48 April 1987 Rs. 7.50 THE ULTIMATE SOLID-STATE MEMORY? Includes information on: • Home work - Table. • Hardware & Components • Checklists • Soldering • Component data & codes. *«&*• SO/', R.p JS3&- • Send a 3 year Subscription for elektor India magazine and get this book free. • Existing subscribers can order this book @30% discount. Please mention Subscription number (or send latest address label from magazine envelope) eIeIctor ElECTRONiCS pvT lid. 52-C Proctor Road Grant Road (E) Bombay • 400 007. Phones: 3674591/369478 Telex No. (01 1 ) 76661 ELEK IN This offer is valid from 01 - 04 - 87 to 30 - 06 - 87 Damle elektor india april 1987 4.03 DO IT YOURSELF LEARN-BUILD- PROGRAM The Junior Computer book is for anyone wishing to become familiar with microcomputers, this book gives the opportunity to build and program a personal computer at a very reasonable cost. to. eklaoR ElEdRONics pvT hd. 52-C, Proctor Road, Grant Road (E), Bombay-400 007. 4.06 el«ktor india apnl 1987 The Indian reprint conies to you from elekter Send full payment by M.O./I.P.O./D.D. No Cheque Please. Packing & Postage free Publisher: C.R Chandarana Editor: Surendra Iyer Editorial Assistance: Ashok Dongre General Manager: J. Ohas Advertising: B M Mehta Production: C.N Mithagari Address: ELEKTOR ELECTRONICS PVT. LTD 52. C Proctor Road Bombay 400 007 INDIA Telex: (Oil) 76661 ELEK IN Overseas editions Volume-5 Number-4 April 1987 Elektor Electronics Standfast House Bath Place High Street, Barnet Herts EN5 5XE U K Editor: Len Seymour Publitron Publicacoes Tecnicas Ltda Av Ipiranga 1100. 9° andar CEP 01040 Sao Paulo — Brazil Editor: Juliano Barsali Elektor sari Route Nationale; Le Seau; BP 53 59270 Bailleul - France Editors: D R S Meyer, G C P Raedersdorf Electronics Technology Design abstracts The ultimate solid-state memory? Remote control in astronomy Software for the BBC computer Op-amp frequency compensation the why and the how 4.21 4.24 4.28 4.43 4.64 Elektor Verlag GmbH Susterfeld- Strafie 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. Estefania. 32-1° 1000 Lisboa - Portugal Editor: Jorge Goncalves Ingelek S.A Av. Alfonso XIII, 141 Madrid 16 — Spain Editor: A M Ferrer Projects MSX extensions-5 SCART adaptor for IBM PC Biphaser Satellite TV reception Junior Synthesizer LCD panel meter Morse call sign generator ••• Information 4.31 4.41 4.38 4.46 4.52 4.54 4.65 In part: Kedhorn Holdings PTY Ltd Cnr fox Valley Road ft Kiogle Street Wahroonga NSW 207b Australia Editor: Roger Harrison Electronic Press AB Box 63 182 11 Daoderyd - Sweden Editor: Bill Cedrum News • News • News • 4.19 New products 4.68 Readers Services 4.74 Info/Data Sheets 4.81 The Circuits are (or domestic use only. The submission ol designs of articles of Elektor India implies permission to the publishers to alter and translate the test and designand to use the contents in other Elektor publications and activities The publishers cannot guarantee to return any material submitted to them. All drawings, photographs, printed cirucit boards and articles published in Elektor India are copyright and may not be reproduced or imitated in whole or part without prior written permission of the publishers Patent protection may exist m respect of circuits, devices, components etc described in this magazine The publishers do not accept responsibility for failing to identify such patent or other protection. Printed At: Trupti Offset 8ombay - 400 013 Ph 4923261.4921354 Copyright © 1987 Elektuur B.V. The Netherlands Guide lines Switch board 4.75 Classified ads 4.80 Index of advertisers 4.80 Corrections 4.80 Selex 22 Automatic flasher 4.56 Voltage divider in BASIC 4.59 Mosquito repellant 4.60 Logic probe 4.62 The membrane model of capacitors 4.62 elektor indie apnl 1987 4.05 PHILIPS PROFESSIONAL OPTOCOUPLERS FROM PHILIPS JSSS FEATURES: * High isolation voltage * High current transfer ratio * Low saturation voltage * Larger dynamic working current range (1 to 100 mA) * TTL compatible A variety of high performance are extensively used in wide optocouplers have been range of telecom, control developed by Philips for instrumentation, computer professional applications. They peripheral and power are ideal for interconnecting TTL electronic circuits and other integrated circuits and worldwide. NEW RANGE OF PROFESSIONAL OPTOCOUPLERS current transfer ratio t(»n (see F.g 1) V c€ = 0.4 V l f = 10 mA switching times = 5V;l Con = 2 mA CNY50-1 0,25 CNY50-2 4.0 CNY57- 0 2 CNY57A 0 4 1.0 100 10 100 4 4 100 44 100 I * These types are available with UL and VDE approval For further information: PHILIPS INDIA (ELCOMA), Bandbox Bldg. 254-D, Dr Annie Besant Rd. Phone: 4930311, 4930590. Telex: (011) 71540, (011) 76049- Worli, Bombay-400 025. eiekjof mdia april 198? 4.07 e k ts available from : precious ® KITS NOW AVAILABLE NAME OF THE PROJECT ISSUE NO NO E 4/5 E 6 E 8 E 9 E 13 E 16/17 E 16/17 E 15 E 15 E 15 E 15 E 18 E 18 E 18 E 21 E 21 E 21 E 26 E 28/29 E 31 E 33 b 36 E 11 EPS NO 83597 80543 78003 83113 9765 84457 84426 9448 9968 9218 77059 9987’/! 84073 84083 1487 9119 9506 82156 85448 77059 85110 9827 84012 PROTABLE EGGTIMER THE STAMP POWER FLASHER VIDEO AMPLIFIER SIGNAL INJECTOR MUSICAL DOORBELL ALARM TIMER ♦ 5V 500mA TTL Power Supply + 15V 250mA Op — Amp power supply * 1 5V 1 A CMOS Power Supply 0-10V. 300mA Variable Power Supply TLEPHONE AMPLIFIER LAMP SAVER I LAMP SAVER II LIGHT DIMMER 7400 SIREN 7 W AMPIFIER LCD THERMOMETER ELECTRONIC DOG SIMPLE SOUND EFFECT TELEPHONE EXCHANGE MAGNETISER V, CAPACITANCE METER PRICE 4500 50 00 80 00 115 00 30 00 105.00 60.00 50 00 70.00 70.00 90.00 125 00 55.00 4500 40.00 3500 50.00 57500 1 80 00 60.00 1200.00 65.00 75000 E 28/29 85447 FAULT FINDING PROBE FOR UPS E 28/29 85431 HI-FI HEADPHONE AMPLIFIER E 28/29 85476 BRAKELIGHT MONITOR E 37 80054 TALK FUNNY E 39 86013 SINGLE TRACE CRT CONVERTOR E 39 9967 VHF/UHF MODULATOR OIGILEX DIGITAL TRAINER E 22 9899 CMOS FUNCTION GENERATOR E 31 80077 LUXUARY TRANSISTOR TESTER E 40/41 86490 RODENT DETERRENT E 40/41 88046 CAR RADIO ALARM E 34 85128 SOLID STATE IGNITION E 16/17 84462 FREQUENCY METER E 45 31111 MW RECEIVER E 46 82178 PRECISION POWER SUPPLY E 47 82162 IN CAR IONISER 60 00 75 00 25.00 210 00 190.00 32500 100.00 135 00 75.00 65.00 450 00 1.150.00 125.00 995.00 235.00 Ordering Information 1. All payments in Advance by M.O.l? O.or D.D. onty- 2. Items will be send by R.P.P. only 3. Price includes Packing & Postage 4 For orders out of Maharashtra state, Please add 1054 as a 'incidental Charges' precious , ELECTRONICS CORPORATION Journal Division 11. Kiln Lane. Ott Laminqton Pd Bombay - 400 007 DESIGNS AVAILABLE UPTO PROTOTYPE STAGE FOR CORDLESS TELEPHONES TELEPHONE INSTRUMENTS AS PER INDIAN P & T REQUIREMENTS GENUINE SSI/PROPOSED SSI UNIT MAY APPLY TO: THE ADVERTISER BOX 0487, ELEKTOR INDIA 52, C PROCTOR ROAD, BOMBAY-400 007 r Please mention elektop when ordering goods or requesting information from advertisers. •v eleKtor india apni 1987 4 09 TRANSISTOR SQUARE MICROPROCESSOR BASE Dynalog will show you the way...... If you think that you have missed the ‘BUS’, or that you are lost on the ‘Data Highway’, it is still not too late. Dynalog will show you the way! Dynalog gives you a strong base for training and development of Microprocessor Based Products. A full range of Training and Development Systems based on almost all popular CPUs like 8085, Z-80, 6802, 6502, 8086, 8088 and 68000 are available from Dynalog Micro- Systems. In addition to these, a wide variety of STD Bus compatible add on cards are also available for system enhancement as well as OEM applications. Dynalog Micro-Systems 14, Hanuman Terrace, Tara Temple Lane, Lamington Road, Bombay 400 007 Tel: 362421, 353029 Telex: 011-71801 DYNA IN Gram: ELMADEV1CE elektor india apnl 1987 4.1 1 ADDMark LINTASSM T 48254 Process Monitoring — TELEPERM 7NG is a two-wire temperature transducer suitable for thermocouple and PT 100 inputs, it offers very accurate (better than 0.5%) conversion and transmission of signals from temperature sensors. Process Interlocking — SIMATIC S5 Automation Systems, incorporating the most up-to-date interlocking and signalling technology, are the only systems to offer comprehensive packages of tried and tested software. Process Actuation Electrical Actuators Type E,F,G, KE6 and ANNA offer the widest torque range and output speeds. These can be coupled with control valves and dampers. For further details, write to: Control and Automation Systems Siemens India Ltd P.O. Box 6597 Worli, Bombay 400 01 8 Siemens India.The automatic choice for automation SIEMENS Instrumentation and Control Systems A new generation of products to meet today’s industrial process challenges Process Control — TELEPERM D compact digital controllers offer a unique single-loop integrity feature along with a complete range of closed-loop functions. eleklor india april 1 987 4.13 IC SOCKETS HIGHLY RELIABLE^ CONSISTENT SUPPLY AT AFFORDABLE PRICES ! £hampmn t\Ea\paxi\cs. Put LM Irf Joint Venture with Champion Electronics Pvt. Ltd. S I 7. MIDC. Bhosari. Pune 41 1 026. INDIA Phone : {0212| 83791, Cable : CHAMPION Tele* 0145 333 MCCI IN. 0145-505 MCCI IN MIDLAND ROSS CORP U S A Socket all your IC’s with CHAMPION IC Sockets and MULTIPLY PROFITS ! t Flats of Spring Contacts are Oriented to Flats of IC Leads. Dual Leaf Spring Contacts. Phosphor bronz or beryllium copper contact material - for total cost vs. performance capability. Buttable end-to-end and side-to-side on 2,54 mounting grid. Kapton 1 M solder wicking barrier - meets Class 1 Dual inverted leaf face- requirements of EIA wipe contacts - for reli- standard RS-486. able connections and Kapton barrier also eosy insertion/extroction prevents IC leads from of IC shorting to PC troces beneoth socket. Wide torget entry into 'r\ Kopion tm*o r4$ni#r«d NodrTxyii of Dupont contact - for easy auto- matic IC insertion. Anti-moisture bosses 'to assist flux removal after soldering. Black gloss-reinforced polyster body meets UL 94 V-O requirements for flame retardoncy. Low profile - 3,94 max. overall height. Ladder frame body con- struction - for improved airflow around IC and visibility of circuit traces on PCB. Tin Plating or Gold plating for total cost vs. performance capability. * For details write to us elektor india april 1987 4.1 5 IRIENT/CE-186 AUTHORISED DISTRIBUTOR: juf ^ $ ’ Jfjiicu ElasirnniBS Pis. Lid. Tc*vn Office 101 Kitchener Road #02-04 Jalan Besar Plaza Sngapore 0820 Tel 2966455 (6 Lines) Fax 2926627 Teiex DEVICE RS 33250 CaWe DEV) TRONIC Factory 6056 MacPherson Road #04-12 Citimac Industrial Complex Singapore 1336 Tel 2886455 <6 Lines) elektor irvdia apnl 1987 4.17 DESIGN ABSTRACTS The contents of this column axe based on information obtained from manufacturers in the electronics industry, or their representatives, and do not imply practical experience by Elektor Electronics or its consultants. PASSIVE INFRA-RED DETECTOR TYPE PID-11 The Siemens Type PID-11 infra- red detector is manufactured from heat sensitive polyvinyl- edendifluoride— PVDF— with all necessary optical and elec- tronic elements on board. Passive infra-red detectors are particularly suitable for observ- ing heat emanating from mam- mals and converting this into an electrical signal— see Fig. 1. They are normally used for the detection of movement (which gives rise to temperature dif- ferences) and are thus ideal sensors in intruder alarms. Although PVDF— which has only been commercially avail- able since the 1970s— has not such a wide range of operating temperatures as other types of material used in heat sensors, such as lithium tantalum oxide, it is much cheaper and easier to manufacture. Moreover, be- cause of the thinner mem- branes used, it has a much shorter response time. Figure 2 shows a Type PID-11 detector removed from its housing. reflector bundles the incoming infra-red radiation. The sensor element is located at the focus of the reflector. The amplifier board, manufactured in sur- face-mount technology, is con- nected to the back of the sensor. The case is made of conductive plastic and serves also as an ef- fective electrical screen. Its di- mensions make it possible for the detector to be installed unobtrusively— see Fig. 4. To compensate for variations in ambient temperature, the de- tector contains a second sensor which is not in the path of the incoming IR beam. Only the dif- ference in signal levels be- tween the two sensors is fed to the amplifier. heat (i nfra-fed) detector it (infra-fed) radiator ; 10 um : 60 W Fig. 1. Schematic representation of intruder detection with a passive infra-red device. Construction The detector consists essen- tially of the following elements: Venetian blind; infra-red win- dow; parabolic reflector; sensor element; amplifier circuits; and case. The Venetian blind pre- vents random light falling onto the window and, together with the reflector, determines the pick-up pattern— see Fig. 3. The infra-red window behind the blind protects the detector from air currents and prevents soil- ing of the sensor element. The Fig. 2. Siemens passive infra-red detector Type PID-11 removed from its case. eleklor India april 1987 4.21 The circuit of the amplifier, at output 4. j to objects moving within the shown in Fig. 5, is based op Figure 6 shows the output volt- | capture area of the detector at three inverters connected as age vs tifne characteristic dur- right angles to the optical axis, linear amplifiers, the first one of ing a sudden change in in- Overloading of the input by which also serves as an im- coming IR radiation. direct sunlight or other heat pedance inverter. sources should be avoided. The two diodes. Ti and T 2 , limit Very high room temperatures against too high inputs and also 1 A nnlication tins reduce the sensitivity of the serve as high-impedance leak- " device, age resistors. Apart from its use in intruder The main technical data are Networks R3-C4 and Ra-Cs sup- alarms, the PID-U will also be gj ven j n Table 1. A supply press low or high frequency im- found suitable for switching on voltage of between 4 and S volts pulse noise respectively. lights, water heaters, electric se ems to be the optimum (see A fourth inverter produces a hanci driers, and door openers. a ] so pj g 7 and p lg 8 ). reference voltage of 11 may a * so usec * as a counter The capture range of the detec- Fig. 3. Radiation pattern ot the or production monitor. The tor depends to a large extent PID- 11 . Urer=[Ub— 0.6J/2 volts ; sensor is particularly sensitive | on the size of the object and on Fig. 4 Dimensions and pinout of the PID-11. Fig. 5. Circuit diagram of the PID-11. Fig. 6. Output voltage vs time characteristic when a warm or cold body is suddenly brought into the capture range of the detector. Fig. 7. Output signal vs supply voltage characteristic. Fig. 8. Current consumption vs supply voltage. Fig. 9. Output signal vs distance of observed body within the capture range of the detector; supply voltage =4.5 V and room tempera- ture = 22 °C. the difference between object and ambient temperatures— see also Fig. 9. Table 2 shows the immunity to some light sources, and is therefore particularly useful for domestic applications. Spurious output signals may oc- cur just after switch-on during the heating up period. Although the PID-11 is splash- proof, its use in the open is not recommended. The recommended range of operating temperatures is -20 °C to +70 °C. Some suggested circuits In the suggested circuits in Fig. 10 and Fig. 11, the PID-11 is connected to a window com- parator serving as signal detec- tor, a timing element, and a drive circuit for the relevant ap- plication. Fig. 10 shows a general purpose circuit, in which the window comparator is formed by OPi and OP 2 . The reference volt- age, Uref, at pin 4 of the