Silicon ChipOctober 1998 - Silicon Chip Online SILICON CHIP
  1. Outer Front Cover
  2. Contents
  3. Publisher's Letter: Millennium bug could lead to huge legal bills
  4. Book Store
  5. Order Form
  6. Product Showcase
  7. Market Centre

This is only a preview of the October 1998 issue of Silicon Chip.

You can view 11 of the 96 pages in the full issue, including the advertisments.

For full access, purchase the issue for $10.00 or subscribe for access to the latest issues.

Purchase a printed copy of this issue for $10.00.
CPU Upgrades: Are They Worth It? SILICON CHIP OCTOBER 1998 $ 50* 5 ISSN 1030-2662 10 NZ $ 6 50 INCL GST PRINT POST APPROVED - PP255003/01272 9 771030 266001 www.siliconchip.com.au PROJECTS TO BUILD - SERVICING - COMPUTERS - VINTAGE RADIO - RADIO CONTROL PLUS: “Connect & Forget” 12V Battery Charger Lab Quality AC Millivoltmeter Replacing Flash Batteries Guitar Limiter October 1998  1 SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.dse.com.au Contents Vol.11, No.10; October 1998 FEATURES   4  CPU Upgrades & Overclocking What’s right for your computer? – by Bob Dyball 16  Electromagnetic Compatibility Testing; Pt.3 Immunity to interference – by Marque Crozman 80  Hifi Review: Dual CS505-4 Turntable Yes, record turntables are still around – by Leo Simpson CPU Upgrades And Overclocking – Page 4. PROJECTS TO BUILD 24  Lab Quality AC Millivoltmeter, Pt.1 New design measures down to below 1µV – by John Clarke 32  PC-Controlled Stress-O-Meter Build it and keep your stress levels under control – by Rick Walters 60  Flash Attack! Adding an external battery pack to your flashgun – by Julian Edgar 66  Versatile Electronic Guitar Limiter Lab Quality AC Millivoltmeter – Page 24 Adds interesting special effects and stops overload – by John Clarke 74  Connect And Forget 12V Battery Charger There’s no danger of overcharging your battery – by Rick Walters SPECIAL COLUMNS 53  Serviceman’s Log Comparing the old and the new – by the TV Serviceman 82  Radio Control The art of the F3B glider – by Bob Young PC-Controlled Stress-O-Meter – Page 32 87  Vintage Radio A short history of spy radios in WW2; Pt.2 – by Rodney Champness DEPARTMENTS   2  Publisher’s Letter 21  Order Form 22 Mailbag 43  Circuit Notebook 58  Product Showcase 91  Ask Silicon Chip 93  Notes & Errata 94 Market Centre 96  Advertising Index Versatile Electronic Guitar Limiter – Page 66 October 1998  1 PUBLISHER'S LETTER Publisher & Editor-in-Chief Leo Simpson, B.Bus., FAICD Production Manager Greg Swain, B.Sc.(Hons.) Technical Staff John Clarke, B.E.(Elec.) Robert Flynn Ross Tester Rick Walters Reader Services Ann Jenkinson Advertising Manager Brendon Sheridan Phone (03) 9720 9198 Mobile 0416 009 217 Regular Contributors Brendan Akhurst Rodney Champness Garry Cratt, VK2YBX Julian Edgar, Dip.T.(Sec.), B.Ed Mike Sheriff, B.Sc, VK2YFK Philip Watson, MIREE, VK2ZPW Bob Young SILICON CHIP is published 12 times a year by Silicon Chip Publications Pty Ltd. A.C.N. 003 205 490. All material copyright ©. No part of this publication may be reproduced without the written consent of the publisher. Printing: Macquarie Print, Dubbo, NSW. Distribution: Network Distribution Company. Subscription rates: $59 per year in Australia. For overseas rates, see the subscription page in this issue. Editorial & advertising offices: Unit 8, 101 Darley St, Mona Vale, NSW 2103. Postal address: PO Box 139, Collaroy Beach, NSW 2097. Phone (02) 9979 5644. Fax (02) 9979 6503. ISSN 1030-2662 and maximum * Recommended price only. 2  Silicon Chip Millennium bug could lead to huge legal bills The saga of the Millennium bug, otherwise known as the Y2K phenomenon, continues to develop. For those that have not been awake for the last two years, the Millennium bug refers to the problem of computer code which describes the year by the last two digits, as in “98” for 1998. When 2000 arrives, computer code that still uses this practice will be unable to distinguish between 1900 and 2000 and so the computer will inevitably crash, planes will fall out of the sky, electricity distribution will stop and so on. At least, that’s what the pundits are forecast­ing. At the present time, there is vast rewriting of old pro­grams in government and large organisations like banks and in­surance companies. Small business organisations, on the other hand, seem to be taking a “She’ll be right” attitude. Well, they have had plenty of warning to check all their systems and make sure that everything works when the year clicks over at the end of 1999. Even at SILICON CHIP we have had to bite the bullet and update our accounting software which was not Year-2000 compat­ible. All manufacturing and importing businesses should also ensure that their suppliers won’t be affected by the bug but that could be a tall order where the supplier is overseas. Apparently though, this same Y2K bug can be a problem in equipment which has embedded microprocessors. This applies to all sorts of equipment ranging from medical equipment in hospitals, security and fire protection equipment, manufacturing equipment and even things like chart recorders and some printers. In some situations, this could be really crucial to the functioning of the organisation and could cost a huge amount to rectify if discovered at the last moment. Mind you, it beats me how equipment (and software) with this defect was sold in the first place. Any company or organisa­tion buying plant and equipment expects it to function for many years without any need for major modifications. If it falls over in the year 2000 it occurs to me that many companies supplying this equipment could be liable to very expensive legal action. After all, all goods sold in Australia are supposed to be of “merchantable quality” which is legalese for having no signifi­cant defects. Something that will fail to operate at the end of 1999 clearly does have a defect. Lawyers will have a field day. So if your company or organisation hasn’t checked out its equipment with “embedded microprocessors”, then you had better get them on the job. Even company directors are in the firing line on this one. If you are a director and your company suffers big losses because of a problem with the Y2K bug in its equip­ment, you will be liable to be sued by shareholders. And if you are a supplier of this “defective” equipment, heaven help you. There is any number of legal statutes that lawyers will be able to use - contract law, the Trade Practices Act, negligence - you name it. At the very least, you need to inform all your customers, past and present, that the product they purchased from you or your company will fail to operate after 31st December, 1999. Leo Simpson M croGram Computers 100Mbps Network Starter Kit Web-Based Training from $9.95 per month* This kit comes with all the A number of courses are “Microsoft Learn about Microsoft Office, Word, Access, hardware components Certified Professional - Approved Study Excel, Windows 95, FrontPage, C++, HTML, required to build a 100Mbps Guides” Internet Explorer, Windows NT and more! network for two PC's as well as a comprehensive installaOver 160 courses on offer *Full details at www.tol.com.au tion manual. All software is part of Win 95/98/NT. The 100Mbps Network Starter Mounts on back plane of a computer but does not plug Even Pentium motherboards are not immune to the Kit provides the most cost-effective solution for users into a slot, it only connects to Year 2000 bug! The Year 2000 BIOS who desire fast throughput at the cost of traditional the power supply. No separate Card solves the problem of progres10Mbps. The kit includes one 4 port 100Mbps Fast case & power supply means sion from 1999 to 2000 as well as Ethernet hub, two 10/100Mbps PCI Fast Ethernet reduced costs, plus everything 21st century leap years. It is an 8-bit adapters, two 5 metre Cat. 5 network cables and softis kept neat & tidy inside the card which provides year 2000 support for motherware drivers for the adapters. computer. boards with a BIOS which only stores the year with two Cat. 11900 100Mbps Network Starter Kit $339 Internet Access Server Internet for everyone! Give all stations on your network simultaneous access to the Internet through this access server. Hardware based firewall ensures your security, while dial on demand minimizes your connect time. It has a built-in DHCP server & includes software to provide clients with their own email address. This pocket-sized Internet-Sharing device provides one communication port (DB25) for Internet access and one RJ-45 port for connection to your 10Base-T Ethernet network. Supports easy Internet connection to your Internet Service Provider via modem or ISDN. Cat. 11294 Cat. 11287 Ethernet Hub Card 5 Port UTP 100Mbps Ethernet Hub Card 5 Port UTP 10Mbps $259 $99 digits. i.e. 97 instead of 1997. Cat. 3359 Year 2000 BIOS Card $129 10Mbps Ethernet 5 Port Hub & LAN Card Video Conferencing Kit Internal PCI Plug & Play 5 Port hub and LAN card does not require external power supply and is a cost effective solution for SOHO users. One port can be used as an uplink port for easy expansion, or used for hub connectivity at the server. A high performance PCI full-motion video/still image capture solution for video conferencing on the net! The kit includes video capture card, CCD camera & VDONet’s video conference softCat. 11295 Ethernet Hub & LAN Card 5 Port UTP 10Mb $109 ware. Ideal for applications such as Video Mail, Video Conferencing Serial Cards We have a large range of serial cards providing either 1, or Full-Motion Video Capture to AVI file format. Video Conferencing Kit $299 2, 4 or 8 ports. Our most popular and versatile single, Cat. No. 3356 dual & four port cards feature high speed 16550 Watch-Dog Timer Cat. 10100 Internet Access Server $499 UARTS, COM 1 to 8 and IRQ 3 to 15. By adding a timing reset instruction to the outer loop of Cat. No. 2297 1 Port RS232 16550 COM1-8, IRQ 3-15 $80 your program, this card will apply a Cat. No. 2239 2 Port RS232 16550 COM 1-8 IRQ 3-15 $99 10/100Mbps 3 Parallel Port Print Server hardware reset to the computer in the Up to three printers can be connected simultaneously Cat. No. 2326 4 Port RS232 16550 COM 1-8 IRQ 3-15 $295 event of a lock up. Utility software The dual port card is now available with 16650 UART with this three parallel port included for DOS, Windows 3.1, Win 10/100Mbps print server. chips with 32 byte FIFO buffers. 95, Windows NT, OS/2 & UNIX. Cat. No. 2333 Two Port 16650 Serial Card $159 Designed with auto-sensCat. No. 17044 WatchDog Timer ing dual-speed capability it Plug & Play PCI models also available. Card $139 auto detects the speed of Cat. No. 17050 WatchDog Timer II Card $349 the network, i.e. 10Mbps or 100Mbps. Built-in TCP/IP Hard Disk Drive Duplicators and IPX protocol support for Windows 95, Windows These hard disk drive duplicators offer a low cost, high Omni-Directional Laser Scanner An affordable, vertically mounted, NT, NetWare and UNIX is provided along with built-in performance solution whether Omni-Directional laser scanner, web management capability and flash memory. DHCP you want high-volume 1 master to 8 drive copying or quick, low volwhich is ideally suited to reading bar server support automatically assigns IP address. ume, 1 master to 2 drive copying. coded products at supermarket Cat. 11293 10/100Mbps 3 Parallel Port Print Server $519 Features include: checkouts. Performance is higher Cat. 11288 10Mbps 1 Parallel Port Print Server $269 • FAT32 compatible than the “Name Brands” with a 24 10/100Mbps Ethernet Cards • Track by track, file by file, whole or partial drive scan line pattern (competitors’ prodAuto sensing either 10Mbs or 100Mbs operation, this copying ucts are 20) & 2,400 scans/sec (competitors’: 2000 PnP PCI Ethernet card uses the Bus Master architecture • Accepts different geometry drives including 2.5” and scans/sec). The depth of field is 300mm & it is availto maximise throughput. 3.5” drives able in either KB wedge or serial models. Cat. 11282 Ethernet Card PCI UTP/STP 10/100Mbps $59 • Copy Win 95 / 98 operating systems in minutes Cat. No. 8521 Bar Code Laser Omni-Direct. KB Wedge $2119 Cat. 11271 Cat. 11272 Ethernet Card PCI BNC UTP/STP Ethernet Card ISA BNC/UTP PnP Jmp $39 $39 100Mbps Ethernet 5 Port Hub Card Cat. No. 6426 Cat. No. 6427 Hard Drive Duplicator Two Drives $2899 Hard Drive Duplicator Eight Drives $6499 Year 2000 BIOS Card Cat. No. 8573 Bar Code Laser Omni-Direct. Serial E & OE All prices include sales tax $2119 MICROGRAM 1098 Come and visit our online catalogue & shop at www.mgram.com.au Phone: (02) 4389 8444 Dealer Enquiries Welcome sales<at>mgram.com.au info<at>mgram.com.au Australia-Wide Express Courier (To 3kg) $10 We welcome Bankcard Mastercard VISA Amex Unit 1, 14 Bon Mace Close, Berkeley Vale NSW 2261 FreeFax 1 800 625 777 Vamtest Pty Ltd trading as MicroGram Computers ACN 003 062 100 Fax: (02) 4389 8388 Web site: www.mgram.com.au FreeFax 1 800 625 777 COMPUTERS: Extending the life of old machines CPU upgrades & overclocking Do current applications and games run like a slug on your PC? If you can’t afford a new machine, consider upgrading the processor and adding some more RAM to boost its performance. By BOB DYBALL If you find that the old grey mare, er PC, is not what she used to be, then maybe it’s time to give it a brain transplant. Although many people don’t realise it, it’s quite easy to upgrade the CPU on most PCs, either to a newer, faster version or by installing an overdrive chip. It’s also sometimes possible to sneak some performance gains by 4  Silicon Chip “overclocking” (ie, running the chip faster than its specified rating). To use a car analogy, a new CPU is akin to fitting a bigger engine, while adding an overdrive chip is akin to strapping on a turbocharger. Over­ clocking is analogous to revving a car engine beyond the redline to extract that extra ounce of performance. And here a word of caution. A CPU transplant alone won’t turn your old 486 into a speed demon. Indeed, the performance boost will probably be far less dramatic than you might expect. That’s because older machines come with all sorts of speed bottle­ necks, including slow hard disc drives, slow graphics cards and slow support chips on the motherboard. It all depends what you are doing. If you are running pro­cessor-intensive applications, then it may be worthwhile spending a couple of hundred dollars to keep your current system going a bit longer. PC/XT - 8088 If you’re still running a PC/XT based on the 8088 proces­sor, then it really is time to get a new machine. This type of machine is really only suitable for running a few basic DOS applications (eg, a DOS-based word processor). The original IBM PC circa 1981 was built around an Intel 8088 CPU running a clock speed of 4.77MHz. The later model IBM PC/XT retained the 8088 CPU, though this ran at 8MHz. Although the 8088 is a 16-bit CPU, it only has 8-bit “data paths”. By contrast, the Intel 8086, on which the 8088 was based, had full 16-bit data paths and was used in some later competing clones with a 30-35% speed improvement, for the same clock speed. After the success of the 8088, some companies tried to introduce compatible CPUs, such as NEC with their V20. In most cases, including the NEC V20, these were not actually clones but “work-alikes”. Engineers in “clean rooms” designed CPUs to do the same job as the 8088 but never having seen the original, were able to say that the CPU was their own work and not the result of direct copying. The V20 was a more efficient chip than the 8088. Indeed, a V20-based system typically outperformed an Intel 8088-based system by 10-30% at the same clock speed. Its performance reign was brief, however. Intel soon countered with the 80286 and the age of the IBM PC/AT had begun. Upgrade Possibilities: this comes under the “why would you both­er?” category. Theoretically, it’s possible to substitute a V20 for an 8088 but obtaining a V20 chip could prove difficult. There’s not much else you can do apart from adding an 8087 maths co-processor (more on these later) but you would really be better off buying a secondhand 386 or 486. The PC/AT - 80286 Intel’s 80286 was the basis of the IBM PC/AT and the first “AT” compatibles. This chip was eventually replaced by the 80386 and companies like AMD soon released competing CPUs to sell against Intel’s 80386SX and 80386DX chips. The 80386SX was a cheaper, smaller version of the 80386DX. It ran the same software as the DX chip but was able to use the cheaper 80286 support chips. Upgrade Possibilities: there’s not too much you can do here, apart from playing with the clock speed. Again, This photo shows a typical 486 “multi-media” machine from the mid 1990s. Such machines are now struggling to cope with the demands of modern operating systems and software but the correct hardware upgrades can extend their useful life in some cases. it’s really not worth playing with an old 286 machine. 