MARCH 1996 £ 2-65 Yaya 11> FINDER: LOWS UAINTALLEL CONTENTS March 1996 Volume 22 PC soundcard as Number 242 ISSN 0268/4519 =e AF analyser — surround-sound subwoofer Design by T. Giesberts part 1 EI PIC RDS decoder Design by U. Nagel PC soundcard as AF analyser Design by M. Ohsmann PC memory chips By our Editorial Sta Nn houseplant buzzer Design by PKersemakers RADIO, TV & COMMUNICATIONS. satellite finder Design by C. Denolle Uris & MEASUREMENT | ‘surround-sound switched-capacitor audio filter subwoofer peng BYE Hist construction . guidelines datasheets index of advertisers new books new products news from the world of electronics next month in Elektor Electronics overview of EMC guidelines reader's letters readers’s services switchboard the INTERNET page i ~ focus on: PC memory chipsEurofighter engineer is 1995 Young Woman Engineer of the Year A twenty-nine year-old senior avionics systems engineer wath British Aerospace, Pamela Wil son, was chosen as the Young Woman Engineer ofthe Year st ‘ceremony held in London in sanvary. Pamela, who already regu: lary ists schools and colleges to explain her interesting ja, told Elekior Electronics at the award presentation: “I want ina electronics because I was interested in how things worked ‘A gladuate member of the IEEIE, Pamela was tunnerup for the 1982 Young Woman Er- gineer of the Vaat Award and this time she was voted by the judging panel top ofthe six f nalists anc was presented the Coveted award for 1995, 2 cheque for £ 750 and a silver rose bow Pamela joined the Miltary ‘Areraft Division of Bish Aer: space Defence at Wharton Aerodrome, Preston. Her two: year graduate training involved placement in equipment engi heering, research and develon- ment and system design. As 2 result of he final training place- ‘ment, she stayec with the Eu: fotighter 2000 cockpit graup Where she was promoted to Avionics Systems Engineer Her work is primarily dofning the man-machine interlace re quirements for the sensors Contral within the EF2000 cock pi Pamela is he eighteenth re cipient of the Award, which is jointly sponsored by the Insti tion Of Electronics and Electr al_Incorporated Engineers (ECIE) and the Caroline Haslett Memorial Trust (CHMT) The aim of the Award 1s 10 highlight electronic and elect cal engineering as @ rewarsing anc worthwnile career for ‘women, Previous winners and Contenders forthe Award have more than proved that women make extremely competent blectronie and electrical engl- al Incorporated Engineer level ~ a career onco almost ‘exclusively male-dominated, Prior to the winner being announced. a video of all six ‘award ‘inalisis at work was shown to representatives trom industry, commerce, academia and media who attended the ceremony whose organizers siil see a crying need for the ‘event in years o come, Runner-up for the 1995 Award and winner of the WISE prize was Audra Gittens, a 26. year-old test engineer with Fobinson Instruments, Run: comm, Cheshire. Audra, a gradu ate member of the IFEIE, was presented with a cheque for 2500. Thirdsprize winner, Rose- marie Hayerott, received a cheque for £ 250. Rosemarie, 27, ia switch maintenance en ginser with Colinot Mobile Communications and is based al the Manchester Switch Gen: tre in Safe. “Twenty two-year-old Yvonne Mortis, an electrical engineer Pa Grade with British Aero: space (Dynamics) Defence in Stevenage recewved the Mary George Memorial Prize—an ad- ditional award given to a young enivant showing particular promise as an Incorporated Engineer. Yvonne recived a cheque for £ 250 and a silver sae. Electronics - a key British industry BBitain's electronics business is leday the fourth largest in the world after the USA, Japan and Germany, and by the year 2000, electronics will be the world's largest industry accounting for some fen per cent of mankina's GOP. Tho alliance of telecom ‘munications, information tech: nology, new digital electronics techniques, and multimedia Communications networks al: lons a wide range of previously disparate industri to con verge, The affilaton of of con- Sumer eleetronies, cable TV ‘and broadcasting, media {reups, information services, Sublishing, sofware, banking. nd retaiing among others has ELEKTOR ELECTRONICS MARCH 1996 been the result. ‘At the same time, the fapialy growing use of elec tronics in Britain and elsewhere ives the industry the status of key industrial sector. AdGi: tionally, , electronics an es sential enabling technology that dives competitiveness land even wealth creation, inthe rest of UK industry, 100. Bo tween 1985 and 1990, the UK's share of elactronics exports by the (G7) major industrialized countries increasea from nine per cent 0 11 per cent, the largest racorded 0, Electronics has become a signiicant force and, with elec- twieal engineering, is now the o> second largest industry in the UK. The industry's sales are worth © 30 billion, and yearly Investment runs at some £1 billion, alongside ABD's £ 2 billion. The worklorcs ‘amounts to some 380,000. ‘As each year passes, ole lwonics assumes more promi: rence in UK manutecturing Value added Increased by eight per cent yearly in the 1880s, and the industry in creased its share of manufac turing output ram five per cent in 1981 to seven por cont in 1990, Electronics in the UK is becoming less concentrates Large enterprises employing ‘mare than 1090 now represent ‘55% of the sector from around 70% in 1980. Nearly 50% of ‘lectronics employment ies in the southeast, although lang and the’ West Midlands ‘generate the highest value ‘added per employee Capital equipment Britain is particularly strong in telecommunications, data pro cessing and software, along: side the provision of capital ‘equipment for civil and detence purposes. Heavy investment in FD demonstrates the indus: trys commitment to innovation with funding averaging six per Cento turnover Electronics’ intemational nature is behind the UK's steady trend towards tho devel: ‘opment of international star: dards, which in turn ereates the ‘opportunities for the UK indus ity to exploit new markets, Hardware and software manu: facturers are also meating in creasingly stringent security standards. The UK industry knows its ‘economic future depends upon is continuously improving te postion in the global market place. The Federation of Elec: tronics Industry (FEI) repre sents the industy where mem. ber companies are developing their response to the challenge and are pursuing policies to ‘maintain their position among the world's electronics front runners. FEI, formed in January 1994 from the sector's two pre vious trade assooiations, brings Uk:based telecommunications, information technology, com- puter services, delence elec Tronics and office products, as wall as electronic components lunder the same banne Federation of Electronics In dustry, Russell Square House, 10-12 Russell Square, London. United Kingdom WO1B SEE Phone +44 171 331 2000. Fax £44 171 $31 2060Differential GPS for land use with RDS By M. Ohsmann The Global Positioning System, PS, criginally developed by the ‘American Delenca Department aware wide satelite supported pin-point navigation system, is ow widely used fr civilian pur poses. Receivers for civil pur poses have been avaliable and Inusesinos hamid 1980s, These recowvers are usotul fora variety ct purposes for example, yacht- ing, inshore fishing, fleet man: ‘agement of road ranspan, anc surveying, The system is based on a {ppsstation receiving the signals ‘fom soveralsatelits, Assuming that the exact time (UTe—uni: versie co-orcinatod)isknoa, ‘the te ofrecelpt ofthe various signals enables the cistance to the diferent satelites tobe cam- puted (hesatalite wansmits utc iso). Since the positions of the satelites are contained in the ltensmited signal thepostionot the receiver ean be established fairly accurately if signals trom atleast tvee satelites wore re: ceived Theexacttmecan becal- tulated atthe receiving station ithe signal ofa fourth satelite is also avaiable. The postion of the satalitesis transmitted the satelite. The postion accuracy is loss than a mata, Mare than 20 salolites ensure that at least ‘x saelites can be received at ‘any point on earth—see Fig. 1 Degraded accuracy ‘Athough theacouracy mentioned oan -08-01 Fig. 1. The GPs is based on 21 NAVSTAR GPS sataltes that circle the earth once ry 12 hours at an altude of roughly £20000 km (12000 miles). Their near-crcular orbits are inclined to the ‘catlieis available to the military andspecialsenices user tis not tothe clan user This caled the standard pasitoning service wth solectiveavalabilty.tmaans thatthe cullanusercannotreckon ‘anaceuracyofbeterthand0-100, mmetves. This degradation arises ‘om the fact thatthe satelite signals do not tell the civilian Useritsexsctposion, butastghty diferent one. This causes @ pe sition enor st tha GPS recover. Moreover, the generated error slowly changes. Improved accuracy The artificial degradation of the accuracy is counteredas follows. ‘AaPsreceiverinanexaciy known Position serves as a reerence This relerence receiver can de- Fig. 2. The sttucture of a DGPS for ln-point navigation. 4° restrial equator at 63 ermine witha ai degree of ac curacy which satelite ies’ to \whatescent.itthenmakes these dataknowntoctherGPsrecevers, these can improve the accuracy ‘of their own estimated position track layout on to ( film, transposing the film copy on to a (positive. sensitive) re board, and etching the superfluous copper. This requires patience (lots) and experience (some), but itis not beyond the ken ‘of most electronics enthusiasts, The ready-made board contains sub-boards for four monitors, since generally more than one houseplant needs monitoring, Populate the board in stages and. check the work after completing each stage The work is begun by soldering resistors Rj-Ry, followed by capaci« tors Cy, Cz and Cg, and then transis- tors Ty and Ty in their relevant posi- tions. Note that, to save space, the resistors are mounted upright. Next, solder the battery-holder leads to the board Insert an HP7 battery into the holder. Use a multimeter, set to the 2V dec range, and check that the voltages at the test points indicated in Fig. 4 (with respect to earth, that uy2. If the potentials at 2 and 3 (collectors of T; and Tz respectively) are about half the battery voltage, it may be assumed that the generator functions correctly. If these potentials are much smaller than U2, the resis- tors are placed incorrectly or theit value is wrong. If the potential is much higher than Uyy2, there is something awry with the transistors (wrongly connected or incorrect type) If the potentials are as stated, remove the battery from the holder is, the ve terminal of the battery) are: D: battery voltage, Lp; 8, @, @ soleer Rs, Ry, C3, T3 and Ty into place, Squeaking crystals The piezo-electric effect occurs when certain materials (crystals) are Subject- ed to mechanical stress. Electrical polarization is then set up in the crystal, whereupon the faces of the crystal become electrically charged. The polanity of the charge reverses if the compression is changed to tension. Conversely, an electric fleld applied across the material causes it to contract or expand according to the sign of the electric field. The effect is observed in all ferroelectric crystals and in all ferroelectric crys tals that are asymmetric and have one or more polar axes. The effect is important because it couples electrical and mechanical energy and thus has many applications for electromechanical transducers. Piezoelectric crystals are used to provide frequency standards and in piezo- electric oscillators. ' Figure (a) shows the principle of a domestic application: a lighter for gas fires. The hammer represents the mechanical force acting on the piezoelectric ‘material. The voltage resulting trom the distortion of the material is indicated by the meter. Figure (b) shows the change of shape of a piezoelectric material when an electric potential is applied to it. When the electric potential is an alternating ‘one, produced for instance by the rapid opening and closing of a switch as in Fig. (c), the material will vibrate in rhythm with the rate of opening and closing of the switch. If this rate lies in the frequency range 20-20000 Hz, the air ois- placed by the vibrating material wil be audible. This is the princiole of opera- tion of the buzzer used in the houseplant moisture monitor. Eloktor Electronics 3198and reinsert the A} tattery into the holder With the multimeter set as before, check that he potential attest point 4 (collectors Ty) is Up? When the multime: ter is momentarily , replaced by the buzzer, this should emit a3 KHz tone If no tone is heard, it may be that its frequency is much too low oF too high this may be caused by incorrect values When all is cor O-w batey tom the rie When all is wel holder, solder the Egil: Varlous fost fx the battery hoe on remaining compo- | points on the peb. See 10 the track side of the rents into place and | fextfor values at these board with double-sided connect the buzzer | POMnts: provisionally, S the multimeter to the 1 mA dc. ran and connect it in series with one of the baitery holder I battery into the hold the current drain is 0.14 ds. Reinsert the rand check that nA. itis appreciably different, check the value Elektor Electronics 3/96 adhesive tape (available from most. stationery and ply shops). The advantage of this tape is that the holder can be removed at a later date if nece Fix the back of the buzzer to the bat tery holder (at the end opposite T, with superglue The probes consist of 15 cm lengths of 1 mm thick insulated cop- per wire, from each of which 2-3 mm insulation is removed at one end and 4 cm at the other end. The 2-3 mm bare ends are soldered to the board. The 4 cm bare ends are tinned to pre pn, Make sure that the probes are straight and equidis tant (12 mm) from each other. Usace After the battery has been inserted into the holder, push the moni upright into the soil, Make sure that the bare probe ends are fully in the soil, but take care that there is suffi ‘ent space between the board and the soil to prevent the board getting, wet when the plant is being watered, Alternatively, bend the probes to allow the monitor to han over the rim of the pot on the outside. Normally, the monitor is pushed into the soil only near the time when watering of the plant is due (a mat ter of experience). With the probes in plant), set P; to.a position where the buzzer is silent, and then adjust it tll the dry(ish) soil (as relevant to th buzzer just operates. Sprinkle water on to the soil (away from the mor ton), wait a few minutes for it to get well into the soil, and check that the buzzer ceases to work. If necessary, readjust Py with the probes ins into a different pot Components list Resistors: AI, RA = 2240 2, Ra = 220k. 5, R6, RB = 100 ke Rr = 47ka Fo = 10Mo Pi = 100k0 preset Capacitors: C1, C2 = 820 pF 3 = 470 pF Chat uF (5 = 100 nF 08 = 10F 63, radial Semiconductors: Dr = 1Na148 TH-13, 75, T6 = BCSSOC Ta = BCS60C Semiconductors Dt = piezo-slectric buzzer, 1.5 V 1 off HP (RA, A) cry battery 1 off holder with leads for battery 11mm thick insulated copper wire for probes (see text) 8 Order no 950118. (p. 70). 