PID-11 determines the centre of the window. The upper window limit, Ui, is given by U. = Uref(Ri/R 2 ) while the lower window limit, U 2 , is given by U2=U,eifR3/(R 3 + R4)) If the output voltage, Ua, at pin 3 of the PID-11 exceeds either of these limits, the open-collector outputs of the two opamps will be about 0.S V, which will trig- ger monostable CPi Relay Ki is then energized for the duration of the mono time, which is de- termined by Ci, Re, and R?. Po- tentiometer R? allows the period to be set anywhere be- 4.22 eiektor mdia apni 1987 Table 1 Technical data Supply voltage, Ub 4.5 V (4-12 VI Output signal, Ua 1.1 V lat Ub~4.5 VI Lens area 300 x 400 mm Current consumption, Ie 0.4 mA (Ub = 4.5 VI Output impedance (CMOS) 2.2 k IUb = 4.5 VI Capture range (see also 7 m Fig. 9) Response time 500 ms Table 2 Immunity to unwanted light Maximum illumination from above from the side 4v230° $h230” Incandescent light <800 lux <3 OOO lux Sunlight < 10 OOO lux <20 000 lurf Fluorescent light r»o effect no effect tween 3 and IS seconds. The upper and lower limits of the window discriminator can be inactivated by St or S 2 re- spectively: the ‘circuit then reacts then only to a negative or positive voltage change— see Fig. 6. If, for instance, Si is open and S 2 closed, the relay will only be energized if a person who has been within the cap- ture range of the detector leaves the area. The circuit in Fig. 11 is intended for automatic operation of stair- case lighting. Compared with that in Fig. 10, the window discriminator here is inverted. Lighting timer SAB0529 is trig- gered via an opto-coupler to ensure that the PID-11 is elec- trically isolated from the mains supply. The timer is set for 63 s. The power supply for the PID-11 may be formed by a bell trans- former: current consumption is of the order of 10 mA. Fig. 11. Automatic staircase lighting control based on the PID-11. Siemens Limited Siemens house Windmill Road Sunbury-on-Thames Middlesex TW16 7HS Many Siemens components are available from Electrovalue 28 St Judes Road Englefield Green Egham Surrey TW20 OHB Telephone: (0784) 33603 Siemens Worldwide Australia Siemens Ltd. 544 Church Richmond Melbourne. Vic. 3121 •& (03) 420 7111. Tx 30 425 Brazil Siemens S.A. Sede Central Caixa Postal 1375 01051 S3o Paulo-SP s (011) 833-2211 Tx 11 23641 Canada Siemens Electric Limited 7300 Trans-Canada Highway P.O.B. 7300, Pointe Claire, Quebec H9R 4R6 s (514) 695 7300, Tx 05822778 Greece Siemens AE Voulis 7 P.O.B. 3601 GR 10247 Athen •s (01) 3293 1, Tx 216291 India Siemens India Ltd Head Office 134 A, Dr. Annie Besant Road, Worli P.O.B. 6597 Bombay 400018 & 4938786, Tx 75142 Italy Siemens Elettra S.p.A. Via Fabio Filzi, 29 Casella Postale 10388 I 20100 Milano a (02) 67661, Tx 330261 Portugal Siemens S.A.R.L. Avenida Almirante Reis, 65 Apartado 1380 P 1100 Lisboa 1 ® (01) 538805, Tx 12563 Sweden Siemens AB Halsingegaten 40 Box 23141 S- 10435 Stockholm ® (08) 161 100, Tx 19880 U.S.A. Siemens Components, Inc. Special Electronics Division 186 Wood Avenue South Iselin, New Jersey 08830 ® (201) 321-3400 Tx 844491 elektor india apnl 1987 4.23 Although the super memories of the future will be predominantly photonic ones, a number of manufacturers are developing massive semiconductor memories for use in the intervening years. Many realize that they cannot make money in the memory market, but need the technology to gain experience for their telecommunications and computer businesses. Most of the manufacturers ment, and a vast market (esti- rising market values. wide. working on the development of mated at close to £1000 million As already stated, the invest- To reduce the risks, a number very large semiconductor during the economic pro- ment costs associated with of manufacturers have decided memories have a 1Mbit DRAM- duction life, i.e., until photonic these super memories are so on a joint venture, for instance (dynamic random access mem- devices become available). high that most manufacturers Siemens and Philips have jointly ory) or EPROM (erasable pro- Most manufacturers, particu- will not succeed in recovering undertaken the development of grammable read-only memory) larly those outside Japan, them from the sale of the a 4 Mbit DRAM (with the aid of m production; the remainder realize that they cannot make devices alone, particularly in the German Ministry of Re- are poised to start production m enough money in the memory view of the prevailing low search and Technology and the the next few months. A number market to recoup their huge in- prices on the depressed com- Dutch Ministry of Economic Af- of them are also already work- vestment costs. They need the ponent market. It is a maxim fairs). Even then, the risks are ‘A 9 ' he „,? e r el0pment ° f 3 ,echnology ' however ’ to gain that component prices drop by high. As an example, the con- 4 Mbit DRAM (for instance, the experience for their telecom- up to 90% during the first 2-3 struction of a new semiconduc- General Electnc-Toshiba-Sie- mumcations and/or computer years of .the device’s life. One tor plant with dust-free working mens-Philips conglomerate). businesses. The Japanese, un- beneficial factor is, however, spaces at Regensburg cost These devices require a state- like the Europeans and that because of the high invest- Siemens almost £60 million. The of-the-art capability, high devel- Americans, believe in a ment costs there are only a price of its UK Communication opment and production invest- strongly growing market and handful of competitors world- and Information Systems’ new 4.24 elektor india apfil 1987 Fig. 2. The rise of solid-state memories {source: IBM). headquarters at Feltham, Middx, has not been disclosed. Even the nature of the project adds to the financial risks, because development and manufacture go hand in hand. When a development team manager so advises or decides, equipment is bought, staff is engaged, or a production pro- cess is modified immediately. If he has made a mistake, that has instant financial consequences. In some instances, it may be necessary to order tools or equipment without knowing for certain whether they will be used at all. The technology The size of the memory cell structures used on these de- vices is of the order of 1.0 to 1.5 pm, so that even dust par- ticles of only 0.5 pm can render the chip useless. Such particles can only just be observed under very powerful micro- scopes. The structures are etched onto the wafer with the aid of light close to the ultra- violet region (A =0.4 (im). When structures become smaller than 0.7 pm, they will have to be etched with the aid of X-ray lith- ography. Typical of the specifications of a Mbit memory are those of IBM's proprietary device: 1 048 576 bits, divided into four blocks of 256 Kbit each, are contained on a wafer with an area of only 80.85 mm 2 . This results in a density of 13 025 memory cells per square milli- metre, so that six of these devices can contain the con- tents of a book of 250 pages— see Fig. 3. The IBM device is a DRAM that needs a supply of only 5 volts. During operation, it requires a power of 0.5 W; in the quiesc- ent mode, only 50 mW. At a tem- perature of 75° C, its access time is 150 ns. Its structures measure 1 pm and it is pro- duced in FET technology. Field-effect transistors are used because their structure is simpler and they are therefore easier to manufacture. Currently, most 1 Mbit memor- ies are DRAMs. since static memories— SRAMs— need more transistors and therefore take up more space. Dynamic memory cells generally consist of one transistor and one ca- pacitor, whose charge has to be refreshed at regular intervals. This is, incidentally, also the reason that DRAMs are slower than SRAMs. It should be noted here that (magnetic) bubble memories have been available for some years in capacities of 4 Mbit (but in wafer areas of about 1500 mm 2 ), while 16 Mbit de- vices are beginning to become available. The production process The production process of a 1 Mbit memory entails no fewer than four hundred stages. The starting point is the wafer, a thin slice of monocrystalline silicon, which has a diameter of be- tween 120 and 150 mm, de- pending on the manufacturer. This wafer is used as the substrate onto which the memory cell structures are fabricated. The memory is designed with the aid of a e?iD (computer- aided design) system, which ensures that the integrated cir- cuit conforms to its physical and electrical requirements. The in- formation obtained from the system is used to produce the pattern on the mask. The mask is a device for shielding selec- ted areas of the wafer. It is either emulsion on glass or an etched thir film of chromium or iron oxide on glass, and is pro- duced by photographic re- duction from large-scale lay- outs. The patterns are cut into the emulsion or film by elec- tron-beam lithography. The substrate is covered with a solution of positive photoresist by spincoating, spraying, or im- mersion. The desired pattern is then produced on the substrate by photolithography (after the elekior india apnl 1987 4.25 solution of photoresist has dried). The exposed portion of the photoresist is depolymer- ized and removed during devel- opment with a suitable solvent, such as trichloroethylene. The polymerized portion remains and acts as a barrier to etching substances or as a mask for deposition processes. When the processing step is com- pleted, the remaining photo- resist is removed with another suitable solvent. The wafer is then etched in a dry process, which ensures uniform vertical edges. The dry-etching process also causes far less track damage than chemical etching as can be seen in Fig. 4. Until recently, interconnexions between circuit components were also etched into the substrate. Nowadays, a number of manufacturers use a different method. The photoresist on the wafer is exposed to light through the mask and then heated: this causes the exposed portions to be preserved. The wafer is then exposed to light without the mask and devel- oped. The portions which were covered by the mask are not preserved and disappear dur- ing the development. Finally a thin film of metal is vaporized onto the wafer, superfluous metal and the remaining photo- resist are removed together. To obtain layers with different conductivity, the silicon is doped with different ions, for instance, by implanting boron ions into the crystal lattice of the silicon. This process, carried out in a vacuum at high electric potentials, is much more pre- cise than the usual diffusion process. A much higher pack- ing density then becomes poss- ible. the switching behaviour becomes more reliable, and cross-talk is much reduced, j The individual transistor elements are then interconnec- ted or insulated from one another by chemical vapour deposition— CVD— of thin layers of gaseous monocrystal- line silicon, silicon oxide, and silicon nitride. In this way, a dielectric of about 15 nm is formed. A number of layers is then put together with the in- dividual metallized surfaces in- sulated from one another by the gettering of quartz (old method) or gaseous silicon nitride (modern method). The latter material protects the wafer 4.26 elektor india april 1967 Fig. 4. Everything looks normal in the right hand part of this picture, but the ultrasound photograph at the left hand side shows serious damage to the interconnecting tracks. more adequately from im- purities. The completed wafer is pro- vided with a protective barrier and cut into individual chips, which are then suitably encap- sulated. As already stated, the entire production process is carried out in a dust-free environment. It is clear in view of the high in- vestment costs that manufac- turers can not tolerate a high rejection rate. All dust particles larger than 0.5 nm can cause short-circuits in the horizontal plane, but these particles are relatively easily removed from working areas. It is, however, much more difficult to cope with dust particles of 0.1 jjm: these are a hundred times as common as 0.5 ^m particles, are invisible, and lead to short- circuits in the dielectric of the memory locations in the ver- | tical plane of the chip. It is worth reflecting on the fact that dust-free production spaces of Class 10 are one hundred times cleaner than any hospital oper- ating theatre. • Final test Before the wafer is cut into in- dividual chips, an electrical fi- nal test is carried out on any two chips simultaneously. In this, 22 needle probes many times form a connexion between chip and test equipment (see Fig. 5). The measurements thus obtained give a clear picture of the qual- ity of the individual memory cells. Subsequently, certain test patterns are run through all the memory cells. Finally, a number of chips are taken at random from a production batch and subjected to a life test in which 100 000 hours of operation are simulated over a period of 30 hours. Manufacturers As already stated, a number of manufacturers already have 1 Mbit DRAMs or EPROMs in production, while others are about to commence fabri- cation. Only one manufacturer, Toshiba, has so far succeeded Fig. 5. During test of the completed chip, 22 needle probes scan all 1 048 576 memory cells at very high speed. (Photograph courtesy of IBM). in producing a prototype of a i Fujitsu produce a 1 Mbitxl 1 Mbit SRAM. The memory cell | DRAM, structures of this device, which is made in CMOS technology, Hitachi produce a 1 Mbit x 1 oi measure only l^m. No fewer 256 Kbit x 4 DRAM in CMOS than 2.2 million circuit elements technology with cell structures have been squeezed onto a of 1.3 They also produce a substrate of only 5.99x13.8 mm. 1Mbit (128 Kx8 or 64 Kxl6) To be sure, the Toshiba device EPROM, is both static and dynamic. Dynamic, because a refresh of IBM has produced 1 Mbit the capacitor charge is DRAMs since April 1986. These necessary, and because each devices are in FET technology memory cell consists of one with structures of 1 pm. All pro- transistor and one capacitor, duction is, however, used in Static, because it does not need IBM computer manufacture, an external refresh controller (this is provided on-chip), and Intel manufactures a 1 Mbit because it has an 8-bit bus. It is EPROM in HMOS-II-E tech- for these reasons that Toshiba nology with structures of 1.4 ^m calls the device a virtual SRAM, and configurations of 128 K x i.e. VSRAM. The device has an 8 bits, access time of 62 ns. As pointed out, all DRAMs need Matshushita produces a 1 Mbit an external refresh controller. DRAM in NMOS technology. Monolithic Memories have therefore developed con- NEC manufactures a 1 Mbit trollers Type 673103 and 673104, DRAM in either NMOS oi which enable several DRAMs CMOS technology with cell with access times less than structures of 1 M m. The device, 150 ns to be controlled simul- which has been available since taneously. the summer of 1986, is ob- Manufacturers engaged in the tamable either as a 1 M x 1 or development or production of 256 K x 4 memory. The company Mbit memories are listed also produces a 1 Mbit EPROM below. in CMOS technology with struc- tures of 1.5 pm. This device is AT&T produce a DRAM of available either whith a 128 K x 8 1 Mbitxl or 256 Kbit x 4 in or with a 64 K x 16 configuration. CMOS technology with cell structures of 1 |im. This OKI produces a 1 Mbit DRAM company started DRAM devel- in NMOS technology, opment in 1984. Fig. 6. This 1 Mbit ORAM from Toshiba is already available in the electronics retail trade. company will also launch a 1 Mbit EPROM later this year which will have structures of 1.0 (im and be available in con- figurations of 128 K8 x 8 or 64 K x 16 Philips is developing a 1 Mbit EPROM, which is anticipated to go into production early next year, as well as a 1 Mbit SRAM. Prototypes of the SRAM are ex- pected towards the end of next year. Toshiba was the first producer of Mbit memories: the first 1 Mbit device was put on the market in November 1985. The device is manufactured in CMOS and is available as an SMD. Its structures are 1.2 pm. This DRAM is available as Type TCS11000C-10 or TC511000-12 (see Fig. 6): the RRP of the latter is £35. These types have access times of 100 and 120 ns respect- ively, and require a supply of 5 V. They are driven in a similar manner as the 256 Kbit devices. The company is also launching a 1 Mbit CMOS EPROM this spring. Siemens is developing a 1 Mbit DRAM and a 4 Mbit DRAM; the former is expected to go into production later this year. This device will have cell structures of 1.2 urn and be produced in CMOS; it will be organized as a 1 Mbitxl or 256 Kx 4 memory. Moreover, the 1 Mbit DRAM will be manufactured as a surface-mount device Texas Instruments is just about commencing production of a 1 Mbit DRAM in CMOS, which will also be available as an SMD (surface mount device). Cell structures are 1 pm. The NEWS • NEWS • NEWS • NEWS • NEW! New Epson printers The new M26XX range of miniprinters, designed for retail and EPOS applications, features improved paper feed mechan- isms, wherein the feed is after the printhead to reduce the possibility of jamming. Other benefits are fast paper feed (20 + lines/second), and a quiet paper feed mechanism, easy-to- change ribbon cassette, and 1-line validation. The new printers are very reliable and need no coin barrier as the paper exit is not vertical. The illustrations show that with conventional paper feed mechanisms paper jams inside the machine if the paper exit is obstructed and on-the-spot at- tention is difficult. With the paper feed roller positioned after the print head, paper jam- NIW MMII W.