80386SX & 80386DX The standard assembler instructions (opcodes, or low level internal programming code) in the 8086, 80286 and 80386 CPUs only included integer maths. This meant that the processors had to do trigonometric, scientific notation and floating point calculations (and even very long integer mathematics) the hard way, using lots of additions, bit shifts and so on. Early in the piece, however, Intel had designed an 8087 chip, called a Numeric Co-processor (or “Co-Pro” for short), to do these calculations. An expensive optional extra, it was nor­mally purchased only by those who had lots of money or a boss who wanted budget spreadsheets completed in a day instead of taking a week! The later 80286 also had it’s own optional Co-Pro, the 80287. Similarly, the 80386SX had the optional 80387SX, while the 80386DX needed the optional 80387DX chip. This is where the 80486DX was such an improvement - it included the CPU and the Co-Pro (or Floating Point Unit) all in one chip. Intel stopped at the 80386DX/33 when it released its 80486DX. Not so AMD, who proceeded to release an 80386DX/40 CPU. This device was cheap compared with the Intel’s new 80486DX/25 and although not quite on par with it maths wise, became a popular chip. Upgrade Possibilities: the AMD 386DX40 is a logical choice here, though often soldered onto the mother­ board. See if you can sal­vage one from an old motherboard, as many people have long since moved away from the 386. Don’t forget to set the jumpers on the motherboard for the new clock speed. By the way, you can make a very cheap and effective print server out of a 386DX40 machine with 4Mb of RAM, an old 100Mb hard disk and a cheap network card. By running Windows for Work­groups 3.11, you can network it to a modern machine running Windows 95/98 or Windows NT. 80486SX & 80486DX Originally released with a 25MHz clock, the Intel 80486DX/25 had quite a performance edge over the 80386DX/33. Initially, the new chip was quite expensive and to counter AMD, the company also released the much cheaper 80486SX. This was essentially an 80486DX without the FPU. Many motherboards at this time came with an extra socket adjacent to the processor and this was intended for an 80487DX co-processor (to go with the 80486SX). Further development saw the release of the 80486DX/33, then the little October 1998  5 Upgrade Processors For Your 486 Machine By Greg Swain Don’t expect an upgrade processor to transform your old 486 clunker. As explained in the main article, there are just too many speed restraints and bottlenecks in a 486 for that to hap­pen. Indeed, depending on the machine you already have, the performance increase may only be marginal at best. So is an upgrade processor worth the bother? Well, that depends on the specifications of your current system and the applications you wish to run. If you’re starting out with an SX or DX machine running at 25MHz or 33MHz, an upgrade processor could be well worthwhile. Owners of SX machines will derive the most benefit because the upgrade processors have an integrated floating point unit. This should provide a worthwhile performance boost when processing complex mathematical functions (eg, for spreadsheets and CAD programs). On the other hand, if you already have a clock-doubled processor, such as a DX2/50 or a DX2/66, the results will prob­ably be disappointing. The CPU will certainly run much faster but the overall system performance will not change much. In fact, if you’re currently running a 486DX/4 processor, you probably won’t notice the changes. A new processor by itself is not the end of the story, either. Adding extra RAM can provide some significant performance increases and this should always be looked at before (and if) you upgrade the processor. How much RAM known 80486DX/50 and the range of 80486DX2 chips. About this time, Intel also gave up its battle to stop other companies from using its numbering system to identify CPUs, registering “i486” as a trademark (pure numbers like 80486 could­ n’t be registered). They also later resorted to words like “Pentium”, instead of following their previous pattern and releasing their next generation chip as an 80586. However, that didn’t stop other companies from releasing 5-something and even 6-something chips (eg, AMD K5 and Cyrix/IBM 6x86). There were relatively few 80486DX/50 chips made and this was due mainly to the external speed constraints that existed at that time. The popular and efficient VESA Lo6  Silicon Chip should you have? If you want to run Windows 95/98, 16Mb is the recommended minimum but this should be increased to 24Mb or more if you frequently have several applications open at the same time. The HyperRace 586 from Hypertec is based on the AMD 5x86 chip and can be directly substituted for a 486 processor. An on-board clock multiplier ensures that the CPU runs at either 133MHz or 100MHz. It simply replaces your old 486SX, DX, SX/2 or DX/2 processor or it can be plugged into the Over­Drive socket, if one is available. And that’s it – there’s no software to install. An on-board clock multiplier is used to either triple or quadruple the processor speed. This multiplier ratio is set (using a jumper) during the installation, according to the system bus speed. For example, if the bus speed is 33MHz, the multiplier is set to x4 (the default) and so the clock speed of the new processor is 133MHz. Similarly, if the bus speed is 25MHz, the HyperRace runs at 100MHz. If the bus speed is 40MHz (for a 486DX/40), the multipli­er must be set to x3 and the processor runs at 120MHz. Note that motherboards running a 50MHz bus (for the 486DX/50) are not supported. The upgrade is easy to install, particularly if your cur­rent processor is in a ZIF (zero insertion force) socket. It’s simply a matter of lifting the lever, removing the old processor and installing the HyperRace 586 in its place. All you have to do is make sure that pin 1 of the processor goes to pin 1 of the socket. A detailed manual steps the user through the entire process and there’s even a troubleshooting process to refer to if you have problems. We installed the unit in a 486DX2/66 machine with 24Mb of RAM and running Windows 95. The machine booted straight up without any problems and cal Bus (VLB) worked well at 25MHz or 33MHz and was compatible with existing expansion cards. However, running a 40MHz or 50MHz bus placed extra demands on expansion cards and motherboards and this was reflected in the prices paid by consumers. By contrast, the 80486DX2/50 CPU ran internally at 50MHz while being clocked at only 25MHz. Similarly, the 80486DX2/66 ran internally at 66MHz from an external 33MHz clock. Oddly enough, the 80486DX4/100 was not really a 4/100 chip but rather a tripled 33MHz chip. In reality, it was an 80486DX3/99 but marketing triumphed over logic! Upgrade Possibilities: if you have an older ISA bus 486 system, try to scrounge a VESA local bus mother­ board from a junked ma­ chine, as the improvement will be dramatic. An ISA bus runs at 8MHz, whereas a VESA bus will run at 25MHz or 33MHz, depending on the CPU speed. As far as the CPU is concerned, there are upgrade or “over­drive” chips available from a number of sources. The original Intel overdrive chips will probably now be difficult to obtain but Kingston and Hypertec both have overdrive chips available that will allow your older 486 to run more like a Pentium – at least as far as the processor is concerned. Having said that, it’s important to remember that the overall speed can still be severely hampered by other slow components in the system. The HypeRace 586 One readily-available processor upgrade is the HyperRace 586 from Hypertec. This unit is based on AMD’s 5x86 chip and includes on-board voltage regulators and an integrated heat­sink and fan to keep things cool. Compatibility Not all machines are compatible with upgrade processors. Some early model 486 computers do not support clock quad­rupling due to limitations imposed by their system BIOS. Appar­ently, this type of BIOS uses CPU-dependent timing loops and if you quadruple the clock speed, there may not be enough time for the system to complete instructions. This problem can sometimes be resolved by upgrading the system BIOS but if you have an old computer, tracking down someone to do the job could prove difficult. all applications worked normally. We then checked the performance using the diagnostic utility (etdiag.exe) supplied on a floppy disc with the unit. This utility calculates Dhry­stones and gave a score of 32,759 before the upgrade and 48,597 after – an increase of 48%. We also ran the Landmark speed tests, recording scores of 225/563MHz before the upgrade and 449/1089MHz after substituting the HyperRace 586. Despite the increased processing power, the machine subjectively felt much the same as before – no doubt because we were upgrading from a DX2/66 and because of other bottlenecks, as previously mentioned. Users upgrading from 25MHz or 33MHz Another alternative, if you have a standard AT case, is to pick up an old Pentium motherboard. If your RAM currently con­sists of 30-pin SIMMs, you’ll need to get some 72-pin modules but these are currently pretty reasonably priced. You’ll probably also want a PCI VGA card – an ISA VGA card will make your Pentium run very slowly, while a VESA VGA card won’t fit into a Pentium motherboard. Pentium CPUs Intel’s successor to the 80486, the Pentium, is available in a range of speed ratings and has picked up a number of com­petitors along the way. Companies such as IBM, Cyrix, Nexgen, AMD and, more recently, Centaur, have all been working at capturing machines (particularly SX models) will probably get much more noticeable speed improvements – particularly when recalculating a large spreadsheet or carrying out some other processor intensive task. The HyperRace 586 costs around $199 and is available from Harvey Norman stores and other retailers. Log onto to www.hypertec.com.au for the address of the retail outlet closest to you. Kingston Turbochip 133 Part of Kingston Technology’s upgrade series, the Turbochip 133 is also a direct replacement for your current 486 CPU. Like the HyperRace 586, it’s based on AMD’s Am5x86P-75 processor and comes with on-board voltage regulators and an integrated fan/heatsink assembly. Unlike the Hyper­Race, however, there’s no option to set the clock multiplier – the Turbochip 133 features a fixed x4 multiplier, making it suitable for use on motherboards with bus speeds up to 33MHz. As before, motherboards running a 50MHz bus are unsupported but there’s not too many of these around. Once again, the upgrade is a snack to install and the manual is well written. You can install the Turbochip upgrade directly into the CPU socket or, if the CPU is soldered to the motherboard, into an adjacent OverDrive socket. There’s no soft­ware to install but Kingston do supply a utility disc that in­cludes the Landmark System Speed Test Ver.2.0. Tested in our 486DX2/66 machine, the TurboChip 133 returned almost some of Intel’s market share. The AMD K5 and Cyrix 5x86 CPUs are cheaper than equivalent Intel Pentium chips but offer slightly inferior performance. Pentium MMX CPUs Just when things were settling down again, Intel released the Pentium MMX. The letters “MMX” probably refer to “Multi Media eXtensions”, since the extra 57 opcodes added to the chip are primarily there to enhance its 2D graphics ability. The MMX has a 32Kb L1 cache, twice that of earlier Intel Pentium CPUs. This alone accounts for the 2030% increase in system speed over a standard Pentium CPU with the same clock speed. The Kingston Turbochip 133 is also used to directly replace a 486 CPU. It features a fixed x4 clock multiplier and can be used on motherboards with bus speeds up to 33MHz. identical results to the HyperRace 586 which is to be expected. This applies to both the Dhrystones measurement and the Landmark speed. At the time of writing, the TurboChip 122 costs around $230. For further information and the address of your nearest reseller, call Simms International on 1800 800 703 (freecall Australia-wide). Summary In summary, the HyperRace 586 and Kingston Turbochip 133 upgrade processors will provide the most benefit to users of older 25MHz and 33MHz 486 machines – particularly SX models. Users of DX2-66 machines and up will probably only experience small overall performance gains. Because DirectX 5 drivers make use of MMX commands, many games run much better on MMX machines, with improved graphics and faster response. Intel has made the MMX standard available under license and both Cyrix and AMD have released compatible CPUs. The Cyrix 6x86MX and AMD K6 both offer MMX support. Upgrade Possibilities: the Intel Pentium MMX is really a very good CPU. If you have one of these chips and you still don’t have enough grunt, check your motherboard’s specs to see if it will support a faster chip – either a faster Pentium MMX or a compat­ible AMD or Cyrix CPU. All these chips use a Socket 7 pin configuration but Intel has now gone October 1998  7 Upgrade CPUs For Pentium Machines Socket 7 based systems. Fortunately, it works be­cause I’m writing this article on a system using an AMD K6-2 and an AGP video card! Cyrix and AMD are both now developing 500MHz and 600MHz chips and shouldn’t be long with these new CPUs. There’s also a new socket coming (called the Socket 370), which will run on a 133MHz bus (faster than current Pentium IIs). Celeron & Pentium II CPUs In addition to the TurboChip 133, Kingston Technology also have the TurboChip 200 and the TurboChip 233 upgrade proces­sors. The TurboChip 200 is based on AMD’s 200MHz K6 MMX proces­sor and is designed for upgrading Pentium 75, 90 and 100MHz systems. It works with system buses running at up to 66MHz and provides clock tripling for the K6 CPU so that it runs at 200MHz. The TurboChip 233 is designed for upgrading Pentium 75MHz and up systems. This upgrade is based on an Intel 233MHz MMX Pentium processor and features a 3.5x clock multiplier to allow the CPU to run at full speed from a 66MHz bus. Unfortunately, neither upgrade is cheap due to the current weak­ness of the Australian dollar. Hpertec also have a upgrade CPU for Pentium systems and this is designated the MXPro200. This unit is intended for upgrading 75MHz and up systems. off on its own with the Pentium II (see below) by using a slot style configuration. Meanwhile, AMD, Cyrix and other chip makers have continued making faster CPUs for Socket 7 and they have been supported by several motherboard manufacturers. Socket 7 boards with bus speeds up to 100MHz (and even 112MHz) are now available and are called “Super 7” mother­boards. AMD’s response to the Pentium II has been to develop the K6-3D, later renamed the K6-2. This chip includes a number of en­hancements, including “3D Now!” which does for 3-D graphics what MMX did for 2-D graphics. The new commands making up “3D Now!” are already supported in DirectX 6, just released from Micro8  Silicon Chip soft, and have the ability to pipeline four floating point calculations per clock cycle, in­stead of just one. It makes short work of the calculations re­quired for 3D sound and 3D graphics effects. If, like many people today, you have already invested in a Pentium system and have an AT case, then