithm to speed up the loading process. As its name implies, this method relies on the use of a microprocessor, ruling out the possibility to use timers with a fixed output period for the generation of the programming pulses. The flow- chart shown in Table 4 shows that the essence of the in- telligent algorithm lies in the raising of Vcc from +5V to +6 V. and the variable length of the programming cycle. The program-and-verify loop can only be left with the byte either correctly programmed, or still incorrect after a 25-pulse cycle. Therefore, with relatively few programming pulses required for a byte to verify correctly, the 5.40 slsktoi .nd,. m., 1987 value of variable x is relatively low. and less time is needed for the address to be loaded. Fol- lowing the variable number of programming pulses, an ad- ditional pulse of 3x ms ensures that programmed databytes are absolutely stable in the EPROM. At this stage, an example might help to illustrate how the algorithm works: A specific byte requires 9 pulses for it to be stored cor- rectly in the EPROM. The pro- gramming cycle thus takes (9xl)+(3x9)=36 ms. Figure 8 illustrates that a pro- gramming cycle can be quite long. In fact, intelligent pro- gramming is not necessarily faster than normal (50 ms), fast-1 (20 ms), or fast-2 (10 ms) timing arrangements, since the worst case cycle duration is 25 +(3 x 25) =100 ms. In practice, however, you will soon find that newly purchased, intelligently programmable EPROMs gener- ally require only the minimum pulse time of 4 ms per address for reliable loading. Returning to the previously mentioned Type 27256, 3 minutes or so then suffice to completely load this device. The intelligent programming methods adopted and rec- ommended by Intel ( intepgent programming, sic) and Fujitsu [Quick Pro™) differ only marginally as regards the dur- ation of the programming pulse, the number of iterations befoie the EPROM is rejected as faulty, and the pulse multiplication factor. National Semiconduc- tor's algorithm, however, is based on the use of 0.5 ms pulses, a maximum iteration of 20. no multiplier, and a V PP level of 13 V instead of the more usual 12.5 V. This MSX EPROM programmer does not support National's algorithm, but none- theless gives good results with their chips. As could be expected, the timing of the programming cycles is interrupt-based and jointly controlled by the CPU in the computer and the CTC in the I/O & Timer cartridge. The control program arranges for timer T2 in the CTC to provide the number of programming pulses required to successfully load a byte into an EPROM ad- dress. Iteration and pulse multiplication are effected in accordance with the flowchart shown in Table 4. Extensive tests have shown that the adopted algorithm gives satis- factory results with the vast ma- jority of intelligently program- mable EPROMs. Although not expressly indi- cated in the flowchart, the con- trol program and the CTC ensure that EPROM data and address lines are stable before any write action can take place. For this purpose, timer T0 in the CTC provides 4 P s long delays as detailed in last month’s instal- ment of this article. Program description An MSX compatible micro can have up to 4 primary slots, numbered ft 1, 2, 3, each with a memory capacity of 64 Kbytes and subdivided in 4 pages of 16 Kbytes. It is also possible for a slot to be expanded, which means that it comprises four sub-slots X-ft X-l. X-2 and X-3. In theory, therefore, there can be a maximum of 16 slots identified as ftft up to and including 3-3. Since the Type Z80(A) CPU is an 8-bit microprocessor, its addressable memory area is 64 Kbytes, that is, four pages, but these can be part of any (ex- panded) slot. It is, for instance, possible for the system to operate with page 0 from slot ft page 1 from slot 2, and pages 2 and 3 from slot 3-2. The absolute address ranges are thus: page 0 = 0000-3FFF; page 1 - 4000-7FFF; page 2 = 8000-BFFF; page 3 = C000-FFFF. Pages can be swapped and switched on and off by means I of particular system commands, which will not be gone into in this article. Page 0 is usually ' reserved for the MSX BIOS (Basic Input/Output system), and page 3 for the system stack l and scratch blocks, variables, the keyboard buffer, etc. At | power-on, an MSX computer in- variably examines pages 1 and 2 in all slots for the presence of (E)PROM-resident programs, j which are immediately started if a particular identification | code is found in the first 16 ad- dress locations. If such an iden- tifier is not found, the BASIC ROM on page 1 is enabled, and the machine boots up accord- | ingly. The control program for the EPROM programmer comes in i the form of a ready-pro- I grammed EPROM Type 27128 (16 Kbytes), available through our Readers’ Services under number 552. This EPROM is in- j serted in the socket on the car- j tridge board for MSX com- puters, described in Elektor \ puters, described in Elektor ! India, March 1986. In the follow- I ing section we will set out how I to correct all add-on units to I make a functional set-up. The programmer software im- mediately runs from page 1 at power-on. After completing the I necessary initialization rou- tines, the program finds out which slot has RAM in pages 1 and 2 for use as the EPROM data area (32 Kbyte maximum size, 4000 -BFFF), it copies part of itself into the highest poss- ible RAM area on page 3, that is, 5.41 ii inserts itself between the stack and string & scratch blocks. After all this has been done, control is returned to the computer's normal start-up pro- cedure, which means in most cases that BASIC will be started. The EPROM software can now be run by typing CALL EPROMx, where x is the car- tridge address area, 0, 1, 2, or 3. The program, when called, automatically selects the appro- priate slot(s) for the RAM buffer, and then switches back to where it came from with the aid of routines on page 3. All switching between RAM and EPROM resident subroutines in the programmer is invisible to the user, and makes it poss- ible for the proposed soft- ware to run on any MSX com- puter equipped with at least 64 Kbytes of RAM. Extensive use is made of vector- addressing, and all keyboard and screen input/output is routed via the BIOS on page 0 To make sure that data for or from the EPROM is not over- writing the system stack, or possibly the RAM-resident por- tion of the control program itself, it is a good idea to check whether there is enough room for your data by typing PRINT HEX$(FRE(0) + &H8000). The address returned should be higher than the top location you need, observing that part of the available memory is used for the string and stack blocks, which extend downwards. Those MSX users in possession of a computer with a disc drive may have to limit the DISK- BASIC workspace somewhat by holding down the CONTROL- key during power-up as a means of telling the sys- tem there is but one virtual disc drive available. Simi- larly, holding down the SHIFT key disables the disc unit altogether. ) Command summary Although the proposed program is extremely simple to use, it is none the less rec- ommended to study this brief summary of the available func- tions, commands and options. After typing CALL EPROMx, you should see the welcome screen. Pass on to the help pages with EPROM data and program information by press- ing any key. You can leaf through the help pages by 5.42 .MOO, India p„y 1987 pressing the appropriate cursor movement keys. The command input screen can be called up at any time by pressing the space bar. The following keys are used during the command input Cursor ' and . select the item you wish to work on. Key H returns you to the help Key P causes the screen con- tents to be dumped to a printer (make sure this is properly con- nected, else you will get a NO PRINTER error). Key T runs a test program that causes all functions on the pro- grammer to be successively enabled with aid of CTC inter- rupts, indicated by the flashing PGM LED. Make sure that jumper J> is not installed, and never run the test with an EPROM inserted in the ZIF socket. The space bar selects the various options for the com- mand items (toggle function). Key S causes the program to start