«. Patwr f..* HmIwIu, ming can only occur at or near the paper exit and so can easily be cleared. The new printers measure 164 x215 x159 mm and accept paper up to 44.5+44.5 mm wide. The M2630 has 17+17 columns, a 7-wire head, and a print speed of 2.4 lines per sec- ond. The M2660 has 21+21 columns, a 9-wire head, and a print speed of 3 lines per second. Epson (UK) Ltd Dorland House 388 High Street Wembley HA9 6UH eteklor India apnl 1987 4.27 REMOTE CONTROL IN ASTRONOMY by Dr Paul Murdin, Royal Greenwich Observatory Light from quasars and galaxies in distant regions carries messages about the beginnings of the universe. But perceiving it is difficult unless telescopes are sited on mountain tops to avoid interference from scattered artificial light. Astronomers have taken the first steps in operating such telescopes from data centres and remote control rooms instead of travelling the world to distant mountains. This is not just a matter of economy: it brings advantages of efficiency and new opportunities. Astronomers who study radi- ations which cannot penetrate the Earth’s atmosphere are used to operating their telescopes remotely. They have had to, for such telescopes have been car- ried on board satellites orbiting the Earth. The International Ultraviolet Explorer (IUE) oper- ated by the European Space Agency tracking station from Madrid, the European X-ray Astronomy Satellite operated from Darmstadt, Germany, and the Infra Red Astronomy Satel- lite with its ground station in the Rutherford Appleton Labora- tory at Chilton, near Oxford, are all satellites in which UK astron- omers have played an active part and to which remote ob- serving techniques have been applied. It is not so clear why a telescope on the ground has to be remotely operated. Until re- cently, most telescopes have been near the astronomers’ bases and could be operated by the astronomers travelling to them. But over the last 20 years astronomers have sought out places on distant mountains as sites for their telescopes, and that is -why we have begun to apply space techniques to ground-based equipment. Why are we putting observ- atories on mountains, in distant countries and in relatively inac- cessible places? Quasars and galaxies which formed in distant regions soon after the origin of the universe emit light which carries messages about conditions in the so-called Big Bang. The light travels to us over times of the order of 10'° years, over ! distances of some 10 23 km. It is j considerably diluted by travel- ling this distance (understate- | ment!) so, to gain information about the conditions in and after the Big Bang, astronomers have to study faint objects. Their light is perceived against a ‘noise’ of background light from contaminating sources such as artificial light scattered by dust and their images are blurred by passage through the Earth’s atmosphere. In this diffi- cult study, maximum infor- mation can be gained only if the faint sources can be seen with the highest contrast, as the sharpest images against the darkest sky. These are the reasons why astronomers have built telescopes in the clear air on mountain tops, far from population centres that emit light and give off smoke. British astronomers have access to telescopes in Australia, Hawaii and, most recently, the Canary Islands, far from the over- crowded industrialized areas of Europe. Relative cost The decision whether to take the telescope’s control and op- I eration system to the user, or the user to the telescope, depends upon the relative cost of travel and that of communi- cations. The UK travel budget for astronomy is over £1 million j per year and, as fuel becomes more expensive relative to the costs of communications band- width, the remote operation of a telescope becomes more cost- effective. Remote operation can also offer certain advantages to the astronomer. If the observ- j atory is very far above sea level it might be advantageous to | overcome the astronomer’s in- efficiency in working at an ' altitude where oxygen depri- I vation can spoil judgement. One astronomer reporting on i his experience at Mauna Kea in Hawaii at 4200 metres is quoted as saying “I confused the co- ordinates and pointed the telescope to the wrong place to take my picture, but that didn’t I matter because halfway through : developing the photograph in i the fixer I realised I had the j darkroom lights on." The simplest observations to make remotely are those which j are repetitive and which generate simple measure- ments. The Carlsberg Auto- No. of transmissions Data 80 2048 x 128-pixel 16-bit images 320 256 x 256 point 16-bit images 10 400 2048 pixel 16-bit graphs 21 600 24 lines x 80 character/ line VDU screens 5 200 60 lines x 132 character/line printer pages Using a 9 6 kbit/s connection between La Palma and the UK, any I of the data in the table (or an appropriate mixturel can be trans- ' mitted during a 12-hour night. matic Meridian Circle (CAMC) on La Palma in the Canary Islands makes observations of this kind. The instrument, jointly built by the Copenhagen University Observatory and the Royal Greenwich Observatory, is a telescope which rotates around only one axis, in a North-South plane. Its purpose is to time stars as they transit through this plane, and to measure their angle of elevation above the horizon. Effectively this measures the positions of the stars, and indeed the planets, including the one on which the telescope is mounted. Construction of a consistent model of the inter- relations of the star positions, and their change from decade to decade, yields information about the motion of the Earth and the dynamics of the solar system and our Galaxy of stars. Traditionally, transit measure- ments have been made by noting the time a star is seen by eye to pass behind a vertical cross hair and measuring its position along the hair. In the new technique used by the CAMC, the star is imaged on to a V-shaped mask which is scanned back and forth". Starlight passing through the mask is read by a photo- multiplier: the phase infor- mation in its output yields the time of transit and the duty- cycle of the signal yields the position of the star along the V. The telescope is automatically operated by two minicomputers which select stars from a pri- ority list held on a disc, position the telescope to catch the selected star for transit, make 4 28 eleklor mdta aqnl 198/ the measurements and reduce the data. They monitor the at- mospheric conditions and cover the telescope whenever it rains; they check for cloud and malfunctions, and they make calibration measurements on a schedule. The efficiency of the telescope is such that it measures the position of 1000 stars per night to an accuracy of 0 2 arc sec (equivalent to measuring the diameter of a 1 p coin, 20 mm, at a distance of nearly 20 km). It constructs in one year a complete catalogue of star positions which would formerly have taken a decade to observe and another decade to reduce. The CAMC operates remotely in the sense that, after it is primed at the beginning of the night, the telescope works without intervention; in fact, the astronomer operating it sleeps some distance away from it. Next step in its operation will be to gain access to its pro- gramme priorities and to the re- duced data from the UK. The longest link yet achieved in remote-control ground-based astronomy is between the Royal Observatory, Edinburgh, and the UK Infra Red Telescope (UKIRT) on Hawaii. Data streams travel by microwave link from the telescope at 4200 metres to the sea-level base at Hilo and then via a chain of packet switched networks across the USA to Scotland. This system, using the relatively low data rates of infrared astronomy based on point-by-point ac- cumulation of data, instead of accumulation of images, has been successfully used for a couple of years. The CAMC and UKIRT generate data at low rates. Before I describe how to operate an optical telescope remotely, let us look at how it is used. Use of telescopes A typical large, professional telescope with a mirror of about 4 m diameter is used by some 100 astronomers per year; their experience varies enormously from student novice to pro- fessional tyro. The telescope is operated from a control desk some 10 m from the telescope by a professional telescope operator: the traditional name for this person is 'night assis- tant’ but the radio astronomy term ‘telescope driver' is. also used. On a prompt from the astronomer, the night assistant causes the telescope to slew to the next star to be observed. A picture through the telescope is presented to the astronomer who then identifies in detail the object he wishes to observe and the telescope is adjusted to point directly to it. The picture is, typically, presented from a television camera viewing the phosphor of an image inten- sifier. On La Palma, the 2-5-m Isaac Newton Telescope oper- ated by the Royal Greenwich Observatory uses intensified television cameras to acquire stars and the pictures can be in- tegrated by allowing charge to accumulate on the target for several seconds before it is read, and/or by averaging suc- cessive pictures in a 512-pixel x 512-pixel x 16-bit memory. Although the picture contains half a megabyte of data, there are usually only a few signifi- cant features in it, so its infor- mation content is much less. It may be that a list of, say 10 stars, including their positions and brightness, is all that is needed to reconstruct the picture. A kilobyte will do for this. After the telescope is pos- itioned accurately, _ it is kept tracking accurately by closed- loop servos to follow the star in its rising and setting across the sky. No data is sent to any remote point in this process, which is all related locally to the telescope. But its performance is monitored by viewing reflected starlight from the en- trance plate of the instrument that is being used to analyse the star, and this image is trans- mitted back to the astronomer. The instrument might be a spectrograph for measuring the wavelengths and intensities of spectral lines in the star. Data is produced by the spectrograph in the form of another image which is read by a detector. It would not be unusual for the telescope to follow a star and for the detector to integrate on its signal for minutes or hours. Not all this integration time is available to transmit previously- acquired information. If the detector is an intensified television system, the signal ac- cumulates in a memory and is available for inspection during the integration. On the basis of a preliminary analysis of a partial integration the astronomer can decide what to do: for example, he may abort because what he wants to measure is not present, or integrate until the signal-to- noise ratio of a feature hidden in the spectrum becomes large enough. The Royal Greenwich Observatory’s La Palma telescopes and the Anglo- Australian Telesope located near Sydney use an Image Photon Counting System (IPCS) to record data. The IPCS fea- Packet switched link Remote operations network for an optical telescope sited at La Palma. elektor india april 1987 4.29 tures image sharpening tech- the only point of issue is safety bottleneck in remote operation | access to this system, which niques and forms images made of personnel and equipment, of optical telescopes lies in data ! provides common data- by accumulating the signals of Altering the equipment con- generation: the dynamic range ; reduction software for analys- individual photons. It is capable figuration and monitoring its in astronomical data, which j ing astronomical images. Once of generating images 2048 x status also requires only a low contains information from the data from the telescope enters 514 pixels in area and at least 16 | bandwidth. Acquiring the star very bright to the very faint, the system, hundreds of man- bits in depth, and produces J field, finely positioning the j raises a problem in the process years worth of astronomical more data more often and with | telescope and monitoring its j of data compression without data reduction software can be more requirement for interne- j position need a higher band- | clipping, and the analogue brought to bear on it, wringing tion; in fact, the IPCS is the j width but image condensation j transmission technique used in the last bit of information from critical test for a remote oper- j techniques are available to j the Kitt Peak video display is the very last photon, ation centre. | present a digest of star field | not suitable. However, over a At the end of the integration the | to the astronomer within the j 12-hour night with a 9 6 kbit/s data are passed into a storage | 60 seconds that is the longest connection between La Palma Programme flexibility medium. Many astronomers | he will tolerate. The Kitt Peak ! and the UK, any of the amounts j Remote operation of telescopes would like, in their excitement, | Observatory 2 • 1-m telescope in of data listed in the table, or an I is stimulated by its technologi- to begin detailed analysis im- | Arizona can be remotely oper- appropriate mixture of them, j cal timeliness, by frequently ris- mediately, and the scientific j ated by what is known as a can be sent. This is just enough j ing travel costs and by the advantages are obvious: dis- | travelling operation station, to operate a productive, moun- | efficiency it brings. It also af- coveries are made when the which use's a video expander to tain-top optical telescope from fords programme flexibility. At adrenalin is flowing. | receive the acquisition field a home station in the UK. present, astronomers are after it has been compressed Once remote operation at a scheduled to use a big ... .. for transmission over telephone central home station is estab- telescope for nominated nights Various problems lines. The unit, part of an fished, one of the next steps and they use it 'come rain or \ I In remote operation of analogue device that generates, is to extend the number of shine'. Eveniftheskyisclear.it telescopes, each of the parts of transmits and receives slow- stations, thereby making it poss- is largely a matter of chance the observing sequence scan pictures, was developed lble to link many universities whether the weather conditions presents different problems to to meet a need for remote sur- into a common programme of are exactly matched to the type be solved within the monetary j veillance by security staff, j astrophysical enquiry, each : of observation the astronomer scale set by the staff travel j Digital compression and trans- j using its specialist knowledge wishes to make; certain particu- budget which would be saved, | mission is even better adapted | to interact with the data and en- larly critical observations may unlike a budget commensurate \ to high-modulation star pic- | sure that the programme sue- need special and infrequent with the launch of a rocket, tures. j ceeds. UK astronomers already ; conditions. It is not practical to Positioning of the telescope is So, remote control of telescopes j have access to a system known 1 house dozens of astronomers within the capacity of a low- from 1000 km away is easy, a as Starlink, which uses nine j on a mountain for weeks at a bandwidth command channel ; simple extension of what is finked computers. Some 90 per | time and move them on and off (even of voice instruction!) and I already done over 10 m. The cent of British astronomers have | the telescope as conditions change, but if they can observe remotely, from their university offices, they can be scheduled flexibly and at short notice whenever suitable weather con- ditions become available. The 4 2-m William Herschel Telescope being built on La Palma by the Royal Greenwich Observatory, is the first telescope to be designed with this in mind. Its particular op- tical design, called after its Vic- torian engineer inventor James Nasmyth, incorporates a mirror which can switch the light beam from instrument to instru- ment at a minute’s notice. At least four instruments can stand by for development as weather conditions and astronomical programmes change. It may be that the next gener- ation of astronomers will look back with amusement and perhaps envy at our present travel to distant, exotic places. After the age in which avionics technology has brought the astronomers to the mountain, in- formation technology will in- Television image of the spiral galaxy M51 in a field of stars. The image is contrast-stretched and clipped so that stars appear as circles sized according to their brightness. Such an image is good enough to identify the nebula and to position the telescope on, say, one of the bright patches in the spiral arms. Its information content is much less than the original half a megabyte of data generated. stead bring the mountain to the astronomers. 4.30 elektor india april 1987 MSX EXTENSIONS — 5: EPROM PROGRAMMER (1) This two-part article describes an advanced EPROM programmer intended to work with an MSX home computer. Fully supported by a tailored software package, this peripheral programmer enables MSX users to read, program and copy EPROMs with a capacity of 2 up to 64 Kbytes. Many of the world's leading semiconductor manufacturers ostentatiously partake in the apparently unending race toward the design of ever faster and more capacious types of EPROM (Erasable Program- mable Read Only Memory). In- teresting as any new device in the series may be, most of us will only consider its practical use in, say, a microprocessor- based system if 1. the one-off price of the de- vice is acceptable; 2. the device operates from a single S V supply; 3. the device contents can be erased conveniently with the aid of an ultra-violet (UV) light source; 4. the pin-out of the device is in line with that of its prede- cessors. EPROMs nowadays come in a wide variety of types, each with its particular access time, power consumption and pro- gramming method. Though fairly exhaustive, Table 1 re- mains but an attempt at enu- merating the most commonly encountered EPROMs. As evi- dent from this list, there is a strong tendency among EPROM manufacturers to use interactive programming and lower programming voltages with increasing device ca- pacity. Thanks to the fast prog- ress in semiconductor tech- nology, even the slowest of Types 2764 and up now feature an access time of 250 ns, while the use of CMOS devices is now common practice to consider- ably reduce power consump- tion and susceptibility to digital noise. The EPROM programmer de- scribed in this article is driven by the MSX I/O and timer car- tridge featured in Elektor India, February 1987. The first part of the present article deals pri- marily with the necessary hard- ware; next month we will dis- cuss the software that has been written for the programmer. Block diagram The proposed MSX EPROM programmer is functionally set up as shown in Fig. 1. Two ports, A and B. on the I/O & timer car- tridge provide the addresses | for the EPROM to be pro- i grammed, while port D is used j to read and write datawords. j Port C drives the control inter- face on the programmer board. By writing the appropriate bit combination to port C, Vcc for the EPROM can be made 5 or 6 V, and Vpp can be made 5, I 12.5, 21, or 25 V. Port C also con- trols EPROM inputs OE and CE as required for the READ, VERIFY, or PROGRAM mode of the programmer. The CTC (Counter/Timer Con- troller) in the cartridge is pro- grammed to drive a software- controlled set-reset bistable in charge of the pulse timing dur- ing the PROGRAM mode. A jumper block is used to ar- 1 range for all programmer sig- nals to be fed to the appropriate EPROM pins. The logic circuits on the pro- grammer board are fed from the computer’s built-in 5 V supply. The programming and supply voltages— Vpp and Vcc— for the EPROM are available from a mains supply incorporated in the program- I mer. Circuit description A quick recap on pinning and signal denotations used for EPROMs in the 27XXX series is given in Fig. 2. It should be | noted that some EPROM manu- facturers— notably Texas Instru- I ments with their 25XX types— deviate slightly from the indi- cated convention. The present EPROM program- mer is not a very complex cir- cuit, as can be seen from Fig. 3. The EPROM addresses are taken from PIO Ports A and B— i.e., from ICi on the car- tridge board. Port A provides the least significant address byte (Ad. . . Ar), Port B the most significant byte (As . . . A(s). As the "smallest" EPROM that can be programmed is the 2 Kbyte Type 2716 (or 2516 from TI), ad- dress lines As. . .A to are con- nected direct to the relevant pins on the EPROM socket. The remaining address lines appear on an extensive jumper block. elektor indie apnl 1987 4.31 Table 1. « • memory programming memory programming Manufacturer Type organization Vpp method note(s) Manufacturer Type organization Vpp method notelsl r AMD AM9716 2Kx8 25 V N National NMC2716 2Kx8 25 V N AM2716 2K x 8 25 V N Semiconductor NMC27C16 2Kx8 25 V N CMOS AM2732 4Kx8 25 V N NMC27C16H 2Kx8 25 V F2 CMOS AM2732A 4Kx8 21 V N NMC27C16B 2Kx8 12.5 V 1 CMOS AM2764 8K x 8 21 V N; 1 NMC27C32 4K x 8 25 V N CMOS AM2764A 8K x 8 12.5 V 1 NMC27C32H 4K x 8 25 V F2 CMOS AM27128 16Kx8 21 V N; 1 NMC27C32B 4K x 8 12.5 V 1; F2 CMOS AM27128A 16Kx8 12.5 V 1 NMC27C64 8K x 8 12.5 V 1; F2 CMOS AM27256 32Kx8 12.5 V i NMC27CP128 16Kx8 12.5 V 1; F2 CMOS AM27512 64Kx8 12.5 V • NMC27C256 32K x8 12.5 V 1; F2 CMOS Fujitsu MBM2716 2Kx8 25 V N NMC27C512 64K x8 12.5 V 1; F2 CMOS MBM8516 2K x 8 25 V N NEC yPD2716 2Kx8 25 V N MBM2732A 4Kx8 21 V N J-PD2732 4K x 8 25 V N MBM27C32A 4K x 8 21 V N CMOS m PD2732C 4Kx8 25 V N OTP • MBM2764 8Kx8 21 V N 1 ^iPD2732A 4Kx8 21 V N MBM27C64 8K x8 21 V N 1 CMOS yPD2764 8K x 8 21 V N 1 MBM27128 16Kx8 21 V N 1 |iPD27C64 8Kx8 21 V N CMOS MBM27256 32K x8 12.5 V 1 iiPD2764C 8Kx8 21 V N OTP MBM27C256 32Kx8 12.5 V N; 1 CMOS MPD27C64C 8Kx8 21 V N CMOS MBM27C512 64Kx8 12.5 V 1 CMOS yPD27128 16K x 8 21 V N OTP HN462716 25 V h PD27128C 16K x 8 21 V N 1 OTP Hitachi 2Kx8 N *iPD27256 32K x 8 21 V HN462532 HN462732 4K x 8 4K x 8 25 V 25 V N N m PD27C256 32K x 8 21 V 1 CMOS HN462732A 4K x 8 21 V N Rockwell R87C32 4K x 8 21 V N CMOS HN482764 8K x 8 21 V N 1 R87C64 8K x 8 21 V N CMOS HN27C64 8Kx8 21 V N i CMOS R27C64P 8K x 8 21 V N CMOS HN482764P 8Kx8 21 V N OTP SEEQ 2764 8K x 8 21 V N 1 HN4827128 16Kx8 21 V N i 5133 8K x 8 21 V N 1 HN27128P 16K x8 21 V N 1 OTP 27128 16K x8 21 V N 1 HN27256 32Kx8 12.5 V 1 5143 16K x 8 21 V N 1 HN27512 64K x8 12.5 V 1 27C256 32K x8 12.5 V 1 CMOS Intel 2716 2Kx8 25 V N SGS/ATES M2716 2Kx8 25 V N 2732A 4K x 8 21 V N M2732A 4Kx8 21 V N P2732A 2764 4Kx8 8Kx8 21 V 21 V N OTP M2764 8Kx8 21 V N; 1 N; 1 P2764 8K x 8 21 V 1 OTP Texas TMS2516 2Kx8 25 V N FI 2764A 8K x 8 12.5 V 1 Instruments TMS2532 4Kx8 25 V N; FI 27C64 8Kx8 12.5V 1 CMOS TMS25L32 4Kx8 25 V N LP P2764A 8Kx8 12.5 V 1 OTP TMS2732 4K x 8 25 V N 27128 16K x 8 21 V N; 1 TMS2732A 4K x 8 21 V N 27128A 16K x8 12.5 V i TMS2564 8K x 8 25 V N FI P27218A 16Kx8 12.5 V OTP TMS2764 8K x 8 21 V 1 27256 32Kx8 12.5 V 1 TMS27128 16Kx8 21 V N 1 27C256 32Kx8 12.5 V 1 CMOS Thomson- ET2716 2Kx8 25 V N 87C256 32Kx8 12.5 V 1 CMOS CSF ETC2716 2K x 8 25 V N CMOS 27512 64Kx8 12.5 V 1 ETC2732 4Kx8 25 V N CMOS 27513 4x16Kx8 12.5 V 1 Paged ET2764 8Kx8 21 V N Mitsubishi M5L2716 2Kx8 25 V N Toshiba TMM323 2Kx8 25 V N M5L2732 4Kx8 25 V N TMM2732 4Kx8 25 V N M5L2764 8Kx8 21 V N TMM2764 8K x 8 21 V N | M5L27128 16K x 8 21 V N; 1 • TMM2764DI 8K x 8 21 V N 1 M5L27256 32K x 8 12.5 V 1 TMM27128 16K x 8 21 V N 1 Mostek MK2716 2Kx8 25 V N TMM27256 32K x 8 21 V 1 Motorola MCM2716 2Kx8 25 V N T C57256 32Kx8 21 V 1 MCM27L16 2K x 8 25 V N LP MCM2532 4Kx8 25 V N MCM25L32 4K x 8 25 V N LP MCM68764 • 8K x 8 25 V M MCM68766 8Kx8 25 V M Abbreviations used in this table: MCM68769 8K x 8 25 V M 1 = interactive programming. N = normal programming 150 ms cycle). FI - fast programming 120 ms cycle). F2 = fast programming 110 ms cycle). The type indications as given may be followed by an access time specification. M - Motorola programming method; not supported by this EPROM programmer. LP = low-power device. The inclusion in this Table of PROMs and EPROMs does not imply their being OTP = one-time programmable device. programmable with the MSX EPROM programmer described in this article. CMOS = complementary metal oxide device. Table 1. A useful list of EPROM types and their technical characteristics 4.32 elektor india april 1987 Hi I / O + timer cartridge Fig. 1. Block diagram of the EPROM programmer for MSX microcomputers. K 2 . where the connections to the EPROM pins can be made as required. All address lines on the programmer board have low-value series resistors to avoid PIO outputs being dam- aged by a defective EPROM. PIO Port D— ie., IC 2 port B on the cartridge— serves to pass databytes to and from the com- puter. As on the address bus, protective resistors have been fitted on the Do. . .D7 lines. All programmer functions are controlled via Port C— i.e., IC 2 Port A on the cartridge. De- pending on the type of EPROM in question, the correct com- bination of function control bits is available at Port C, bits A 0. . ,A 6 . Port C bits 0 and 1 select one of four programming voltages, S, 12.5, 21 or 25 V. One section of dual two-to-four decoder IC? translates the bit pattern of Ao- Ai into a low level atjhe corre- sponding TY0...1Y3 output, which then causes one of four voltage determining networks to be connected to ,the refer- ence input of voltage regulator IC?. Each output on ICa drives two open-collector (OC) TTL buffers; one to enable current to flow through the associated resistors RiJ-Rs*; R25-R35; R 26 -R 36 or R22-R37, and one to drive the associated Vpp indication LED. Example: writing V PP 0=1 and V PP 1=0 to the programmer causes ICj to activate output TYT, LED Ds to light and IC? to output 21 V with Ra-R 26 -R 36 and the resistance of the OC output of buffer N 7 determining the output voltage. The operation on Vpp regulator IC? will be reverted to in due course. As some of the more recently introduced EPROMs require Vcc to be raised from 5 to 6 V during interactive program- ming, provision has been made to enable the computer to sel- ect either one of these voltages as appropriate. Port C bit 3, via inverter Ne, selects either R 33 -Rn, or parallel network R 33 + R 11 //R 31 + R32 + RoC(N 5 ) to determine Vom of ICe. The former condition is brought about by the output of Ns being high (A3 = 1 ; Vcc =5 V); the latter by the output being low (A 3 = 0; Vcc= 6 V). LEDs D3 and Da clearly show the presence of the currently selected value of Vcc. Port C bit A 2 switches V PP on or off, and bit A a does the same for Vcc. Port C bit As determines the logic state of EPROM input OE (output enable), which must be low during READ operations. A two-LED indication, D12-D13, shows the data direction, ie., from the computer to the EPROM (PROGRAM), or vice versa (READ, VERIFY) __ Port C bit As controls CS (chip select) of the EPROM. Diodes Dis-Di? and pull-up resistor R30 are an AND gate to ensure the correct driving of the CE/PGM pin on EPROM Types 2532, 2564, 2732, 27256 and 27512. As with OE, V PP , A 11 ...A 15 and PGM/PGM, signal C3 appears on jumper block IG to ensure that every EPROM pin is driven with the appropriate logic level. Port C bit A? is the only one set for operation as an input line. The control program running on the MSX computer checks for the presence of a logic 0 on this line, whose state is con- trolled with push-button Si. Depression of this switch causes the program to halt and return to the main menu. Pressing CRTL and STOP on the computer returns control to the BASIC interpreter. Pull-up resistors have been fit- ted on all Port C control lines to elektor india aprll 1987 4.33 Fig. 3. Circuit diagram of the EPROM programmer, which is driven by the I/O and timer cartridge featured in our January issue. ensure a correct bit-configur- ation at power-on. The Vcc and Vpp supplies on the program- mer board are essentially ident- ical circuits based on the well- known Type L200 regulator. Port C bits A 2 and Aa, when high, cause Ta and T 3 to be driven hard by OC buffers No and Nib, respectively. In this manner, the current sense input of the associated regulator chip is pulled to ground, causing the IC to turn off its internal output driver. The "hard shut down" arrangement is simple and ef- fective to ensure the absence of overshoot on the Vpp and Vcc lines. Fig. 4 shows this quite evi- dently. The output voltage of Vpp regulator IC 7 was pro- grammed to step from 5 to 25 V with Vpp-off intervals between successive steps. The test was carried out with a Type 2732 plugged into the programmer. The Vcc and Vpp supplies are short-circuit resistant and can supply 100 and 50 mA, respect- ively, as defined with Rj (ICs) and R 7 (IC;). Decoupling capacitors Ci and C 3 -C 11 afford protection against spurious voltage transients on the Vcc and Vpp lines. Both supplies have an on/off indicator to enable users to spot defective EPROMs at a glance. The 5 V supply for all logic cir- cuits on the programmer board is taken from the cartridge via Kt pins 21 and 22. This means that the computer actually feeds both the cartridge and the programmer from its inter- nal 5 V supply. As already ex- plained in the article about the MSX cartridge, users should be well aware of the capacity of this supply, and take every pre- caution not to overload it by connecting the peripheral boards. It will be recalled that the current source capability of the standard MSX slot is 300 mA. The programmer and the cartridge can be expected to draw a total of 100 to 