look closely at the Super 7 alternative to the Pentium II system. You can use most of your existing hardware and just upgrade the motherboard and the CPU. Although originally developed for Pentium II systems, AGP video cards can also be used on many of the newer Super 7/Socket 7 systems. AGP (advanced graphics port) wasn’t originally intend­ed for older Intel made a completely radical change when it developed the Pentium II, abandoning the older “Socket 7” CPU for a new CPU with a straightline slot connector. Indeed, it looks more like a Nintendo cartridge than a CPU. Pentium II motherboards have a different footprint, or “form factor”, to the older AT motherboards. Because of this, the new “ATX” style mother­ board needs a different case than that used for the Socket 7 motherboard. The newer ATX case will often take an older AT mother­ board but not vice versa. In an attempt to attract budget buyers, Intel has also recently introduced the “Celeron” – basically a Pentium II but without the expensive integrated L2 cache. Although L2 cache can have a dramatic effect on some games and applications, it makes little difference in other cases. Indeed, the maths ability of the FPU in the Celeron is the most important factor as far as the game Quake is concerned and there is little difference between playing this game on a Celeron-based machine and one with a standard Pentium II. Upgrade Possibilities: if you have a Celeron CPU, you will need to ensure that you can run a Pentium II. This is not always possi­ble, so check before you buy. Either way, try to buy a system that gives you the ability to substitute a faster CPU later. This way, you can buy a good system and save money by not buying the latest CPU (which is usually over-inflated in price because it is the latest). You can then wait until the faster processor moves a few rungs down the pecking order, by which time it will be vastly more affordable. Pentium Pro Designed (and priced) with high end servers in mind, the Pentium Pro is basically a Pentium CPU with an on-chip L2 cache that runs at the full CPU speed. As a result, a 200MHz Pentium Pro can outperform a Pentium II 233 CPU for some applications. Note that in the Pentium Pro, the L2 cache is on the same chip as the CPU, not “closely coupled” as in the case of the Pentium II. Upgrade Possibilities: until recently, users of Pentium Pro CPUs have had to resort to running a dual or quad CPU system if they wanted extra performance. Of course, this required a motherboard that sported the extra CPU sockets. However, Intel has now released the promised MMX upgrade path for Pentium Pro users, by way of an overdrive chip. In addition to supporting the new MMX instructions, the new over­drive CPU has a 32Kb L1 cache and a “closely coupled” 512Kb L2 cache, again running at the same speed as the CPU. Systems running a 150MHz or 180MHz Pentium Pro CPU can now go to 300MHz using the new overdrive CPU, while, 166MHz and 200MHz systems can go to 333MHz. Overclocking As the name suggests, overclocking involves running a CPU at a speed faster than it was designed for. To explain, the speed at which the processor runs is set by the two factors: the system bus speed and a multiplier (or ratio) setting. These are usually set by jumpers on the mother­ board. Some common bus speeds are 50MHz, 60MHz and 66MHz, while the multipli­er settings generally range between 1.5 and 3 (eg, 1.5, 2, 2.5, 3). Many recent motherboards can provide even higher bus speeds (eg, 75MHz, 83MHz & 100MHz), as well as higher multiplier settings. This can be checked out by referring to the manual. In practice, this means that if we have a 133MHz processor (for example), the bus speed will be set to 66MHz and the multi­plier to 2 (2 x 66 = 132). Note that, in this case, the processor runs at twice the bus speed. Similarly, a 75MHz processor will run on a 50MHz bus with a multiplier of 1.5, while a 200MHz processor will run on a 66MHz bus with a multiplier of 3. Overclocking involves changing the bus and/or multiplier settings on the motherboard to bump the CPU Table 1: Common Bus Speeds & Multipliers Multiplier 1.5 1.75 2 B us S peed 50MHz 2.5 3 3.5 4 4.5 175 200 225 CPU Speed (MHz) 75 87.5 100 125 150 60MHz 90 105 120 150 180 210 240 270 66MHz 100 116 133 166 200 233 266 300 75MHz 112 130 150 188 225 263 300 337.5 83.3MHz 125 145 166 208 250 290 333 375 100MHz 150 175 200 250 300 350 400 450 This table lists the common bus frequencies and multipliers, together with the resulting CPU speeds. The speeds highlighted in yellow are the multiplier/bus speed combinations commonly recommended by the CPU manufacturers but you can try other combinations if you wish to experiment with overclocking. up to the next speed. For example, a 120MHz processor could be overc­locked simply by changing the bus speed from 60MHz to 66MHz, so that it runs at 133MHz. What about the pre-Pentium chips? In the case of earlier 8088, 80286 and 80386 CPUs, it’s sometimes possible to tweak the speed up a little but the performance gain is modest (so modest that you wouldn’t bother, in fact). With 486 and later chips, the reliability is usually better and the performance gains can be more worthwhile. There are a few risks, however. First, if you’re not careful, you can easily destroy the CPU due to overheating. That’s because the faster a chip goes, the hotter it gets and increasing the clock speed will run a chip closer to it’s limits. A heatsink and a cooling fan should be fitted to the CPU if you intend over- A processor upgrade is easy to install if a ZIF (zero insertion force) socket is fitted to the motherboard. You just lift the lever to release the old processor. clocking the system. If these items are already fitted, you may have to improve the cooling by fit­ting a high capacity fan. The second risk involved with over­ clocking is instability. This is more critical with Windows 95 than Windows 3.11 but the only way to find out is to try it and see. You will need to consult your motherboard manual to set system bus and multiplier speeds. If you have a 486SX/25 or DX/25 system, you could try running it at 33MHz. Similarly, if you have a 33MHz processor, try running it at 40MHz. If you are running a 486DX2/50, it will sometimes be possible to get it to run as if it were a 486DX2/66, simply by running a 33MHz bus instead of the original 25MHz bus. As discussed above, a 486DX/4 CPU actually triples the bus speed. In some cases, you may be able to run a 486DX4/100 system as a 486DX4/120, simply by setting the bus to run at 40MHz in­stead of 33MHz. With Pentium and later CPUs, it can get a little more com­plicated. Here you will find clock rates and multipliers that have one set of rules for one CPU and a completely different set of rulers for another. Note also that some chip manufacturers included “overclock protection” on their CPUs. If this is the case, try changing the bus speed instead of the multiplier. On the other hand, some CPU’s such as the IBM/Cyrix MX-PR333 almost beg to be overclocked. The PR333 can run at 2.5x and 100MHz (250MHz), 3x and 83MHz (250MHz), October 1998  9 Tom’s Hardware Guide at www.tomshardware.com/overclock contains some excellent advice on overclocking CPUs, including a step-by-step guide. A number of other web sites also offer useful advice on this subject 3.5x and 75MHz (263MHz), or 4.0x and 66MHz (266MHz). Despite running at 250-266MHz, this CPU performs as if it were a 333MHz chip – hence IBM/Cyrix’s PR rating. Be warned that a 75MHz (or higher) bus might cause system instability if you are using EDO RAM (especially the 70ns type). That’s because EDO RAM was designed to operate at a maximum bus speed of 66MHz, where­as some types of SDRAM can run at up to 100MHz. In some cases, you might be able to get around this problem by altering the DRAM wait state values in the system BIOS. Be warned also that you might encounter problems with older expansion cards if you increase the system bus speed. The PCI bus runs at half the system bus speed and too high a speed could cause problems with some older cards. By the way, using a high bus speed and a low multiplier to set the CPU speed will give faster results overall. That’s be­cause the data throughput on the PCI bus will be much higher. For example, using an 83MHz bus and a x2 multiplier (= 166MHz) will give better performance than using a 66MHz bus and a 2.5x multi­plier (if the system will operate on an 83MHz bus, that is). Conversely, overclocking the processor by decreasing the bus speed but increasing the multiplier will provide only margin­al benefits (if any). A good example here is if you overclock a Pen- Weigh The Risks Before Overclocking! The information on CPU over­ clocking included here is intended as a general guide only and you should carefully weigh up the risks involved before attempting to over­clock your system. If you are unsure as to what you are doing, the best advice is “don’t do it”. In particular, readers are warned that overclocking could lead to system instability and data loss and could void any warranties. It could also cause the CPU (and 10  Silicon Chip even the motherboard) to fail due to overheating. At the very least, you should back up all your data and the registry (system.dat and user.dat) before making any changes. Note that none of the CPU manufacturers recommends overclocking. Finally, you make any modifications at your own risk. Silicon Chip Publications Pty Ltd disclaims any liability for any data loss or damage that may result from readers experimenting with overclocking. tium 133 to 150MHz by increasing the multiplier from 2 to 2.5 and decreasing the bus speed from 66MHz to 60MHz. Sure, the processor will run faster but the memory and other parts on the motherboard will run slower due to the decreased bus speed. As before, heat is your biggest enemy. If the system is OK when it’s cold but becomes unstable after it’s been running for awhile, try fitting a bigger cooling fan to the CPU and improving the case ventilation. Table 1 shows a list of common bus frequencies and multi­pliers, plus the resulting CPU speed. The speeds highlighted are the common combinations recommended by the chip manufacturers. You can try other combinations for yourself but remember – the risk is entirely yours. Don’t be over-ambitious when it comes to overclocking. You might be able to get the system to operate reliably at the next highest speed setting but that will probably be the limit. Precautions There are a few precautions that you should observe. First, watch out for static electricity as it can damage motherboards, add-on cards, CPUs and RAM chips. If you change the CPU, make sure that the new device is installed correctly, with pin 1 going to pin 1 of the socket. Some CPUs will be damaged if you plug them in the wrong way around. Be careful not to bend the pins of the CPU as it’s difficult to straighten them properly afterwards and all too easy to break them off. Running a CPU at less than its specified rating won’t hurt it but over­clocking can overheat a chip and damage it. Make sure that the CPU stays cool and keep an eye on the temperature while testing new clock speed/multiplier combinations if you intend overclocking. A larger cooling fan will be required in most cases. Don’t forget to double-check the CPU voltage jumpers if you replace the CPU. This won’t apply to 486 and earlier chips but is very important for Pentium or later CPUs. Note that MMX type chips usually run from dual supply rails. Finally, there’s lots of advice on overclocking available on the Internet. Check out Tom’s Hardware Guide at www.tomshardware.com/overclock, SC for example. Silicon Chip Bookshop SUBSCRIBE   AND GET   10% OFF SEE PAGE 21 Guide to Satellite TV Installation, Recept­ion & Repair. By Derek J. Stephen­son. First published 1991, reprinted 1997 (4th edition). This is a practical guide on the installation and servicing of satellite television equipment. The coverage of the subject is extensive, without excessive theory or mathematics. 383 pages, in hard cover at $55.00. Servicing Personal Computers By Michael Tooley. First published 1985. 4th edition 1994. Computers are prone to failure from a number of common causes & some that are not so common. This book sets out the principles & practice of computer servicing (including disc drives, printers & monitors), describes some of the latest software diagnostic routines & includes program listings. 387 pages in hard cover at $90.00. Video Scrambling & Descrambling For Satellite & Cable TV By Rudolf F. Graf & William Sheets. First pub­lished 1987. This is an easy-to-understand book for those who want to scramble and unscramble video signals for their own use or just want to learn about the techniques involved. It begins with the basic techniques, then details the theory of video encryption and decryption. It also provides schematics and details for several encoder and decoder projects, has a chapter of relevant semiconductor data sheets, covers three relevant US patents on the subject of scrambling and concludes with a chapter of technical data. 246 pages, in soft cover at $50.00. The Art of Linear Electronics By John Linsley Hood. Pub­lished 1993. This is a practical handbook from one of the world’s most prolific audio designers, with many of his designs having been published in English technical magazines over the years. A great many practical circuits are featured – a must for anyone inter­ ested in audio design. 336 pages, in paperback at $70.00. Digital Audio & Compact Disc Technology Produced by the Sony Service Centre (Europe). 3rd edition, published 1995. Prepared by Sony’s technical staff, this is the best book on compact disc technology that we have ever come across. It covers digital audio in depth, including PCM adapters, the Video8 PCM format and R-DAT. If you want to understand digital audio, you need this reference book. 305 pages, in paperback at $90.00. Surface Mount Technology By Rudolph Strauss. First pub­lished 1994. This book will provide informative reading for anyone considering the assembly of PC boards with surface mounted devices. Includes chapters on wave soldering, reflow­soldering, component placement, cleaning & quality control. 361 pages, in hard cover at $99.00. Radio Frequency Transistors Principles & Practical Applications. By Norm Dye & Helge Granberg. Published 1993. This book strips away the mysteries of RF circuit design. Written by two Motorola engineers, it looks at RF transistor fundamentals before moving on to specific design examples; eg, amplifiers, oscillators and pulsed power systems. Also included are chapters on filtering, impedance matching & CAD. 235 pages, in hard cover at $95.00. Guide to TV & Video Technology By Eugene Trundle. First pub­lished 1988. Second edition 1996. Eugene Trundle has written for many years in Television magazine and his latest book is right up date on TV and video technology. 382 pages, in paperback, at $55.00. Electronics Engineer’s Reference Book Edited by F. F. Mazda. First published 1989. 6th edition. This just has to be the best refer­ ence book available for electronics engineers. Provides expert coverage of Your Name__________________________________________________ PLEASE PRINT Address____________________________________________________ _____________________________________Postcode_____________ Daytime Phone No.______________________Total Price $A _________ ❏ Cheque/Money Order  ❏ Bankcard  ❏ Visa Card  ❏ MasterCard Card No. Signature_________________________ Card expiry date_____/______ Return to: Silicon Chip Publications, PO Box 139, Collaroy NSW, Australia 2097. Or call (02) 9979 5644 & quote your credit card details; or fax to (02) 9979 6503. Prices valid until 31st October, 1998 all aspects of electronics in five parts: techniques, physical phenomena, material & components, electronic design, and applications. The sixth edition has been expanded to include chapters on surface mount technology, hardware & software design, semi­-custom electronics & data communications. 