executing your set of com- mands. Always make sure that the command screen shows what you want before pressing & Key I enables the storing of BASIC programs in EPROM. The software automatically arranges for the correct in- italisation of the memory begin & end, and, EPROM begin & end addresses. Link addresses are automatically adapted to enable the BASIC program to be run from EPROM With reference to Fig. 9, these are the various parameters you need to define before the pro- grammer does what you want it to do: EPROM TYPE and PRO- GRAMMING VOLTAGE: con- sult Table 1 or the relevant help page and use the space bar to select the appropriate EPROM type: notice that EPROM BE- GIN & END change in accord- ance with the holding capacity of the relevant EPROM type. It is possible to program part of an EPROM by keying in the rel- evant hexadecimal address range. The program accepts en- tries in hexadecimal only, and produces an error message if you try to define an impossible [ address range, or if the EPROM BEGIN & END entry is not in accordance with the MEMORY BEGIN & END entry. Example: you want to program the first half of a Type 2764 (8 Kbyte): EPROM BEGIN = 0000; EPROM END = 0FFF; MEM- ORY BEGIN 4000. MEMORY END - 4FFF. BLANK CHECK should be a | fairly well-known facility; it checks, with the aid of EPROM BEGIN & END, whether the j specified address range con- | tains only bytes FF. indicating } that data can be loaded there. PROGRAM speaks for itself. This function uses both EPROM BEGIN & END and MEMORY BEGIN & END VERIFY checks whether the EPROM contents and the RAM buffer contents are the same, and evidently uses EPROM BE- GIN & END and MEMORY BE- GIN & END to determine what , address ranges are to be com- READ AND RUN CHECKSUM loads the data from the EPROM into the buffer and adds the values of all bytes to produce a 16-bit checksum. DISPLAY MEMORY offers the user the possibility to load the EPROM contents into the com- puter for examination on the screen (hexadecimal and ASCII format, 8 bytes per line, preceding address). You can not alter the displayed bytes. PROGRAM MODE simply selects normal, fast-1, fast-2, or intelligent programming as ap- propriate for the specific type of EPROM. Consult Table 1 or the relevant help page. ADDRESS COUNTER at the lower end of the screen is a 16-bit counter that keeps track of the EPROM location cur- rently read from or written to. The RESULT line at the bottom of the screen can be used to display the following messages (H returns to the help pages): ADDRESS ERROR is a general message to tell you to re-do the EPROM BEGIN & END and/or the MEMORY BEGIN & END entry before pressing S again. BLANK reports that the stated address area contains only bytes reading FF. The EPROM area is not copied into RAM. NOT BLANK reports that one or more bytes in the specified EPROM area do not read FF. The address counter displays the first address encountered, and the program is halted. READING. COMPLETED speaks for itself. The contents of the EPROM are available for examination with DISPLAY MEMORY. For modification, you will probably want to resort to BASIC or a suitable utility package. VERIFIED reports that the verification routine has com- pleted without finding errors. VERIFY ERROR indicates that one or more differences exist between the contents of the EPROM and that of the RAM. The address counter displays the first incorrect address en- countered, and the program is halted. REPROGRAMMABLE indicates that a verify error was found, but the relevant byte is reprogrammable, i.e„ any of its bits reads logic 1 when it should be logic 0 Logic low levels in EPROMs can only be changed into logic high by ex- posing the chip to a dosage of ultra-violet light. NOT PROGRAMMABLE re- Table S Port C ci EPROM READ | VE RIFY | 1 2716 0B 08 + Vpo 1 2732 OF i 0C + Vpo 08 + Vpo 08 + Vpo [ 48 + Vpo 27128 27256 • 27512 2516 2532 j oc + Vpp OB | pons that the address indicated by the address counter can I not be loaded correctly, even after applying 25 programming j pulses (see Table 4, intelligent j prgramming only). | EXECUTI ON STOPPED is displayed in response to the pressing of the RESET switch on the EPROM programmer. DEVICE I/O ERROR indicates that the computer is not receiv- ing interrupts from the car- tridge, which is possibly set to ! the wrong I/O address. NO PRINTER is a message that speaks for itself. IL LEGAL COMMAND ORDER informs you to re-do the YES/ NO setting of one or more com- mands. Note that it is allowed to chose YES for BLANK CHECK, PROGRAM and VERIFY; the program performs these steps in the correct order, without the J need for intermediate com- I mand starting with S. As already noted, it is advisable j to think well before pressing the S key and so start the program. If you get an error report, do not get into a panic, but study the command screen to trace the fault and under- stand its nature. Once you have | worked with this EPROM pro- I grammer for some time, you will notice that it is highly user- j friendly and easy to get going { with the aid of the help pages. I which are instantly available at | the pressing of key H. | If you do not know how to program an EPROM which is not included in Table 1, simply begin with the lowest program- ming voltage. 12.5 V, to see if anything happens to the con- tents of the device; you can not damage it in this way. provided you do not select intelligent ! programming, as this causes the Vcc line to be raised to 6 V during the programming cycle | In conclusion of this section, a | few more tips. When an EPROM is stated to be pro- grammable in the normal (50 ms) mode, it is worth while to try out the effect of selecting fast-1 or fast-2 programming to save time. If you want to docu- ment the program settings for a specific EPROM, it is a good idea to use the screendump op- 1 tion for the recording of the checksum and other relevant data. Remember that a Type : 27512 (64 Kbyte) EPROM must be programmed in two 32 Kbyte I passes. Press CONTROL-STOP | to return to MSX BASIC, and type CALL EPROMX to run the programmer again. Use an assembler or a machine language utility package to write bytes into the RAM buffer for loading into an EPROM, but make sure that data is not over- written by stack or buffer usage of any program you run in com- bination with the EPROM pro- grammer software. Keep in mind that running BASIC programs that use PLAY commands require the com- puter to be reset and hence the EPROM programmer software to be re initialized. This is because the proposed program locates its jump tabie and vari- able map in the voice queue area. In more general terms, do not use the EPROM program- mer software before you are sure that there are no other pro- grams, or remnants thereof, still around somewhere in the com- puter's memory. The best way to avoid trouble is to reset the machine with the EPROM car- tridge inserted. Finally, Table 5 shows the con- trol words for the various EPROM types. These 7-bits words are specific to the EPROM type to be dealt with, and can be used by anyone contemplating the writing of his own version of the control soft- Getting started Commence with fitting jumpers B. D, E and I on the EPROM car- tridge board, then mount EPROM ESS 552 in the 28-way socket. Plug this cartridge into a slot of the MSX computer, and plug the I/O & timer cartridge either in a remaining slot, or in the one provided on the EPROM cartridge board. Con- nect the EPROM programmer I to the I/O & timer cartridge via the 50-way flat ribbon cable, and you have the system ready for use- see Fig. 10. Please note that it is not possible to use the add-on busboard for MSX com- puters. in conjunction with the j timer & I/O cartridge. Do not | yet fit an EPROM in the ZIF socket, switch on the power, ; and call the program on com- | pletion of its initialisation. After viewing the welcome and j copyright screen, go to the command screen and run the ' built-in test routine prior to vyorking on any EPROMs. If all LEDs on the programmer’s front ' | panel can be seen to go on and off at regular intervals, there is good reason to assume that I the hardware and software I functions satisfactorily, and it is high time to set the system to work on any EPROM that you may have available. AR We regret that we can not pro vide information on the use of this EPROM programmer with computers other than those in the MSX series. Previous articles on MSX exten- sions have appeared in the fol- lowing issues of Elektor India: February 1986 (I/O bus, digit- izer, I/O port); March 1986 (EPROM cartridge board); April 1986 (add-on bus board); February 1987 (I/O and timer cartridge). 