250 mA, but it is none the less wise to actually measure this with the aid of a regulated supply, be- fore connecting the boards to the MSX slot. The programming pulses for the EPROM are obtained from S-R (set-reset) bistable ICs. Two units in the CTC in the cartridge are programmed to operate in the TIMER mode. When started, timer output 0 (TOO) is set up by the software to pro- 1 vide a 4 ^s delay to ensure 4.34 elektor India apnl 1987 word for each programmable EPROM will be given in next month's instalment. For now, the jumper configurations on K 2 are given in Table 2. Construction The EPROM programmer is constructed on ready-made, thrOugh-plated circuit board 87002— see Fig. 5. It is a fairly densely populated board, but its completion should not pre- sent too many problems if the soldering is done with care and precision. To save board space, all resistors, except R 31 ...R 37 incl. are mounted vertically. The L200 regulator chips can do without heatsinks. It is essential to start the fitting of parts with those at the track side of the PCB. i.e., all LEDs, the jumper block, and the ZIF socket (consult Fig. S, these parts are shown in dashed lines). Depending on the en- Table 2. K2 2716 2516 • 2732 2532 2764 2564 27128 27256 27512 pos. signal 1 n.c. 2 CE 1 CE 3 PGM 1 CE/PGM 9 PD PGM 4 Vpp 1 Vpp w^r~ 5 OE fir 6 CE ■ CE 7 8 9 OE pz I 51 1 Qi An 9 An r 9 An 1 9 An CE/PGM 9 C5 PGM fl CE/PGM 1 CS PGM 10 Vpp 9 OE/Vpp 1 OE/Vpp 11 Vpp 1 Vpp 12 An 1 An 1 An 13 CE/PGM 9 PD/ PGM I PD, PGM 14 Air ■ Al2 9 9 Vcc ■ Vcc 1 Vcc 1 Vcc 1 Vcc 9 i mm 1 ri 1 n.c. 1 sl 17 OE ■ nc I ■ n.c. I n.c. 1 S2 9 Vpp 1 n.c. 1 n.c. ■ n.c. | n.c. 1 Vpp 9 1 PGM I PGM m Vpp 1 Vpp i Vop ■ \/pp m 1 A, 2 1 An 1 Au 1 Air m ■ n.c. ■ An 1 Au ■ A ” 9 A 14 1 Ai. ■ Al4 9 An 8~Ais 9 n.c. that EPROM data and addresses are stable before activating the PGM /PGM line. CTC output TO0 is also applied to the CLK input of the second timer in the Z80-CTC package. This timer is started at the first zero count of timer 0 and its output period is of the order of 0.5 ms, since the programmed divide factor is 7x256. The third timer in the CTC is programmed to operate in the counter mode and counts a variable number of 0.5 ms pulses at its input. In next month's final part of this article we will revert to the practical use of the software-driven PGM pulse generator. For now, it is readily seen that the timing out of the third counter causes the S-R bistable to be reset. The 0 and 0 outputs of ICs are made available on jumper block K 2 (PGM; PGM), and LED Dio pro- vides an indication for the presence of the programming pulses. The EPROM programmer has a built-in power supply of con- ventional design delivering the raw input voltages for the Vcc ; and Vpp regulator circuits. There are a few rather import- ant considerations for this supply, and these will be dis- cussed in the following section. I In conclusion of this circuit ] description, it is seen that the bit-configuration of Port C lines Ae. . .A6 is specific to the type of EPROM plugged into the ZIF socket. The appropriate control Fig. 4. The turn-on and turn-off characteristics of the L200 in pos ition ICt ensure the absence of overshoot on the Vpp line. Table 2. Two blocks of jumpers suffice to select a wide range of EPROMs. fllefctor india apnl 1987 4.35 74LS05N A . SKK8425 HT PC74HCT139P 541130VK 0200 SP tyOTUMl 8S3SFS SN7407N rad sided board. A Textool ZIF (zero insertion force) socket is undoubtedly the best to choose, and it is conveniently soldered onto two 14-way socket terminal strips soldered onto the PCB. A 28-way wire- wrap socket is also a feasible way of bridging the distance between the board surface and the enclosure top lid. The mains parts, S 2 , Tr., Tr 2 and Fi are not accommodated on the programmer board, and should, therefore, be fitted with the usual care in dealing with wires and terminals at mains potential. These parts will have to find their way in some of the left over space in the enclosure. Resistors R 31 ...R 37 incl. are preferably mounted in a 14-way IC socket to facilitate exchang- ing any of them should this be required to reach the appro- priate regulator output voltage —see Fig. 7. As you may have noticed from studying Table 2, the connec- tions on K 2 are made in blocks of three and four. This has been so arranged to make it possible to use only two "large” jumpers for the setting of all possible configurations. Simply glue together three and four jumpers and you will have very little dif- ficulty in finding the correct configuration on K 2 for your closure type you have in mind for this project, two corners of the board may have to be cut as shown on the component over- lay. The connecting leads of the ZIF socket, K 2 and the LEDs must be left long enough to enable the devices to protrude from the enclosure top lid. Think well before mounting the ZIF socket; it is not easy to remove once soldered onto the double- Resistors I + 5%) Ri;R2;R6;Rii;Ris;Rh;R2s;R27; Ris;Rj9 = 1K0 Rj-3R9 R*;Rs;Rti;R2i;R24 = 6K8 Rs = 22R R? = 8R2 R. = 15R Rio;Ri2;Ri*. . .Ri$;R2>. . .Rjo; R«o...R72 = 12K Rit — 1 MO R2o = 39K R 22 = 680R R2s = 1K5 R2. = 820R Rji = 15K R32;Rj3= 1K2 • Rm = 1K0 * Rjs = 220R ' Rs6 = 68R * Rjr = 56R ' Capacitors: ' Ci;Cj;C 5. . ,C#:Cio;Cri = tOOn C2;C. = 220n C» = tOOp; 16 V radial Cu=470g; 25 V axial Ci j = lOOOu 25 V axial Semiconductors: Bi = B80C1500 Di. . Du incl. = LED red Dm... Dir incl. - 1N4148 ICi -74LS05 IC 2 .IC 3 = 7407 IC* = 74HCT139 ICs = 4027 IC»;ICr = L200 Tt. . .T 9 incl. - BC547B Te = BC557B Miscellaneous: Ft =63 mA delayed action. Ki = 50-way male PCB edge connector (angled typel. K 2 = 50 way male PCB edge connector (straight tvpel. 7 off jumpers for K 2 . Si = push-to-make button. Sr = SPST mains switch. Tri;Trr=12 V; 1.2 VA mains transformer. * Suitable ABS enclosure, e g. OKW 9409111. PCB Type 87002 (see Readers Services). 28-way ZIF socket for EPROM (e.g. Textool 28). Fuseholder for Fi. ' See text Fig. 6. Suggested lay-out for the programmer’s top panel. 4 36 elektor mdia aprtl 1987 Fig. 7. Close-up of the set of voltage determining resistors R 31 . . . R 37 . specific EPROM type. In order to make for a neat ap- pearance of the completed pro- grammer, its top lid can be lettered as shown in Fig. 6. The connection of the programmer to the I/O & timer cartridge is made in a 50-way flat ribbon cable. A clearance should be cut into the relevant side panel of the programmer enclosure for the connection to Ki. Testing and setting up As all essential functions of the programmer have one or two LEDs to indicate the current state, the testing of the com- pleted peripheral can be done largely with the aid of software. Plug in the I/O & timer car- tridge into the MSX slot, but do not yet connect it to the programmer, whose internal supply must first be tested. Switch on S 2 and measure Vs. and Vs 2 . It is very important that VS 2 is less than 40 V under all circumstances. If necess- ary, use another set of mains transformers to prevent dam- aging IC-. Now connect the programmer to the cartridge, and switch on the computer, which ' should boot up as normal. Check for the presence of + 5 V on 2. the programmer board, and measure the voltages at the test points indicated in the circuit diagram. Proceed with keying in the test program shown in Table 3. It will allow you to see each LED on the programmer to go on and off upon the pressing of a particular function key. This is what the whole set-up should do if the circuitry functions cor- ; rectly: 1. At power-on, these LEDs should light (default state): V PP = 5 V (De); Vcc= 5 V (Dj); DATA IN (D, 3 ) POWER (D,). Running the test program causes the MSX function keys to do the following 10 ....................... TEST PROGRAM .TPROMMER 20 ' 30 '=================== ======================================= address -area 40 A= 3# 1 6 50 DA= 4+A : DB= 5+A : DC- 8+A : DD= 9+A 60 CA= 6+A : CB= 7+A : CO10+A : CD= 1 1 4 A 70 T0 - 1 2 «-A : T1-13+A : T2=14+A : T3=15 ON : NEXT 210 ON STOP GOSUB 280 : STOP ON 220 A= 1 : B= 1 : C = 2S5 : D= 1 230 '== = ========~====~=======~** a ===================-======= execution loop 240 OUT DA. A : OUT DB . B : OUT DC . C : OUT DD.D 250 IF INP (DC) < 128 THEN 280 : ' ===-========*=n*a*r*«x«= reset pressed 7 260 GOTO 240 270 '============ 3 " B3 ============="====-=======3====*-a==x= on stop routine 280 STOP OFF : GOSUB 450 : OUT DC . C : END 290 'Rotate address line hiqh =*"= 3 =========*======**======== kev 1 routine 300 KEY ( 1 ) OFF 310 IF A- 1 28 THEN A=0 : B=1 : ELSE A = A«*2 320 IF B n 128 THEN B=0 : A=1 : ELSE B=B«2 330 KEY ( 1 ) ON : RETURN 340 'Rotate datallne h l ah ==============*==x================= kev 2 routine 350 KEY ( 2 ) OFF : IF D=128 THEN D= 1 ELSE D=D*2 : KEY ( 2 ) ON : RETURN 360 'One proqram pulse of 50 ms ===-===*=*=================_= kev 3 routine 370 KEY ( 3 ) OFF : OUT T2 . &B0 1 0 1 0 1 0 1 : OUT T2. 100 : OUT T1 . &B0 0 1 1 1 1 0 1 380 OUT T1.7 : OUT T0 . &B000 1 0 1 0 1 : OUT T0 . 0 KEY(3> ON : RETURN 390 ' Vcc chanqe ~ = = = = = = aa:i * n ** :x * B * aB ** = = = = = = = = = = = = = = = = = = = = = = = kev 4 routine • 400 KEY ( 4 ) OFF : C=C AND 8 : C«C + 8 : C=C AND 8 410 C = < I NP (DC) AND 247) OR C : KEY ( 4 ) ON : RETURN 420 ' Vpp Chanqe = »s»**«» = = s«««««««»«*3 = **s = = * = = B :x = = = B = Bs a B a «a kev 5 rout lne 430 KEY ( 5 ) OFF : C»C AND 3 : C-Ol : C=C AND 3 440 C = ( I NP (EXT) AND 252) OR C : KEY ( 5 ) ON : RETURN 450 'Reset = » = = --■ = « = « = = = = = = = = = = = = = = = = - = = = = = = = x = nxxB». = = = = = = = = kev q routine 460 KEY ( 6 ) OFF : C=255 : KEY(6> ON : RETURN 470 'Chip enable =========================~b*x===== ========== kev 7 routine 480 KEY ( 7 ) OFF : C=C AND 64 : 0064 : C=C AND 64 490 0(1 NP (DC) AND 191) OR C : KEY ( 7 ) ON : RETURN 500 'Output enable ========== ====== a == a *3==================== kev 8 routine 510 KEY ( 8 ) OFF : C=C AND 32 : 0032 : OC AND 32 520 0(1 NP (DC) AND 223) OR C : KEY ( 8 ) ON : RETURN 530 'Vcc on/off =========================xx================== kev 9 routine 540 KEY ( 9) OFF : OC AND 16 : 0016 : OC AND 16 550 0(1 NP (DC) AND 239) OR C : KEY( 9) ON : RETURN 560 'Vpp on/off =================*=========3B»============== kev 10 routine 570 KEY (10) OFF : OC AND 4 : OO 4 : OC AND 4 580 C = ( INP (DC) AND 251) OR C : KEY(10) ON : RETURN 87002 -I - T3 Table 3. This test program uses the MSX function keys to check the correct operation of the pro- grammer board. (RESET aborts the pro- gram): Fl= drive successive address lines high; F2 = drive successive datalines high; F3= pulse PGM/PGM for 50 ms (temporarily fit jumper J1 at output of N.); F4 = toggle Vcc = 5 or 6 V F5 = step Vpp =5; 12.5; 21; 25 V; F6- force default functions; F7 = toggle CE; F8= toggle data flow direction (OE); F 9 = turn Vcc on/off F10= turn V PP on/off Measure Vcc and V PP during this test to see whether any one of R32. . R]7 needs adapting to enable ICe and IC7 to output the correct voltages. Adapt R33 if the Vcc supply fails to output exactly + 5 V, then check the +6V level by pressing F4; slightly adapt Rj; if necessary. Measure all four values of V PP to see whether the stated re- sistor values in the Rs< . . . R37 j positions result in the correct output of IC7. Make small changes at a time to all voltage determining resistors, and if possible use high stability types to get Vcc correct to within 0.1 V, and V PP to within 0.5 to 1 V. Next time The concluding part of this article will be published in next month’s issue. As already stated, we will then concentrate on the software for the programmer. It is our intention to make this available to you in the form of a programmed EPROM Type 27128, which should be plugged into the EPROM socket on the Add-on Cartridge Board for MSX computers, de- scribed in the March 1986 issue of this magazine. AR elekior india april 1987 4.37 BIPHASER by W Teder A sound effects unit that can add a new acoustic dimension to a wide variety of musical instruments. There are various ways of ob- taining the well-known phasing or flanging sound effect. Most phasers use phase-shifting net- works, bucket brigade delay lines, selectively activated L-C networks, comb-type filters, or the like. The present circuit utilizes phase shifting, but has none of the drawbacks gener- ally associated with this type of Phaser, since provision has been made to obviate the troublesome amplitude-modu- lation effect caused by selec- tive filtering at relatively low Phaser speed settings. Where this effect is still tolerable— and often expressly sought after— with the rhythm guitar, it all but ruins the sound of numerous solo instruments, whose par- ticular sound is not in any way embellished by appreciable volume variations. The use of a Phaser based on the periodic shifting of, say, two stop-band filters results in a very lively ef- fect with input signals relatively rich in harmonics, e.g. those of an acoustic rhythm guitar. The same phaser, however, is practi- cally useless with a solo-instru- ment. since the played notes are subject to variations in am- plitude, rather than in timbre. When analysing the correlation between phasing effects and pitch of the input sound, it is noted that relatively high fre- quency components in the in- put sound typically require modulation with a correspond- ingly fast phase modulation signal. Similarly, the best effect for low input notes is obtained with slow phase modulation. The foregoing considerations have been taken into account in the design of this biphaser, so named because of the use of two phase shifting circuits, each with its individual centre frequency and phase modu- lation speed control. These two circuits can be operated in parallel with two phaser speed settings to bring about a very good phasing effect without undesirable amplitude-modu- lation of the input signal. The circuit as presented here is but the minimum set-up of a ver- satile phaser unit whose con- trols offer a considerable var- iety in output sound. For those who wish to experiment a little further, there are interesting possibilities to extend the cir- cuit to individual needs, as will be seen in the following sec- tion. Circuit description The circuit diagram in Fig. 1 shows that the biphaser con- tains the usual building blocks of an effects unit. The mono or stereo input signal is raised in amplifier At and fed to two phase delaying circuits via R13 and R 3 «. The upper series of opamp-based all-pass filters is dimensioned for a relatively high centre frequency, while the lower series covers most of the lower part of the AF spec- trum. Notice that the delay lines are identical but for the four frequency-determining capaci- tors, C6-C9 (high cascade) and Cu-Cu (low cascade). The cir- cuits around opamps An and A12 are virtually identical, tunable oscillators which out- put a filtered triangular signal to the gates of the associated line of FETs in the delay chain. Sufficient phase shift is ob- tained from both filter lines by controlling the resistances at the + input of the opamps, i.e., the resistance of the FET drain- source junction. Presets P3 and Pa enable a precise adjustment of the bias voltage on the gate line. The FETs in this circuit are selected for matching charac- teristics, to avoid the syn- chronicity of the opamp sec- tions, and hence the final sound effect of the phaser, being im- paired. The output signals of the PM oscillators are inte- grated with the aid of Rm-Cm (high) and R13-C15 (low) to obtain sinusoidal control signals for the FETs. Three-way switch Si selects the output of either one, or both, phase shifting lines. Mixing of the original input signal with the phased signal is ac- complished by R28, R49 and Rj. Opamp A2 is the output buffer of the biphaser. The