7By our editorial staf Close on the heels o' FOCUS ON: yf _ the processor (CPU), the ery. s is one of the impor- ponents in a PC. Developments in the software industry over the last cou- ‘: ple of years have made aoe amounts of memory necessary to run an operating system like MS-Windows. Today, — skavis Dravis rD0 RAMs sins, there are so many dif- SIPPS ..., there is a such bewildering number of memory options for the ferent types Of mem- computer thot itis hard to keep them “ory available that it is "tate. mer mporanty tose the right type if you want to extend no longer easy to the memory in your computer. know exactly which In principl ple. Currently, two types of memory are available: static RAM (SRAM) and dynamic RAM (DRAM). RAM stands for random access memory. In a static all very sim= roductory overview —kAM, each memory cell consists of a erefore presented in this article. flipflop (comprising a few semicon- ductors) which is at one of two logic states, and so capable of ‘remembering a value, The dynamic RAM has a much simpler construction: a capacitor is charged or discharged by a fild-f- fect transistor (FET) or an ordinary transistor. The major disadvantage of this type of memory is, however, that the memory contents has to be re- freshed frequently (every few mile liseconds) because the charge con tained in the capacitor disappears slowly because of leakage. This leak- PC MEMORY CHIPS age is significant because capacitors in a DRAM have a capacitance of less than 0.1 pF This problem calls for a more complex type of drive. Be- cause charging and discharging ‘capacitor takes more time than mak- ing a semiconductor switch on or off, static memories are much faster than dynamic ones. The latest SRAMS boast access times of 10 to 20 ns (nanoseconds), while values of (60 to 70 ns are common with state- ‘of-the-art DRAMs. Because of the more com- plex structure of the SRAM, this vice is typically larger and more pensive than its dynamic count. part. That's why the mass memories in today’s computers consist mainly of DRAM. SRAMs ate only used for {ast intermediate memories (cache) SECOND-LEVEL CACHE Since the introduction of the 80386 processor, PC motherboards have a certain amount of fast cache memory that forms a buffer between the fast Elektor Electronics 3/96processor and the much slower (but relatively expensive) DRAM memory: Cache memory these days comes in towo variants: asynchronous and syn- chronous SRAM. With asynchronous, SRAM, the processor has to wait for the associated data to appear at the ‘outputs after it applies an address. With synchronous SRAM, an address is supplied, and then the data also ap pears after some time. However, the SRAM then immediately sends the dlata atthe next three locations, with- out actually requiring new addresses. This trick results in faster timing, The latest in cache memory {s the pipelined burst cache RAM, which, is based on the principle of the syn- chronous SRAM. The pipelined ver- sion however has an extra buffer latch’) at its outputs, which enables a new address to be supplied while dlata are still being read from the out. puts, This allows access times of be- tween 4 and 8s to be achieved, so that even the fastest Pentiums can process cache memory data without wait loops. SRAMs come in many dif ferent physical shapes and sizes. Until recently, SRAMS used to be housed in ordinary DIL cases. Nowa- days they are often implemented as surface-mounted devices (SMDs), and are soldered directly on to the board. The latest development is the cache module, which is plugged into a socket on the motherboard. Such a module simplifies exchanging the cache memory considerably MILLIONS OF CAPACITORS As already mentioned, the large user memory in your PC consists of dy namic RAMs, During the course of computer history, the capacity of RAMs has grown steadily. Way back in 1970, the first 1-Kbit DRAM ap- peared on the market. Today, 256-Mbit chips are being developed and tested in semiconductor laboratories. With DRAMS, too, the physical appearance has changed considerably. Whereas the first computers had lots of memo, ry chips in DIL (dual-in-ine) plastic or ceramic cases, today’s computers con- tain almost exclusively SMDs ‘To make changing the memory con- figuration on the motherboard a little tasier for the user, modules have been developed. These are small printed-cir- cuit boards which contain a number of memory chips. Initially, there were two types of module: the SIPP and the SIMM. The difference is easy to see: 2 SIPP has connecting pins, while 2 SIMM has connecting copper pads (contact fingers) on the board. Mean while, SIPPs seem to have died out, and today's PC motherboards allow only SIMMs to be fited. SIPP-to-SIMIM adaptors are still available, however. Originally, SIPPs and SIMMs were invariably 30-pin modules. Today, however, an increasing number of SIMMSs is of the 72-pin type, which is identified as the ‘PS/2 SIMM’. The in SIMM and its 30-pin prede- cessor have different memory struc: tures, as will be seen further on in this article MEMORY ORGANISATION Ina PC Since the introduction of the Intel 8086 CPU, the bus width of processors has shown a steady increase. Limiting out: selves to the CPU generations which are important in this day and age, itis seen that the 80386 and 80486 feature tan external bus width of 32bits, and the Pentium, one of 68 bits. Add to Fig. 1. Three different packages used for ‘memory ICs: the good old Dik. case as you probably know it from ‘many other ICs; the 30-pin SIMM and the 72-pin SIMM. Elektor Electronics 3196ae Rah that the fact that an ordinary SIMM has a width of eight bits (sometimes nine if a parity bit is available a PS?2 SIMM, a width of 32bits, and you will have ne problems under- standing why a PC motherboard and must contain a certain number of memory modules, Such a group of Fig. 2. The processor bus width also deter- mines the logic width (in bits) of the memory banks. This is an ‘important fact to keep in mind when you start thinking about extend- jing the memory of your PC, modules with a total width equal to the CPU's external bus width is called a bank An 80385 or 80486-based ‘computer has to be fitted with at least four SIMMS to allow the full bus width Of 32 bits to be addressed in If you want to increase the memory of such a PC, that is only pos- sible by adding, at least four SIMMs However, if P52 SIMMs are used, itis possible to work with one module at a time, because the module has a width of 32bits, This is illustrated diagram ‘matically in Fig,2. With a Pentium, the ‘memory bus width should be 64 bits, so that you are forced to work with ‘multiples of two P52 SIMMs (Pentima motherboards usually contain 72-pin memory slots only) DRAM tyres Until recently, there was just one (generic) type of DRAM for computer use. The only specification that mat- tered in the choice of this device was the access time, This specification is one oper- ation. ‘Memory requirements for PCs have increased dramatical- Iy since the arrival of MS-Windows 3.1, and even more so when Windows 95 was released. Eight megabytes seems to be the minimum amount of memory a PC should have to be abie to use these operating systems at all. But memory is expensive, however, and some Software houses recognized the market potential of a product which enables the amount of memory to be ‘increased artiicially These programs are generally referred to as RAM doublers. How do they work? To begin with, a RAM doubler provides much better resource management. Useful as that may be with Windows 3.1, it is not normally necessary with Windows 95. Secondly, these programs do exactly what their (generic) name implies: they increase the amount of memory available, Use it made of the fact that Windows uses a swap file. Basically, that is a file LA” fon the hard disk which is used to store data if it does not fit in the user memory any more. Because a hard disk is much slower than a memory, the use of a swap file tends to slow the computer down. Most RAM dou- ber software compresses data before writing it to the hard disk, $0 that less space is taken up (the process is similar to the file compression techniques used by packiunpack programs such as PKZIP or ARJ). Good! On the down side, howaver, compressing and decom- pressing data is a software overhead which takes time, and requires a dedicated bufier area to be reserved in the user memory. The net profit is, well, marginal, if we are to believe the test results published in various com- puter magazines, Here, too, the rule is: nothing beats real RAMS. None the less, those of you struggling with Windows 3.1 or a small hard disk may like the resutts of programs such as SoffRAM, RAMDoubler or MagnaRAM. Most of these programs may be obtained at prices of 25-odd pounds. Elektor Electronics 3198,usually printed on the device as a suf- fix tothe type number, For most 80386 and 80485 based machines, 6011s or 7Unns is a good choice. With modern PCs, however, the external bus clock is so high that a couple of wait slates have to be ‘thrown in’, not for amusement, but to allow data to be processed reliably even if 60-ns DRAMs are used (which are cur rently the fastest types around) Because DRAMs faster than 60 ns are still difficult to produce in volume quantities, the manufactur cers have come up with a couple of tricks to make their DRAMS faster, at least as far as a number of tasks are concemed. Today's magic word is, EDO-RAM. The abbreviation stands for extended data out. A special out- put register enables data to remain available longer at the output of the RAM. This allows a new address to bbe supplied while the data with the previous address is still being read. In practice, this approach offers a speed increase of between 10 and Wper cent as far as data exchange is, concerned between the CPU and the memory. EDO-RAM is normally a little more expensive than ordinary DRAM. The price difference is, how- ‘ever, expected to disappear before long. The control electronics on the motherboard must be capable of dri ving EDO-RAM. Consequently, you can't normally mix EDO-RAM and regular RAM on a PC motherboard In spite of its higher (appar: ent) speed, EDO-RAM is no real sub- stitute for a second-level cache. The latter remains essential for fast inter- mediate processing of memory data Meanwhile, our highly esteemed memory chip manufacturers have come up with yet other variants like bbust-CAS DRAM, syrchroniows DRAM, RAMbus DRAM and multtank DRAM, None of these is, however, ready for ‘wide application in computers. ABOUT THE PARITY BIT Since the introduction of the first IBM. PCXT all ‘compatibles (clones) have used a memory which contains 2 par- ity bit. The parity bits an extra bit that represents the checksum of the other eight bits. It allows the CPU to check the integrity of the data contained in the memory Unfortunately, memory ICs can develop so-called soft errs. These ‘errors are caused by alpha parts from Elektor Electronics 3/98 1M x 32 (4MB) 1M x 32 (4MB) 41M x 32 (4MB) 41M x 32 (4MB) 41M x 32 (4MB) 2M x 32 (8MB) 2M x 32 (MB) | 2M x 32 (8MB) 2M x 32 (8MB) 2M x 32 (8MB) 4M x 32 (16MB) 4M x 32 (16MB) 4M x 32 (16MB) 4M x 32 (16MB) 4M x 32 (16MB) 8M x 32 (32MB) 8M x 32 (32MB) 8M x 32 (32MB) 8M x 32 (32MB) 8M x 32 (32MB) radioactive Kaliumyy which is present inside the IC enclosure. In the mean, time, the production of the relevant materials is under much stricter con- trol, and the density of the chips has increased appreciably. These two facts have reduced the occurrence of a soft error to about once in ten years, as- suming that a modern system is used in a normal way. The risk of a soft terror occurring is considered so small that you may reasonably question the need for a parity bit. Moderns SIMMs, whether 30 or 72-pin types, may be purchased with or