5.43 Local Area Networking The proliferation of personal computers (PCs) as a business tool has driven the need for a distributed processing environ- ment where many microcom- puters can share expensive peripheral devices, such as printers and hard disk drives. The capability to network equipment also enables users to share files and programs and to centralize backup facilities and procedures. Local Area Networking has two main requirements. It must be implemented in VLSI, to sim- plify design and lower the overall "per node" cost of con- nection to a network. Second, the LAN must also run standar- dized software and conform to an industry standard, so that end users can interconnect equipment from different ven- dors without worrying about protocols. The IEEE 802.3 standard (Ethernet™) has gained wide acceptance by both large and small companies as a high- speed (10 megabit/second) LAN. However, because of its cable requirements, it can be relatively expensive to imple- ment. In response to this drawback, Thin Ethernet, also known as Cheapernet™ — was developed. Thin Ethernet uses less expensive coaxial cable and features a "node-inte- grated" transceiver. Thin Ethernet maintains full compati- bility with Ethernet's 10 mega- bit/second data rate. Another network sponsored by the IEEE 802.3 committee is StarLAN™, a 1 megabits-per- second implementation that features a "star" configuration. Each node is connected to another central hub in point-to- point fashion. Continuing development of LAN interface chips has driver, the LAN connection cost per node down to new levels, mak- ing networks affordable at all business levels. Because of its cost-effectiveness, the personal computer connection segment of the LAN marketplace is forecast to grow faster than any other segment. According to ; Dataquest, revenues in 1990 will top US$ 528 million. Revenues in 1985 totalled USS 181.7 million. The installed base of networked PCs will be 3.7 million in 1990, up from 438,000 in 1985. Current Status The decision by 3Com and Novell to port their LAN operating stems to National Semiconductor's DP3890 Net- work Interface Controller marks the first time a semicon- ductor supplier has taken an ac- tive role in making network software standard with their chips. This makes it easier for designers to use the chips in a network, rather than having to write software themselves. For original equipment manufac- turers (OEMs). DP8390 compati- bility with 3Com’s 3+ network software and Novell's Advanced NetWare means an easy path to LAN design for IBM-compatible PCs. OEMs can use National Semiconductor's tool kit con- taining DP839EB LAN evalu- ation boards and 3Com 3+ network software to develop networking and workgroup computing products. Or they can use Novell’s development kit and the DP839EB. or the DP8390 LAN chip set, to design local area networks 3Com and Novell are respon- sible for setting "de facto" stan- dards in the PC LAN industry. 3Com is the leading vendor of LAN add-on boards for PCs, with a 19 percent share of the market, according to Dataquest. Novell's NetWare, with 60,000 installations, is the most widely used PC LAN operating system. It supports 35 local area net- work systems. including 3Com’s Etherlink and Etherlink Plus. AT&T's StarLAN and IBM's PC Cluster and Token-Ring Network. Support from two predominant LAN suppliers reflects the emergence of the DP8390 as the standard LAN chip set of choice among system design- National's Local Area Network Chip Set Focusing specifically on the IEEE 802.3 local area network standard encompassing Ether- net. Thin-Ethemet (Cheaper- net). and StarLAN compatible networks, National designers developed three integrated cir- cuits: an Advanced Network In- terface Controller (DP8390 NIC), a Serial Network Interface (DP8391 SNI) and a Coaxial Transceiver Interface (DP8392 CTI). The chip set was the first complete VLSI implementation to meet the entire IEEE 802.3 standard. Its availability makes National Semiconductor well positioned to provide the rap- idly expanding PC LAN market with its cost-effective chip set. In particular, the DP8392 was the first monolithic chip im- plementation of a cable transceiver. The high level of in- tegration saves users a signifi- cant amount of board space. In fact, the network chip set is the only one that fits on a short-slot PC card. 5.44 APEX Microtechnology s latest power operational amplifier, the Type PA73M. offers oper- ation up to * 30 V: outputs of up to 5 A: a high-efficiency class C output stage; and MIL STD-883C screening Further details from APEX s UK rep resentatives: Pascall Elec tronics Ltd • Saxon House • Downside • Sunbury-on Thames TW16 6RY. | A new MSDOS emulator for Arcom's 80188 processor based development path. The package, called APPCOM. is supplied in an EPROM by Unit 8 • Clifton Road • Cambridge CB1 4WH. The DP8390 NIC features two 16-bit DMA channels that deliver all the data-link layer functions required for data packet transmission and recep- tion. The DP8391 features a patented digital phase lock loop for most reliable data reception. The DP8392 CTI im- plements all driver, receiver, jabber and collision-detecting functions required by the IEEE 802.3 cable transceiver. In ad- dition, the DP8392 exceeds the. one million hour MTBF re- quired in the 802.3 specification for transceivers. Illustrating National Semicon- ductor's technological breadth, three distinct process techno- logies were used in fabricating the chips: microCMOS for the DP8390, a high-speed oxide isolated bipolar process for the DP8391 and a junction-isolated bipolar process for the DP8392. The DP839EB evaluation board, containing the chip set, plugs into any IBM PC-compatible computer and incorporates all of the components required to provide a LAN interface to Ethernet or Thin Ethernet networks. The entire LAN chip set and evaluation board are all cur- rently in production. Ethernet is a trademark of Xerox Corporation. Cheapernet is a trademark of National Semiconductor Cor- poration. StarLAN is a trademark of AT&T Bell Laboratories. (Source: National Semicon- ductor) SCHOOLS EQUIPMENT FOR TOMORROW’S SCIENTISTS by David Tawney, MA FlnstP* From magnifying glasses to microscopes, callipers fo computer interfaces, pipettes to pH meters: the range of equipment in the catalogues of educational laboratory suppliers is vast and continually changing in response to technologi- cal and educational de- velopments. The introduction of micro- computers such as the BBC B and the RM 380Z to schools in the United Kingdom has been rapidly followed by the develop- ment of computer inter- faces whose purpose is either to enable data from school laboratory ex- periments to be captured, processed and displayed, or to control simple de- vices. Similarly, the rapid adop- tion by industry of biotech- nological techniques has led Britain's major sup- pliers to sell kits by which these techniques can be simulated in schools. How- ever, not all new equip- ment is stimulated by recent technological inno- vations: an educational concern to introduce science and technology to pupils aged five to 11 has led to much recent in- terest in construction kits. There are several reasons why the range of edu- cational laboratory equip- ment is so wide. First, science is taught in British schools over the age range of five to 18. Sec- ondly, as the emphasis is on giving pupils hands-on experience, suppliers have learned how to provide equipment that schools can afford in quantities sufficient for classroom work. Some of it is, of course, for demonstration by teachers but much is meant to be used by pupils working in groups Using Griffin & George progra of two and three. Pupil practical work is a cor- nerstone of education and