effect bypass circuit essentially con- sists of an optional footswitch, T9, and a network of electronic switches, ES1-ES4. Since the footswitch (if used) carries a direct voltage, rather than any AF signal, its connection can be made in a fairly long, unscreened two-way cable. One possible extension of the biphaser is the fitting of two phasing depth controls, P5 and 1 Ps, at the outputs of Ae and Am, 4 38 elektor tndia april 1987 ■y|HgS| ■ I— T Salnn^WSiftVaili ■■iibmbbM “1*1 “III A1 , A2 = IC1 = TL072 ES1...ES4 = IC2 = 4066 A3...A6 = IC3 = 4136N A7...A10 = IC4 = 4136N At 1 , A12 = IC5 = TL072 T1...T8 = BF245C (sal) BF256C (set) *# optional loot switch BF24SC / B BF256C \ S 1 G 0 BC547B B I S B 8702B-1 Fig. 1 At the heart of the biphaser are two individually modulated phase delay lines. elektor mdia apnl 1987 4.39 g«gj|iBng3 gM respectively— see Fig. 2a. Alter- natively, the two potentiometers can be replaced with a single stereo type as shown in Fig. 2b. The wire links at one end of Rn and Rj 9 enable both phase shift- ing lines to be driven from a single PM oscillator. A further, more radical, extension of the circuit could involve the con- struction of additional phase delay lines, each dimensioned for a specific pass-band, and controlled by an associated os- cillator. If you consider trying this out, remember to use matched FETs only, else the ef- fort is useless. The biphaser is powered from two 9-V batteries or a small sym- 00-|f-0 Fig. 2 Possible extensions of the biphaser. Fig. 4 Perform this current source test to select FETs with matching characteristics. metrical mains supply. The positive and negative supply rails are adequately decoupled with C 20 -C 26 to prevent any likelihood of noise or hum be- ing picked up. Current con- sumption of the unit is of the order of 40 mA on each 9 V supply rail. Construction and setting up There is virtually nothing to say about the construction of this effect unit. Hardly anything can go amiss if you stick to the Parts List and the component overlay shown in Fig. 3. The AF input and output of the phaser, as well as the foot switch input, are best made with insulated jack sockets, as customary with ef- fect units. The enclosure must, of course, be quite sturdy, and it is recommended to use one of the smaller types of Eddystone diecast boxes, the top lid of which can be used to fit the footswitch and the speed con- trols. Alternatively, the biphaser can be incorporated in a mains- operated, remote-controlled ef- fects unit, together with a fuzzer, a reverberation/echo unit, and the like, which can all be controlled from a set of footswitches on the stage. The completed unit requires no alignment other than setting presets P 3 and P« for an accept- able phasing rate at a minimum of distortion. This is best done with the aid of an oscilloscope and an AF sinewaVe generator set to about 1 kHz at 1 V PP . Con- nect the generator output to either one of the phaser inputs, and use the scope to monitor the phaser output signal. Adjust Ps and P« for optimum ampli- tude modulation, ie., the FETs should operate over the full ex- cursion of the sinewave, without appreciable offset and/or clip- ping. Remove the sinewave in- put signal and use a voltmeter to check whether all inputs and outputs of the opamps in the phase delay lines are at about 0 V with respect to ground. Finally, Fig. 4 shows how to sel- ect FETs for nearly identical characteristics with the aid of a simple test circuit. The FET under test is connected as a current source, and the drain- source voltage is monitored to find devices dropping the same voltage across the drain re- sistor. D SCART ADAPTOR FOR IBM PC As a growing number of colour TV sets come with a SCART input, many owners of an IBM PC will have toyed with the idea of using a SCART compatible set as a CGA-driven, medium- resolution, RGB display. Well, here is the adaptor circuit to do just that! Medium and high-resolution RGB monitors with TTL- compatible monitors are gener- ally recognized as costly devices. It is not suprising, therefore, that many an owner of an IBM PC or PC compatible starts wondering about driving the video and sync circuitry in a modern colour TV set with the TTL signals from the CGA (colour graphics adaptor) in the computer. After all, the resol- ution of the typical TV set should be adequate for the 320 x 200 pixels from the CGA. Considerable difficulty, how- ever, arises from the fact that the CGA composite video out- put supplies a NTSC signal (American TV standard), rather than a PAL signal as required for most European TV sets. The solution to the above prob- lem van be found in the use of the SCART input on the TV set; vyhat is required is an add-on interface to convert the TTL levels from the CGA outputs to SCART levels. The vertical synchronization and the hori- zontal centring adjustments in the TV set will need to be slightly re aligned to obtain a stable image from the com- | puter. When the TV set is to remain suited for normal broad- cast reception, it is suggested to fit a separate set of image ad- justment controls aligned for the IBM video standard. A simple switch then makes it easy to select the appropriate setting. Circuit description The TTL-to-SCART level con- verter is shown in Fig. 1. Those readers wishing to familiarize themselves with the SCART standard and its technical characteristics, are advised to read SCART adapter, in Elektor India. October 1985. In the proposed circuit, the level conversion is essentially from digital (0=0 V; 1=5 V) to analogue. Three identical level shifters, based around Ti...Te «' , «»tuor india apnl 1987 4.41 provide the SCART-compatible TV set with correctly rated R, G, and B signals with two intensity levels, selected with the I out- put from the CGA. With presets Pi, P 2 and P 3 set to about the dentre of their travel, a logic high I input causes the ana- logue colour outputs to vary from 0.3 V to 0.6 V, while a logic low I input gives an output range of 0 V to 0.3 V. The toggle voltage should be set at the same level for all three buffers, ie., Pi, P2 and P3 should be ad- justed for identical wiper pos- itions. The final alignment of the intensity ratio depends on your personal taste, and some time should be spent in turning the presets for best colour repro- duction on the TV screen. Transistors T? and Ts together form the synchronization mixer- buffer-inverter. The CSYNC signal is used to drive the CVBS (composite video, blanking, synchronization) input of the TV set via SCART pin 20. When you use a standard, male-male, T1, T3, T5 = BF451 , T7 = BC547B T2, T4, T6, T8 = ZN2Z19 D1...D3 = 1N414B SCART cable between the adaptor and the TV set, the CSYNC output is applied to connector pin 19. As the proposed adaptor circuit comprises only very few parts, it is conveniently built into the computer enclosure. The supply voltage can be taken from CGA pin 7, as shown in the circuit diagram. The current consumption of the adaptor is of the order of 150 mA; should this exceed the capability of the CGA board— they come in various forms and are often slightly different from the original IBM version— a separate wire may be run to the +5 V bus line on the mother- board, or a Type 7805 regulator may be used to provide the supply for the SCART adaptor board. Finally, Figures 2 and 3 sum- marize the connection between CGA and computer monitor, and the pin assignment of the SCART connector, respectively. D IBM and IBM PC are registered trademarks of International Business Machines, Inc. NTSC = National Television System Committee. PAL = Phase Alternation Line. Resistors l±5%): Ri;Rs;R*;Rn;Rir;Ri6“2K2 R 2 ;R.;Rio = 100R R3;R7;Rii = 39R Rr;R#;Ri2;Ri7;Ris = 47R Rii = 2K7 Rti-56R R20-68R Pi;P 2 ,Pi = 2K5 preset Semiconductors: Ti;Tj;T» = BF451 Tr;Ti;T.;T. = 2N2219(A) T r = BC547B Di;0r;Di = 1N4148 Miscellaneous: Ki * 9-way sub-D plug K: = 21-way angled SCART socket We regret that no ready made cir- cuit board is available for this protect. Fig. 2 Connection between CGA and RGB computer monitor. SCART connector pin function 1 audio 01 2 audio ir 3 audio 01 4 audio gi 5 B Input 6 audio ir 7 B Input 8 9 G input 10 11 G input audio output (R) audio input (R) audio output (L) audio ground B Input ground audio input (L) 8 Input G input ground 13 R Input ground 14 15 R input 16 fast blanking 17 CVBS ground 18 last blanking ground 19 CVBS output 20 CVBS input 21 connector shield nominal amplitude 0.5 v lm . 0.7 V„„ in 75R 0.7 V.„ In 75R 0.7 V in75R 3 V 1 V pp in75R 1 V pp in75R Fig.1 Only a handful of commonly available components are needed to make this TTL-to-SCART adapter for the IBM micro. Fig. 3 Pin assignment and voltage level convention of the standardized SCART connector. 4.42 elektor mdia april 1 987 SOFTWARE FOR THE BBC COMPUTER — 4: ANALOGUE CIRCUIT DESIGN Anyone who has ever designed any kind of electronic circuit knows that this should essen- tially involve the following steps: 1. enumeration and classifi- cation of object circuit func- tions, and the definition of the minimum performance level; 2. finding the appropriate building blocks to realize the set funcitons; 3. make an on-paper design of the interconnected blocks; 4. building a test set-up with various measuring points readily accessible; 5. using measuring equipment to verify the required per- formance, make corrections, and locate the critical sections in the circuit; 6. returning to step 3, or poss- ibly step 2, to re-assess the functioning of the various cir- cuit sections, until the test set- up functions satisfactorily. If only it were that simple! In practice, circuit design involves a rather more complex process, which is one of continuous feedback, re-dimensioning, the replacement of complete cir- cuit sections, and a good deal of awareness in spotting ever better components for a par- ticular function. Time and again it will happen that target techni- cal characteristics prove unat- tainable because of component specification or cost, but also because the designer is at a loss how to get the most out of a specific circuit. It is then necessary to first build this par- ticular section for closer analysis with the available test equipment. Textbooks are con- sulted, calculations are made, and the circuit is re-worked un- til one finds its performance to be adequate. Although often relatively simple circuits, filters and amplifier stages, or a combi- nation of these, are notorious for their rather unpredictable in- circuit behaviour. Calculating 1 their performance is one thing, making them function as re- quired is quite another. Ob- viously, the dimensioning of these circuits in a test set-up is a tedious and time-consuming task, which requires due atten- tion to be paid to all variables in question, and, more import- antly, the way these interact. The widespread use of the microcomputer has brought to many designers the possibility to simulate circuits under devel- opment. This means that the actual building of the circuit in- volved can be done with con- fidence after the computer has made a prediction about the relevant technical qualities. Just : how well-founded that predic- i tion is depends on a great many factors, such as the precision of the calculations, the number of component parameters taken into account, and the "aware- ness” of the program that com- ponents are never ideal. Until a few years ago, computer- assisted design (CAD) was only possible on professional com- puter systems (mainframes), mainly because of the speed and complexity of the parameter calculations per- formed in recursive programs. SPICE was one of the first pro- grams for linear circuit analysis to become available for use on , mainframes. As designing a cir- j cuit on a computer is in fact making a theoretical analysis of the dynamic characteristics on the basis of available compo- nent specifications, it is readily seen that any programming session initially entails the defi- nition of in-circuit junctions, called nodes, connecting reac- tive networks, active compo- nents, etc. After running a considerable number of matrix- | comparison routines,, the com- puter is able to analyse, for in- stance, the frequency charac- teristic of the circuit in ques- tion. The complexity of the calculations, the size of the parameter library, and the re- quired precision all determine the amount of computer mem- ory required, and the final com- putation time. If the CAD pro- gram provides for the possi- bility to closely simulate the actual behaviour of compo- nents, the results obtained are very useful for testing in a real circuit. In the following sections we will discuss two programs for linear circuit analysis, and take the opportunity to show you how these can be used to re- duce design time by having the computer do the necessary thinking before the user is con- fident about putting a practical version together. Analyser II Analyser I and II are BASIC pro- | grams available for many home computers. Analyser I is the simpler of the two, offering less freedom of component selec- tion as compared with version II. Graphics presentation is not available with Analyser I. Suitable for running on any BBC Model B or Master computer, with or without a second pro- cessor installed, Analyser II is based on the use of BASIC types I, II, or IV. Modifying the program as required is a rela- tively simple matter; for in- stance, we could readily re- place the time & date input routine by one that reads the relevant data from the Master’s built-in RTC. Depending on the amount of memory in use for other programs, Analyser II can Circuit diagram, design data and object response (II) of the filter under test. elektor mdia april 1987 4.43 Number One Systems Linear Circuit Analysis Program ANALYSER II (0 1984 < CIRCUIT NAME DE-EMP 15 Jan 1987 Component list: TEST RESULTS Frequency 10.00k Gain ( dB abs) -5.22m Phase (deg) -1.58 14.13k -10.41m -2.23 19.95k -20.74« -3.15 28.18k -41.26m -4.44 39.81k -81.84m -6.25 56.23k -161.41m -8.75 79.43k -314.89m -12.16 112.20k -602.21m -16.64 158.49k -1.11 -22.19 223.87k -1.95 -28.43 316.23k -3.20 -34.46 446.68k -4.83 -39.03 630.96k -6.73 -40.97 891.25k -8.65 -39.72 1.26M -10.37 -35.56 1.70M -11.70 -29.58 2.51M -12.61 -23.17 3.55M -13.16 -17.38 5.01M -13.47 -12.68 7.08M -13.63 -9.10 10.00M -13.71 -6.45 The component values and circuit nodes are neatly summarized before Analyser II starts printing the results of the simulated sweep Number One Systems Linear Circuit Analysis Program ANALYSER II (C) 1984. CIRCUIT NAME DE-EMP 15 Jan 1987 GAIN G,« PHASE P,+ ANY TWO I Gain(abs) 10.00kHz 11.89kHz 14. 13kHz 16.79kHz 19.95kHz 23.71kHz 28.18kHz 33.50kHz 39.81kHz 47.32kHz 56.23kHz 66.83kHz 79.43kHz 94.41kHz 112.20kHz 133.35kHz 158.49kHz 188.36kHz 223.87kHz 266.07kHz 316.23kHz 375.84kHz 446.68kHz 530.88kHz 630.96kHz 749.89kHz 891.25kHz 1.06MHz 1.26MHz 1.50MHz 1.78MHz 2.11MHz 2.51MHz 2.99MHz 3.55MHz 4.22MHz 5.01MHz 5.96MHz 7.08MHz 8.41MHz 10.00MHz Phase (deg) 100® 200m 300m 400® 500m 600m 700m 880a 900® 1.0 l.l . + * • • G P . * + 6 P . * ♦ G P * + GP *+ P6. + *. P 6 . + * . P . G . ■ • . ♦ * P . 6 . ♦ .« P . 6. * P .6 t P .6 + . * . p 6. + . * p. 6 . +. * p 6 . * .6 P . f. ♦ 6 . P. . # . + . 6 . P t ♦ . . G . P . * .6 . P .