63 chapters, soft cover at $160.00. Audio Electronics By John Linsley Hood. Pub­lished 1995. This book is for anyone involved in designing, adapting and using analog and digital audio equipment. Covers tape recording, tuners & radio receivers, preamplifiers, voltage amplifiers, power amplifiers, the compact disc & digital audio, test & measurement, loudspeaker crossover systems and power supplies. 351 pages, in soft cover at $75.00. Understanding Telephone Electronics By Stephen J. Bigelow. Third edition published 1997 by Butterworth-Heinemann. This is a very useful text for anyone wanting to become familiar with the basics of telephone technology. The 10 chapters explore telephone fundamentals, speech signal processing, telephone line interfacing, tone and pulse generation, ringers, digital transmission techniques (modems & fax machines) and much more. Ideal for students. 367 pages, in soft cover at $55.00. ✓ Title Price ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ Guide to Satellite TV $55.00 Servicing Personal Computers $90.00 Video Scrambling & Descrambling $50.00 The Ar t Of Linear Electronics $70.00 Digital Audio & Compact Disc Technology $90.00 Surface Mount Technology $99.00 Radio Frequency Transistors $95.00 Guide to TV & Video Technology $55.00 Electronic Engineer's Reference Book $160.00 Audio Electronics $75.00 Understanding Telephone Electronics $55.00 Postage: add $5.00 per book. Orders over $100 are post free within Australia. NZ add $10.00 per book; elsewhere add $15 per book. TOTAL $A October 1998  11 SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.dse.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.dse.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.dse.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.dse.com.au EMC Explained Pt.3: Immunity To Interference Having discussed the testing of equipment for electro­magnetic emissions, we now focus on the flip side of the coin – immunity. Immunity is the measure of how susceptible a device is to electro­ magnetic emissions from external sources. By MARQUE CROZMAN* Immunity is particularly important in equipment used in hospitals and aircraft. Notices or announcements are usually made to the public banning the use of electronic equipment in these places, for fear of having an effect on instrumentation or monitor­ ing equipment. Unfortunately, equipment making minute measure­ments can also be particularly sensitive when it comes to suscep­tibility. We also get annoyed when we buy equipment and find that other devices that we own affect them. Most people have had the situa­tion when they have tried listening to an AM radio sitting next to a home computer. Radio stations can be completely swamped by the noise the computer puts out and the same goes for the noise radiated by TV sets. Is it the fault of the radio, the PC or the TV set? Really, it is an overall responsibility: the PC and TV should A great deal of EMC immunity testing is performed in semi-anechoic chambers to keep extraneous signals from affecting the test results. Here a log-periodic antenna is being set up for a range of tests. 16  Silicon Chip not radiate and the radio should not to be susceptible to noise outside the frequency band you are listening to. How do you control susceptibility? The trade-off is to either control the environment in which the equipment operates which is an impossible or expensive option or to design the equipment to withstand these effects. An example of a device where the environment is controlled rather than the machine is MRI (magnetic resonance imaging) in hospitals (typically used for brain scanning). The room in which it is located is effectively a Faraday cage to prevent external fields getting in and stopping internal fields getting out. From the outset, it must be noted that Australia and New Zealand do not require testing for susceptibility for the purpos­es of C-Tick compliance. They only require compliance for emis­ sions. The upside is that much of the effort taken to reduce emissions also helps in making the device less susceptible. European CE compliance and most military standards do re­quire compliance testing. CE implements its immunity requirements via a suite of standards that are outlined in the accompanying panel. These follow the same format as the emission standards, with a generic standard covering all products, then product family standards and finally, specific product standards. Each of these standards calls up other standards that out­ l ine the tests to be carried out and methods by which to do so. For instance, *Marque Crozman is a design engineer with Innotech Control Sys­tems, in Brisbane, Qld. Phone (07) 3481 1388. EMC testing is also done at open area test sites where ambient signal strengths are low. This site in Victoria is operated by EMC Technologies Pty Ltd. EN61000-4-3 defines the levels and methods for testing RF immunity. Susceptibility – sources of interference Susceptibility testing attempts to subject the product to the worst sorts of interference that it might experience in the real world. Sources of this interference can be the static electricity that we build up on days when the humidity is low and which discharges when we touch something metal or the switching tran­sients from contactors or relays such as a washing machines on its spin dry cycle. Or it can be RF fields generated such as when we talk on a mobile phone and the mains- borne interference that gets into the product if it is mains-powered. Susceptibility testing is hard to quantify, since the func­tioning of the product has to be evaluated in determining whether it has passed or failed. The criteria for this is provided by the manufacturer in his specification for the product and through requirements in the various standards. Before testing for immunity, the manufacturer has to detail how the product operates and what would be construed as a pass or failure of the product for each test. This information (in the test plan) is then used during the testing and is included in the test report as to how the device Fully Accredited Testing for performed. Any anomalies or effects during testing are noted and included in the test report. Different tests have different performance criteria for the device under test. The standards outline three criteria and each test then specifies one of these criteria of operation to deter­mine a pass or failure: Criteria A states that the apparatus must operate as per the manufacturer’s specification for the duration of the test. Criteria B states that the apparatus must operate as in­tended after the test and that during the test no change of operational state or stored data is allowed. A possible example of this would be interference to a monitor. The picture may tear or distort but it is not allowed to change resolution or screen settings. It would also not be allowed to forget user settings. Criteria C states that temporary loss of function is al­lowed, if this is recoverable – either by itself or by manual operation of the controls. The monitor may lose control of the picture being displayed and may crash the controlling firmware, so long as this is recoverable, either by itself (watchdog reset) or by manually switching it off and on again. For all operational criteria, the apparatus is not allowed to become dangerous or unsafe. Testing susceptibility – electrostatic discharge The immunity test for static electricity is the Electro­static Discharge Global Markets EMC Technologies' Internationally recognised Electromagnetic Compatibility (EMC) test facilities are fully accredited for emissions, immunity and safety standards. >> NATA endorsed reports for ALL electrical products covered by the and regulations >> Accredited Competent Body for TCF approval >> Anechoic chamber for accredited immunity testing. Field uniformity ensures that your products are not "over tested". DC – 1000MHz <at> 100-200V/m, 1-18GHz & >50-100V/m >> Open Area Test Site (OATS) available for hire for DIY testing >> Low ambient OATS in Melbourne, Sydney and Auckland (NZ) >> RF testing DC – 40GHz >> CE (Europe), FCC (USE), VCCI (Japan) EMC Technologies Melbourne T: +61 3 9335 3333 F: +61 3 9338 9260 E: melb<at>emctech.com.au Sydney T: +61 2 9899 4599 F: +61 2 9899 4019 E: syd<at>emctech.com.au Auckland T: +64 9 360 0862 F: +64 9 360 0861 E: auklab<at>emctech.com.au Visit our website: www.emctech.com.au October 1998  17 Fig.1: a simplified schematic diagram of an ESD gun. Fig:2: this is the voltage waveform used for electrical fast transient (EFT) testing. or ESD test. This sets out to simulate the conditions when it is touched by a statically charged object or person or a static electricity discharge is made in proximity to the device. Based around standard IEC1000-4-2, the test involves performing discharges to and near the equipment under test (EUT). The equipment is tested inside a screened room on a wooden table. On the table is a coupling plane. The EUT sits on the coupling plane but is isolated (usually via Mylar sheet) from it. Discharges are carried out with an ESD gun with controlled energies and voltages. Fig.1 shows a simplified circuit for an ESD gun. Table 1 lists the discharge levels for the various tests. The contact tip of the gun has the same dimensions as a typical finger. Contact discharges are then carried out to all parts of the device where you can get a finger into or touch. Points where a discharges are made (to the metal parts of the case or where insulation breaks down) then get 10 contact discharges to that point to 18  Silicon Chip make sure that any effect can be seen. The generic standard EN55082-1 specifies 4kV for contact discharges, although other standards call for different voltages. Air discharges are then carried out with a different tip in the close proximity of the device. EN55082-1 calls for air dis­charges at 8kV. The device passes the test if the conditions for criteria C are met; ie, no permanent damage to the device and normal opera­tion can be restored. Electrically fast transients Very fast transients/bursts such as those generated by switching (in- Table 1: Voltage Levels For Discharge Tests Level 1 2 3 4 Contact Test Voltage Air Discharge 2kV 4kV 6kV 8kV 2kV 4kV 8kV 15kV terruption of inductive loads, relay contact-bounce, etc) normally cause trouble by coupling into equipment wiring. These generally only occur spasmodically but especially affect devices with microprocessors and/ or logic. Spikes can sometimes cause false triggering of gates and corruption to software. Testing conditions are specified as criteria B – devices are allowed to falter but must not change state or mode and no cor­ ruption of stored data is allowed. This test is carried out in a screened room. 3-metre cables are connected to the EUT in the same fashion as for the radiated emissions test. Mains and data/interconnection cables are treat­ ed separately, with separate levels of severity. The device is set up on a wooden table in a similar way to the ESD test and the generator is located on the floor of the screened room. The device is powered up via a mains outlet on the test generator. The waveform shown in Fig.2 is injected directly into the mains cable leading to the EUT. Tests are carried out in both positive and negative polarities and in turn to each supply lead and to the protective earth. The length of the test is nominally one minute per polarity per lead. Test generators are normally computer-controlled and have a program that carries out the polarity control, test timing and output switching. It also includes a substantial filter to stop noise getting back into the mains supply. Mains injection levels vary from 1kV to 4kV. Some genera­ tors run initial tests at lower levels to avoid damage to the EUT if protective counter measures have not been taken in its design. This generally affects the equipment enough to demonstrate that more work needs to be carried out before being exposed to the full force of the test. Tests on data/interconnection cables are made via a capaci­tive clamp which is 1m long and is connected to the test generator by a high voltage coax cable. Cables are laid in the clamp and the cover pulled across. The EUT is now powered directly from the mains and the test repeated. Lower severity levels are used for this test. As an example, the light industry generic standard EN 50082-1:1992 requires mains injection levels of 1kV, with data cable levels of 500V. Again, tests are carried out for a nominal time of one minute and in both polarities. Voltage surge As with EFT, voltage surge tests phenomena commonly found on the mains, although these are slower. These are overvoltages or overcurrents caused by electrical faults, heavy load switching and lightning. The setup is the same as for EFT, however data and signal lines are not tested. The tests are defined as being 1.2µs/50µs (rise time/fall time) voltage or 8µs/20µs current waveshape surges. At least five positive and five negative surges are applied at a repetition rate no faster than one per minute. This allows the protection devices time for recovery. As with EFT, the EUT is expected to meet the requirements of criteria B to attain a pass. Voltage dips & interruptions The VDI test applies only to mains-powered equipment. It sets out to simulate the sort of interference caused by mains faults, power distribution switching and heavy load switching in the supply grid (that cause dips or brown-outs in the supply). The setup and methods of testing are the same as those for EFT. Dips and short interruptions are initiated at any phase angle of the input voltage, to a level of 0%, 40% and 70% of the nominal voltage for a duration of 0.5 to 50 periods. Short term variations are made to a level of 40% and 0% of nominal voltage (ie, no voltage at all) for one second at the test level. The product is expected to meet the requirements of criteria C. Radio frequency immunity This would have to be the longest of all the suscep­tibility tests. There are actually two methods for testing RF immunity (RFI): conducted and radiated. It really depends on the standard being applied as to which one is used. The assumption is that a product could be subjected to a constant RF field from nearby radio and television transmitters, mobile phone towers and the like. As a result, criteria A is chosen: that the equipment must operate as intended during the test. Different standards call for different RF levels but gener­ally speaking if the EMC Immunity Standards EN50082 Part 1:1997 Generic immunity standard, part 1: Residential, commercial and light industry environment. Scope All apparatus intended for use in the residential, commercial and light industrial environment – both indoor and outdoor, for which no dedicated product or product-family emission standard exists. Equipment in this environment is considered to be directly con­nected to the public mains supply or to a dedicated DC source. For the purposes of testing, the equipment is considered to be operating normally; ie, fault conditions are not taken into account. Tests Enclosure: •  IEC 1000-4-2 Electrostatic discharge Air discharge: 8kV Contact Discharge: 4kV (10 discharges to preselected points of normal contact) •  IEC1000-4-3 RF immunity Severity: 3V/m; 80MHz to 1000MHz AC mains: •  IEC 1000-4-4 Electrical fast transient burst testing 1kV Bursts of 5ns/50ns (rise time/fall time) pulses at a repeti­tion rate of 5kHz with a duration of 15ms, applied in both polarities between power supply terminals (in­cluding protective earth) and a reference ground plane. Data and signal lines: •  IEC 1000-4-4 Electrical fast transient burst testing 500V bursts of 5ns/50ns pulses at a repetition rate of 5kHz with a duration of 15ms and period of 300ms, applied in both polari­ties via capacitive coupling clamp. EN55104:1995 Electromagnetic compatibility – immunity requirements for house­hold appliances. Tools and similar apparatus – product family standard Scope All apparatus intended for use in the domestic environment, including toys and tools. This standard is the counterpart to EN55014. For the purposes of testing, the equipment