so equipment must be strong and relatively inex- pensive Higher education equivalent The high degree of specialization by British pupils who stay on after 16. in many cases to pre- i Je scientific instruments pare for a course at a uni- versity or polytechnic, is sometimes criticized but it has advantages. The level reached by these pupils is typical of first or even second year university students in some countries and so equipment that is useful for higher edu- cation is produced in the quantities that schools need and at prices they can afford. Britain does not have a centralized educational system and schools are given extensive choices in the courses they provide for their pupils There are eight area boards offering iexaminations for the more academic pupils, and while procedures ensure comparability between these examinations, syllabuses do vary which increases the range of equipment needed. In the 1960s and 1970s, there was considerable in British schools, much supported by money from the Nuffield Foundation or from the government- funded Schools Council. This stimulated corre- sponding innovations in equipment, another reason why the range is so wide. In the late 1970s and early 1980s, most of the new courses were revised and these revisions, together with restrictions on money tor education caused by (ailing school rolls and the worldwide slump, have eliminated from the sup- pliers' catalogues any- thing that has proved unpopular. Much of what is left — still very extensive — has been refined over at least a decade. British schools tend to buy their laboratory equip- ment from three main general suppliers: Griffin & George (GG), Philip Harris (PH) and Irwin-Desman (ID). However, a number of specialist firms are also used. For example, inter- faces for coupling equip- ment with microcomputers to capture data or dem- onstrate control are of- fered by all three of the major suppliers as well as the specialists. Many functions An example is the measurement module sup- plied by Educational Elec- tronics (EE) which enables data from the outputs of a range of instruments — Hall probe for measuring magnetic fields, pH meter and so on - to be re- corded and strikingly displayed in several forms on a television monitor. A range of sensors is being developed to go with this and other computer inter- faces. One of the most interesting recent devel- opments using micro- electronics is the GiPSI (Griffin Programmable Scientific Instrument). There is concern that much of the more sophisti- cated equipment used in education spends much of its time on the shelf and is used only when its turn comes round in the sylla- bus so this instrument has many functions and will measure current, voltage, resistance, magnetic field, 5.46 .Mao. India ma, 1987 pH, light levels, and so on. The function wanted is selected by connecting a module containing an ap- propriately programmed read only memory (ROM) and fitting overlays over the control panel makes it easy to use. Another current growth area is electronics teach- ing kits. There have for many years been small components of electronics in some school physics courses but such physics teaching has recently been modernized and separate school elec- tronics courses developed The emphasis has shifted from simple introductions to semiconductor diodes and triodes to a systems approach to digital elec- tronics and to operational amplifiers. There are currently many approaches to teaching electronics embodied in kits. The equipment for one very popular course, "Micro-electronics for All", intended tor 11 to 13 year olds but in fact used for older pupils as well, is available from Unilab (U). Ideas underlying micro- electronics — or infor- mation technology as it is sometimes called — are learned through solving simple control problems. Other kits drawing interest are the Independent Schools Micro-electronics Centre (ISMEC) kits avail- able from Griffin & George, Philip Harris, and Unilab. Unilab specializes in elec trical and electronic equipment for education at competitive prices such as power supplies, meters, radiation counters, signal generators, and so on, all items that can, of course, be obtained from the general suppliers. Move to plastics It is easy to look just at re- cent major developments and target that the bulk of purchases made by edu- cational establishments are for consumables, notably glassware and chemicals, both supplied by Griffin & George and Philip Harris. Another company that specializes is BDH Chemicals. A devel- j opment over the last few | years has been the slow acceptance by schools of plasticsware in place of glassware. Early examples of plasticsware stained too readily but recent prod- ucts are more satisfactory and stand up to pupil use | much longer than glass, j Many of the top pan bal- ances bought during the boom in science edu- cation in the 1960s and 1970s are now wearing out and schools are replacing i them, as funds permit, with i electronic balances with digital displays. These are very quick to use so that fewer are required for a class. Griffin 8r George. Philip Harris, and Irwin- Desman all supply bal- [ ances but there are also j several specialist suppliers, ' notably Oertling. There are several ranges of microscope and specialist firms such as Prior have j suitable instruments for the I educational market. Re- cently, biologists have shown interest in kits for environmental studies containing meters that measure pH, conductivity, | temperature, light level, ] and so on. An example is an enzyme kit which pro- vides insight into the in- dustrial use of biotech- j nology. A recent growth point has been equipment for pri- mary school science. The educational emphasis is on using what can be j found in the home and the classroom with the minimum use of special ; equipment but some is j needed, such as simple kitchen-type scales, magnifiers, thermometers, construction kits, and so j on. Specialist primary ! school companies such as E.J. Arnold and Osmiroid have equipment suitable j for primary science edu- cation. Checking for safety I The School Science Set- vice provides information and consultancy on school science equipment and safety for the majority of British schools. Its task is to examine and test equipment and make recommendations to teachers. Copies of its reports can be obtained overseas through the British Council or through subscribing to the service as an overseas associate. Frequently the service is obliged to be critical of certain products but sup- pliers usually make modi- fications in the light of criticisms. E.J. Arnold Ltd. Lockwood Distribution Centre, Parkside Lane, Leeds, West Yorkshire, England, LS11 5TD. BDH Chemicals Ltd, Broom Road, Parkstone, Poole, Dorset. England, BH12 4NN. Educational Electronics, 28 Lake Street, Leighton Buzzard. Bedfordshire, England, LU7 8RZ. Grift in & George Ltd, Bishops Meadow Road, Loughborough. Leicester- shire, England, LE11 ORG. Philip Harris Lid, Lynn Lane, Shenston, Staffordshire, England, WS14 OEE. Irwin-Desman Ltd, 294 Puriey Way Croydon, Surrey, England, CR9 4QL. Oertling. W. & T. Avery Ltd, Smethwick, Worley West Midlands, England, B66 2LP Osmiroid, E.J. Perry Ltd, Gosport, Hampshire, England, P013 OAL. Prior Scientific Instruments Ltd, London Road, Bishops Storttord, Hertfordshire, England, CM23 5ND. Uni tab Ltd, Clarendon Road, Blackburn, Lan- cashire, England, BB1 9TA. •David Tawney is director of the Brunei University-based School Science Service of Britain's for the Provision of Science Equipment (CLEAPSE). Hot ICs - no need for fear l( is perfectly normal for ICs particularly bipolar digital ICs such as TTL. to become very warm in operation. These ICs draw considerable power which is finally dissipated as heat. An example is the common. TTL 1C 74145, Typical dissipation for this device is 215 mW and approximately 360mW maximum; this is in the quiescent state with unloaded outputs. When these are loaded the dissipation is even higher Since the area of the 1C package is relatively small, the operates perfectly even at 70°C. When the computer is installed in a housing, care regarding the temperature rise of ICs, the data sheet should be consulted; an 