withouta parity bit, or with a simulated one, Incidentally, there are many PC motherboards around these days that do not use the parity bit at all. If that is the case, it 21 Empty, 1M x 32 (4MB) 2M x 32 (8MB) 4M x 32 (16MB) 8M x 32 (32MB) Empty 1M x 32 (4MB) 2M x 32 (8MB) 4M x 32 (16MB) ‘8M x 32 (32MB) Empty ‘1M x 32 (4MB) 2M x 32 (MB) 4M x 32 (16MB) 8M x 32 (32MB) Empty 1M x 32 (4MB) 2M x 32 (8MB) 4M x 32 (16MB) __8M x 32 (32MB) makes no sense to fit memory: tes with a parity bit. bv any ease ing memory modules witha par seems unnecessary if you have a fait ly recent motherboard FINALE: SOME PRAG- TICAL THOUGHTS Memories for modern come in many shapes and si ing the user a har time to sl computer? Is it ordinary DRA pacity per module, and how connector locations do [have able? What is. the processor width? How many modules install at a time? Is a parity 6 quired? By taking a few lo sions it should not be too di find the right memory ext ‘your computerIn audio engineering there is often arequire- ment for a fil- ter that either passes or blocks cer- tain frequen- cies or fre- quency ranges. This require- ment is met by the fil- ter described in this article: the type of filter Filter responses: sae 40 He ~ 25 KHz response required is ar pass pare 5 20 He ~ 20 kHz available at the flick of Fe ee Smptcaton 1 gh mami Switch. outputvoltage "PPE Sn, .76VVImns sine wave) Me ies tage The use of a special fil- (eit impedence as ter IC from National output offset 240 v, 2W Semiconductor in con- junction with switched Capacitors make the construction and align- ment, even for high- order configurations, straightforward. Although the filter is intended primarily for test and measurement, it may also be used for signal enhancement and noise suppression in short-wave receivers and for experimenta- tion in electronic music. Powor requirement The filter fulfils the same function in _measurement technolog ‘wave receivers: it limits the desired fre quency band and suppre ‘outside that band. It then ope a band-pass filter. If, however, band below a certain frequency, say, 50 Hz, has to be suppressed, it function: Js filer. Where high-frequency. s must be suppressed, the filler can be made to work as a low-pass type. If a frequency, $3 50 He or 100) He, within nd has to be suppressed, the filter can op- erate as a notch type. All these types of filter function may prove useful in ‘experiments in electronic music. hi noise oF whist Tue Fitter IC The filter is bas IC from on a special filter tional Semiconductor. ains two second- Aand B, which, 4, Each section is designed variable filter that can funetio a discrete desig i the simplicity of the with which the fre circuit and the quency can be re order functions. The pinout and circuit of the 20-pin IC are shown in Fig, 1. Pin 4(17) is the input of the device. Pins 1 (20) and 2 (19) are the low-pass and band-pass fur to the provides an all-pass, ‘outputs respectively, Pin 3 (18) ther output which, accordi switching ve high-pass or notch charac pin: 6) and 6 (16). In the present Circuit itis arranged to provide a high: pass characteristic istic at in 15 isthe analogue-circut earth to which the non-in of all internal op tence points ofthe erting terminals nps and the refer are strapped. The power su logue digital and 8, Eloktor Electronics 3198STATE-VARIABLE FILTER A state-variable filter is based on two integrators and an inverter Ifthe grator resistors are variable, the filter can easily be tuned. Unfortunately, it is impossible to build a variable resis- tor into the present IC. Therefore, the switched-capacitor principle is used, which provides a variable filter fre- quency, fo, that is dependent on the external clock supplied to pin 10 (11). The equivalence of a resistor and. switched capacitor is shown in Fig, 2, In case of a resistor, the current, I, is determined by the applied voltage, U, and the resistance, R. In case of 3 ca- pacitor the capacitance, C,, fulfils the same function as the resistance R. When the switch connects C, to the input voltage, LI, the capacitor is being, charged. When the switch position is, changed, the eapacitor provides (part of) its charge to the outpul. Thus, in cach switching period a certain charge is transferred from input to output. In this way, an average current, I, flows that is determined by the input voltage, the capacitance of the switched capacitor and the switching, frequency, f,. Thus, the higher the switching frequency, the higher the ‘current, This means that a switched ca- pacitor can replace a variable resistor The linear relation between the switch- ing frequency and the equivalent ad~ mittance of the setup is noteworthy. ‘A small flaw in the setup is that a tiny part of the switching signal is pre- sent in the output signal. In case of a sinusoidal input signal, the output re- sembles a digitally ri ted sine wave, that is, it looks like a stepped ‘waveform. The amplitude of the steps is proportional to the signal voltage, Fortunately, the frequency of the in- terfering voltage is much higher than that ofthe input signal. Thus, it will be virtually undetectable over 2 wide range of filler frequencies, In a practical circuit, integrated ‘©Mos switches are used, Depending on the voltage at pin 12 the clock for ICy must be 50-100 times hhigher than the wanted filter frequen= yf With +5 Val pin 12, the rmulti- plication factor is X50. This arrange- ment ensures that the internally gen- trated noise ofthe filter isa minimum, ‘Moreover, the clock at pin 10 (11) must not exceed 1.5 Mhz, which means that with a multiplication factor of *100, the upper frequency of the filter ‘would be limited to 15 kHz. Note that the factor x50 must not be taken too literally: it has a tolerance of 2%; in the prototype, it was 494. Elektor Electronics 8/07 ‘Top View THE FILTER PROPER With the possible ence tion af the switehed-ca- pacitor principle, the funtion of of the circuit in Fig, 3 i straightforward Circuit IC operates as a bipolar input buffer that is provided with an input protection network, Ry-Dy-D> “The voltage drop across Ry is com- pensated by the amplification of Which is set to L5 by Ry and Ry, Ca pacitor C3 stabilizes the op amp, since 15834 is internally compensated only for amplifications exceeding x3, The offset of IC, is compensated with Py. ‘while its output is taken directly to the input, pin 4, of Cy ‘The low-pass (UF), band-pass (br) and high-pass (Hi?) outputs of filter section A of IC; are applied to ganged switch array $ySz, which can also se- leet notch fiter Wand octave filter 0 With the switches in positions tr, Lr and or, the outputs of filter section A are applied to the input of iter section B. That is, the tw filter sections are cascaded (fourth order filter) to pro- vide a steeper slope of the response The outputs of filter section B are applied to ‘output buffer [Cy via the switeh aera. ‘Output socket Ky is preceded by low-pass section, Ray-Cox, which, at least at higher val- ues of fy Suppresses any residual clock 2 signal at the output With the switch array in postion x, a notch filter is con- structed in summing amplifier IC, from a combination of the high-pass and low- poss functions of fi- {er section A in IC, O=O are ees through the frequen- cy-dependent switch, ae The output of IC, is applied to the output butfer via the switch array: The second se tion of IC; is then not used: its input Ry. grounded via the switch, resis ‘The band-pass characteristic of the filter is relatively narrow. For cases in Which a wider frequency band is needed, an octave-flter has been pro- vided. When the switch array isin po- sition 0, the high-pass output of filter section A in IC; is applied to section B via the switch array, and then taken from the low-pass output (pin 20). At the same time, Sj, ensures that the clock to the first section is half that to the second section. This arrangement [provides an upper limit ofthe Hi char acteristic which is an octave Tower than that of the low-pass characteristic, so that the frequency band between these two points is available at all times. REMAINDER OF crRcurr ‘The output of IC, has a small offset voltage which nor- mally does not affect the operation. The larger part of the circuit enclosed by the dotted lines is the dod penertog witch ie Neu Coty . LD » ey? = Cs 25)ap is based on IC,. Although the manufac- turers’ data sheet of this cmos device ndicates that it can work up to 500 kHz, it operates with reasonably steep edges to well over 1 Mhz. when, itis arranged as an astable, Resistors Ryg and Ryo, in conjunction with Ds and Dy, provide a duty factor of 50%, hich is required for optimum opera- tion of IC, IF necessary, the factor can, be trimmed to exactly 50% with Py. Capacitors C-Cy and Cay-Cay, Which are switched with Sia, and Py enable three overlapping frequency ranges to be set: 1-10 kHz, 10-100 kHz, and 0.1-1 Mhz, corre sponding to filter frequencies of 20 Hz. to.20 kez, TD ensure troublefree oscillating of Cy at high frequencies, the device is switched between -3 Vand +5 V. Re- sistors Ra) and Roo revert the oscillator -voltage to-T1- level referred to ground, wh The clock is applied to pins 10 and 1 of IC} direct and via Sy. Half the clock frequency required for iter section 4 of IC, when the oc- tave filter function is selected is pro- vided by ICs,,a bistable connected as binary counter, Socket Ky has two funetions. To standard operation a frequency meter may be connected to it to aid in setting the filter frequency which is gq of the measured dock. In the fourth position of the range switch, ICy is disabled via the reset input (Sq,) and isolated from the re mainder of the circuit by Sqy- An ex ternal clock with known frequency at ‘Tm level can then be connected to Ki IF the signal is higher than rit, it will be limited by network Ri-DyDy ‘The orvoff indicator Li, Dy, and its series resistor, Ry, are connected be- tween the -ve and +ve supply lines to ensure symmetrical loading of the power supply. The power supply is a traditional design. The specified transformer is 26 short- sponse curves of the individual filter types. These curves have been nor= malized and are obtained with the minimum deviation from fy in all cases. xsi Fig. 6. Normalized 6 transfer functions of the M10.The Radio Data System (RDS) is now well estab- lished among VHF- FM broadcasters aiming at providing additional information grammes. In addition to up-to-date traffic information, RDS also brings you, on an LCD ‘screen, alternative fre- in their cars. A PIC16C84 micro- controller allows a simple and pretty compact RDS add-on decoder to be built based on only two integrated circuits. from the traffic-re- lated functions aimed at the car radio market, other RDS functions such as transmitter ide The RDS decoder presented here was developed to function as a simple ex- tension to any (we hope) domestic FM stereo receiver, The circuit dis plays the station name, time (with atomic accuracy from some transmit: ters) and radiotext on a 216 or 140. character liquid The printed ¢ 32 quencies, time infor- ns mation and a feature i tie! a called Radiotext. The n ec ig ra latter in particular ma? yas” a is on the rise invous are tind ynage these days, with ‘patie annem wo 105 onset”? an increasing simle® ve yoo mole cure number of stations 4 ee omole? a onan _ Offering this adci- 1 Om ay FM a ssyaesons tional information to * ena matches the size of a 2 [etencsathome and = Apart. acerLCD unit THE ciRcuIT RDS processor ICI contains every- DS subcarrier filters as emodulator The ree thanks to a Elektor Electronicscoder as simple as possible. supply voltage of SV is required. to power the RDS decoder. The de- coder supplies the familiar set of RDS output signals: RDS CLOCK {pin 12), RDS. DATA (pin 13), RDS QUALITY (pin 14) and ARI (Autofahrer Radio In formation, for Germany only) (pin 15). ‘The TDA7330 may be used with a 4332-MHz or an 8.664-MHz quartz exystal. The first option is selected by leaving pin 18 open-circuited, the lat- ter by strapping pin 18 to +3V. Here, an oscillator frequency of 4.332 MHz. is used, which doubles as the elock for the PIC, allowing this device to work without its on-chip oscillator Design information on the TDA7330 is given on the Elektor Electronics Datasheet elsewhere in this issue, The input of the demodulator chip. is supplied with the multiplex (MPX) signal, which is ‘stolen’ from the input of the stereo decoder in the FM receiver. Only two of the previously mentioned RDS outputs are used here, namely the RDS data signal (RDDA, pin 13) and the asso- ciated clock (RDCL, pin12). The clock frequency equals 57,001028 Hz ie, 187.5 Hz. It is used to produce an interrupt at processor pin RBO. [n the interrupt routine, the processor reads the data signal at its RTTC pin {pind of the PIC), The liquid crystal display (LCD) is operated in four-bit mode, in which Eloktor Electronics IC socket. The display is se- cued atthe solder sie ofthe board with the aid of short wires or header/socket con- nections, so that the solder sides face one another. Al though the noise emission of the decoder/display unit will be lovy it does no harm to fit the project in a screened metal enclosure. In most cases, however an all plastic case will be equally suitable The board has a cinch socket via which the muiplex signal js applied. Before taking the RDS decoder into use, you should make sure the board is correctly populated, Also check the use of the jumper {o2e above), Foran inital test power the decocer from a di rect