is considered to be operating normally; ie, fault conditions are not taken into account. Apparatus is classified into four categories: Category I: Apparatus containing no electronic control circuitry Category II: Mains powered appliances containing electronic control circuitry with no internal frequency higher than 15MHz Category III: Battery powered apparatus containing electronic control circuitry with no internal frequency higher than 15MHz Category IV: All other apparatus within this scope Tests Enclosure: •  EN61000-4-2 Electrostatic discharge testing Air discharge: 8kV Contact Discharge: 4kV (10 discharges to preselected points of normal contact) •  ENV 50141 RF immunity testing Severity: 3V; 150kHz to 230MHz Conducted AC mains: •  EN 61000-4-4 Electrical fast transient burst testing 1kV bursts of 5ns/50ns (rise time/fall time) pulses at a repeti­tion rate of 5kHz with a duration of 15ms, applied in both polarities between power supply terminals (in­cluding protective earth) and a reference ground plane. •  EN61000-4-11 Voltage dips and interruptions Dips and short interruptions initiated at any phase angle of the input voltage, to 0%, 40% and 70% of the nominal voltage for a duration of 0.5 to 50 periods. Short term variations to a level of 40% and 0% of nominal voltage and to recover from it and one second at the test level. •  EN 61000-4-5 Surge testing At least five positive and five negative surges, at a repetition rate no faster than one per minute, of 1.2ns/50ns (rise time/fall time) voltage or 8ns/20ns current waveshape surges at levels of 2kV (line to earth) and 1kV (line to line). Data and signal lines: •  IEC 1000-4-4 Electrical fast transient burst testing 500 bursts of 5ns/50ns pulses at a repetition rate of 5kHz with a duration of 15ms, applied in both polarities via capacitive coupling clamp. October 1998  19 PCs are well-known sources of interference and here a system is being set up on a wooden table at an open area test site device is for domestic or light industry the levels are 3V of conducted RF and 3V/m radiated RF. Heavy indus­try levels are 10V conducted and 10V/m radiated. As a general guideline, devices that have microprocessors are tested using the radiated method and those that don’t are tested via the conducted method. Household devices therefore are mostly tested using the conducted method. Testing for conducted RF immunity is straightforward. RF is coupled into the mains supply powering the device under test. The point at which the RF is injected also has a hefty filter to stop RF getting out into the mains supply. This test is carried out in a screened room with the EUT on a wooden table with the injection device on the floor. The equipment is monitored throughout the test for its adherence to criteria A as RF is stepped through the frequency bands to be covered. This can take several hours. For the domestic equipment standard (EN55104:1995), 150kHz to 230MHz is tested. The radiated method of testing is far more complex. A semi-anechoic chamber is required to stop the formation of standing waves that can be of much higher magnitude than the test re­quires. These are shielded rooms with either ferrite tiles or special RF-absorbing foam-like structures Even wheelchairs need to be tested for electromagnetic com­pliance. Here a wheelchair is being set up on a wooden table prior to testing. 20  Silicon Chip covering the internal surfaces of the chamber. A log-periodic antenna is set up and is driven by a signal generator and amplifier to generate a field in the room. The EUT sits on a wooden table, three metres from the antenna. The test band covers 80MHz to 1GHz (IEC1000-4-3). Before any equipment is placed in the field, the field is calibrated. The calibration procedure involves placing a wooden frame into the space where the equipment will sit. The frame consists of a matrix of 4 x 4 wooden rails. The field strength is then measured at each of the 16 points on the matrix. The field-strength meter is coupled to test equipment outside the room by means of a fibre-optic link. This builds up a table of calibra­tion data that then controls the output power of the RF amplifi­er. The standard requires that 9 of the 16 measured points must be within +3dB of the required level for each frequency to be tested. Once complete, the calibration ensures that the space in which the device is to be placed has a uniform field. A previous standard only required that a field strength meter be placed in close proximity of the EUT and that the power of the RF amplifier be wound up until the field strength read the level being tested to. It also did not require the use of a semi-anechoic chamber. The standard was changed since this caused non-repeatable re­sults, with the equipment being sometimes subjected to standing waves well in excess of what was called for. The equipment is then tested in the calibrated section of the field and monitored either through a portal or via a video monitoring system. Like the conducted RF test, criteria A is used. This also takes a few hours to complete with the test generator stopping at each frequency and radiating the EUT for a time before moving on to the next. For equipment to pass these tests requires protection to be designed into the product from the start. Band aids to existing products by placing suppressors on the terminals or the like normally don’t work. SC Acknowledgment: all photos by courtesy of EMC Technologies Pty Ltd, Victoria. ORDER FORM BACK ISSUES MONTH YEAR MONTH YEAR PR ICE EACH (includes p&p) TOTAL Australi a $A7.00; NZ $A8.00 (airmail ); Elsewhere $A10 (airmail ). Buy 10 or more and get a 10% discount. Note: Nov 87-Aug 88; Oct 88-Mar 89; June 89; Aug 89; Dec 89; May 90; Aug 91; Feb 92; July 92; Sept 92; NovDec 92; & March 98 are sol d out. All other issues are currently i n stock. $A B INDERS Pl ease send me _______ SILICON CHIP bi nder(s) at $A12.95 + $5.00 p&p each (Australi a only). N ot avail abl e elsewhere. Buy five and get them postage free. e & Get Subscrib ount On c is D A 10% r Silicon e th O ll A e rchandis Chip Me $A SUBSCRIPTIONS  New subscription – month to start­­____________________________  Renewal – Sub. No.________________    Gift subscription  GIFT SUBSCRIPTION DETAILS RATES (please tick one) 2 years (24 issues) 1 year (12 issues) Australia (incl. GST)  $A135  $A69.50 Australia with binder(s) (incl. GST)**  $A159  $A83 New Zealand (airmail)  $A145  $A77 Overseas surface mail  $A160  $A85 Month to start__________________ Overseas airmail _____________________________  $A250  $A125 **1 binder with 1-year subscription; 2 binders with 2-year subscription YOUR DETAILS Your Name_________________________________________________ Message_____________________ _____________________________ Gift for: Name_________________________ (PLEASE PRINT) Address______________________ _____________________________ (PLEASE PRINT) Address___________________________________________________ State__________Postcode_______ ______________________________________Postcode_____________ Daytime Phone No.____________________Total Price $A __________ Signature  Cheque/Money Order   Bankcard   Visa Card   Master Card ______________________________ Card No. Card expiry date________/________ Phone (02) 9979 5644 9am-5pm Mon-Fri. Please have your credit card details ready OR Fax (02) 9979 6503 Fax the coupon with your credit card details 24 hours 7 days a week Mail order form to: OR Reply Paid 25 Silicon Chip Publications PO Box 139, Collaroy 2097 No postage stamp required in Australia October 1998  21 MAILBAG GST could raise prices Most of your readers will be aware that electronic compon­ents, tools and test gear carry a 22% sales tax and that the current government is talking about a GST of perhaps as low as 10%. Great, electronics is going to get cheaper? No! The 22% is applied to the wholesale price, giving the retailer’s “cost price”. To this they add their mark-up to cover all their other costs and turn a profit. This mark-up is typically 100% but can be much higher on small parts. If the mark-up is exactly 100% then the tax component is equal to a 9.9% retail tax) so a 10% GST will see similar prices while a higher rate will cause a significant rise. However, current wholesale sales tax rates vary (starting at 12%), as does mark-up, so many items will increase in price while a few will drop. Many items are currently exempt. These include fasteners, “building materials” like electrical fittings, PVC pipe and fittings, magazines and trade journals, all of which will rise. The “S” in GST is for Services, meaning you will pay 10-15% more every time you pay someone to do something, every time you use a telecommunications service, etc. This includes the elec­tronic service industry and it will have a severe negative ef­fect. The price of large TVs, VCRs and sound systems which have a fairly high tax component due to being taxed at the “luxury” item rate of 32% on the wholesale price will likely drop somewhat. Meanwhile a repair job, currently not taxed (except for any parts used) will rise by 10-15% due to the tax added to it. This reduction in the difference between the price of repair and replacement increases the likelihood of replacement. This means electronic service businesses will either lose business or will need to cut prices to an unrealistic level. Either way, staff will probably be cut or the whole business close. The costs of setting up for the new tax are also very high, with $8000 being one suggestion, plus a modern 22  Silicon Chip computer if one is not already owned. It should be noted that the largest economy in the world, the United States, does NOT impose a tax on services, its state retail sales taxes applying only to goods. Of all the countries that have introduced such a tax, only three have seen no rise in the rate and don’t believe anyone who promises not to raise the rate here. J. Sortland, Hornsby Heights, NSW. Vibrators for car radios still available I have just been reading the “Serviceman’s Log” in the September 1998 issue and I was interested to read of his experi­ences with the old Delco car radio from the 1955 Cadillac. What a pity he didn’t know that “plug-in” replacement “solid state” vibrators are still available from Antique Electronic Supply, PO Box 27468, Tempe, Arizona, USA 85285-7468. Their 1988 catalog has them listed at $US24.95 (about $A43.77) each and they are avail­able in 3 and 4-pin, 6V and 12V, negative and positive earth types. He could have saved himself a lot of time and bother! The OZ4 rectifier valve also seems to be available from the same source, although I don’t hold any grudges against someone replacing rectifier valves with semiconductors! One thing he should be aware of: the “new” diodes will increase the B+ voltage so a dropping resistor should be installed to return the B+ to its original voltage, otherwise some of the valves (and some components) will be operated beyond their maximum ratings. In our R-390A Communications Receiver, for example, replacing the 26Z5W rectifier valves causes the B+ to rise from 265V DC to a whopping 290V DC! T. Robinson, Woodend, Vic. Some uses for old computers I decided to send in a couple of suggestions for uses of old computers, as called for in the editorial in the September 1998 issue. My first suggestion is that surplus computers could be used as email machines. Email doesn’t require a lot of computing overhead – it’s only text based, so there’s one use for old computers for you. For that matter, you could even use them for web browsing. There aren’t many computers that can’t keep up with even the speediest modems available today. An old 386 with an early web browser should be fine for most web browsing. Anyway, this has diverted me (partly at least) off the track of my main suggestions. These were classrooms and gifts to developing countries. Since computer skills are becoming a re­quirement in a large number of professions these days, it is advantageous for school kids to start acquiring skills as soon as possible. Not all public schools have the budgets of private schools and so can’t afford to buy the latest Pentium machines for all students. However, an old computer would be a lot better for a school student to learn on than nothing at all. You can run typing tutors on them – the keyboard of an old machine is pretty much the same as the keyboard of a new machine. You can also run simple DOS-based spelling software, etc. Just getting children familiar with computers and how to operate them is a very positive step. And once again, you could run a basic web browser and teach them WWW activities (need to watch that one of course!) My last suggestion is to give these old computer systems to developing countries, possibly for their classrooms. As you’ve stated, it seems ridiculous to throw a perfectly working computer on the scrap heap. So why not collect them and send them to developing nations where they could have similar educational benefits as mentioned in the previous paragraph, rather than letting them add to our landfill problems? Maybe you could get our foreign aid department, AusAID, to throw a small amount of money at a project to collect old comput­ ers, employ someone to put all the pieces together in working units, then ship them off to some of our needy neighbours. It’s a win-win situation! We reduce our landfill, improve ties with our neighbours and they get a valuable educational tool. S. Stringer, Ainslie, ACT. More uses for old computers I bet there would be a lot of people out there interested in accepting old PCs otherwise destined for the tip, myself included. These come in to their own when connected to hardware where processing needs are not too great but where timing issues can complicate matters under Win95. A dedicated Pic/Eprom/Gal programmer comes to mind. A custom burglar alarm, CNC controller, data acquisition, diagnostics and measurement, home brewing and networking with Linux are also possibilities. Maybe you don’t even need an entire PC. Embedded controller boards aren’t cheap but an old motherboard may do the trick. You can ditch the monitor for an LCD character display or even dis­able the display and keyboard in BIOS if you don’t need them. Many flavours of DOS will run from a floppy (no HDD required) or can be embedded and the application could even be placed in an Eprom. Software development is cheap/free with all the 8x86 tools out there. It’s usually only the drives etc that need voltages other than 5V, so it may be possible to substitute a simple linear supply. There are even plans on the Web for turning an old floppy drive into a robot! J. de Stigter, Frenchs Forest, NSW. Technical Aid To The Disabled I welcome the opportunity to introduce a unique charitable organisation called Technical Aid To The Disabled (TAD). Our mission statement is “to improve the quality of life of people with disabilities and those caring for them, though the applica­tion of technology”. We provide three very needed services: (1) the Custom De­signed Aids Service – developing custom-designed aids where commercial equipment is not available, utilising the skills of volunteers; (2) the Information Service – providing information on aids and technology to people with disabilities, rehabilita­tion and engineering professionals; (3) the TAD/Gale Computer Loan Service – providing computers to people who have a disabili­ty. No other organisation in Australia provides these services. The Computer Loan Service is funded by the R. A. Gale Foun­dation. It lends used computers and ancillary equipment to people who have a disability throughout NSW. This service has enabled hundreds of people who have a disability to become more independ­ent with the use of a computer. At present there are over 400 computers out on loan. The service is dependent on businesses and the general public for donations of equipment. We accept donations of whole systems, part systems, components and peripheral devices. It is a wonderful recycling process. Last year alone we received over 2000 pieces of equipment. Clients received a 486 or