1C with a maximum dissipation of 10 mW for instance, should not exhibit noticeable temperature rise. The Microcomputer as a source of interference operates with relatively fast logic ICs. such as Schottky digital signals have rapid-rise slopes which produce harmonics extending far into the VHF/UHF region This cause interference, and not The problem is not restricted to home made microcomputers; some commercially built microcomputers, particularly teaching and experimental system, can unfortunately be classed as sources of electro- magnetic pollution. The only mcrocomputer in a (metal) screened housing with an earth connection; it may also be necessary to fit a mains RF- suppression filter. Scree ned (coaxial) cable should be used for connections between the computer and peripheral apply to all digital equipment This particular topic receives full attention in Junior Computer Book 2 (to be available shortly), but there is no harm in whetting the appetites of our readers even if it is a little prema- How can the Junior Computer display words? Normally speaking, data and address information is displayed with the aid of the monitor routine SCANDS. This involves one of the hexadecimal numbers, 0 . . F, in each display. Where texts are concerned, text display on die Junior Computer As we know, the display of the Junior Computer is suitable for displaying both numerical and hexadecimal data. By utilising a seven segment alphabet it is also possible to display written texts. If the text is to be static, a total of six letters are available. If, however a longer message is required, this may 'run' along the display rather like the electronic news display at the top of tall buildings (dynamic text). from an idea by U. Seyffert however, the monitor routines are no good. What is needed is the subroutine SHOW with the addition of a special look-up table which contains the corresponding seven segment pattern for each individual letter. Table 1 provides a survey of letters and figures together with the corresponding data which has to be entered into port A for them to be displayed. This table has been partly based on sugges- tions made to us from one of our readers. Obviously, letters which include diagonal lines (such as K, M, N, Q, V, W, X and Y) will have to be adapted to the horizontal and vertical set up of the display segments. Experience has shown, however, that the eye and the brain soon become accustomed to this. Now for a short program that will allow a six letter word to appear on perma- nent display. A good example would be the word 'Junior' as indicated on the prototype of the Junior Computer in the front cover photograph of the May 1980 issue of Elektor and Book 1. The program, JUNIOR, is listed in table 2. Here the modified SHOW routine will be called SHOWDS and the look-up table that holds the information relating to the display of any particular character is called TXT (text table). The Y index register acts as the display counter and text index. The value con- tained in the Y register increases from 00 tc 05 as an index for the particular character to be displayed. As soon as the value in the Y register becomes 06, 5.47 the Y register contains a delay value which determines the length of time that each display is actually lit. For this reason the previous value contained in the Y register (display counter/text index) must be saved in the address location TEMPY (0004) before the jump to the SHOWDS subroutine takes place. The function of the X index register, on the other hand, is the same as it was for the SHOW routine: it acts as a display digit switch by way of port B. In other words, the information contained in the X register (08. 0A, 0C, 0E, 10 and 12 consecutively) is passed to port B data register to turn each of the displays on in turn. Text on the run . . . A stationary text is all very well, but it does tend to get a little monotonous after a while. A much more interesting possibility would be to update the displayed text every few moments. In this manner whole sentences could be displayed instead of just single words. This can be accomplished with the aid of the program JUNTXT shown in table 3. The effect is very similar to that of an electronic news display. It is an expanded version of the earlier program JUNIOR (table 2). Page 03 is used to store the actual text which can, there- fore, be up to 256 characters in length 5.48 elakto Table 3. JUNTXT 0200 A9 7F 0202 80 81 1 A 0205 A5 00 0207 38 0208 E9 05 02OA 85 02 BEGIN 020C A9 00 020E 85 01 DSTIME 0210 A9 6F 0212 85 03 OISMPX 0214 A2 08 0216 A0 00 ONEDIS 0218 84 04 021 B 18 021 C 65 01 021 E A8 021 F 20 39 02 0222 A4 04 0224 C8 0225 CO 06 0227 F0 02 0229 DO ED TMECHK 022B C6 03 022D DOE 5 022F E6 01 0231 A5 02 0233 C5 01 0235 B0 D9 0237 90 D3 SHOWDS 0239 B9 00 03 023C 8D801A 023F 8E82 1A 0242 A0 7F DELAY 0244 88 0245 1 0 FD 0247 8C 80 1 A 024A A0 06 024C 8C82 1A 024 F E8 0250 E8 0251 60 LDA = 7F STA-PADD LDAZ-NUM SEC SBC # 05 STAZ NUMCOR PA0 . . . PA6 are outputs contents NUM (00001 to accumulator C- 1 NUMCOR = NUM minus 05 STAZ-NUMVAR LDA = 6F STAZ-DISCNT LDX a 08 LDY =00 STYZ-TEMPY TYA CLC ADCZ-NUMVAR TAY JSR -SHOWDS LDYZ-TEMPY INY CPY =06 BEQ TMECHK BNE ONEDIS DECZ-DISCNT BNE DISMPX INCZ-NUMVAR LDAZ-NUMCOR CMPZ-NUMVAR BCS DSTIME BCC BEGIN LDA-TXT, Y STA-PAD STX-PBD LDY = 7F DEY BPL DELAY STY -PAD LDY =06 STY-PBD INX INX RTS establish text display time start from Di 1 display counter (Yl = 00 store display counter Y to accumulator C- 0 A *-Y + contents NUMVAR (0001) accumulator to Y retrieve state of display counter increment display counter have all 6 display been accessed? if yes, move on to time check if not, next display time up? snd of text? see JUNIOR program TXT = 0300 (table 41 text index = Y + contents NUMVAR 01 23456789ABCDEF 0300 7F 7F 7F 7F 7F 7F 07 0B 2F 23 01 7F 20 7F 02 7F 0310 01 6F 07 0B 7F 07 0B 06 7F 61 63 2B 6F 23 2F 7F 0320 46 23 48 0C 63 07 06 2F 7F 3F 7F 03 63 1 1 7F 03 0330 23 23 0A 7F 24 xx xx xx xx xx xx xx xx xx xx xx 0000 (NUMI - 34 — enough for the average length paragraph! Again, this program uses the subroutine SHOWDIS, only this time the text table (TXT) is located at address 0300 and although the Y register is still used as a display counter it is no longer used as a ’text index directly. Instead, the particu- lar section of the text to be displayed is calculated by adding the instantaneous value in the Y register to the contents of address location NUMVAR (0001), The value contained in NUMVAR will be constant for the period of time a certain text is on display (the actual duration can be adjusted by modifying the contents of location 021 1). As soon as that period of time is over the con- tents of NUMVAR are incremented by one: the entire text shifts one location to the left and the right hand display shows a new character. When the contents of NUMVAR are greater than the contents of location NUMCOR, we will have arrived back at the be- ginning, as this means that the entire text will have been displayed. This is because the contents of NUMCOR are 05 less than those of location NUM. The latter (location 0000) is where the user must store the low order byte of the last memory location of the text table. In other words, if the last character of the text message is stored in location 0332, the value 32 is stored in location 0000 (NUM). Table 4 provides a sample text which can be displayed on the Junior Com- puter with the aid of the program JUNTXT as given in table 3. The text contains a message for Junior Computer Book 1 owners. A text should always be preceded by at least six blank spaces (7F), so that the beginning and end of the message are clearly separated from each other. M 5.49 selex-23 SOUND OF THE SEA It is a well known fact that j | the sound of the roaring surf of the oceans is the most satisfying sound in our environment. Those who have experienced the magic of this sound in an otherwise calm surroundings will immediately agree. It is quite a fantastic feeling to sit on the beach, close the eyes and listen to the sound of the seal The body and nerves which have been subjected to tremendous stress of the day to day life get relaxation from this sound and derive renewed