voltage the) source supplying for anything be tween 9 and about 12V, for tor IG, a programmed — imple, your only D4 through D7 are | piGanda standard benchtop regu: used. Data lines D0, D1, D2 LED module lated de. power and D3 are, therefore, not tised. RBI, RB2 and RBS are the con- trol signals for the display. Jumper [PI selects the ICD type used: JP1 open = 2x16characters; PI closed = Lxd0-characters. The contrast fon the LCDs adjusted with preset PI Voltage regulator 1C3 allows the de- coder to be powered from a supply voltage higher than 5 V, which may be available in the FM receiver. Current consumption of the decoder is small: er than 15mA. Diode DI acts as a supply reversal protection, and at the same time creates a an aur bias of -0.7 V which is applied to the lower terminal of preset P1. The use of a (small) negative voltage en- sures that older LCDs, too, can be made to produce sufficient contrast Because the PIC processor combines everything that belongs in a com: puter on a single chip, it emits only low interference levels, reducing the main sources of noise to the LCD. connections, which should be kept as short as possible. Actually, that is why the LCD is fitted directly on to the decoder board CONSTRUCTION AND CONNECTION ‘The single-sided printed circuit board, 4s simple to populate. If you go for ‘minimum cost give a least the PIC an, 33 supply. You will also need a properly working FM radio with ‘line’ outputs. This may be ‘an FM stereo tuner or a mono radio with a tape recorder output. With ‘most high-end FM stereo tuners, the ‘multiplex signal is suppressed to the ‘extent that the RDS decoder will not be able to detect even the smallest trace of it at the receiver's line out: puls. On most down-market FM ra: dios, however, as well as on many mono receivers with 2 line (tape recorder) output, it is definitely ‘worth trying the audio outputs! With stereo tuners, it makes no difference whether you use the left or right line output. itis, however, essential to have the receiver tuned to a strong, signal from a station of which you are certain it transmits RDS. ‘On being switched on, the RDS decoder produces an identification which remains on the LCD until a valid RDS signal is detected and processed. The identification may be used to adjust the LCD contrast Once a valid multiplex signal is re- ceived at the input of the decoder, the display shows the station name, for example, BBC RI, almost immediately after the power-on ident has disap: peared. After a further minute or so, the fime also appears on the display, next to the station ident. When youCepacitors: C1.C2,03,C10 = 100nF Ch = ame C5 = 10250 67.08 = 270F C9 = 100uF 25v (08 = 100F Inductor: Lt = 100, choke Semiconductors: Di = 14002 G1 = TDA7330 (SGS-Thomson) 1C2-= PICI6CB6 (programmed, ‘order code 868505-1), 103 = 7aLos Miscollaneous: Ki = POB-mount cinch socket K2 = LCD, 2x16 or 1340 characters. retested type: Sharp LM 16421 (16) KS = 2way PCB terminal block, ‘grammed PIC for this project: order Code 260050 (ses page 70) 190000000000000) O are receiving a trafic information sta tion, the display shows a (dash) at the far right. This changes to a’” when traffic information is transmitted The lower display line shows the radiotest (if available), which scrolls, Up to two messages of 64 characters may be transmitted, These are joined to give one 128-character text, which is displayed sequentially If you are unable to achieve us- able results from the receiver's line ‘output(s), there is no alternative but to locate the MPX signal inside the tuner, at the input of the stereo de- the block dia- gram of a typical FM tuner, Fig. 5, this point is found near the FM de modulator, where the demodulated FM multiplex signal ‘before de-em phasis’ is available, The actual point to tap is easily found using the re- ceiver’s service documentation and/or your oscilloscope. If have neither of these, itis also possible to find the MPX signal by trial and error, Starting from the receiver's coder. As shown cinch output sockets, trace the wires or PCB tracks that lead to the stereo decoder IC. Once you have located this IC, the first thing to do is make sure your are tuned to a station of which you are certain it transmits RDS. Next, carefully touch each of the decoder IC pins with a probe connected to the RDS decoder input. Because of the processing time needed by the RDS demodula- tor, you may have to keep the probe A connected at least 30seconds to each pin before data starts to appear found, the station name is displayed after a short while, followed by ra diotext (if transmitted). Having found the proper signal for the decoder's MPX input, the tunit may be installed and wired per manently, either inside the receiver, fon the display. Once the right pin or as a set-top extension. If you choose to fit the decoder in the re- ceiver case, you will need to locate a suitable take-off point for the 9-V supply voltage. Hopefully, that is not too difficult to find in the receiver. The signal connection between the tuner and the decoder is best made in thin screened cable. If the supply voltage is also taken from the receiv er, the cable screening may be con nected to ground at one side only the receiver side! This is necessary because of the polarity reversal pro- tection diode, D1, which would oth- cerwise he shorted ot, causing lower contrast on the display. TEST VALUES AND TROUBLESHOOTING The circuit diagram, Fig. 1, gives you a number of d.c. test values as well as points at which active signals can be measured. The relevant signals, measured on our prototype, are shown in Fig.3. While measuring the direct volt: ages in the circuit, you should note that pin marked ‘0 on K3 does not represent ground. This is actually al-0.7V because of diode Di (whose anode is at ground poten- tial). For measurement purposes, ground may be taken from, for ex- ample, the negative terminal of C9, the cinch socket. In addition to the operating voltage (5 V) and the ref- erence voltage (approx. 2.2V at pin2) you n sure the logic levels at pins 14 (RDQ) and 15 (ARI) using a digital multimeter: RDQ (designated QUAL in the datasheets) indicates the qual- ity of the received RDS signal. This pin supplies a logic high level when an RDS signal of sufficient quality is received. RDQ is low when the RDS input signal is too small, oF missing, The ARI pin goes high when an ARI signal is being received, and low when only an RDS signal is avail able. The level of the ARI pin in un- determined when neither RDS nor AR is being received. In the UK, the yy also want to mea Eloktor Electronics 3198Signals at the measuring points indicated in the circuit ‘diagram: (a) stereo multiplex signal; (b) filtered 57-kHz signal; (c) RDS data signal (upper trace) and recovered 1187.5-Hz bit clock signal (lower trace); (d) 4.332-MHz oscillator signal. ARI system is not used The signals at the other measuring points, A through D, are shown in Fig. 3: the multiplex signal (Fig. 3a), the filtered 57-kHz signal with RDS. and ARI (Fig.3b), the recovered bit, clock and data signals (Fig 3e) and, nlly, the oscillator signal (Fig. 34) The levels supplied by pin 14 (RDQ/QUAL) and pin 15 (ARI) are easily evaluated with a multimeter to give a go/no-go verdict on the op- eration of the RDS demodulator chip. Ifthe levels measured do not make sense, you may have hit upon a station that does not transmit RDS. If you are sure, however, that you are tuned to the right station, then a ow level at the RDS output (and possibly at ARI also) indicates that the tuner either does not supply an MPX signal (wrong internal connec tion), or the MPX signal is too small The latter is unlikely, however, be- cause the TDA7330 boasts a sensitiv ity of mV for RDS signals and 3 mV for ARI signals (sce datasheet ex Eleitor Electronics 3198 tracts elsewhere in this issue). With- out the help of an oscilloscope, you may have to solve this problem by trying out another receiver, or find ing another point in the receiver to tap the unfiltered MPX signal When ICI reports an RDS signal (pin 14 goes high), the LC display should at least indicate the station name after a few seconds. If the power-on message does not appear at all, the first thing to check LCD contrast setting. If another dis- play is used than the one mentioned in the parts list, it could just be that the connections are turned 180 de- grees (this is the case particularly with older LCD models). When the display does not show a dark line at oyany setting of PI, and with all ICs removed from the decode should simply try to reverse all con. nections by 180° 5 When the display works all right, and the TDA7330 reports RDS re- Like any stereo decoder, the RDS ception at pin 14, the only remain- | ggmodulatoridecoder has to be sup- ing source of trouble is really the os- | feet with the stereo MPX cillator, The oscillator signal at pin 16 | Signal, which is available at the output may be checked with an oscillo- | of the receiver's FM demodulator. oa scope, just like the signals at pin3 (RDS data) and pin6 (RDS bit clock). Assumi are okay, and the display is working, and 1g that these signals. may be damaged connected the right way the device, Although it rarely hap- RDS demodulator wi around, it is time for a thorough pens in practice, it can not be ruled Elelor Elecironics Ost 1993. check on all solder joints and PCB out that the CMOS controller has Radio Data System (RDS) dearer Fe Ectrmis May 1993, tracks around the PIC on the board. become the victim of a static dis- age Date Sytem (RDS) decoder El If you can’t find an error, the PIC charge or overheating, (ooo, Electronic January 1958, ‘Radio Data System (RDS) infor grams or even GPS correction data ‘mation has been transmitted by The Radiotext (RT) as displayed by the present decoder ‘mary VHE-FM radio stations since 1988. RDS signals are conveyed at a subcamier ie quency of 57 kHz. This tre- RADIO DATA SYSTEM mies quency is phaseocked to the pilot carrer of 194He Gict9 kite © 37 kh), Figure? shows the theoretical o- ‘greney spectrm ofthe multiplex signa a tansrted by 2 VAPEM radio ston broadcasting ROS. ROS daa mmoclated using a tachnique known #8 doublesidebard Suppressed carrer (DSSC), which results in two clearly visible sidebands whose bandwidth Is roughly equal 10 the binary RDS data rate of 1187.5 bits per second (avg). The bt clocks recovered rom the suBcarrr by dividing & by 48. That also explains why a frequency of 1187.9 Hee avalabe atthe clock output (ADCL, gn 2) ‘of the RDS demodulator IC TDA7330, provided, of | Course, a valid ROS signals being recenved | Basically, the ROS signal consists of binary data which ‘are transmitted serially. This data is organized in blocks of 26 bits each. Each block contains @ 16-bit dataword and a 10-bit checkword (Fig. 2a). The high redundancy of the checkword affords security of the deta even under very adverse receiving conditions such as in a moving car. Four 26:bit blocks togethor form a 104-bit group. In a group, the blocks are sim- ly identified by their order, i.¢., Block through Block 4 (Fig, 2b). Blocks and groups are transmitted continuously, 80 that the transmission time per group | equals 87.5ms at the above mentioned baudrate. Each ROS group is of a certain type. RDS supports 15 group types. The following information is contained in ail groups: programme identification (P1), programme type (PTY) and tratfic (TP). The rest of the information Conveyed in the group is variable, and depends on the ‘group type. The information type list 's long, and extends from datalclock info (Type 4 Group), through internal infor ‘mation on the radio station (Type 6 Group), right up to @ transparent data channel (TDG, Type 5 Group) which is intended to convey small computer pro- i 36 may be ascertained by exchanging, Fletor Elects May 198% Radio. Data. Systems, Fedor Eh July August 1991 Rado Data System (RDS) decoder strnies February 1991 electrically, which is contained in type-2 groups, and consists of yp to 84 characters which supply any information the broad- casters feel free to throw at you, usually related to the ‘currently transmitted programme. AT supports spaces, enabling the fext fo be easily read from a 2x32 display. Inpprinciple, itis also possible o transmit texts of any length in successive lines (64-character chunks). In practice, however, that is rarely done. Apart from the low transmis- sion rate thet can be achieved (no more than about 20characters per second, provided every alternate ‘group contains FT), the main objection against using Jong texts is probably that the broadcasters are painfully ‘aware of the relatively low number of people who are willing and able to receive and read radiotext. Hopefully, that situation changes for the better following the publi- cation of this article 36e10bH AE tonH | se+t0HN 6410 locket | sleck2 | locks | aocks Eloktor ElectronicsAn ordinary PC with a 16-bit sound card is turned into an excellent audio-frequency mea- suring system, at a very ‘small outlay, just by running some clever software. Applications of the PC-controlled AF analyser include mea- ‘suring frequency and phase response of amplifiers, filters and, in particular, loudspeaker cross-overs. But that’s not all because the analyser also enables you to determine the impedance of loud- ‘speakers, the induc- tance of coils, or the Capacitance of capaci- tors. A noticeable fea- ture of the analyser is its wide measuring range, for instance, from 50H to 100” for inductors, and 100 pF to 