better IBM or compatible PC. We provide the client with IBM PC DOS 7 and Windows 3.1 (donated by IBM and Microsoft). The software provided includes shareware and special software. This service could not exist without volun­teers. At present there are 32 volunteers who sort, test, repair, package, tutor and deliver the systems. The majority of the volunteers are retired engineers and computer technicians who come to the service once a week and utilise their skills. The Computer Loan Service is always in need of more equip­ment. At present we need IBM compatible 386s and above, mice, monitors (VGA and SVGA), Inkjet printers, sound cards, compon­ents, cables, batteries. Also, in order to extend and improve our service, we need to recruit more volunteers to install, tutor and support clients. We also need people to deliver computers to clients and to collect donated equipment. If people would like more information about the services we provide, or would like to donate their obsolete equipment, or are interested in becoming a volunteer, please contact TADNSW: John Travis, phone (02) 9808 2012; fax (02) 9809 7670. TADVIC: Martie Nash, phone (03) 9853 8655; fax (03) 9853 8098. TADSA: Julie Sullivan, phone (08) 8261 2922; fax (08) 8369 1051. TADWA: John Weedon, phone (08) 9379 3733; fax (08) 9317 2833. TAD Queensland: Maureen Beny, phone (07) 3216 1733. By the way, we are having a TADDAY on 14th November from 9am to 1pm. You can donate and deliver old computer equip­ ment to TAD, 227 Morrison Rd, Ryde NSW or Somerville Road via James Craig Road, Container Terminal, Glebe Island, NSW. J. Trifunovic, TADNSW, Ryde. Donate old computers to schools I write with regard to your editorial in the September 1998 issue of SILICON CHIP. Yes, there is a need and use for 386/486 vintage computers. I am an electronics, systems technology, robotics teacher (secondary) who could use such machines. My primary use would be to run the many programs that run quite happily on Windows 3.1 (but need the colour capabilities) – pro­ g rams such as LEGO Control Lab, PC logo and the many other logo programs, Intellecta and the Softmark interface project from the last issue of SILICON CHIP, to mention but a few. My secondary uses would be to teach computer repair and upgrade skills to students. I believe some local schools are already into this activity. To emphasise the need, this year my school received a dona­tion of 4 x 486 computers, without monitors, mice or keyboards from Alcoa. Setting up the computers with monitors, etc, broke our budget and we may be able to find enough money with P & C help next year to provide software and site licences. I am at a reasonably well-off secondary school in Perth but there are many other schools not so well off, especially the smaller primary schools (and country schools) who could also put such computers into productive classroom educational activities for their students. If you have such computers to give away, please contact your local schools. M. Callaghan, Maddington, WA. October 1998  23 AC Millivoltmeter measures down to one microvolt How do you measure the extremely low noise signals in modern audio equipment? You can’t use a digital multimeter or an oscilloscope because the signals are just too small. That’s why we designed this AC Millivoltmeter which is capable of measuring noise levels down to below one microvolt. 24  Silicon Chip LEFT: simple switching is a big feature of this new AC Millivolt­meter. It will measure audio noise signals down to less than one microvolt (1µV) and shows the result on a digital readout. Pt.1: By JOHN CLARKE A N AC MILLIVOLTMETER is a  vital piece of test equipment if you want to measure the performance of audio equipment. For example, it is used in conjunction with an audio signal generator to measure signal to noise ratio, frequency response, sensitivi­ty, power output, channel separation, crosstalk, signal levels and amplifier gain. To measure some of the latest pieces of audio equipment you will need an AC Millivoltmeter which can measure to very low levels, indeed. For example, a typical CD can have a signal-to-noise ratio of as much as -104dB with respect to 2V. To verify that ratio, you need an instrument that can measure down to 12.6µV. Similarly, our latest Class A Amplifier, which we published in the July & August 1998 issues, has a quoted signal-to-noise ratio of -113dB with respect to its 15W (into 8Ω) output power. This noise measurement corresponds to a reading of just 24.5µV. These are just two typical examples but they demonstrate that if you want to measure modern audio equipment, you need an instrument which can measure down to just a few microvolts. Unfortunately, such instruments are very expensive and can run into many thousands of dollars. Our new AC Millivoltmeter has been designed specifically to address the measurement problems associated with modern audio equipment. It can measure up to 200V RMS which means that you can use it to measure very high power amplifiers, 100V line levels in PA equipment or even signal levels in ultrasonic equipment. At the other end of the scale, for low level noise measurements, it can measure down to less than one microvolt (1µV). And it’s a wide bandwidth instrument too. The upper limit of its frequency response is above 200kHz while at the low fre­quency end, it is -3dB down at 5Hz. Noise filtering When measuring audio noise signals, it is usual to add in some form of filtering, so that you are not measuring wideband noise. For this reason we have provided a 20Hz to 20kHz bandpass filter which rolls off noise frequencies above 20kHz and below 20Hz. This filter is used to obtain the “unweighted” signal-to-noise ratio measurements in audio equipment. Alternatively, “A” weighted measurements are often used for noise measurements. This type of filter is used because it dupli­cates the sensitivity of the ear at very low sound levels and so we obtain a better idea of how loud the noise will sound to us. “A” weighted measurements should always be compared with the un-weighted values. If the “A” weighted measurement is sub­ stantially better then you can usually assume that there is a fair amount of mains hum in the noise. Fig.1 shows the shape of the A-weighting filter used in our AC Millivoltmeter. Features Our prototype is housed in a plastic instrument case which measures 260 x 190 x 80mm. It is powered from the 240VAC 50Hz mains supply and is switched on at the front panel. A Neon indi­ cator within the switch shows when the unit is on. Other controls on the front panel include three rotary switches, for the input attenuator, the filter selection and dB/V modes. There is one toggle switch for selecting “Earthed” or “floating” measurements and there is a potentiometer to set the reference level for ratio measurements such as signal-to-noise or crosstalk. Finally, there is a 31/2-digit LCD panel meter which displays readings in Volts, mV or in dB (decibels). The attenuator switch has six ranges, from 2mV (full scale) to 200V, giving a 20dB level change between successive ranges. Actually, while the measurement ranges are labelled 2mV, 20mV, 200mV and so on, the panel meter is a normal 31/2-digit display so that the maximum readings on these ranges are actually, 1.999, 19.99 and 199.9mV respectively. Measure mode The Measure switch selects between Volts and dB. In dB mode, it displays a relative reading with a 0.1dB resolu­ tion. The dB position is used for noise measurements and indi­cates the number of dBs (decibels) the noise is below a preset level. This preset level is adjusted using the “dB Set Level” control. Two insulated BNC sockets are provided on the front panel, one for Main Features •  Measures AC volts in six ranges from 2mV to 200V RMS (20dB steps) •  Measurement in dB from +40dB to more than -60dB below each range •  31/2-digit LCD panel meter •  Flat, 20Hz to 20kHz and A-weighted filters •  Simplified switching •  Oscilloscope output •  Input overload protection •  Overload indication October 1998  25 AUDIO PRECISION SCFREQRE AMPL(dBr) vs FREQ(Hz) 10.000 01 SEP 98 11:22:36 FLAT 0.0 20Hz-20kHz -10.00 A-WEIGHTING -20.00 -30.00 -40.00 -50.00 10 100 1k 10k 100k 200k Fig.1: this diagram shows the frequency response of the AC Millivoltmeter in Flat mode, for the 20Hz to 20kHz filter and for the A-weighting filter. The Flat response at 200kHz is -0.2dB. These measurements were taken at the moving contact of switch S2. the signal input and one for a buffered version of the measured signal which can be fed to an oscilloscope. Assuming you are measuring a 1V sinewave on the 2V range, then the oscillo­ scope signal will be 100mV RMS or 282mV peak-to-peak. On the dB setting, the same 1V signal is about 100mV peak-to-peak and this will be constant over a wide range of input signals. What this means in practice is that the oscilloscope can be set to 100mV/div and won’t need changing for most measurements. Overload indication When reading Volts, the digital panel meter will indicate overrange automatically by displaying a “1”. An overrange indica­tion suggests that the attenuator should be moved up a range. For dB measurements, there is an overload LED indicator which begins to glow when the Millivoltmeter circuitry is about to clip. Taking measurements with the overload LED alight will give incor­ rect readings. Again, you should switch up to the next range if the overload LED is on. The circuit is protected from serious overload at the input with a fuse. For example, if you apply a 200V signal to the Millivoltmeter when the attenuator is set to a much lower 26  Silicon Chip range, the excessive current flow will blow the fuse. Earth/float switch Earth loops can be a real problem when making audio meas­ urements and this is where the Earth/Float switch comes into play. Normally, the entire Millivoltmeter circuit is not connect­ ed to mains earth but if an oscilloscope is connected, the Milli­voltmeter will be earthed by the scope lead. Or it would be, if not for the Earth/Float switch. In the Float setting, the oscilloscope’s signal lead shield is connected to the metal front panel of the Millivoltmeter and hence to the mains earth. In the Earth setting, the oscilloscope signal earth is connected to the Millivoltmeter signal earth. But if no oscilloscope is connected, the Milli­ voltmeter floats, regardless of the setting of the Earth/Float switch. Block diagram Fig.2 shows the block diagram of our new AC Millivoltmeter. The signal to be measured is applied to the attenuator which effec­tively divides the input signal level to a 2mV (full scale) output regardless of the input. For example, with 200V applied, the attenuator divides the signal down by a factor of 100,000 to obtain 2mV. Following the attenuator is an amplifier with a gain of 34 and this is followed by a second amplifier with a gain of 29.4 and we obtain 2V with a 2mV input signal. This second amplifier has a gain adjustment which is used to calibrate the instrument. The signal from the second amplifier is applied to the input of two filters, a 20Hz-20kHz bandpass filter and the “A” weighting filter which is a passive network followed by an ampli­fier. Filter switch S2 selects either the Flat (unfiltered), 20Hz-20kHz filter or “A” weighted signal and passes it to the Voltage Controlled Amplifier (VCA). Assuming for the moment that the Millivoltmeter is set to read Volts, the control input is grounded with switch S3a and the gain of the VCA is 1. The 2V full scale signal is then sent to both the oscilloscope output divider (divide by 10) and the precision rectifier. Output from the precision rectifier is applied to both the error amplifier and the Volts input for switch S3b. The signal path from the error amplifier is provided for dB measurements. We will bypass this section for the moment and describe how the voltage measurement section operates. Voltage reading When switch S3b is in the Volts position, the DC output from the precision rectifier is applied to inverter IC7a. This simply changes the sign of the DC voltage from -2VDC full scale to +2VDC full scale. The non-inverting input to IC7a is connected to an offset control to compensate for the offsets due to the various op amps in the circuit. Op amp IC7b acts as a level shifter to offset the output of IC7a so it can drive the panel meter input. This is required because the panel meter is designed to operate from a separate battery supply which is floating with respect to the input voltage. We will discuss this aspect later. Finally, the digital panel meter has its decimal points controlled by IC8. This selects the appropriate decimal point depending on the setting of the attenuator, S1a & S1b. dB readings When the Millivoltmeter is selected to read dB, the circuit operates in a Fig.2: block diagram of the AC Millivoltmeter. It uses a liquid crystal display panel meter and a logarithmic VCA (voltage con­trolled amplifier) to minimise range switching. much different manner. The control input to the VCA is disconnected from ground via S3a and instead error amplifier IC6a compares the precision rectifier output against a reference voltage at its non-inverting input and its output drives the control input to the VCA. The VCA is what its name implies; ie, its gain can be controlled via an input voltage. The circuit now operates in a feedback loop whereby the error amplifier controls the gain of the VCA so that its output after rectification and filtering equals the value of the refer­ence. With this control loop the output from the VCA is always the same regardless of input level, provided the VCA can provide sufficient gain. In fact, the gain of the VCA can be adjusted to such a level that noise generated in the Millivoltmeter circuitry can be amplified sufficiently to equal the reference voltage. “Well,” you might say, “so what?” “How does this allow the Milli­volt­ meter to measure in dB?” Well, the VCA that we are using has a special feature that allows the conversion of gain into dB via the control input which is logarithmic. This is ideal since dB scales are also logarithmic and the VCA has a gain specification of 30mV per dB. All we need to do is scale the 30mV/dB to something more useful and we can directly read gain changes in dB on the panel meter. And this is what op amp IC6c does. It divides the input by 3 to obtain 10mV/dB. The amplifier is also offset by the set level potentiometer (VR4), which can adjust the amplifier output level without changing the 10mV/dB calibration. The set level pot is provided so that the dB reading can be set to zero initially so that changes in signal level can be read directly in dB from the meter. Switch S3b selects the dB signal which is applied to IC7a. The following stages then operate as previously described. Overload indication for the dB signal reading is provided using comparator IC6d. This monitors the control voltage of the VCA and drives the overload LED if the VCA is attenuating the signal to such an extent that the input amplifiers are clipping. Circuit description The full circuit for our new AC Millivoltmeter is shown in Fig.3. Its performance relies on low noise op amps IC1 and IC2 and on the special low noise and logarithmic voltage controlled amplifier IC4. The Millivolt­ meter’s noise performance is mainly due to the use of an OP27 or LM627 op amp as the first stage of amplification. These op amps are very quiet in terms of noise and their 700kΩ input impedance allows us to use a relatively high input impedance for the attenuator. The input signal is AC-coupled to the attenuator via a 1µF capacitor. In conjunction with the 110kΩ impedance of the atten­uator, the 1µF capacitor rolls off frequencies below 1.4Hz. The attenuator comprises a series string of resistors with tappings to divide by 10, 100, 1000, 10,000 and 100,000. The step between each range changes by a factor of 10, or a ratio 20dB. Signal from the attenuator passes through to the non-inverting input of IC1 via two 47Ω resistors and fuse F1. The fuse is included as a protection for IC1’s input. If a 100V signal, say, is applied to the input