force and energy. Unfortunately most of us can enjoy this pacifying experience once in a while, during holidays, For those who are deprived of this PtfU , 3t ' m / * • •* * '•* k : • • .is- ; luxury, we present here a small circuit which can generate the 'Sound of the Sea'. The circuit can be built from just a few components and imitates the sound of the sea in an excellent manner. This can also be used as background for a session of viewing your slides of a holiday on the beach. As it is already indicated, the circuit must produce the sound of the sea. This is done by the part A of the circuit block diagram shown in figure 1. In order to imitate the rising and falling of the roaring surf, it must have a control for the sound. This control is provided by the blocks B and C. Block B is an astable [ multivibrator which produces a rectangular pulse train, with a non symmetrical duty cycle. , From the pulse train, block C generates a saw tooth waveform with a rapid rise and slow fall. Both these signals are fed to the input of an amplifier block D in such a way that the signal from A is amplified by block D with amplification proportional to the signal coming out of block C. The rising and falling of the sound is created by this sawtooth waveform. 2 ©— 5.50 selex circuit. Figure 2 shows the noise source This is an unusual connection for a transistor. The NPN transistor T1 is connected in j a reverse manner using I only the base emitter junction. The collector is left unconnected. A transistor connected in such a manner behaves in a very noisy way. The intrinsic noise of a semiconductor device is a complex phenomena and will not be discussed here. The reverse biased base emitter diode behaves some what like a zener diode. A reverse current flows through this diode and resistance R1. The noise component in this current is connected to the next stage via capacitor Cl. Figure 3 shows the next stage which is an amplifier controlled by an astable multivibrator. T2 is the amplifier of block D. T3 and T4 form the astable multivibrator. The potential divider R2 and PI give the proper bias voltage to the base of T2 through R3. The setting of potentiometer PI decides the minimum volume of the sound of the -tb □ □_ 5.51 selex In absence of the signal from the astable multivibrator, the sound would be a continuous noise tone. To convert it into a rising and falling roar of the surf, the control signal is fed to the base through R6. The AMV (astable multi— vibrator) formed by T3 and T4 produces a rectangular wave as shown in figure 4. The frequency of this control signal is about 1 /8 Hz. This low frequency is required for the most realistic effects. The C4 (R7+P2) combination produces a sawtooth wave | from the rectangular signal, j The sawtooth wave is a I result of charging and discharging of capacitor C4. | During the OFF period of j T3, The capacitor C4 | charges through R8 and D1. During the ON period of T3, | the charging can no more i continue, but discharging can take place through R7 and P2. The values are so selected that by the time C4 I is discharged, the next charging cycle starts again. ( As R7 and P2 form a | potential divider, the signal j fed to base of T2 depends on setting of P2. In | technical language, the C4 I and (R7+P2) combination is I said to be an integrator | which integrates the signal at the collector of T3. | The sawtooth signal is I superimposed on the constant DC level set by PI | at the base of T2 and the resulting voltage looks like the waveform shown in figure 5 (bottom part). This voltage at the base of T2 controls the amplification factor for the noise signal being amplified by T2. Thus the output of amplifier T2 rises sharply and falls slowly, similar to the real The Construction The complete circuit can be accommodated on one small SELEX PCB. Component layout is shown in figure 6 As the circuit layout is a bit crov/ded compared to other simple SELEX circuits, the placement and soldering should be done carefully. The soldering sequence is as usual — jumper wires, resistors, diodes, capacitors, trimpots. transistors and finally the soldering pins or lugs for the external connections. Pay proper attention to the polarity of electrolytic capacitors. An important point to note that its collector is not connected anywhere. It should not.be left floating around on the board but should be cut off near the transistor casing itself T1 should preferably be BC 107 and may need some trials for selecting a good' noisy one. To select T1 by trials, the circuit of figure 2 can be connected to the Tape or Pickup input of the | preamplifier of the Hi-Fi ' system. If this gives a soft noise output through the | speakers, the transistor has j good noise properties. After | selecting T1 the circuit 1 canbe assembled and then the output of T2 should be 1 connected to the Hi-Fi ' system through the output ] capacitor C3. A 9 V miniature battery | pack is enough to power the circuit, as the current drawn I is between 2-5 to 4.3 mA. However, an ON/OFF switch must be provided. A shielded cable must be used j for connecting the circuit to the Hi-Fi system, so that the | 50 Hz hum is reduced. The shield wire can be | connected to the signal ground. Adjustments Two trimpots PI and P2 are provided for adjustments. The adjustments are 1 interdependent and should , be done as follows: Both the sliding contacts of I PI and P2 should be fully turned towards the earthed | terminal initially. Now PI is slowly rotated till a soft noise is heard. P2 is then 1 adjusted to get the periodic J rising and falling of the I sound. PI can be once again adjusted to get the j desired volume for the | sound of the sea. I Now, close the eyes and j relax, imagining that you | are already on the beachl selex TOUCH KEYS Touch keys, sensor switches, touch switches. TAPs Touch Activated Programmer), these are many names for the touch keys. The principle df opeiation is the same for all. Elektor magazine had developed and published the first touch switch project almost fifteen years ago. Since then there have been many variations and developments and the touch keys have replaced the mechanical keys switches in many sophisticated products. Just touch with a finger and without any "click— clack" the switching operation takes place quickly, safely and quietly. Principle The sensing surface of the key consists of two conductive surfaces separated by an insulator. The insulator must have an infinitely high resistance or it can even be an airgap. If these two surfaces are now touched simultaneously with the fingertip, the resistance between them drops below 500K. The exact value depends on various factors like the skin resistance of the individual, the size of the touched area, pressure exerted and even the humidity of the skin. To understand the working principle, you can carry out a small experiment as follows: Connect two coins to a multimeter with the help of crocodile clips. Set the multimeter in Megaohms range. Keep the two coins near to each other with a small airgap between them. Now touch both of them together with the fingertip. The meter now reads a value less than approximately 500K. The value falls down further if the pressure is increased or if the fingertip is moistened. The principle is thus very clear: the skin has a finite resistance and this resistance appears across the two sensor surfaces of the touch key when it is touched with a finger. When the key is not touched, the resistance between the two sensor surfaces is very high. This means that a current can flow between the two surfaces when touched. The two possibilities are shown in figures 2 and 3. Figure 2 shows a two stage sensor. When the key is touched, a current flows into stage A, this activates the output stage C and the relay is energised. But the relay remains energised only as long as the key is touched. Figure 3 shows a three stage sensor. When the key is touched, a current flows into stage A. this triggers a flip flop stage B and the flipflop activates output stage C. As the flipflop is latched, the stage C remains activated and relay remains energised even after the finger is removed from the key. To de energise the relay the key must