100,000 uF for capacitors. Sea art sound card like the Soundb 16 VE will set you back less than 70 pounds, yet it contains, among, others, a stereo 16-bit analogue-to- digital converter (ADO) for 4i-kH4z Sizer chip capable of generating complex. waveforms. No wonder you may want to make all this high- joa bit more than pro- nd hobbyists lyser sta alike is the versatile 40 sd in this two- le. This first part deals with bili system, ff the mea- surement while next month’s second and final part tack- les the operation of the software de- veloped for the project. It will also discuss a small add-on box for mea. surements, The box contains a cou. ple of resistors and switches, and turns the PC in a multi-purpose im: pedance measuring station For EXAMPLE To begin with, the main technical fea- lyser system w tures of the discussed by amples a couple of ew EckFrequency response of an amplifier The graph in Fig. 1 is the result of a1 measurement performed on the audio power amplifier in a low-cost car radio. A non-reactive load was driven. The frequency response was recorded over a range from 20H to 20 kHz. It is seen that frequencies around 100 Hz get an extra ‘boost The ‘bass’ control was turned up for this measurement. Clearly, the fre- quency response starts to drop at about 5 kHz. The phase response is not very linear either: Impedance graph of a loudspeaker ‘The results of an impedance mea- surement on a loudspeaker system are shown by the graphs in Fig. The measurement was performed on a co-axial drive unit (ie., one with the tweeter located centrally in the woofer/midrange speaker) which was claimed to have anomi- |) nal impedance of 4 over the fre= guency range from 20 Hz to 20 kHz. The graph shows the real part (re- = sistance) and the imaginary part (re som actance) of the loudspeaker’s com same plex impedance. The marker is set at 122Hz, where the first self-reso- nance of the drive unit occurs. Where the real part goes through a maximum of about 209, the imagi- nary part shows a zero crossing. The 19.00 ome second self-resonance point occurs at about 5kHz, where the imped ance rises fo nearly 7. As you can see, the measurement system is ca- pable of capturing all essential de- tails of an impedance response. Inductance measurement on an RF choke igure 3 shows that the value of an RF choke rated at 40H can still be | BS name eee measured with confidence. Obvious- SG GT Gath. RS, eS ty, such a measurement calls for the (20 nase $0) et rn highest frequencies the analysers ca- sense som pable of producing, in the example, eee igen the measurement was carried out aie 240 to over a frequency range from 5 kHz t0 aes loin kHz, The measured inductance (approx. 40H) is shown in the top pare ~ part ofthe Mustration. The lower part vow se shows the resistance of the choke, ao Which amounts to about 19 (at a et measurement error of about 102). me, You guessed it the AF analyser is an excellent help when you have to wind RE coils yourself. But low-fre quency coils foo, can be measured . without problems, for example, those )40-uH RF choke — in loudspeaker filters which have val ecked out. tues in the millchenry range. Elektr Electrons 396 wt ~~MEASUREMENT OPTIONS OF AF ANALYSER Frequency response (level and phase) of amplifors + fiters loudspeaker cross-over fiters control loops Impedance measurement on two-poles + loudspeaker parameters + fter inductors (electrolytic) capacitors + resistors: ‘input/output impedance of fiters Inductors, even smaller than 50 cH MINIMUM PC REQUIREMENTS FOR AF ANALYSER PC: 486DX80 with VGA colour ‘graphios Sound card: 16-bit SoundBlaster VE (Creative Labs) Software: AF-Analyser program (DOS), published by Elektor Electronics MAIN TECHNICAL DATA 2 Frequency range 15 Hz to 20 kHz * Freely selectable start/stop frequency ‘Sweep: logarithmic or linear, up to 1,000 ‘measurement points Marker function for accurate indication of measured values On-line help function % Readout in dB, degrees, Q, F, H (auto- ranging) + 16 bit theoretical resolution Measures gain, phase and frequency response Impedance R: 1210 1M C: 100 pF to 0.1 F Li 50uH to 0.1H Measures equivalents of two-poles RYIO, R||L, A+l and A+C SOUNDBLASTER, Compatipte & Co. The measurement software was writ. ten for a SoundBlaster 16VE from Creative Labs. This card features 4 kHz. 16-bit stereo sound sampling, The software should also run on 10% SoundBlaster compatible cards, preferably those with the Cre ative Labs chip set. Regrettably, many low-cost clones use a different chip set, so you are well advised to buy the original SoundBlaster. At a current street price of 70-odd pounds, the SoundBlaster 16 VE is a good investment, and should not blast’ too large a hole in your hobby budget! The software, called AF-ANALYSER, is available on diskette through our Readers Service (see page 70). AF ANALYSER (AFA.EXE) is simple to install, Once you have verified that the diskette you received from us is okay, you start by copying all files on the diskette into a suitably named subdirectory on the hard disk. Next make sure that the SoundBlaster card is properly configured (address, DMA channel and interrupt ~ see the README.TYT file) Start the program from hard disk by typing AFA. It will first attempt to address the sound card. If that fails, an error report is produced, Very likely, your sound card is then improperly configured. Next, the program attempts to secure addi tional DMA compatible memory of fered by the operating system. If that fails as well, another error re port is produced. In some cases, you may have to modify your memory setup. The author uses the analyser system without problems on an 80- MHz 486DX with 4 Mbytes RAM Once the initialisation is finished, a picture appears on the monitor, and you are ready to start your first ‘measurement. Incidentally, the soft ware alters the settings of the mixer parameters used on the sound card The next important point ta con- in is the wiring between the SoundBlaster card and the mea sured circuit. The connector pinouts and signal functions are shown in Fig. 4. For an initia test, connect the SIGN.-OUT (signal output) socket to a loudspeaker, pressing the x’ key on the ke board, The loudspeaker will pro- duce a signal whose frequency rises. ‘and start a sweep by 42 The measurement system generates its test signals with the aid of the EM sound synthesizer chip on the sound card. The audio power am- plifier, which is also contained on the card, allows a peak-to-peak sig nal wel of about 750mV to be achieved across a load impedance of BQ. In other words, the syst suitable for running checks on loud speakers at low power levels. The fact that you do not know the ab solute level of the signal applied does not distract from the useful- ness of the me: makes no difference for the m: surement principle used (Fig. 5) The sound card having a stereo rarement because it A-to-D converter, one channel is tused as a reference channel which receives the get ly. The reference channel enables the software to accurately measure the level and the phase of the generator tor signal direct- signal. The generator signal is fed through the measured object, for in stance, an amplifier ora filter, whose output signal is applied to the other channel on the sound card The software uses the signals on the reference channel and the mea- surement channel to compute the gain or attenuation caused by the measured object, as well as the phase shift. The measurement makes clever use of the A-to-D con- verter and a phase-sensitive rectifi er. A series of individual measure. ments is performed in the selected frequency range, and the r displayed on the computer screen in the form of a plotted graph. The user may view two graphs at a time on the monitor, The parameters shown by the individual graphs are selected via the keyboard. With measurements on amplifiers, for ex ample, itis possible to view the ac- tual value and the phase shift of the measured gain, Similarly, with im. pedance measurements the real and the imaginary parts may be dis- played separately, or the discrete values of an equivalent RL combi nation. The measurement parameters are easily changed via the keyboard Likewise, the complete bundle of settings is readily stored on the hard disk. Finally, an on-line help function is available, and may be called up at any time to obtain a short descrip. tion of each available command. Elektor Electronics 3196SIGNAL GEN. a Fig. 4, Pinouts and ‘connecting diagram of the SoundBlaster VE 16 sound card. Your FIRST FREQUENCY RESPONSE MEASUREMENT As your first measurement object we st using an RC bandpass filter a shown in Fig, 6, connected to the rel evant sockets on the SoundBlaster card (Fig. 4). The default parameters used by the measurement software are okay for this initial test. All you have to do is press the x key (for ex- ‘ecute). That launches the frequency 5 FM synthesis chip Soundblaster speaker output Elektor Elect ai96 measured object 3.5-mm 16-bit soundcard jack plug LINE IN ‘SPEAKER OUT. Jnr and phase response measurements, the results of which appe screen after a while After the measurement, the dis- play shows two graphs as illustrat. on the ed in Fig.7. To enable individual values to be read off accurately, you use the + and ~ keys to manoeuvre the marker. In this way you capture, for example, f, at 1,950 Hz, or an at tenuation of ~9,774B at the reso ‘nance point. Both values come very close to the theoretical ones. By pressing the keys correspond ing to the letters enclosed in square AF analyser software A-D converter A-D converter reference measurement ‘channel channel aa | Soundblaster Tine input 4 43 '960053-14 brackets, itis possible to modify the relevant parameter in the software That may be necessary at times to improve the way new measure ments are matched to given circum- stances. At this point you are ready to do any AF gain and/or phase shift mea surement you like. It should be ¢ by now how the measured object is connected to the sound card. If nec: essary, an attenuator should be used to prevent overdriving the mea sured object ar the sound card. Cit= cuits with a low-impedance output graphics Fig. 5. Schematic rep- resentation of the ‘measuring principle.LIMITS OF THE FREQUENCY RESPONSE MEASUREMENT ‘No measurement resull makes sense if you neglect, or do not ‘know, the limits of the measurement system used. With fre- quency response measurements, the practical limits of the AF analyser are determined by a number of extemal conditions, each of which has an effect on the accuracy of the measure- ment, Although sudden irregulanties like offscale peaks in the graphical representation of a frequency response curve are Sufficient waming that a serious measurement fault exists, itis still useful to be aware of the main problems you may encounter. The first source of trouble may be the FIM synthesizer ascila- tor on the sound card. Fram about 15 Hz, this oscillator supplies @ triangular waveform instead of a sine-wave, This is owing to the fact that the oscillator is of the ‘sam- pling’ type. To keep the resuling error as small as possi- ble, certain measures are taken by the software so that only the relevant basic frequency of the signal, not any of the harmonies, is employed for the measurement. None the less, resulls of measurements above 15 kHz should be taken with a pinch of salt. = Second probiem: the A-to-D converters must not be over- driven. Many measurements on amplifiers will therefore require an attenuator pad to be inserted in the measure- (less than about 10 k2) can be mea sured directly. When a high-imped- ance output is used (more than 10 kO), impedance matching should be provided by a small preamplifi- cr with a high-impedance input NEXT MONTH Next month’s second and last instal- ment of this article will guide you through the operation of the AF analyser software. We also present a small add-on box for measurements, on Don’t worry, the box contains only a handful of resistors and a switch, To- ‘gether with the AB analyser hardware and software, this gives you a fairly advanced impedance measuring in. strument Fig. 7. Results of a frequency ‘ment channel. If you look at the attenuation graph above the frequency response, it is safe to say that the A-D con- verter is probably overdriven when levels greater than 0.48 ‘occur, ie., when the signal level in the measurement chan- nel exceeds that in the reference channel. + Special attention should also be given to the input imped- ance of the SoundBlaster card. The fact that this input does not have a particularly high Impedance should be faken into account when measuring on high-impedance signal sources. If necessary, insert @ preamplifier with a high input impedance in the measurement channel. Obviously, the frequency response of this preamplifier may. also be checked using the system, ~ A further limitation to keep in mind is caused by two con- tending factors: one the one hand, the 16-bit resolution of the sound card, and on the other, the high (electrical) noise level which exists in a PC. Theoretically, a 16-bit converter achieves a signal-to-noise ratio of 95 a8 at full drive. In a very ‘noisy’ environment such as a PC, howev- er, the present analyser system can measure attenuations: down to about 70.dB. Not a bad value, mind you, consia- ‘ering that better results are almost impossible fo obtain rom simple sound cards. - The final sources of interference which must be mentioned here are stray capacitance and resistance, whose effect is especially