and the attenuator is set to the 2mV position, the excessive voltage swing will cause diodes D1 and D2 to clip the signal to October 1998  27 28  Silicon Chip Fig.3: the circuit can be broken up into a number of sections. First, there is the attenuator followed by two op amps with a combined gain of 1000. These are followed by the filter stages and the VCA (IC4) which greatly simplifies the switching re­quired. There is a precision rectifier which provides the DC signal measured by the panel meter. +15.6V and -15.6V and if the current is high enough, the fuse will blow. The diodes are fast recovery types to ensure that their capacitance does not affect the frequency response of the ampli­fier under normal operating conditions. IC1 provides a gain of 34, as set by the 3.3kΩ and 100Ω feedback resistors connected to pin 2. A 39pF capacitor in paral­lel with the 3.3kΩ feedback resistor rolls off the high frequency response above 1.2MHz. This prevents oscillation of the amplifier but the rolloff frequency is sufficiently high not to affect the frequency response of the amplifier up to 200kHz. Op amp IC2a functions in a similar manner to IC1 with the only difference being the gain adjustment provided by trimpot VR1. This is adjusted so that the combined gain of IC1 and IC2a is 1000. This is part of the calibration procedure. Filter stages IC2a drives the filters. Op amp IC3a is a 20kHz low pass filter followed by IC3b as a 20Hz high pass filter and they combine to provide the required 20Hz to 20kHz bandpass response. The filters are Sallen-Key, alternatively called Voltage Con­trolled Voltage Source (VCVS) types. Both filters are set with a gain of 1 in the passband. The “A” weighting filter is a passive RC type which has a loss of 3dB at 1kHz which is compensated for by op amp IC2b to give an overall gain of unity (1) at 1kHz. Switch S2 selects the filter output and it feeds IC4, the VCA, via a 10µF bipolar capacitor. In conjunction with the 18kΩ input resistor, the 10µF capacitor provides a low frequency rolloff (-3dB) at 0.88Hz. IC4 is an SSM2018 VCA made by Analog Devices Inc. It has a dynamic range of 117dB, .006% distortion at 1kHz and unity gain, and a control range of 140dB. Its gain is varied by October 1998  29 Specifications Input impedance.............................................................110kΩ unbalanced AC reading accuracy...............................................................................2% dB linearity.............................................................. 0.5dB over 60dB range Flat frequency response........................ -3dB at 5Hz and -0.2dB at 200kHz 20Hz to 20kHz filter response...............................-3dB at 21Hz and 21kHz A-weighting response.................................................................... see Fig.1 Noise floor................................ 64dB below 1mV with 20Hz to 20kHz filter; 68dB below 1mV with A-weighting Oscilloscope output................................... 200mV at full scale volt reading; nomi­nal 100mV P-P on dB setting the voltage at pin 11. The 100kΩ resistor between pin 12 and the +15V supply rail sets the bias level for the output at pin 14. This bias can be selected for class A or class B operation. Class A gives better distortion while class B provides better noise perfor­mance. We opted for class B operation in order to obtain the better noise performance. The output of IC4, pin 14, connects to the precision recti­ fier and the oscilloscope attenuator. The attenuator consists of the 62kΩ and 6.8kΩ resistors which divide the signal by a factor of about 10. The 1MΩ loading of the oscilloscope input will not affect the signal level or the performance. Precision rectifier IC5a & IC5b are connected as the precision rectifi­ er. These op amps effectively remove the 0.6V forward voltage drop of the diodes so that very small signals can be rectified without error. For positive signals, the output of IC5a goes low to reverse bias diode D3 which effectively disconnects it (IC5a) from the summing junction of IC5b. Diode D4 and the 3.3kΩ resistor between pins 6 and 7 of IC5a limit the negative swing of IC5a. So posi­tive signals are fed to IC5b via the 20kΩ resistor and its 22kΩ feedback resistor gives a positive signal gain of -1.1. For negative signals, diode D3 conducts and IC5a acts as an inverting amplifier with a gain of -1. This gain is set by the 3.3kΩ input resistor to pin 6 and the 3.3kΩ resistor from pin 6 to the cathode of D3. This inverted signal is summed in IC5b via the 10kΩ 30  Silicon Chip resistor from the cathode of D3 to the pin 2 input. Negative signals are also fed to pin 2 of IC5b via the 20kΩ resistor. Since the signals across the 20kΩ resistor and the 10kΩ resistor are equal but exactly opposite in value and the 10kΩ resistor is exactly half of 20kΩ, the net result is a negative signal gain of -1.1. So positive signals applied to the full wave rectifier will have a gain of -1.1 and for negative signals the gain is 1.1. Thus the rectifier output goes negative for both positive and negative input signals. A 10µF capacitor across the 22kΩ feedback resistor of IC5b results in a negative DC voltage output at pin 1 which is propor­tional to the input signal. Since IC5b has a gain of 1.1, the DC output is actually proportional to the RMS value of the signal, provided the input is a sinewave. (Form factor of a sinewave is 1.1; RMS/average = 1.1). The precision rectifier output is fed to the voltage meas­uring circuitry via switch S3b and to the error amplifier, IC6a. This has a gain of -100 and compares the rectified signal against a reference voltage at its pin 3 non-inverting input. The output of IC6a then drives pin 11 of IC4. The circuit operates in a feedback arrangement whereby the gain of IC4 is continually adjusted by IC5a so that the voltage from the full wave rectifier equals the reference voltage. The reference voltage at IC6a’s pin 3 input is derived from REF1, an LM336-2.5V. Trimpot VR2 sets the maximum reading for signal-to-noise ratio of the Millivoltmeter. REF1’s -2.49V is inverted with op amp IC6b to give +2.49V. This positive and negative reference and trimpot VR4 provides an offset adjustment for op amp IC6c, to zero the dB reading on the panel meter. The control input to IC4 at pin 11 is logarithmic at 30mV/dB and we use this to provide the dB measurement mode. IC6c attenuates the output from IC6a to 10mV/dB. Overload indication The voltage applied to pin 11 of IC4 is monitored by com­parator IC6d. If the voltage reaches about +0.8V as set by the 18kΩ and 1kΩ divider at pin 5, the output of IC6d goes low to drive LED1. A voltage above 0.8V means that the VCA is attenuat­ing the signal to such an extent that the input stages are over­loading. Switch S3a connects pin 11 of IC4 to ground (0V) for Volts measurements. This fixes the gain of IC4 at 1. S3b is the second pole of switch S3. It selects between the output of the full-wave rectifier for Volts measurement and the output of IC6c for dB readings. The signal from the wiper of S3b is fed to inverter IC7a. This inverts the voltage and dB readings so that the panel meter will show the correct polarity for dB measurements. VR5 provides offset adjustment for IC7a to ensure that the digital panel meter shows 0V when no signal is present. It is adjusted to compensate for the offsets produced in IC5a, IC5b, IC7a and IC7b. IC7b level shifts the output of IC7a to allow for the input offset of the digital panel meter. This occurs because the panel meter runs from a separate supply rail. Digital panel meter The digital panel meter has differential inputs IN(-) and IN(+) and requires a 9V power supply between its Batt (+) and Batt (-) terminals. Its IN(-) is fixed at 2.8V below Batt (+). This 2.8V is the reference for the meter so that it reads accu­rately. The only way to use the panel meter when a separate 9V supply is not available is to make sure that IN(-) is kept at 2.8V below Batt (+). We do this by pulling IN(-) up via a 1kΩ resistor. This feeds current into the internal reference of the panel meter which allows us to draw some current out without starving it. The 10kΩ resistor All the circuitry is mounted on two PC boards with virtually no off-board wiring. We’ll give the full construction details in next month’s issue. from pin 5 of IC7b to the IN(-) input biases the op amp output to the IN(-) voltage. When the output of IC7a is at 0V, then the IN(+) is at 2.8V below Batt(+). Since this is the same as IN(-) the meter will read zero. Thus the circuit fulfils the requirement, keeping the IN(-) input as is but applying an offset to the IN(+) input which is equal to the voltage at IN(-). Pin 7 of IC7b provides 2V at full scale for the meter but since the panel meter requires 200mV at full scale this voltage is divided down by a factor of 10 by the 10kΩ and parallel con­nected 910kΩ and 100kΩ resistors. Decimal point switching Switches S1b and S3c and IC8 provide decimal point switch­ing for the panel meter. S1b is the second pole of the attenuator switch S1a. Position 1 (2mV) and position 4 (2V) are connected together, position 2 (20mV) and position 5 (20V) are connected together and position 3 (200mV) and position 6 (200V) are also connected together. When S3c is in the Volts position, the wiper of S1b can apply 9V to either the A, B or C input of IC8. The other inputs are held low via the 10kΩ pulldown resistors. IC8 can be regarded as a 3-pole 2-way switch controlled by the A, B and C inputs. When the A input is low, pin 12 (ax) is connected to “a” at pin 14. Similarly, when B is low, then bx connects to “b”. Finally, when C is low, cx connects “c”. If one input is high, then the “y” terminal for that switch pole is connected with the “x” wiper open. For example, if we have the A input high, the ay terminal at pin 13 connects to “a”. The ay, by and cy terminals all connect to the inverted backplane signal (BP-bar) from the panel meter, while the ax, bx and cx terminals connect to the backplane (BP). A decimal point can be switched on by connecting it to the BP-bar signal while the BP will switch it off. While switch S1b controls the decimal point selection when the meter in Volts mode, it is effectively switched out of cir­cuit when S3c is changed to the dB position. In this setting, the A input of IC8 is pulled high to connect the BP-bar signal to the “a” terminal at pin 14. This switches on decimal point DP3. Power supply Power for the Millivoltmeter is derived from the 240VAC mains via a 30V centre-tapped transformer supply which is recti­ fied with diodes D5-D8 and filtered with 1000µF capacitors to provide a nominal plus and minus 21V DC supply. 3-terminal regu­lators REG1 and REG2 provide regulated +15V and -15V supplies for the op amp circuitry. The 9V rail for the panel meter and IC8 is derived from the +15V via a 470Ω resistor and zener diode ZD1 which is bypassed with a 100µF capacitor. Next month, we will describe the construction, setting up and testing SC of the AC Millivoltmeter. October 1998  31 Feeling stressed or calm? The Stress-O-Meter can help you devel­op methods to lower your stress rating. By RICK WALTERS Don’t blow your stack. Keep your temper in check with the Stress-O-Meter This little fun project interfaces with the games port on your PC. You put your finger in the side of the case and it measures your pulse rate. It also measures your skin resistance. The computer then calculates your “stress level” and displays it on the screen. Are you highly stressed? Find out with our Stress Meter. 32  Silicon Chip S TRESS IS QUITE INSIDIOUS. It builds up gradually and you tend not to notice it but the people around you certainly do. They see when you’re about to bite the carpet or crawl up the wall. If you can measure stress and then calm yourself down, it will make you a more pleasant person and maybe you will live a lot longer. Of course no machine can calm you down directly. It takes medicine, or your awareness of the problem, to do something about it. Once you have measured your stress level, you can begin to carry out procedures to reduce it. These are outlined later. The Stress-O-Meter consists of a small black box (well, it is black, isn’t it?) containing the electronics, which is connected to the printer and games ports of a PC-compatible computer. The PC must have a VGA monitor. Sensors to monitor your skin resistance are connected to two fingers of one hand and your pulse rate is read from your other hand’s index finger. This information is fed to the computer by the black box and your pulse rate is calculated and displayed, along with the stress value. The latter value is updated every 10 seconds, to allow you to monitor your progress. An audible tone is generated which rises and falls in frequency as your stress level increases or reduces. Circuit details The circuit of the stress meter is shown in Fig.1 and while it does not look too complicated, there are a number of concepts to be covered. Two connections must be made from the black box to the PC, via the parallel port and the games port. Power for the circuit is taken from the games port and it also receives the signal for the skin resistance. The circuit to measure pulse rate consists of IRLED1, PD1, IC1 and IC2a. IRLED1 is an infrared light emitting Fig.1: the circuit measures your pulse rate and Galvanic skin resistance (GSR) and this information is processed by your computer to produce a stress display. October 1998  33 Interfacing To The PC Games Port The games port on a PC has provision for two joysticks and four buttons. Each joystick consists of two variable resistors, one handling the X-direction, the other handling the Y-direction. In this project we only use the X-direction input, identified in Basic as STICK(0). When this resistor is varied, the count (generated by a quad 555 timer on the games card) varies from 0 to 255. If you have an older machine this will probably be the case. Some of the newer sound cards are fitted with a MIDI port, which appears to double as a games port, and they seem to only be capable of 100 counts. You can test this by wiring a 250kΩ potentiometer across pins 1 & 3 of the 15-way D-type games plug and diode, which is positioned above PD1, a photodiode which is sensitive to infrared radiation. When your finger is placed between these devices it reduces the light falling on the photodiode but as the blood pulses through your finger, the amount of transmitted light varies. from Basic (GW or Q) type 10 PRINT STICK(0): GOTO 10 then running the program. Pressing and holding down the Ctrl (control) key while pressing the Break/Pause key will get you out of this loop. As you vary the resistance, the value of STICK(0) will change from zero at minimum resistance to some maximum value before jumping back to zero. A value slightly less than the resistance that gives the maximum count should be used in paral­lel with the LDR instead of the 240kΩ shown on the circuit. If you cannot reach 200+ counts the card is not suitable for this project. Older-style games cards should work and can probably be purchased from your local computer store and from flea markets. This change is amplified by op amps IC1a and IC1b which have an overall DC gain of unity, which means that the DC voltage at pin 7 of IC1b will be the same as that across PD1. This is due to the inclusion of the 10µF capacitors in series with the 4.7kΩ feedback resistors. Your (typical) pulse rate is around 80 per minute which is a frequency of 1.33Hz, so we need an amplifier with lots of gain at this frequency to amplify the small variations detected by PD1. The AC gain is 47.8 (220kΩ/4.7kΩ +1) for IC1a and the same for IC1b, giving a total AC gain of 2286. Now you can understand why we needed a DC gain of unity. If the DC gain was also 2286, a level of only +2.2mV at PD1 would cause the output from IC1b to sit at +5V (2.2mV x 2286) and it could never amplify any input signal bigger than, say, 1.5mV. The output of IC1b is fed to the non-inverting input of IC2a which is used as a comparator. Its inverting input, pin 2, is held at +4V by the 47kΩ and 12kΩ resistors. When pin 7 of IC1b exceeds this level, pin 1 of IC2a will swing to the +5V supply. The 2.2MΩ positive feedback resistor between pins 1 & 3 squares up this input signal; ie, makes the slow input transition into a rapid output transition. This pulse signal is fed to the PC via the printer port and then processed by the software. The software (subroutine 3000) records the period between six consecutive pulses and computes the average. This value is then displayed on-screen beneath the stress chart, as your aver­age pulse rate. Your initial stress value is twice your measured pulse rate. So if your pulse rate is 90, the stress value is 180. Galvanic skin resistance (GSR) Fig.2: these scope waveforms show the process of turning the square wave output from IC4 into an approximate sinewave. The top trace shows the square wave output at pin 3 of IC4, the middle trace shows the triangle waveform and the bottom waveform is the sinewave approximation. 