# + . .6 .P „t ♦ .6 P . -45 -40 -35 -30 -25 -20 -15 -10 -5 Attenuation and phase shift of the de-emphasis filter as functions of the input frequency (Analyser II) 4.44 elektor indie april 1987 handle up to 30 nodes and 100 components. It computes both the amplification and the phase shift of the object circuit over a user-defined frequency range. Provision has been made for the presentation of a linear or a logarithmic scale, while the fre- quency can be stepped in various increments to ensure the necessary resolution. In- itially, the results of the test sweep are presented in the form of a table, but the menu allows the dumping of a graph on a printer, which need not necessarily be Epson compat- ible, as Analyser II outputs data by means of standard characters. Group delay as well as input and output impedance calculations are also available to the user of Analyser II. The program is remarkable for its ability to take into acount that various components have parasitic reactances invariably present at terminals. A transis- tor, therefore, is not considered an ideal switching device, but rather a complex network of re- sistance and capacitance. Similarly, any inductor’s parasitic capacitance is fully observed in filter response calculations, while the differ- ence between the use of, say, a 741 or a LF356 opamp in an amplifier design becomes evi- dent from the program output data. Analyser II also enables users to enter additional com- ponent parameters taken from data sheets. The computer's disc facilities are used to create an extendable filing system to hold component data, which are then instantly available for trying out in a particular circuit simulation. FETs, transformers, inductors, chokes. . ., Analyser II has got them all stored on difc and ready for use in j various ratings. The available sweep band extends from 0.01 Hz to 1.1 GHz. Unfortu- nately, the program does not provide for the analysing of DC settings in the circuit. However, any change in the bias con- dition of, say, a transistor is recognized by Analyser II, which promptly corrects the stray capacitance figure to en- sure a faithful simulation of what would undoubtedly happen in practice: a different frequency response! In conclusion. Analyser II is an efficient and user-friendly program that will require a minimum of effort on part of the user to familiarize himself with the extensive range of com- mands and options available. A slightly unusual feature of the program is its presentation of the frequency axis in the sweep curves, but this merely requires some getting used to. The documentation supplied with Analyser II is an excellently de- tailed 25-page manual. AC Circuit Analysis Program This program from Markie En- terprise can only run under BASIC II because of its direct calling of routines in the BASIC ROM. It is. however, possible to run this program on the Master computer by loading BASIC II from disk into SRAM, and selec- ting the language as the default by means of * CON. LANG (the manual supplied with AC Cir- cuit Analysis does not mention this trick). This program can handle a maximum of 372 components. The calculated results appear in the form of a curve on a MODE 0 screen. Unlike Ana- lyser II, AC Circuit Analysis is written specifically for the BBC computer, and makes good use of its function keys. The menu comprises a HELP file which makes it easy to determine one’s whereabouts in the program, and provides an in- stantaneous overview of available program options. Un- fortunately, the package does not comprise a simple card- board template for quick refer- ence to the various functions called up by the function keys. A regrettable fact about AC Cir- ■ cuit Analysis is its limited range of components that can be used in the simulation. There are. for instance, only three transistor types to choose from. Whether or not this works out to be a serious disadvantage depends mainly on the applications you have in mind; for work on passive filters, the program is probably hard to beat. Unfortu- nately. the addition of compo- j nately, the library can not be | extended by the user. | The graphics offered by this program are of excellent qual- ity, giving the impression of working with an oscilloscope- like instrument. It is possible to store a complete screen onto disk, while a screendump can be made with the aid of, for in- stance. DUMPOUT 3. The printer need not be Epson- | compatible. A practical case 1 In order to compare the per- formance of the two previously introduced programs, we had them analyse a number of cir- cuits, simple as well as fairly complex ones, including the EBU de-emphasis filter incor- porated in the Elektor Indoor Unit for Satellite TV Reception (see Elektor India, December 1986). The printouts in this article should give you some idea of the results obtained with the linear circuit analysis pack- j ages. Althougn the filter under test is a fairly complex type, both pro- grams did not require too much time to calculate the response. Therefore, the effect of chang- ing a particular component is obvious the moment the print- out is available. What does the parasitic capacitance across the inductor do to the filter response? How can the roll-off be made sufficiently steep with- ' | out causing too great a phase shift? What is the expected overall attenuation of the filter, and how should it be termin- ated? Does it have any spurious pass-bands which could lead to oscillation in the amplifier connected to the filter output? These are but a few of the vast range of questions that can be answered by studying the out- put plots from the CAD pro- grams discussed in this article. Both the makers of Analyser II and AC Circuit Analysis will no doubt be able to supply you with more information, so just write to them at the addresses given below. Analyser II costs £51.75 incl. of VAT, and is available from Number One Systems Limited • 9A Crown Street • St Ives • Huntingdon • Cambridgeshire PE 17 4EB. Telephone: (0480) 61778 (IBM, Spectrum, and Amstrad versions also available). AC Circuit Analysis costs £60.00 incl. of VAT, and is available from Markie Enterprises • 17 Percy Road • Shepherds Bush • London W12 9PX. St I AC Linear Circuit Analysis in action. 'Note that the curves are written with the printer switched to 1 graphics mode. elektor mdia april 1987 4.45 SATELLITE TV RECEPTION: YOUR QUESTIONS ANSWERED by J & R v. Terborgh With the growing interest in domestic reception of signals from geostationary TV satellites, but with many aspects of the subject still surprisingly hard to find in various publications, this article is a round-up of questions, simple ones and complicated ones, and associated answers, clear and to the point. The reception of satellite TV services is a subject encom- passing so many aspects of electronics, mechanical engin- eering, applied telecom- munications, and other fields of interest, that it is not surprising to have baffled quite a number of readers, both those who are actually in the process of building the Elektor IDU, de- scribed over the past few months, and those who take a general interest in following any publication that has something to do with the pres- ent subject-matter. But to begin with, a few points must be made expressly clear. 1. Depending on the specific aspects raised in the ques- tions, these— and the associated answers— are dealt with in separate sections in this article. 2. The following convention ap- plies concerning references to earlier articles on satellite TV reception in Elektor Elec- tronics: [A] : Satellite TV reception, September 1986; (B) : Indoor unit for satellite TV reception, parts 1, 2 , and 3\ Oc- tober 1986, November 1986, and January 1987. 3. The answers to all questions are necessarily short and to the point. In many instances, further information can be found in the publications men- tioned at the end of this article. The system set-up Q. The only suitable location for my dish forces me to use some 25 m of fairly expensive downlead coax, which in- troduces an attenuation of 11.5 dB at 1750 MHz. Willthisim- pair reception? 4.46 elektor india april 1987 A. It certainly will. In general, cable losses between the LNB and the IDU should not ex- ceed about 4 dB. Long runs of low-loss— ie., fairly rigid- coaxial cable tend to be costly as well as cumbersome to in- stall permanently, requiring quite a bit of digging and dril- ling before the signal is availa- ble at the IDU input. A possible solution to your pro- blem is the fitting of the IDU RF board (see [Bj; part 1) into a wa- terproof, 1 temperature regu- lated enclosure as close as pos- sible to the dish stand. A length of inexpensive, multi-way scree- ned cable can then be run to the home, along with the base- band output cable, made in RG58 or TV coax. Do not forget, however, to lay out the tuning voltage circuit for a relatively low output impedance, in order to prevent hum and noise being picked up (steer clear of mains wiring!). Q. I intend to use an older type LNB which requires to be fed with 18 V but not over the downlead cable. Any modifi- cations required in the IDU? A. Regulator ICs can be re- placed by an 18 V series regulator circuit based around the L200 or 78GU, provided it is fed with a separately obtained input voltage of about 24 V. Remove Li and run a separate supply cable from point +LNB on the PSU/vision/sound board to the relevant connection on the LNB. The IDU design 0. Why have you not used the Type AT1020 and AT3010 modules from Astec? These units are specifically made for satellite TV reception and come ready-made, requiring no ad- justments whatsoever. A. The main disadvantage as- sociated with these devices is the limited IF range of the converter module AT1020, which is designed to accept the LNB IF range of 950-14S0 MHz, according to the satellite stan- dards used in Northern America, where IDUs were originally designed for the 500 MHz wide 4 GHz downlink band. In practice, the use of these modules in Western Europe makes it impossible to receive transponders broad- casting above 10 GHz + 1450 MHz = 11,45 GHz. In Table 2b in [A], you can see what this means for ECS-1 . . . The AT3010 610 MHz IF ampli- fier/demodulator provides a 3 dB bandwidh of only 26 MHz. which is expected to give diffi- culty in proper reception of the future DBS services, which will operate with 36 MHz wide downlink channels. Q. What about the funny round arrows at the polarization selector switch on the IDU? A. Circular polarization— see [A]— offers a number of tech- nical advantages over conven- tional, linear polarization. Fig. 26 shows the essential dif- ferences between these systems. Linear polarization is either horizontal (H) or vertical (V) with respect to the earth plane, caus- ing the 'A A probe inside the waveguide input of the LNB to have to be positioned as re- quired for reception of the rel- evant transponder. Circular polarization is either clockwise (cw) or counter- clockwise (ccw), and requires a specially shaped waveguide-to- PC board coupler. At present, satellites only transmit linearly polarized signals, and LNB feeds suitable for cw/ccw operation are, therefore, still fairly uncommon units. If it is recalled that , polarization of downlink signals is essentially a method of allow- ing two transponders to operate at about the same frequency without causing interference at the receiving station, circular polarization has the following advantages over linear polarization: 1. co-channel station discrimi- nation is typically IS dB better; 2. downlink signals are less severely affected by Faraday rotation in the atmosphere; 3. depending on the construc- tion of the LNB feed, the dish illumination, and hence the dish efficiency, is slightly im- proved. It should be noted here that the use of a round LNB feed does not necessarily mean that the system can receive circularly polarized signals only: a round waveguide of specific diameter does nothing to the polarization of the incoming wave, and is, therefore, often used with steerable H-V polarizers to enable the LNB to be rotated over 90°, using a bearing ring around the feed, and a small, remote-controlled servo or stepper motor to select horizon- tal or vertical reception. Fundamentals Q. Why do satellites not transmit in AM, so that private reception is possible with a conventional TV set, without the need for a special FM demodulator? A. Transmitting an amplitude- modulated TV signal re- quires highly linear operation of the transponder power out- put stage, which must conse- quently be biased for class A or class AB operation, resulting in a_ relatively low overall ef- ficiency. From about 5 GHz on- wards. sufficient transmitter power for satellite TV services can only be obtained from travelling wave tubes (TWTs), which require to be operated in Class C at very high acceler- ation voltages to output a car- rier power level of the order of 10-30 W at an acceptable efficiency— which is extremely important in view of the limited battery power available in the craft. FM offers the following advan- tages over AM: 1. with several carriers trans- mitted by a single transponder, there is less Fig. 26. Linear (H/V) and circular tcw/ccw) polarization. likelihood of unacceptably high intermodulation products from the power output stage; 2. with a suitably dimensioned combination of pre- and de- emphasis, the obtainable S/N ratio for both vision and sound is higher at a given receiver C/n input ratio; 3. no power is wasted in the process of modulating the , carrier; 4. vestigial sideband sup- pression is entirely irrelevant. The fact that an FM TV system typically occupies a greater bandwidth than an AM system : is of no consequence what- j soever in view of the vast \ capability in this respect of the : centimetre-wave bands accom- i modating satellite TV uplinks and downlinks. Q. Iam utterly confused by the use of terms relating to the ' system band width. Is it true that a single satellite TV channel oc- cupies a greater bandwidth than all short-wave bands together? A. Yes. There is nothing mys- terious about the output bandwidth of 27 to 36 MHz re- quired for each transponder in the satellite; it is merely the already high frequency of the modulating signal that causes the wide output spectrum. In fact, TV transponders are gen- erally operated at a remarkably low modulation index, m’ : m’ — Af/fmv where Af is the maximum in- stantaneous deviation from the carrier, and fmv is the maximum frequency in the modulating signal causing that deviation. With the still widely used peak- to-peak deviation of 13.5 MHz, Af is of course 6.75 MHz, while fmv is usually about 5 MHz (it will be recalled that we are dealing here with a composite colour video signal). The modulation index, m’ , thus works out at only 1.35. Note that sound subcarriers in the base- band spectrum are disregarded for the moment, in order not to complicate things unnecessar- ily. In theory, it can be shown that the RF output signal from an FM transmitter contains an infinite number of harmonics whose amplitudes decrease as they are further away from the car- rier. Without going into the complex mathematics of FM at low values of m’ , some 98% of the total RF energy produced by the transmitter is contained in a bandwidth, BW, written as Carson’s rule: BW=2(m' + l)fmv With the previously mentioned system parameters, this gives BW = 23.5 MHz, exclusive of elektor india apnl 1987 4.47 sound carriers, which can be expected to occupy a further 5 MHz or so. With a tendency on part of transponder leaseholders to use relatively large values of deviation (up to 28 MHz PP ) so as to improve the attainable S/N ratio at limited RF power, there is, at present, increasing pressure on receiver manufac- turers to give up the widely used 27 MHz bandwidth stan- dard (for Af= 28 MHzpp, BW works out at 38 MHz). 