noticeable with impedance measurement on AF components. More about this phenomenon in next ‘month's concluding instalment. SIGNAL GEN. er. NP. MEAS, Elektor Electronics0 is actively involved and with access to may use his or her er information about s, technical software er interesting products. ‘the PC you have almost access to information which are ‘hot’, up to \d of a staggering size. to time, however, the Internet user has a 3m finding exactly those which have interesting is a new regular fea- : magazine, intended to you about Internet sites it hold on-line information ‘of interest to electronics enthusiasts like you ‘we kick off by presenting (world-wide web) pages from ar Harris Semiconductor, The ad- ‘ttp:/wwewsemicharris.com Itisako possible to down- related software, and the site con- information about recently re- and about-to-be-released Harris, fs In practice, this informa- ‘more up-to-date than the Published by Har- the www site allows submit technical queries to the design engineers at Harris ‘Semiconductor. A totally different Internet site Is the one from HiTools Inc, who are also based in the USA. Hiools are suppliers and manufacturers of micro- processor development sy tems, The www pages found athttp/wwwhitex.com bring. masse you, among others, lots of demo software like an 8051 and an 80166 simulator. A highly interesting option of- fered by these pages is the link to the so-called Chip Di- sectory, which enables you to trace the function and origins fof unknown components, ‘The actual search operation is quite simple |s all you have to do is enter the type number of an unknown IC, and the program does the rest. If available in the database, the information appears on your screen after a short while. summary ofthe information at this www site may be found on the Development Tiuls CD-ROM. This CD-ROM. contains a number of HTML (Hypertext Mark-up) Lan- guage) files which give the user information on 44000chips and 150 manufac- turers, The CD-ROM ako can- tains an 8051 and an 80166 simulator, an 80166 debugger and text versions of compilers. Allis of Frequently Asked Quas- fions (FAQs) about the 8051, 6S11C1 ancl 68k families may help to reduce the comple ty of developing application for these processors. AS a mat- ter of course, the Chip Dieto- rib aso included on the CD- ROM. Although the CD con- tains an HTML browse many users will prefer their own web browser like Netscape, Mosaic or Internet Explorer Tortunately, that i possible! The author of Chip Dictory i Jaap van Ganswijk who can be contacted at Ine novative Design, PO. Box 3215, NL- 2601 DE Delft, The Netheslands, tel (G31) 152132638 fax (+31) 152140244, eur) A electronics on-line Found awe ste hich other render of Eek tor Electronics should knoe about? Let us Joe the name andl URL by telephone fx oF ceil elektr@ eure.) If este sin ded useful and interesting, wel pai i the limelight sing this column, Elektor Electronics 3196CMe ee ADVANCED STOCK CONTROL AND KITTING PROGRAM Foe the frst time in a market rinated by large exponsive ac counts-oriented packages, Num ber One Systems have launched sock conf progr fete ral world simple uncomplicated Drawing on the experience of ‘ineering entrepreneurs and de sign depaetments already using Number One Systems range of ictronc CAD programs, Stock is designed to help smaller manufacturers and RED depart iments with the day-to-day prob: lems of managing stock. Stocki is spectically tailored for 2 small to mediom sized proc tion or development environ: ment it not an adaptation of ‘Stoctt hols complete records of parts lists, stock levels and con ponent suppliers, atomatcally alcabing Stock when a new pro duction batch i equi. A val able tal Kiting fachty allows diferent production scenarios 0 be evaluated before a definitive DOUBLE-SPEED Portable Add-ons recently intro duced the bunch of CD_ Mobile an external doublespeed CD. ROM drive and Type-ll PC Card terface designed for notebook raters. With four unique power options compute mins, andanl, or rechargeable AA bat teres) the CD Mobile gves total Aesiblity of use, Since the CD Mabie fered, it can be used anywhere nb hattery-pow waking ita truly portable sol Aa retail price of £299, CD_Mo- bile i an inexpensive way 0! adiing CD-ROM functionality to tvsting notebooks. The fw note books with a CD_ROM drive builtin tend to be prohibitively expensive internal drive always draws cur ent it wll place a oad on the nachine’s battery in adiion, a builtin CD-ROM drive adds toa notebooks bulk. As the CD_Mo bole a separate ut tcan be un plugged andthe PC Card Inter ne remove weno in se ducing both power conssmption and weight. The CD_Mobile can production plan is adopted. The ability to handle fractional quan lites allows cable Feng, con sumables and labour to be i ‘dudes inthe eaeulations, Stock. its equally ap plicable to batch manutactare 39 any other industries. Priced at just £14500, Stockit is far easier fo use [sand cheaper than many account-based even prints enquiry sheet ‘sent to suppliers for quotae stock contol programs intended tions for large-scale m The structure of Slockit enables new werstomigate est | See ing systems gradually, fone assembly” at time. The benefits of Stockt can be realised Iu Stock Pet) ‘quickly without wait wD ing for every single Prd part to be entered rod By maintaining on | RCE up-to-date cost for | apna each part, Stockit | Op makes t simple tod be enoagh wakes it simple to de fdas termine product man: ey facturing costs and org order values. Stockit eed sao calculate a cur any rent stock valuation | (Medtagatiog in a matier of sec Pra nds. Cae While designed with the electronics ind ‘umber One Systems, Harding Way, Somersham Road, St. ves, Huntingdon, Cambs PEI7 4WR. Tel (01480) 461778, fox (01480) 494082, at (or Production) tu Order 2828222223288 223) Of users, The IDE-based interface is. wth al gooxl PC Cards flly host-swappable The CO_MobileS CD-ROM drive is fully MPC2 compliant, It also uupports Video. CD and CDs in addition to CD-ROM modes 1 and 2, audio CDs, multises Photo. CD, and CDA, The beautifully styled CD_Mo- bile has. full control panel and 3 rulti-function LCD screen, The tunit cam be used as 2 super quality personal CD player, al Towing you to listen to. your favourite album while compiling, spread-cheet on the train, Th CD-Mobile may’ al be use swith recent in-car stereo systems That feature ¢ ne input, brainy cosette/radio if a commonly available adaptor is used Portable Add-Ons (UK) Ltd, Surrey Technology Centre, 40 Oceam Road, Guildford GU2 5¥H. Tel. (07483) 1440777, fax (01483) 452304, e-mail: ‘[email protected]. Eleitor Electronics 3/96_ Adjusting a satellite TV dish is often com- z pared to finding a needle in a haystack. Although the coarse setting of the dish is fairly easy to find by looking at the direc- tion of nearby dishes in the street, you may need a lot of patience, mechanical twid- ___ dling and shouting to and from the TV room before the first pictures appear. The prob- lems are often aggravated by the fact that you are standing on a ladder or ‘dangling’ from a balcony. The instru- ment described here is simply connected to the LNC output, and does away with the need to actually see the received picture. All you have to do is _ watch the needle of a ~_ moving-coil meter, and peak the dish adjustments for the highest reading. Battery-powered and portable! Despite the ad- vice given by dish installers to seek professional help when it comes structive, to yourself. The dis were much larger and, consequent- ly, much more difficult to erect and point at the satellite. The techno- P logical advances made in LNC de- 452 and construc brought satellite TV rec THE VARIABLES Getting tuned to those exciting satel. of elim 0 prove lite TV channels s all a matt inating unknown varia for p things we do not know ‘when we unpack the box. QI. Receiver tuning: you may be at ‘any’ frequency between 950 MHz Elektor Electronics 3196and 2,000 MHz, oF at ‘any’ of up to channels @ LNC polarization: this may hor- izontal, vertical, left-hand or right- hand circular 03. LNC band selection: you don't {know in which band you are (Eu- {elsat/Astra/Astra-1D/DBS/Telecom). (Ob Re-modulator tuning: what are ‘you actisally reeeiving on your TV Set? Note? this is not a problem if ‘you are'using.a SCART link between | theteceiverand the TV ‘G5, Dish position, horizontal plane azimuth’): the theoretical range is 180°? .Q6. Dish postion, vertical (‘eleva on’ the theoretical ange is 9° Nowthat boks pretty dawating. $0, let's eliminate the variablesone 6H. The first three points are ensil solved by. consulting satellite chan- Se-Rel overviews which are published = montily by specialized magazines stich as Whai Sateliteyand matching the information given with that pre- sented in the user maitual that came with your receiver. re pre-tuned.te~Astr, Eulelsat and sesame = the-Raning system Elektor Electronics 3196 HITTING THE BEAM Assuming that your receiver and TV are set properly, a big problem still exists in the fact that you can't see the TV sereen from where you are on the roof or balcony. So you need a helper and/or a pair of walkie-talkies. If nei- ther is available, the only solution is to drag the TV screen within sight. Apart from the obvious problems and perils caused by hauling even a portable TV set and a compact sat re- e 0 the roof, this solution will not necessarily give the b sults When you first hit upon the beam transmitted by the satellite the effect fon the TY screen is very sudden and switeh-like, quite unlike the more ‘gradual effect you may be accustomed, to from tuning to terrestrial television stations. The suddenness is caused by {hwo factors: (1) the use of FM modula tion and (} the small pointing angle of the dish. Once the signal strength ex ceeds the so-called FM detection threshold, the TV signal changes sud- denly from almost invisible to crisp and clear, The same with the dish angle: change ita little and you lose the picture almost instantly, although some ‘sparklies’ may appear ifthe sig- nal strength is just above the threshold (7 to. 94B SIN), Uniortunately, the fact that the picture is fuee from sparklies, does not mean that you have hit upon. the right dish position. You may be- ‘come painfully aware of this during heavy rain or snowfall, when sparklies, appear, and i's not Star Trek or the lat cst computer animation on the screen! SIGNAL STRENGTH Although the quality of the received picture is fine for an initial adjustment ‘of the dish, you should not miss out ‘an those few extra dBs which are noc- essary to prevent sparklies under ad- verse weather conditions. The only way to squeeze out these dBs from the 53) Sf Oe AZIMUTH «By sooo system is to judge the signal strength, with a dedicated meter rather than the ‘TV set or monitor, Peaking the dish ad- justments for the highest signal strength as indicated by 2 meter will give you the best possible reception, And some headroom in case of adverse weather conditions. Some satellite TV receivers do offer a signal strength meter function, butt is difficult to use in most cases, being aimed at the pro- fessional installer. Alternatively, the si nal strength may be indicated by an on-screen horizontal bar, accessible through the setup menu. Unfortu- nately that's fancy stuff, and not much ruse either if you can’t see the TV screen from the roof! ALL IN HAND ‘The present instrument consists bast cally of a two-stage high-gain amplifi- er and a signal rectifier. The input of the amplifier is connected to the LNC via a short length of coax cable. The signal strength is indicated on a cassie ‘moving coil meter ‘The circuit diagram of the Satel- lite Finder is shown in Fig.1. Let's first look at the power supply. The ircuit and the LNC are powered by 1 pack of four series-connected 4.5. volt batteries. The 18-V battery volt age is fed to the LNC via switches SI, $2 and choke L1. Most, if not all, currently available LNCs are capa- ble of operating at 18V. Whether that voltage selects horizontal, ver= tical, right-hand or left-hand circu- lar polarization in the LNC is large- ly immaterial because most satellites transmit enough signals in one po- larisation plane for the tester to de- tect and respond to. Do make sure, however, that the LNC receives the desired band at 18V. Note that switch SL must remain closed in the(rare) case of an LNB which is not powered via its coax output. The battery pack voltage also ar- tives at the input of a voltage regu- lator, IC3, which supplies a stable 12 rail for the amplifier blocks, ICL and IC2, and the logarithmic meter driver, TH. Current consumption of the circuit is only about 75 mA. De- pending on its make, the LNC may draw anything between 100mA and, 500 mA. An LED, D2, is used as an ‘on/off indicator on the instrument. ‘The output spectrum of most NCs extends from 800MHz to about 2,000 MHz. The signal levels are generally high, given that most LNCs have a conversion gain in ex cess of 6008. That is still not enough, however, to enable a simple (passive) rectifier to be used in com- bination with a moving-coil meter Hence, more gain is required, for which ICI and IC2 are responsible. Each of the type MAR8 monolithic drop-in amplifiers supplies a solid 15 to 208 of gain. The MARS are cas caded, and operated at their typical supply voltage of about 7 V, which is applied via series resistors R5 and Ro. Note that the supply voltage of ICI is adjustable with pot PI. That is done to enable the gain of the first amplilier tobe