be touched again, so that the flipflop resets and stage C is deactivated again releasing the relay. Practical Design A practical circuit is shown in figure 4. The functional blocks A. B and C can be easily recognised in the diagram. The first part of the circuit, consisting of transistors T1, T2 and T3 and the resistors capacitor and sensor corresponds to block A. If the sensor is touched, a current flows to the base of transistor T1 . Transistors T2 and T3 amplify this current and due to the collector currents of all three transistors passing through R6 and R5 a sufficiently large voltage drop is developed across R6. This brings down the voltage at U1 to almost zero level. Transistors T1, T2 and T3 are all connected in such a manner that they give maximum possible amplification. Information about this type of connection (known as Darlington Connection) has already been given in SELEX The middle portion of the circuit enclosed with a dotted line in the diagram corresponds to block B which is the flipflop. For proper understanding of the functioning it is assumed that T5 is conducting and T4 is open. Now when the sensor is touched. U1 drops from 9V to OV. This is transferred to the base of T5 via D3 (at point U5). The transistor T5 stops conducting and drives T4 into conduction. This condition is retained even after the finger is removed from the sensor. touched, the jump in voltage at U1 from 9V to OV is connected to the base of T4 via D2 and now T4 stops conducting and drives T5 into conduction again. Figure 5 shows all the voltage at various points U1 to U6 in the circuit. Voltage at U6 is used to activate the last stage C which drives the relay. Stage C consists of transistor T6. Voltage at U6 which is connected to the base of T6 via R1 2 switches the transistor ON and OFF depending on whether it is OV or 9 V. when U6 = OV, a current flows through R13 and R12 which develops a positive voltage at the base of T6 and T6 goes into conduction. When U6 = 9 V no current can flow through R13 and R12 and T6 is cut off. The relay contacts can be used to switch on any device connected through it. Construction The complete circuit of figure 4 can be assembled on a double size SELEX PCB (80 x 100 mm). Component layout of the circuit is shown in figure 6. The layout shows two connections in dotted lines between points A— B and C— D, and a connection in solid line between A— D. Connections A— B and C— D are to be used if the complete circuit of figure 4 is assembled. Connection A— D will be used if only the blocks A and C are constructed without using the flipflop circuit of block B. The flip flop will not be required if the switch has to close only for the period when the touch key is touched with a finger, figure 7 shows an assembled PCB as per the layout of figure 6. A 9V miniature battery pack is used as the power supply. The current consumption of the circuit is less than 3 mA when relay is not energised. Any 9V battery eliminator can also be used as the power supply, but this needs a change in the value of Cl . It should be increased to 10 #iF. The relay contact can be 5.54 .i.uo selex connected in parallel with the existing ON/OFF switch of the device that is to be controlled by the touch key; for example a Hi — Fi amplifier system. The scheme of this connection is shown in figure 8. The connection will be similar for any device. Only precaution to be taken is that the rating of the relay contacts must be suitable for the application. The touch key is universally applicable, if it is properly mounted in a suitable casing and the relay contacts are made available over sockets as shown in figure 9. Key Tip Construction of key tip can be done according to ones own creativity. The important feature to be remembered is that the two conductive surfaces must be separated by an insulator or air gap. The gap should be so small Ithat it can be easily bridged by the tip of a finger. Two ideas are illustrated in figures 10 and 11. One uses a banana socket with a small lug covered with insulating sleeve inserted from behind. 5.55 selex This type of touch key is very easy to install, as the banana socket comes ready with threading and matching nut. Only thing you have to do is drill a hole on the panel and mount the touch Another type of touch key construction shown in figure 1 1 uses decorative nails, drawing pins and washers of suitable diameter. This type of construction is very difficult because it requires accurate drilling and soldering. The washer perfectly concentric and an insulating material must be provided between them if the head of the nail is larger than the internal diameter of the washer. The mounting surface also must be of an insulating material as both the parts are directly mounted on it. Two pins must be accurately soldered onto the washer as shown in figure 1 1 . After inserting these through the panel they can be fixed with adhesive on the back side of the panel. In case the touch key is also assembled directly on the PCB, these pins can be directly soldered on to the PCB instead of fixing with adhesive. 5.56 f,W PRODUCTS • NEW PRODUCTS • NEW SCOPE/METER This system whch is IBMPC supported is designed to make calibrations ot oscilloscopes with bandwidths upto 1 GHZ. (a) Sensivities are as low as 40 mV to 200V/div. with outputs ol * DC and square wave of 1 0HZ' 100HZ. 1 KHZ and 10KHZ. (b) Automated or manual (c) Fully programmable (or automated testing. The Automated meter calibration system — Lets you calibrate on extensive range of voltmeters, ammeters, ohmmeters and multimeters. More information : M/S THE EASTERN ELECTRIC | AND ENGINEERING COMPANY I PRIVATE LIMITED. Regd Off: Gy an Ghar, Plot 434 A, j 1 4th Road, Khar. Bombay 400 052 Tel: 537210 LCR METER Ando Electric Co. Japan, offers the AG-431 1 digital LCR meter This unit is designed to make measurements as close as possible to actual user conditions for L. C&R components, semiconductors, complex components, electronic Measurement can be made at 31 spot frequencies from 100 Hz to 100 KHz and atleast signal levels between 1 mV to 5 V.R. L.C.D.Q.G. ESH B. IZI-0 and deviation (absolute or percentage) of component value from a programmable value are measurable. The instrument incorporates an automatic off- set ' Zero'' adjustment function and a high resolution mode to measure minute fluctuations in parameters. Other features include remote control or use of AG -431 1 as part of an automated measuring system through the use of the optibnal GP-IB plug- in interface and use of an external measurement frequency A test fixture and test leads help to give quick positive measurement capability. An external d c bias can also be applied through the instrument to the component For further details contact: MURUGAPPA ELECTRONICS LTD Agency Divn Parry House III floor 43 Moore Street Madras 600 001 COMPONENT MARKING No special tools are required for these Machines. Specifictions: Component Sizes: Body length: .250" - 2.75” Body dia : 0.80" - 1 -0" Maximum Imprint: Depending on type with Rubberplates 2.14" x 1" circum. Cycle rate: Upto 3200 cycles/hr Weight: 115 1b (52.3 Kg) For further information please M/S KELLY CORPORATION 1413 Dalamal Tower Nariman Point Bombay-400 021 POWER FACTOR METER Riken Instrumentation. Chandigarh, have developed a wide range of Power Factor Meters which are claimed to be compact, handy and light weight while they are simple in operation. These meters available in MINOR, MAJOR. CLIPON and PANEL types The 1 248-68 and are type tested j For further details write to RIKEN INSTRUMENTATION 181/32 Industrial Area Phase I Chandigarh - 160 002. SLIDE SWITCH IEC" have now introduced new Slide Switch with a rating of 2 Amperes. 