0- / am under the impression that the quality of reception offered by my receiving system is slightly improved as its gets colder outside. Why is that? A. Refer to Fig. 6 in (A) to see that the noise figure, FdB, of your LNB is a function of its noise factor and the ambient temperature; the curve shown is relevant to To= 17 °C, but the inset calculations make it quite evident that Tr, and hence Pntsys), decrease with lower values of To . It goes without saying that the final S/N figure is improved accordingly. Q. With reference to Satellite TV reception in the September 1986 issue, lam able to follow all the calculations from system noise to the theoretical S/N formula, (14). Yet 1 am intrigued by the origin of the constant, x, given as 147.3 dB for 36 MHz system bandwidth. A. Formula (14) is a purposely simplified evaluation of the standard S/N calculus reading S/N{video. rms) = 101og,o[3/2 (Afpp/fmv) 2 BW/fmv) + C/n+13.2 [dB) (14a) in which S/N(video, rms)= weighted, effec- tive signal-to-noise ratio at the output of the receiver’s FM vi- sion demodulator; Af PP = peak-to-peak deviation resulting from modulating the FM transmitter with fmvj fmv = highest video frequency in uplink & downlink baseband spectrum; BW= theoretical bandwidth of transponder’s output spectrum; C/n= theoretical carrier-to- noise ratio at the input of the re- ceiver's FM vision demodu- lator— see (12); 13.2= the effect of pre- emphasis and r.ms. weighting to CCIR Report 637-1. The use of (14a) with parameters Af P p= 13.5 MHzpp, fmv = 5 MHz, BW=36 MHz, and C/n = 9.66 dB results in S/N(video. rms)= 101ogio(78.74)+9.66+ 13.2 dB S/N(video, rms)=41.8 dB. From this it is seen that (14) is a slightly too optimistic S/N cal- culation, yielding the so-called unweighted quasi-peak value. Formula (14a), obviously more complex than (14), is the more authoritative of the two, as it is given by the EBU in Literature Reference [5], Dish location and adjustment Let us consider the following chicken-and-egg problem, which has puzzled many con- structors of the IDU: to be able to line up the dish aerial, one needs a fully operative receiver; to be able to align the receiver, one needs to have the dish ad- justed to "see” the satellite. Practice does it! With a few helping hands available at the time of positioning the dish, you will find that this is not nearly as difficult as it may seem at first sight. In fact, by studying the following questions and answers, sufficient insight can be acquired to be able to go round the majority of difficult- ies encountered while lining up and tuning in. Q. 1 can not decide on a suitable location for my dish in the garden. Can you give an approximate indication of the maximum height of obstruc- tions, given a specific angle of elevation? A. The answer to this question is best given in the form of the formula h=k + d siner or d=(h-k)/sinor where h = height of obstruction in line-of-sight path to satel- lite; d = horizontal distance be- tween dish and obstruc- tion; k = safety margin; 1 metre is recommended; a = angle of elevation for the dish. Especially with trees, due ac- Table 5. >L I ST 10 REM azimuth and angle of elevation for geostationary satellites 20 DIM Orb<6> : RESTORE : MODE 3 : REM 24x80 text only 30 H 2 180 /PI : REM rad-deg conversion 40 FOR X=0 TO 5 .READ Pos% : Orb ( X ) =Pos% : NEXT X 50 R=6371 :ALT= 35822 : REM See EE September 1986 60 PRINT"*** Longitude and orbital position WEST of Greenwich: PRECEDE BY MIN US SIGN **«" : PRINT 70 INPUT'Longitude of location 7 “LO : LO=LO/H 80 I NPUT “Latitude of location 7 "LA : LA=LA/H 90 GOSUB 1000 100 B=LO-SAT 110 A2 I = 1 80+H* ATN( TAN( B) /S IN( LA) ) 120 AZ I = I NT < AZ 1+0.5) : PR I NT : PR I NT “Azimuth = “;AZI; M degrees " ; 130 W$=" West of South" :Ef=“ East of South” :S$=" straight South" 140 IF AZ 1=1 80 THEN PRINT" = ";SS:GOTO 170 150 IF AZ I < 180 THEN PRINT " = ”;180-AZI;" degrees ";Ef:GOTO 170 160 PRINT "= “ ; AZ I - 1 80 ; ” degrees ” ;WS 170 ELE=H*ATN< ( COS ( LA ) *COS C B > -R/ < R+ALT) ) /SQR< 1 -< COS ( LA ) ~2*COS < B ) ~2 ) ) ) 180 I,F ELEC1 THEN PRINT"Satel 1 1 te below hor l zon" : GOTO 70 190 PRINT’Elevation = " ; I NT ( ELE+ 0 . 5 ) ; ” deg rees " : PRI NT 200 GOTO 60 1000 PRINT'Which satellite 7 ":PRINT 1010 PRINT" 1 = INTELSAT V Fl/7 ( FRG) *60 deg. E" 1020 PRINT"2 = EUTELSAT 1 F- 1 (ECS-l) +13 deg. E" 1030 PR I NT “ 3 = EUTELSAT 1 F-2 (ECS-2) *07 deg. E M 1040 PR I NT" 4 2 INTELSAT IV A F2 (NORDIC-1) -04 deg. W" 1050 PR I NT "5 = TELECOM F-l -08 deg. W (not in CSS band)" 1060 PRINT"6 = INTELSAT V F4 (UK/US) -27.5 deg. W" 1070 PRINT-7 = other satellite- 1080 PRINT: INPUT'Select 1-7 >"N 1090 IF N>= 1 AND N<=6 THEN SAT=Orb ( N- 1 ): SAT=SAT/H : RETURN 1100 IF N=7 THEN INPUT'Orbi tal position of satellite > "SAT : SAT=SAT/H : RETURN 1110 GOTO 1080 S000 REM geostationary arc; orbital positions East to West 5010 DATA 60. 13, 7. -4. -8. -27. 5 >RUN *** Longitude and orbital position WEST of Greenwich: PRECEDE BY MINUS SIGN *»* Example: Dundalk, Ireland Longitude of location 7 -6.5 Latitude of location 7 54 Which satellite 7 1 = INTELSAT V Fl/7 (FRG) 2 = EUTELSAT 1 F-l (ECS-l) 3 * EUTELSAT 1 F-2 (ECS-2) 4 = INTELSAT IV A F2 (NORDIC-1) 5 = TELECOM F-l (F> 6 = INTELSAT V F4 (UK /US) 7 = other satellite Select 1-7 >6 Azimuth = 205 degrees = 25 degrees West of South Elevation = 26 degrees 86082-4-T5 ♦60 deg. E ♦13 deg. E ♦07 deg. E -04 deg. W -08 deg. W (not in CSS band) -27.5 deg. W 4.48 elektor india april 1987 count should be taken of their growth and their leafing. Q. / live in Dundalk, Ireland, and I have a complete satellite reception system. Iam, however, at a loss to understand how the dish is to be pointed at, say, Intelsat VF4. Do I have to turn it 27.5° west of south? If so, at which angle of elevation? What is the difference between azimuth and orbital position? A. In [A] it was already stated that there is a complex re lationship between the terms raised in your question. Given the longitude and the latitude of the terrestial location, and the orbital position (OP) of the satellite, the azimuth, expressed as an angle y with respect to the geographic north, and the as- sociated angle of elevation, «, (see Fig. lb in [A]), are obtained from the trigonometrical equations y=180+arctan [tan(Lo-Op)/sin La] o=arctan cos La cos(k)-Op>r/(r + ah I l-cos ! La cos 2 (Lo-Op) ] where Lo = longitude of location; La= latitude of location; Op = orbital position of satellite; a & r = see (1) in [A], A pocket calculator providing the stated trigonometric func- tions should be set to its degree mode, and longitudes as well as orbital positions west of the Greenwich meridian should be entered with a preceding minus sign. It should be borne in mind that the result of the azimuth calculation is an angle expressed in degrees with respect to the geographic north, so that east, south and west correspond to 90°, 180° and 270°, respectively, similar to •the indication on a magnetic compass. Depending on the specific terrestial location, there is a difference between the geographic and the mag- netic north, making a compass only suitable for finding the ap- proximate satellite position, not the final azimuth. None the less, a good quality compass will soon prove indispensable dur- ing the setting up of the system, as will be seen further on in this article. Table 5 is the listing of a univer- sal dish positioning program based upon the previously given trigonometrical calcu- lations. Though written for the Acorn and BBC micros, the Fig. 27. Using a compass to find the approximate azimuth for the dish aerial (example). program should not be too hard to patch for other computers and their specific BASIC syntax conventions, while graphics ap- plications may be added as re- quired. Since it was deemed useless to have the computer present the resulting angles with, say, 9-digit precision, lines 120 and 190 use the INT(x+0.S) instruc- \ tion to attain a precision of ±0.5° for azimuth and angle of elevation, respectively. At the end of the program are 6 orbital positions given as DATA items and put into an array called POS% by the READ loop in line 1 40. Selection of item 7 from the list of satellites enables establishing the aerial position for services yet to be commisioned— e.g DB satellites, see Fig. 9 in [A]. With the positioning angles calculated and noted on a piece i of paper, you are now nearly ready for the first practical at- tempt at receiving the satellite. First, however, consider the fol- lowing points: A. Your location should be within the satellite’s foot- print. Calculate the expected C/n ratio as set forth in [A]; if this works out as lower than . Fig. 28. Using a plastic protractor ’and a plummet to set the angle ' of elevation (example). +8 dB, good reception will be very difficult, if not impossible, even if all equipment is known to function satisfactorily. Very good reception requires a C/n ratio well in excess of 14 dB. B. The dish location should of- fer an unobstructed line of sight to the relevant satellite. Go to the planned dish site and use the compass to find south, ie. the needle should register with the N indication. Stand with your back to the north and im- agine a horizontal line, starting from the compass pivot, across the calculated azimuth value on the dial, straight to an orien- tation point well removed from your position— see Fig. 27. This point may be any fairly high, well discemable object, such as a tree top, a building, a neighbour’s aerial mast, a lamp post, a traffic sign, etc. Straight above this point, a considerable area of the sky should be vis- ible, i.e. there must not be higher objects further towards the horizon. In western Europe, most satellites can be received with angles of elevation of the order of 20° to 35°, i.e. they are sufficiently high up in the sky to ensure a line of sight path with the dish mounted on a post in , the ground. However, in densely built areas, it may be necessary to raise the dish well above the ground to ensure a clear view in the appropriate direction. In view of both the inaccuracy of most types of compasses, and the difference between the magnetic and the geographic north, it is recommended to first adjust the aerial elevation as shown in Fig. 28. Make sure that the protractor is held exactly parallel with the dish axis and read the angle of elevation, which is the same as « shown in Fig. lb in (A], With a sufficiently heavy plummet, and in the absence of gusts at the time of adjustment, the angle of elev- ation can be set with an accu- racy of about + 1°. Owners of an offset dish or a Polar Mount system (see [A]) can not make use of the above procedure, and should consult the dish supplier for positioning in- ; structions. , Never attempt receiving a satellite without having at least an idea of its whereabouts in the sky; it is a waste of time and rightly comparable to finding a needle in a haystack. alektor tndia aprll 1987 4.49 Upon reaching the requisite angle of elevation, provisionally lock the relevant dish adjust- ments). If the dish has a hole at the centre of its reflective sur- face, look through it to check whether the LNB feed is exactly on the dish axis, ie. the feed aperture should offer optimum illumination. Unlock the aerial azimuth ad- justments) and make sure that the dish can revolve freely around its mounting system, without any change in the set elevation. Use the compass as explained to roughly determine the azimuth, and use the IDU SCAN facility as detailed in the section Aerial positioning unit in Part 3 of [B]. Turn the dish very carefully across the ex- pected azimuth range; as the 3 dB directivity of a 1.5 metres dish is only about 1°, aiming it at the satellite is in no way com- parable to adjusting, say, a UHF TV aerial! Consult Fig. 29 if you are still unsure about the differ- ence between « and y. Once you have managed to see the first synchronization bars, it is a relatively simple matter to peak all dish and LNB feed con- trols for maximum S-meter deflection. Spend some time in finding the correct focal point for the LNB input, and see whether the polarization can be optimized by rotating the feed over a small range. Depending on the angle of elevation, there is a polarization offset angle to be taken into account. Es- pecially with a smaller than 20°, it is well worth trying to estab- lish the correct polarization off- set, which may amount to +45° as viewed from the front of the dish. You will probably find that manual adjustment of the dish soon becomes a routine job, and spotting various satellites within 5 minutes or so can be done with the help of two or three orientation points at a familiar location, and a few simple notes as a guide in set- ting the two dish angles plus the tuning dial indication on the IDU for a specific transponder. Miscellaneous matters and the future Q. Apart from ECS-I and Intelsat VF10, are there more satellites transmitting TV pro- grammes? 4.50 elektor india april 1987 Fig. 29. To line up a dish aerial, the required azimuth and angle of elevation must be set separately. A. Yes, there are. You may try ECS-2 at OP 7° E, which transmits three EBU newsfeed channels operated at pre- scheduled times and intended to provide unedited news flashes to many of Europe’s national TV broadcast organiz- ations. These transponders are also used as two-way relay sta- tions carrying technical instruc- tions for camera crews during important international events, such as sports, games, con- ferences, etc. (Eurovision Ser- vice, co-ordinated from EVC, Brussels). Also on ECS-2 is the VISNEWS newsfeed channel, and Televerket Norway, which transmits in C-MAC. The Nordic-1 satellite at OP 4° W beams down Sveriges 1 and II in C-MAC. If your location allows a wide view towards the East, you may try Intelsat VF12, nicknamed Copernicus, at OP 60° E, which is above the Indian Ocean. This satellite carries four German TV programmes, and can be re- ceived with very good quality, provided the dish elevation can be reduced to about 10° (average value in the UK). If you are the fortunate owner of an outdoor unit comprising a Polar Mount and a steerable polarizer, it is highly interesting to spend an afternoon or so in 30 GEOSYNCHRONOUS SATELLITES - 135* E SAKURA A2 (1983) japan - 130* E SAKURA B2 (1983) japan 118* E PALAPA B3 (1985) .NOONIS.A PALAPA 82 (19B4| mdomsu r ue* e p. r 113* E F ° 1 r- 1I0*E BSE 2 BS-2A (1984) 1 BS-28 (1985) japan INTELSAT IV F8 179* E INTELSAT IV FI 174* E AUSSAT III (1965) Australia 164* E AUSSAT II (1964) AUSTRALIA 160° E AUSSAT I (1964) Australia 158° E 143° W SATCOM V us 139* W SATCOM IR (1983) us — 135* W SATCOM I. GALAXY I (1963) us - 131° W SATCOM MIR us - 127° W COMSTAR IV TELSTAR l»C ( 1984) us - 123* W WESTAH V us 119* W SATCOM ll SPACENET I (1964) us 116* W ANIK Cl CANADA 114° W ANIK D2 Canada 113° W ANIK C2 'canaoa 109* W ANIK B canaoa • — 106* W GTE II (1964) us ^ — 104* W ANIK Ol Canada „ 103* W GTE I (1964| US — 102* W ILMUICAHUA (1985) m**ico I 100* W SBS I us • 99* W WESTAR IV us 97* W SBS II us 96* W COMSTAR II TELSTAR MIA (1983) us • ~~ — 94* W SBS III us PALAPA B1 (1983) noon«Sia 106* E- ISCOM I INDIA 105* E - INSAT IB (1983) inoia 94* E GORIZONT USSR 90* E RAOUGA USSR 86* E- PALAPA A 2 indoncsia 63* E PALAPA A1 inooncsia 77* E- INTELSAT IVA f 6 INTELSAT V F5 63* E INTELSAT V F12 60°E GORIZONT uss« 53* E - n 74*W GALAXY II (1963) us i 70* W SPACENET II (19641 us i ° ^ 53* W INTELSAT IV F7 • MALLE Y l (19*6) un 31* W-, INTELSAT IVA F4 34 5* W . INTELSAT V F2 J WESTAR l WESTAR II " 1 1 Advanced westar n d9»4i us *— 8 3* W SATCOM IV us 85* W ILMUICAHUA (1965) m**.co J 7* W COMSTAR ill TELSTAR IIIB (1964) us 91* W WESTAR III AOVANCEO WESTAR I (1964) us RAOUGA USSR 35* E - . ARA8SAT (1986) 26° E ARABSAT (1986) 19 B E ECS-2 (1964) (sa 15* E ECS- 1 (1983) csa 10* E o' TELE X (1968) SWEDEN 5° E -79* W ADVAN to* W TELECOMM I (1983) TELECOMM II (1963) FRANCO G'Rman 5* W SYMPMONIE I SYMPHONIE II FRANCO 1 GERMAN 14* W GORIZONT USSR 16 5* W INTELSAT IVA FI 19° W TV-SAT A3 (1986) GERMANY L-SAT (1967) ESA LUX SAT (1967) LUXEMBOURG TDF 1 (1967) FRANCE — 21 5* W INTELSAT IVA F2 24 5* W INTELSAT V F3. SIRlO