reduced within a reas sonable range to prevent overload- ing of the meter circuit It should be noted that the MAR-8, unlike its family members, has an input and output impedance which is higher than 502. The actual value depends on the frequency. Fortu- nately, thatis not a problem here be- cause of the high signal levels, and because the there is plenty of gain to compensate small mismatches. Low noise is not a point, either! More on MAR amplifiers may be found in the inset and in the Refer- tence at the end of this article. The rectifier is formed by a Schot- thy diode, D1. If difficult to obtain, the HP2800 may be replaced by a lower-spec type such as the BATS2. ‘The meter driver is conventional and based on a FET (T1) which en- sures a partly logarithmic meter re sponse. The meter is nulled with pot P2, Absolute readings are not in order, so a scale is not really neces sary on the meter. All we are after is 2 welldefined peak in the meter reading, and that’s exactly what the circuit will give you, provided you are able to point the dish at the satel lite 4* BUILDING UT The Satelite Finder is best built on a small printed circuit board of which the design is shown in Fig. 3. Because the board is not available ready- made through the Readers services, you will have to make it yourself, or have it made. The board 1s double- sided, but not through-plated. The top side of the board acts as a ground plane. I's a wonderful fact that this cir- cuit has no adjustable inductors or esoteric parts to contend with, al- though it operates in the gigahertz frequency range. The only inductor in the circuit is a small choke, Li which consists of three turns of 03. mm dia (30SWG) enamelled copper wire (e.c.w,) through a 3-mm long, ferrite bead ~ see Fig.4. The only problem’ with the MARs is that they are, well, tiny! ‘The MARs are fitted at the under- side of the board. Check the orienta tion: the dot on the device marks the RF input. The two electrolytic capac- itors, C1 and C15, are mounted about 2mm above the board surface to en: able their negative terminals to be soldered to the copper ground plane at the top side of the board For proper screening, the com: pleted board must be fitted in a small metal case. The case is best made from a 30-mm wide strip of tin plate which is bent around the board edges (see photographs). Do not sol- der the seam before you have deter- mined the position of the feedthrough capacitors which are fit ted in one of the side panels. The solder eye of the feedthrough ca- pacitor should be at the outside of the case. The collar is soldered all around to the side panel using a ‘medium-power sodlering iron. De- pending on what you have avail- able, the RF input socket, KI, is ei- ther a BNC" (lange) type or an ‘F socket, Although the latter is cheap- cr, it may be more difficult to get hold of. Ibis also less suitable for fre- {quent connection and disconnection, so we really recommend using a BNC socket. You will also need to make a short coax cable to connect Elektor Electronics 3198the tester to the LINC. This cable has an F plug at the LNC side, and a BNC plug at the side of the tester. The flange of the BNC socket is ei ther soldered or screwed to the tin- plate panel, and the centre pin is red directly on to the RF input pad on the board. If necessary, re- move the PTFE collar around the centre pin with a sharp knife. The completed tin-plate c built into a diecast enclosure (which may also contain the battery pack). The front panel holds the moving. coil meter, the two LEDs and the controls (two pots and two switch es). The controls are connected to the respective feedthrough capaci- tors via short wires. If an external battery pack is used, itis connected to the tester via a low-voltage dec adaptor plug and socket PRACTICAL USE Easy, as far as the tester is con: cerned! Connect the tester to the LNC with the home-made cable. Flick the two switches, the LED should light. The meter may show some indication already. Check that the indication varies if you turn pot PI, Null the meter by turning pot P2 with the LNC switched off ($1), The meter needle will show a marked peak if you hit upon the satelite beam. Once you have found the initial dish position, reduce the 3198 5o gain by turning P2, and then adjust the dish again. In this way, you will be able to obtain the best possible re sults from your outdoor unit, es Reference: ‘Using the MAR-x series of very wide- e integrat circuits (MMIC), by Joseph Cart or Electronics October 1992. band monolithic microwa\ Ey Fig. 4. A close look at the Satelite inden Notice Bi = 2haz Ri2 = 8200 the feedthrough 2 4702 potentiometer pagent, KO potentiometer Capacitors: Gt = 100uF 25V radial (€2.06.616.017 = fescthrough capacitor nF cit.cr2c18 = INF G7 = 10pF ©8,C9,C10 = apF3 (C13 = 100nF C14 = 100 C18 = 224F 25V racial C19 = ape. Inductor: Li = 3 turns 03mm dia. ec. on ‘mm ferte bead ‘Semiconductors: Dt = HP2800 or other RF Schottky diode (2.9. BATE2) D2 = LED 1h = BF245B. IC1,\C2 = MARB (Mini Circuits K1/= antenna socket (with ange) {see tex) 81,82 = onfoft switch BTt = four 4.5V battones Mt = moving coil meter, 1mA ts. 55reMARkable wideband RF amplifiers MAR devices from Mini Circuit Laboratories are remark able because they make AF amplifier design accessible to anyone. MARS are cheap, easy to use and pretty fault: tolerant. These wonderful fourpin devices match 60-2 input and output impeciances without external impedance transformation circuity (which is normally a nightmare to beginners). They are basically silicon bipolar monolithic Is in @ two-transistor Darlington configuration, intemally laid out for ultra-low stray inductance and capacitance. Depending on the exact type, the overall gain of these devices is typically 13 to 33. dB between dc. and about 2GHz (yes!). Some types are optimized for low noise, ‘others, like the MAR8, for very high gain. In the basic circuit for a wideband amplifier based on a MARx device, the RF IN and RF OUT terminals are pro- tected against d.c. by coupling capacitors C1 and C2. Also traditional is the supply decoupling consisting of a tantalum electrolytic capacitor (C4) and a ceramic 100- nF type (C3). The amplifier is powered via a series resistor which determines the operating voltage, as shown by the formula, The operating current for the device used may be found in the datasheets, and will be of the order of 40 to 80 mA. Elektor Electronics 3196somennerg 2001613 6 As RDS Demodulator + Filter TDA7330 RF and Video Manulacterer: SGS-Thomson ‘The TOAT30 is an RDS (Radio Data System) ‘demodulator. The IC includes a 57-kHz switched ‘capacitor input band pass fier, a bitrate clock recovery citcult, DSB demodulator circu, bi- phase PSK decoder, ferential decoding cir Cults, ARI identification and signal quay aut- pts. The data and clock output signals (RODA, ROCL) can be processed further by a suitable microprocessor Absolute maximum ratings: Supply Voltage: v Operating Temperature Range: ~40 0 +85C Features: = High-performance, stable 57-KHe fier = Adjustmentfee fier, ‘no extemal components = RDS demodulation \wthout external components = ARI output ($0 indication) = RDS signal quality output = 4.322-Me quartz crystal oscilator (8:8684-MH2 optional) = Low-naise mixed bipolar/CMOS technology Block Diagram PSILELKIrONRS ELECT WEES DATASHEET _ 03/96 Pin Connection — er Nonarering conga nt seats te COUT [ie eu 50 [oon T_T apa ooo) — Test ava ated) —— 14 estaba ot used) SCOUT Osea cunt (OSCW | Ocedaor opt 157] stow ot ag) ‘OGL [ROS eek oot ROOK [AOS aaa put (QUAL | Sal uty ont igh = 206) Jounin aR | gh = ROS + AR or A oy ow = RDS (ny alin when ea peer) Voc [Sunt voane 7 [12 [Test ota i ash 18 | FSEL_—|Femancy seer "este in (open » rari lsat a Wt = Test ose). 25 | POR ese wp resin Gta). a] RF Amplifiers MAR-x RF and Video RF Amplifiers Manuacturer:to-Crouts CAMini-Circuits ‘The MAR-cascadable amplifier series isa family ‘of silicon bipolar monolithic integrated circuits fabricated with nitride sef-aignment, fon impan- tation for precise control of doping and passiv sation to achieve high retablty These devices, Priced from below one collar in volume quant, exhibit excelent unit-to-unit uniformity and are ideally suited as 50-ohm ampiter ulding blocks. The MAR devices are simple to use if the proper boaré layout is used, along with proper grounding, biasing and stops to minimize parasites. Mini Circuits, BO, Box 350166, Brooklyn, New York 11235-003. Applications: = Low-power transmitter © Boost signal for improved detector efficiency = Multistage ampiir chain ‘= Butfer ampli for oscilators = bsolator Features: = Easy to use, 60-0 inpuvoutput Smooth response aver the band, no extemal resonances = Easy for printed-circut designs, one input and one output ~ Can operate as low as 6 Vdc Low impedance, less susceptible to Eh BILEKTOLR ELECTRONICS DATASHEET _ 03/96 Case: ww05 [Fron Gordeae ELEKTOR RDS Di Fil Taras0 BILE TOL ELECTRONIC ELECTRONICS RF and Video DATASHEET _03/96 RF and Video DATASHEET cane a ees og = 5M Ta ene Output (148) Input ‘Symbol Parameter (oo ane) Si T0000 18s t WS +10 Noo Supply Voltage 38] +5] iia. +5 2) 2 + 4 | +8 as [ a +5 . Ts #5 21a 3 [Sn Sera nce d= Se FF leat {= a D208" ea {= 57 ke (RDS + ARI) A [lester Pd ope) Wav] a¥ [si2v] 98 a vy z = Om nara csc-renweecso:, | 8 sans i) i el Fose | Oscilato Frequency wh 4 332-4dHe-Ouare) | iwi) =) 60" 2a fay = es 0664) ost wna) — [| 0 | emir wana) 80 =o | 0 RDS Dsecon Seerey — me [arpa fei on Sty waar | 2 | 2) oot OS ek ne 3 == bm at VIE apa OY Voge Dah Pat fr ROS t = Tho pase none dtr as APh = -BE + O41 + a], whe ki Be onbanp hae aiteene ae | taney Goeuton of Re= Wee ~Val M2) Yee Sua tae spite to Re IM Ye age ate sip aaa of UE uses was cent fa wee The esi of Re oben 7 eas = YE RELL 0 | Ri 5 = ase] 35 [<0] “Hc wiapato mo) er | 3198 Elektor ElectronicsWe surround-sound subwoofer Most surround- sound installations use loudspeaker boxes of modest dimensions so as to avoid making them too obtrusive for the usual living room. The consequence of this is a limited bass response, whereas especially the low frequencies can pro- vide impressive effects with a good surround-sound sys- tem. To counter this drawback, the bass response can be enhanced with the subwoofer described in this three-part article. | Drive unit 300 mm (8 in), e.g. Monacor (SPH-300TC); KEF: Rado Shack (40-1024); Parts Express (295-240) Dimensions of box (660% 406x420 mm (26%16%16 "ie in) incl. legs | Votume of box about 65 Inet Type of box bass reflex Nominal impedance 8 Q.per channel Efficiency 88 dB WT art Frequency range 45-105 Hz Surround sound, the popular craze of the past few years, c vide an impress sound and pi is used in conjunction with a Tv set. Good-qual- ity spatial sound is provided by a ‘number of loudspeakers (usually five) Tocated in front of and behind the lis- tener(s) Five loudspeakers present a problem, of course, in that they take Lup a lot of space in the ave room. To keep the space them toa minimum, the loudspeakers are often fairly small. Moreover, in feconomy-price systems, cost is impor tant, too, and this also tends to keep the boxes small, Unfortunately, small loudspeaker boxes are detrimental to good bass re- production. On the surface, this may not seem such a terrible thing in an audio-visual system until itis realized that the low frequencies contain spa- tial information. Moreover, we per- Technical data Loading ‘max 250 W per channel 62 part 1 ceive low frequencies not only via our ears, but also through our entire body and this causes good low-frequency reproduction to give that adcled feel ing of reality to the sound, All this makes it clear that the importance of Jow frequencies must not be underes- timated. ‘The reproduction of low frequen- cies requires the displacement of large volumes of ait This in turn means that a large low-frequency drive unit (woofer) should be used. But such a lunit must be contained in a large en- closure to enable it to reproduce low frequencies effectively. And this is ‘where the crux of the matter is: most living rooms just do not have the space for such a large box. Tn the subwoofer described in this article an attempt has been made to find a compromise between the con- tradictory requirements just outlined, It uses a large (300 mm) drive unit 3/96housed in a modestly-sized enclosure fof 65 1 The enclosure is designed in the form ofa side table with the drive unit fitted between the legs so a to ‘make it (virtually) invisible. The vole lume of the enclosure i not really large enough for very low frequency reproduction, but a solution for this ill be published in next month's n= This consists of an active cor~ network and associated am- that bring the -3 dB point Jown to 20 Hz. This article describes ive version of the subwoofer can be used without any dif- with existing apparatus. Is fre- ey range extends from about tefficiency of the unit pro- xi match with the (smaller) peakers. paced on a 300 mn (8 in} LB0NTC drive unit, but such as KEE Radio Shack Pants Express 295240) performance as well VTC isa relatively imex- with a fait lage magnet a volume of around (2 make it suitable for use enclosure. idspeaker is to be used o system, it should have for both channels. This ther two drive units or 9 th dual voice col should channels via a suitable fil- d with the simulation pro- calc, and aims to arrive at a mise between a (relatively) the pipe tical resonator) tuned to 23 Hz ‘overall frequency response is own in Fig. 1. The -3 dB point is at 45 