250V AC/DC. This Slide Switch is available in single pole, on-off sequence with insulation resistance of J 100 M ohms and can withstand high voltages upto 2KV. The switch has a bakelite body with brass terminals and red ABS ) operating knob The terminals contacts or solder contacts. The mechanical life of more than I 20000 cycles and electrical life over 10000 cycles. For further information, write INDIAN ENGINEERING COMPANY Post Box 1 655 1 Worli Naka Bombay 400 018. CABLE TIE One piece Power Cord Cable Tie provides positive holding of natural or black colour for cable bundle diameter upto 51 mm. Maintain holding strength over I a temperature extreme of -20°C to ♦ 95°C For further details please contact: SURESH ELECTRICS & ELECTRONICS Post Box No. 9141 Calcutta-700 016. MECO DPMs MECO has just introduced 3 new series of light weight Digital Panel Meters featuring slim profiles with large display GM-135A B 3/1 digit LED type), GM-035 A/B 3 Vi digit LCD type) and GM 0-45 A/B 4’/j digit LCD type). These 3 new series feature automatic polarity switching. automatic zero function: over range indication and built-in hold function. The LED Modules operates on 0-5V power supply while the LCD Module operates on 0-9V power supply. Decimal point selection, high input impedance (more than 100 M ohms) are additional features. models include bias current 1 pA typical and lOpA maximum: measurement precision of 0.1% * 2 digit for 3'/; digit and 0.05% temperature coefficient of 25 ppm/C for reference voltage, operating and storage temperature ranges 0-50C and minus IOC to 60C respectively and sampling speed of 2.5 For further information, write to: MECO INSTRUMENTS PVT. LTD. Bharat Industrial Estate. T.J. Road. Sewree. Bombay 400015. 2W PRODUCTS • NEW PRODUCTS • N£ A I.G.E, (India) Ltd. introduces Computer Numerical Controls in technical collaboration with UNITERRUPTIBLE POWER SUPPLY PROFILE' has introduced an 50 HZ, consumes 10 W power The TV also works on 12 V DC adopter. A VCR sockt is provided for VCR viewing. The IGE will manufacture GE's Mark Century 1050 HLX 2 axis lathe control and Mark Century One for less complex machine | tool applications, including j ballistics, accoustics engine ! analysis, crash testing, high voltage power life failures, data , stream quality and accuracy, j biophysical research and many analysis based on multiple analog signal inputs. Data acquisition is accomplished by the Gould DASA system with a unique, instrument quality, mutli- channel signal digitiser. This front end sub system accepts upto eight analog signals (150 mV to ± 500 V) expandable for j upto 1 1 2 channels and samples j them simultaneously at predetermined rates from 500 Hz to 1 /3 MHZ. I.G.E. (INDIA) LIMITED Nirmal Niraman Point Bombay - 400 021. TARGET MARKETING DBS Executive Centre Rahe/a Chambers Nariman Point Bombay — 400 021. The use of common time base, clock accuracy of 0.01% and individual channel 8 bit A/D For further details, contact: HARESH G. NATHAN I. TAURUS ELECTRONICS. 13. Bussa Udyog Bhavan. Lower Ground Floor. T.J. Road, Sewree (W) near Sewree Bus Terminal, Bombay ■ 400 01 5 DIP REED RELAY PLA series DIP reed relays ar now available with 1 C/0 and 2C/0 contacts as well, besidf KEYBOARD SWITCH M/s. Darshana Industries has Profile Tactile Keyboard Switch. It is a 12mm x 12mm.. Four terminal S.P.S.T. N/O Switch MINIATURE SWITCH | SWITCHCRAFT now offer a miniature toggle switch type ' 202. D.P.D.T. rated for 2A-250N A.C These switches can be used for electrical S electronic applications, in telecommunication, electronic data processing etc. The insulating body is made of sockets PLA series DIP reed relays are available with keytop with an acrylic cover. Legends may be Hot stamped. Engraved or stuck on to the For further details, please aieiectic strengtn. contact are made of copper, silver plated • and terminals are solder lug contacts capable of switching 10W/VA at 0.5 amps, and 100V Max Salient features also include high speed switching and excellent input to A.8.S top and the acrylic cover snap fitted on top. The switch is available with Silver/gold plated contact terminals LARSEN S TOUBRO LIMITED ( Instrument Division) Venkataramana Centre, 8th floor. for electrical life at full capacity load for 25.000 operations, the overall dimensions of the switch behind the panel are 13x12.7x14.8 mm. the output isolation characteristics . ** & 563 Anna Salai. Post Box 6093 Teynampet Madras 600 018 MAKE-YOUR OWN TV KIT Taurus Electronics has mounting is on 6 mm dia threaded bush. n ■ oi For further information contact M/s. Darshana Industries introduced 14" B & W MAKE- YOUR-OWN TV Kit Model 14220. It is a low priced SKD Kit with cabinet. Picture Tube, all accessories packed In a JL rm For further inforamtion write to: M/S SAI ELECTRONICS Pune 411 013 with built in connectors The unit has a highly sensitive VHF Tuner with Channel Coverage 2 1 2 and gain at 60 db. It has a liiivflP Tdim INDUSTRIES) Thakor Estate. Kurla Kirol Road. Vidyavihar (West) Bombay-400 086 DASA SYSTEM A data acquisition and signal analysis system (DASA) developed by Gould Inc.. USA. system Only BEL make picture Peak Autio output is 2 W with an earphone jack for private For more information contact: SWITCHCRAFT 24 Pankaj Phones: 5131219/5136601 5.62 Mk.o. mdia m»» 1987 & Toubro Limited (L&T) It is listening. The TV when assembled works on Ac 22G-V. Vakola Bridge. Santacruz IE) Bombay-400 055. classified ads. advertisers index New Elektor kits assembling service and unmodified kits repairing facility avai- lable. Please write for details to: RACHANA RADIOS. Laxmigani Guna 473001 (M.P.) Wanted Knowhow for uninterrupted power supply and SMPS' outright purchase or royalty. Write in con- fidence MARATHEY TARALIKA. 216. Bhalchandra Road, Bombay - 400 019 For Printed circuit boards. Capacitors. Snapping clamp. Any type of press part. Art work. Layout, designing also under- taken. Contact Shiv enterprises P Bhagat Marg.Tukaram nagar, Ayre Road. Dombivli (East) 421 201 KITS -Radio remote control Call bell Rs. 1 50/- Radio remote control music light Rs. 150/- Ask project list with 60 paise stamps Super Electronics. Shivaji Nagar 8arsi - 41341 1 CORRECTIONS Precision power supply February 1987p. 2.49 Cl . C2 is 1 0OOu/25 V and R 22 is 0. 22 /3W as shown in the COMPUTERSCOPE-2 February 1987 p. 2.51 Hard copy of the screen image may be 1. Write the screen contents into a disk 2. Use a printer with an RS232 interface 3. Use the Electron interface on the BBC Figure 10 of the article is wrong in several areas and should be replaced by new Fig, 10 shown here. True-RMS meter January 1987 p. 1.30 The correct signal assignment for the contacts on Sec is: S6C contact a = Dp 2: S6C contact b Dp 1; S6C con- High power AF amplifier July 1986 p. 7-18 should be the i the SK39 as ACE COMPONENTS 5.64 ADVANCE VIDEO LAB 5.58 ANANT ENTERPRISES 5.70 APEX ELECTRONICS 5.64 APPOINTMENTS 5.03 5 68 BMP MARKETING 5.66 CHAMPION ELECTRONICS 5.65 COMTECH 5 70 CREATIVE DATA SYSTEMS 5 76 CYCLO 5 12 DEVICE 5 69 DYNATRON 5 12 5 63 5 66 ECONOMY ENGINEERING 5 63 * ELECTRONICA 5 16 ENGINEERING SYSTEMS 5 06 IEAP 5 66 INSTRUMENTS CONTROL DEVICES 5 66 KLAS 5 64 LEADER ELECTRONICS 5 16 LOGIC PROBE 5 61 MAYAN INDUSTRIAL 5 70 MECO 5 07 MX ELECTRONICS 5 10 5 11 NARMADA VALLEY 5 64 NCS ELECTRONICS 5 16 PHILIPS 5.13 PRECIOUS 5.08 5.09 5.15 SAI 5.59 5.61 5.63 SAINI 5 14 SIEMENS 5 38 5 39 SMJ ELECTRONICS 5 02 TESTICA 5 14 TEXONIC INSTRUMENTS 5 14 TRIMURTI ELECTRONICS 5 58 UNLIMITED ELECTRONICS 5 58 VASAVI ELECTRONICS 5.06 VISHA ELECTRONICS 5.75 YABASU 5.12 5.72 .i. R.N No 39881/83 MH/BYW-228 UC No 91 ON IMPACT- 1 Impact- 1 is a unique combination of Hardware and Software designed for the first time in India for learning Process Control Applications. It is an 8085A based system with on board ADC/DAC, Timer/Counter, Interrupt Controller, RS 232C Serial Port, Centronics Parallel Port, ASCII Keyboard Interface, Cassette Interface, 48 Parallel I/O lines, EPROM Programmer for 2716 to 27256 with optional Fast-Intelligent Mode, Hex Keypad with 6 digit LED display, STD Bus on card edge and six 28 pin sockets to take memory upto 64K total. Powerful FIRMWARE is given in a 16K EPROM and is supported by comprehensive Documentation. In short, Impact-1 has everything that is required for training in process control applications and that is why it greets you with the message ‘Pro Con’ at power up! Want to know more about Impact-1? Write to us today for the colour brochure of Impact-1 and other 8085 based systems. O Creative Data Systems 14, Hanuman Terrace, Tara Temple Lane, Lamington Road, Bombay 400 007 Tel: 362421, 353029 Tlx: 011-71801 DYNA IN Gram: ELMADEVICE LEARN PROCESS CONTROL