Hz, which, considering the small ‘box volume, is pretty good. The -3 dB point is low enough to allow the sub- ‘woofer to be used as a passive unit with most existing systems. Saeieaee 396 Tue FILTER Since the design aims at keeping, the casts as low as feasible, the (passive) filter has been kept as simple as possible, which, in the case of a suib- woofer, is nat as easy as it may seem, “The impedance characteristic of the drive unit is shown in Fig. 2. The two ‘voice evils are connected in parallel to ‘obiain a reliable curve (which means that for each coil double the imped- ance value must be taken). The curve ‘shows two peaks, The lower one at about 10 Hr results from the bass 1e- flex alignment (which, by the way, is exactly in Tine with the 23 Hz res- conator), The second peak, just above 51) Hz, is caused by the resonance fre- {quency of the drive unitin the box. Normally, filtering of a subwoofer starts at around 100 Hz or Slightly lower to ensure good matching with the standard slereo Joudspeakers. A passive filter, however, has the drawback that it func tions properly only if it is terminated into a pure resistance. Ik the cut-off point were chosen at 100 He, the 52 Hz peak would create a problem: the resulting overall curve of a theoretically computed filter would not be usable. To solve this problem, the impedance curve of the drive unit has to be corrected. This is often ef- fected by connecting parallel across its terminals (for each channel) an RLC network with the same resonant fre- quency. Unfortunately, at such tow fre- quencies, the values of the necessary inductors ancl capacitors are such that they result in physically large (and ex- pensive) components. ‘The solution in the present design consists simply of shunting the ‘voice col with a resistor This does nol totally ‘eradicate the peaks, butoon 13 Th adapted as appropriate} The box consists of six rec angular sheets and a stift- ening crosspiece which are sy firmly fixed together w suitable heavy-duty glue At one side are the apertures for the drive unit and acoustical resonator, e resonator consists of a 365 mm long piece of 80min dia, PVC pipe aval: Fig. 3. The filter has: able from a builders mer- flattens them sutf- Been iii uncnaies | MeSBtOPRS eorets The four tonana Sito pros -hedope sockets for connecting the Reraing he o PERRMMORCE ROARED ct a in om oF he i io CRM oe ofthe ae parce fcuency. re. (eet aeae Raters eign sponse is close to 105 Hz requirements, To keep the number of components small, the filter isa second-order type consisting of conductor L, and capac itor C, (see Fig. 3). The resistor in se- ries with the capacitor damps the LC circuit to some extent, The effect of the filter is shown mn Fig. &. Although the high cutoff point is about 105 Fz, the response will enstre a good match to most small loudspeakers, BUILDING THE BOX to rest on four 50 mm high legs with the drive unit fit- ted at the bottom facing the floor of Alter the glae has dried thorough: ly and the material has been sand-pa- ppeted, the bos can be given a final coat to individual taste The box is half filled (up to the cress piece) with suitable loudspeaker wadding, but take cave that the open- ing of the pipe remains reasonably free of it The filter components are available from a specialist audio/hifi retailer or Note that some retailers stock general- purpose filter boards. Serew the completed filter into the box and wire it upsas shown, Take care not to interchange the plus and minus connections to the two channels. The cables to the drive unit must be termi: nated into cable clips to avoud the ne cessity of having to solder i the drive unit terminals Place the resonator in postion, make the connections to the drive unit make sure that the connections to the + terminals match, otherwise the drive unit does not work). Finally, place a strip of draught-exctuding tape Under the rim of the drive unit and screw the unit fo the bos. Some constructors (or their wives) may find it aesthetically pleasing to place a sheet of glass, marble or simi Jar material on top of the box ta sve it the look of a side table. The passive subwoofer is then ready for use. It may be connected in parallel with the existing stereo speak- ers, It will work most satisfactorily when its efficiency of about 888 W-l nt corresponds roughly to that of the existing loudspeakers and itis placed in close proximity to these Note that if you want the active ver~ sion, which will be described next month, you do not need the passive filter; the box remains the same The prototype bos is made from a goad electronics shop. The inductor ean 28 mm thick mediumedensity chip- isa 15 mF type with a.56 mm ferrite board (MDF), but, as in some eases core, preferably an H1QSé from IT. The may not be poss: Capacitor is bipolar type be to obiain this, lgeliThe frequency |) with smooth terminals 22 mm thick chip- response of the foud- The filter compo- board niay be Speaker and fitter nents may be glued to a used (note that the sear cot ‘small sheet of wood,chip- dimensions in Fig. | board, or prototyping board ‘Speakers. - top Frequency = asin Eektor Electronics 3196 4Drive unit revisited During the design of the loudspeaker, & thor- ough search was made for a 300 mm dive Unit at a reasonable price (to keep total costs down to not more than 80-80). Of course, such an economy price unit cannot be expect. ed to be perfect. And, indeed, inthe testing of the SPH-S00TC unt, it appeared that the pare- meters stated by the manuiacturer did. not agree with our own measurements. Fortunately, the deviations were beneficial to the box dimensions. Also, there was a kind of rustling noise at large cone movements. This was suspected in the frst instance to be caused bya loose cone or air leak, buta sec- ‘ond example exhibited execty the same noise A detailed investigation showed thatthe cust hhood inthe cone (the convex cap that closes the upper side of the cone) was the culprit ts ‘ateral is fairy sof, so thet at large cone ‘movements it begins to vibrate at fs own (higher) frequency and thus causes the rusting noise. This deficiency is easily negat- 0 by spraying the dust hood a couple of times with 2 sultable plastic spray or applying 2 few layers of a sultable cone impregnator This makes the cap more gid so that isnot set into vibration at large cone movements. The Parts Express unit appears 10 be rather ‘more rugged than the Radio Shack and is rather cheaper Elektor ElectronicsLetters Component availability Dear Editor recently decided to build a decent nif amaltier ‘and chose 5 design by Mr Giesberis (September 1995). have several problems which need to be resolved wiih. | foal, should not have arisen in ‘a magazine for he amateur ‘constructor who wauld have a limited number of sources for ‘components 4. 10000 uF; 50 V capacitors. ‘The Poe layout requires a 30 rm dia. component (a 95 mma. type wl foul other components on the ree, But is ‘ore commonly avalaole. ‘ter much searching, | ound twee manutactuers: El DDublier and Rulbycon. The cis tnoutors for Elna seem to stock only the 35 rm dla, siz (atthough Eina do eth 80 mm ‘and 35 mm); the Dubie iter requires @ £280 minum otdar. The Rubycon items are stocked by C&CD of Lane End, Bucks (01494 882848), part ‘ngs SOUSP 1000040 (0 mm long and 5.02 A nipple) ‘end SOMXR f0000MC50. (0 mm ong and 4.08 A ti ple). Nols tha the professional Maplin catalogue advertises 3 30 mm iter, but this is typo: {should be 35 mm, 2, i081 output devices, These ‘wansistors are single source ‘rom Tostiba and are not com: ‘monly stocked by their tran- chised distibutos. A norvran chised distributor, Stentite (0124 643 6399, contact Jenny Skinner) isthe only stocks in the UK according to Toshiba, 3. 96 resistors. are some of ‘he designs by Mr Giesberts so cntcal that ony esistons tram the E96 range can be used? May | rete to Unesey-Hooe, Dougies Sei, Bengt Oisen, of al, whse stalo-otthe-at ni _mpifier design (EWM) seems only fo use the E24 range of resistors 4. Fischer Sks¢ heat sink The parts list stats that the thermal characteristic ofthe heat sink required is 0.8 KW" ‘The phoiogranh in the Sept. 95 article shows @ heat sink where the height i Pat the width of ‘about 75 mm. inthe Fischer catalogue No, 12, page A2t the graph clearly shows that fr 8 stock lenath of 75 mm the thermal characterietc is about 1.18 KW. Which is corect, OS KW or 15 KW? 5, Serow terminals. Where, toh where, does one gel the Fee mounted screw terminals 00 ‘he main ampiier Pca tor the supply and speaker leads in the us? 6, Single or twin power supply? It's surprising that ‘consideration has not been ivan fo the power supply con Situction and ratings with a sutably rated ree (twice the current and twice the number of reservoir capacitors?) fora stereo unt 7. Terminal pin hole size. ‘What size are the holes for the terminal pine which are used. for connecting wires to ard {vom the rot? The holes used are too latge for commonly available [Ver0} pins which are ‘nommally 1.04 mm, Where cant ‘soutee the correct size push pin? 8, Tone controls | recently enquired abul fone contro ci ls for use with his ampifier and received the curt reply that tone contols are ‘outs though they re some fashion accessory and are no used in rear hifi ampiiers. First, (an no doubt many other readers) stil have older recor: Ings which nged judicious adjustment af the bass d treble response to compensate for the p00 quality recording, Secondly, Ido not have a spe Clalyfeic-out listening room but a small ving toom in which anything larger than my nsF101 loudspeaker would be out of place. Thirdly, fam in my lato fies and in common with ‘everyone (younger hich purists, please note), my hearing Frequency range has reduced as | have got older, so do sometimes need to "weak’ the treble a tio bi. Fourthly, the statement that tone controls are not part ofthe hii sane is ‘based' on the assumption that the recording engineer has cor ried out histher task to periec- 4® ion and that the equipment set in the stualo is beyond reprosch Last. but nol feast. object strongly to some nif fol lower of fasion tating me now | snauié listen fo mymosty classical and some As}music! By tho way, would be a fal lacy to assume hat reduced hearing frequency range ‘equates to recuced sensitty to distortion ‘Teel that information on the sourcing ¢! suitable compo: ens in the Ux s sadly lacking. “The assumption by the cont entals thatthe hobbyist in the Luchas acoass to instant know 2000 of wher to get ary com- ponent dreamt up by the {designer seoms not fa stort of arrogance. el that Elektr Elec- Ironies should set the standsra here by issuing quidetines to Continutors to use compo- nis that are cornenonly va able worldwide, The magazine ‘ould suggest that destaners ‘only choose components from ‘list Eloktor Eleetronice lined distibutars that are ‘avaliable forthe hobbyist both ‘nthe Continent, inthe uk anes inthe usa and that specialist components are source ident fied by the contrioutr. {Alan Beaman, Bracknell 1.42 Many electronics rota fs, cluding our own adver 7, are provided wih 2 pre view: incl. parts ist ofthe mag- azine content about three weoks before he publishing ‘ata ofthe magazine fo enable them 10 abla stock of pats used, The preview states the manufacturer or other source ot Any less common device. (Mr Giesberts comments as fl- ‘ows: 9. "When metched transistors 70 used, 195 olerance compo: rents are necessan, olherwise the use of matched compo- ens woul make no sense. ‘The £24 ranges, theoretical, 8 28sfolorance series. In prac tice, howaver, E96 wil be sup- led ever! the same vale is found in ho E12 serios. Is live that 1% resistors may be found in the £24 sens: Unfor tunatey, kits onen contain 5% types, obviously since these _are cheaper It's alvays poss! ble to use E24 hypes even if £96 types are specified a8 ong 28 one is propared to accapt a ‘sight degradation of quay’ 4a Fischer catalogues up Yo 1998, a wrong graph was ferinted with me data fore, #85 heatsink From the 1994 Catalogues onwards, is stated that a x85 heat sink is 75 mm high and has a thermal resist- ‘ance of about 1.2K W"!. The Ry value indeated in our parts list should nave corraspondted to tha value. In practice, a dou: led tnermal resistance dos no ‘nar, because the design is based on a woreicase stv tion. Nevertheless, tis recom ‘mended to use a heat sink that ‘meets ine specified thermal resistance raing 5. "A possible solution, beter than the orginal one, isto ter- ‘minate tne cable into a.cable eye and secure it wih a screw rect atthe copper side* 6. "To prevent eerhing foops (Which cause hum, 3 mono Supply's recorimendad. For ‘anyone who is after qualty, me ‘possible cross-ofecis that a ‘stereo supply may have on the ‘we channels the wiring (a ‘tar configuration i not idea!) and particulary te radiation of tne supply abies ara good reasons to make use of mono supply. Infact is batter to build wo mono ampilfiers 7, "is beter to make use of soldering eyelets, which nor ‘mally have a wedge” Finaly, we apologize forthe ‘curt reply you go! to your query ‘on tone controls. We hily share ‘your leelngs that in many ea806 fone controls are not only raquired, but are essential Nevertheless, some amplifiers We publish have no tone con: trol and these are aimed espe- cially af readers who de not fee! tone controls necessary, (Esitor) Electronic magazines Doar Editor—By al moans ‘make your publication avail- abioolectronicaly for he ‘no‘ak wearing minority (Let Jers! ~ FR Fatt, Dac 1998). ELEKTOR ELECTRONICS MARCH 1998