34  Silicon Chip Your skin resistance, usually referred to as GSR (galvanic skin resistance), is then measured by IC2b, which is connected as a unity gain inverter. The GSR electrodes are connected to two fingers on one hand, as mentioned earlier, and the resistance across them forms a voltage divider with the 100kΩ resistor connected to the +5V rail. For convenience, let’s say your GSR is 100kΩ. In this case, the voltage fed to IC1b will be +2.5V. Pin 5 of IC2b is connected to the output of IC3, an octal buffer. This is configured as a digital-to-analog converter (DAC). IC3’s inputs, D0-D7, are connected to the corresponding parallel port data lines which are capable of outputs between 0 and 255 (8 bits, 28 = 256). The output bits of IC3 are summed in what is called an R.2R ladder. A digital input of zero will give an output of 0V, while an input of 255 will give an output of +5V, with intermediate digital inputs giving corresponding analog outputs. Some readers may wonder why the R.2R ladder could not simply be connected to the parallel port outputs directly. Well, it could have but because the logic levels from the computer may not be exactly 0V (low) and may be considerably less than +5V (high), it is better to use IC3 to ensure that the DAC outputs do range between 0V and +5V. The software now begins to output counts (subroutine 4000) until the output of the DAC is close to the GSR voltage. In other words, the GSR voltage at pin 6 of IC2 will be equalled by the DAC output voltage at pin 5. If you look at the circuit you will observe that the output of IC2b (pin 7) is fed to IC4 and a 47kΩ resistor which feeds IC2b’s output voltage to LED1. This LED shines on a light dependent resistor, LDR1. Its resistance depends on the amount of light falling on it. In darkness, its resistance is around 10MΩ and with a bright light it is around 300Ω. A 240kΩ resistor is wired in parallel with it. What actually happens is that the DAC output is increased until the STICK(0) reading equals the stress value and this STICK(0) value is then used as the reference for all future stress readings. (If you are not familiar with STICK(0) read the GAMES PORT panel). Audible indicator As well as the screen display, the circuit also provides a tone output which reflects the value of your skin resistance. The output of IC2b is proportional to your skin resistance and it is used to control the VCO section of IC4 which is a 74HC4046 phase locked loop. We are only using the VCO (Voltage Controlled Oscil­lator) of this chip. A VCO varies its output frequency in sym­pathy with its input voltage; so the higher the input voltage, the higher the frequency. The operating frequency is set by the .039µF capacitor between pins 6 & 7 and the value of the resistance from pin 11 to ground. Hence, you can use VR1 to set to the initial frequency to suit. Some people like it high and others like it low. The output frequency from pins 3 & 4 of IC4 is a square wave and the last Parts List 1 PC board, code 07111981, 107 x 77mm 1 plastic box, 150 x 90 x 50mm 1 57mm 8Ω speaker 1 25-way “D” male PC mount rightangle connector (CN1) 1 15-way “D” male solder-pin connector (CN2) 3 8-pin IC sockets (IC1,2,5) if required 1 16-pin IC socket (IC4) if required 1 20-pin IC socket (IC3) if required 2 knobs to suit VR1,VR2 2 4mm banana sockets 2 4mm banana plugs 12 PC stakes 4 M3 x 6mm countersunk screws 4 M3 x 6mm round-head screws 4 M3 x 20mm tapped spacers 4 small adhesive rubber feet 250 x 18mm self-adhesive Velcro 2 aluminium foil strips, 210mm x 36mm 1 100kΩ linear potentiometer (VR1) 1 1kΩ logarithmic potentiometer (VR2) 1 floppy disc with Stress.Bas software Semiconductors 1 TL072 dual op amp (IC1) 1 LM358 op amp (IC2) 1 74HC573 octal buffer (IC3) 1 74HC4046 phase locked loop (IC4) 1 LM386 audio amplifier (IC5) thing you need if you are stressed is to listen to a harsh square wave. The 4.7kΩ and .068µF capacitor shape the square wave to an exponential triangular waveform, which is then shaped to a sinewave approximation by the 4.7kΩ resistor and diodes D1 and D2. The filtering process is shown in the digital scope waveforms of Fig.2. The top trace shows the square wave output at pin 3 of IC4, the middle trace shows the triangle waveform and the bottom waveform is the sinewave approximation. No, it isn’t perfect but it’s definitely better than listening to a square wave. The sinewave is then fed to VR2 1 LTE4208 IR LED, DSE Cat Z-3235 or equivalent (IRLED1) 1 5mm red LED (LED1) 1 LTR536A photodiode, DSE Cat Z-1956 or equivalent (PD1) 1 ORP12 light dependent resistor, Jaycar Cat RD-3480 or equival­ent (LDR1) 2 1N914 small signal diodes (D1,D2) Capacitors 2 100µF 16VW PC electrolytic 3 10µF 16VW RBLL (low leakage) PC electrolytic 2 0.22µF MKT polyester 4 0.1µF MKT polyester 1 .068µF MKT polyester 1 .039µF MKT polyester Resistors (0.25W, 1%) 1 3.3MΩ 9 20kΩ 1 2.2MΩ 1 18kΩ 1 1.2MΩ 1 15kΩ 2 1MΩ 1 12kΩ 1 240kΩ 7 10kΩ 2 220kΩ 4 4.7kΩ 4 100kΩ 1 180Ω 2 47kΩ 1 10Ω Miscellaneous Tinned copper wire, hookup wire Software Availability The software for the Stress-O-Meter (Stress.Bas) is available on a 3.5inch floppy disc from Silicon Chip Publications. Cost: $10 (includes p+p). which sets the signal level to the power amplifier IC5. This can be turned right down if you don’t need it or don’t like it. As your stress level reduces, the frequency you initially set with VR1 will lower (your GSR will increase as you relax causing the output of IC2b to fall), although this may be at such a slow rate that it will be imperceptible. Let’s now summarise what the circuit and the software actu­ally do. First, the circuit measures your pulse rate and displays it on-screen together with your stress value which is twice the pulse rate. From then on, your pulse rate is ignored unless you restart the October 1998  35 Fig.3: follow this diagram when assembling the PC board and wiring up the case. The two leads marked CN2 go to the D-connector for the games port on your computer. program by pressing “R” (without the inverted commas) on the keyboard. The circuit then measures your skin resistance and the DAC output counts up to match that value. IC2b’s output 36  Silicon Chip is used to vary the operating frequency of the VCO in IC4. You initially set IC4’s frequency with trimpot VR1 and as your skin resistance changes it will cause the frequency to vary from the initial measurement but usually only by a small amount. If you have managed to reduce your feelings of stress, by slow deep breathing etc, your lowered heart rate This photo shows how the assembled PC board is mounted on the lid of the case, using spacers and machine screws. The speaker can be glued to the lid using silicone sealant or contact adhesive. will only be taken into account if the program is restarted. Software subroutines The software is segmented into small subroutines which should make it easy to follow. Subroutine 1000 defines a function to centre text, a function to clear to the end of the current line, the different screen colours which can be displayed and various values which are used in the program. By defining all these values here this is the only place where a value has to be altered if a change is needed, instead of going through the whole program searching for and altering each value. If you wish to use the second printer port (LPT2) change line 1600. Lines 1230 to 1360 draw the opening screen and write the SILICON CHIP logo. Subroutine 2000 draws the coloured stress blocks. As we have previously mentioned subroutine 3000 computes your pulse rate. Because of the long time constants used in the pulse amplifiers (220kΩ + 4.7kΩ and 10µF), it takes several seconds for the circuit to stabilise. Normally, with your finger removed, IRLED1 is shining on the photodiode and the output voltage at pin 1 of IC2a is +5V, causing both 10µF electrolytics to charge to this voltage. When you insert your finger between IRLED1 and the photodiode, the output voltage drops close to 0V but it takes time for the elec­trolytics to discharge to this level. As they discharge, the pulse reading becomes erratic. When you press the spacebar, you hear the beep which is sounded each time the comparator output swings high. When these beeps become regular you press the spacebar again and the next five pulse durations are recorded. These are averaged on line 3230 and the value is written to the screen on line 3250. In subroutine 4000, the digital output to the printer port is ramped up, starting from count 60. This is done because IC2b is not operating in a linear mode at this time. Input pin 6 has +2.5V applied and pin 5 is at ground (0V). Since IC2 is operating without a negative supply, the first 50-60 counts will not change the output. As we ramp up the digital count, the voltage applied to pin 5 of IC2b (and thus the output voltage at pin 7) is also increas­ing. This voltage supplies Resistor Colour Codes  No.   1   1   1   2   1   2   4   2   9   1   1   1   7   4   1   1 Value 3.3MΩ 2.2MΩ 1.2MΩ 1MΩ 240kΩ 220kΩ 100kΩ 47kΩ 20kΩ 18kΩ 15kΩ 12kΩ 10kΩ 4.7kΩ 180Ω 10Ω 4-Band Code (1%) orange orange green brown red red green brown brown red green brown brown black green brown red yellow yellow brown red red yellow brown brown black yellow brown yellow violet orange brown red black orange brown brown grey orange brown brown green orange brown brown red orange brown brown black orange brown yellow violet red brown brown grey brown brown brown black black brown 5-Band Code (1%) orange orange black yellow brown red red black yellow brown brown red black yellow brown brown black black yellow brown red yellow black orange brown red red black orange brown brown black black orange brown yellow violet black red brown red black black red brown brown grey black red brown brown green black red brown brown red black red brown brown black black red brown yellow violet black brown brown brown grey black black brown brown black black gold brown October 1998  37 a current to LED1 which will illu­ minate LDR1, causing its resistance to fall and consequently the count from STICK(0) to fall. Read the panel titled “Interfacing To The Games Port” if you don’t understand this. Once it reaches PCAL, the pulse calibration count, this subroutine is completed. The final step is subroutine 5000 which continuously takes the average of 10 GSR readings over 10 seconds and moves the indicator on the stress graph to reflect these GSR changes. The output from the games card is not rock steady, changing by a couple of counts up and down with a fixed resistance input. So a variation of one or two counts in successive stress readings should be ignored. What matters is the overall trend. Putting it together Most of the parts, including the DB25M connector, are mounted on the PC board. Note the plastic tube fitted to LED1 and LDR1 (see text). All the circuitry for the StressO-Meter and the 25-pin D-socket is mounted on a PC board measuring 117 x 77mm and coded 07111981. The PC board should be inspected for shorts between parallel tracks or where tracks run between IC pads. Also check for any hairline open circuit tracks or undrilled holes. Any defects should be fixed before installing the parts. The complete wiring diagram for the Fig.4: you need some self-adhesive Velcro and aluminium cooking foil to make the GSR electrodes. 38  Silicon Chip How To Enter Basic Listings Listings are a series of instructions which Basic exe­cutes, one by one. As it functions by interpreting key words any misspelt or illegal instructions will cause the program to stop or malfunction. Any text that appears after an apostrophe (‘) is only there to help you (and me) understand the logic behind the program; it is ignored by Basic. There is no need to type in the apostrophe or any following text to get the program working. QBasic If you have QBasic, you can load it by typing QBASIC. The listing can now be typed in and saved by using the mouse and clicking on FILE then SAVE, naming it STRESS. As you enter the program, any errors Stress-O-Meter, includ­ing the wiring inside the plastic box, is shown in Fig.3. When assembling the board, fit the link, the PC stakes and then the diodes and resistors. Then insert the capacitors, followed by the ICs. IC sockets are optional. Make sure that the electrolytic capacitors are installed with correct polarity. A reversed electrolytic can work for a while, gradually getting hotter and hotter, until it explodes with a bang. To have the capacitor’s case whistle past your ear or worse, hit you in the eye, is no fun. The smell also lingers for quite a while. More often the electrolytic doesn’t explode but expires with a stream of hot, smelly electrolyte. Of course, I have NEVER installed one backwards, it’s only other people telling me what happens that lets me pass on this information. (That’s his story! ...Ed.) A short length of black heatshrink sleeving should be slid over LED1 and LDR1 should be pushed up close to the LED. A piece of Blu-Tak or similar material should then be used to block both ends of the sleeve, as any ambient light falling on the LDR will upset the testing. Mount the photodiode (PD1) on the copper side of the PC board so that its leads sit against the PC board and it protrudes through the large hole with will be flagged and you will be able to correct them. QBasic will also ask you to save the file before you exit the program. Most line numbers are unnecessary; the only ones required are the first lines of each subroutine (1030,2030, etc). GW Basic If you have GW Basic but have not used it before, or don’t have a directory named BAS, then from the root directory C:\ type MD BAS (then press Enter). This will create a directory named BAS. Now type CD BAS and you will be in the basic directory you have just created. Now type GWBASIC. If you get the message ‘file not found’ it means that your DOS directory, where GW Basic should be located, its face parallel to the board. Don’t poke its leads through the PC board holes, as your finger can touch them when you are taking a pulse measurement; just solder them to the pads. If you must poke them through, cut them off flush with the PC board and cover them with a piece of electrical tape. Sleeve the lead that is not connected to earth so that it doesn’t short out to the track it runs across. IRLED1 can be mounted so that it is above your finger or, as the photograph is not in your path statement. The simplest solution is to type COPY \DOS\ GWBASIC.EXE and then press Enter. This will copy it to the BAS directory. To run a GW Basic program, change to the BAS directory (CD BAS) then type GWBASIC NAME, in this case GWBASIC STRESS in upper or lower case. If you are entering the program and get sick of typing, just type RUN 5 (then press Enter) to save what you have typed in so far (as long as you have typed in line 5). The comments which applied to QBasic regarding apostrophes and subsequent text also apply here. You can exit from GW Basic by typing System and then pressing Enter. shows, slightly offset so that the light is entering your finger at around 30 degrees from the vertical. You may need to experiment with the position while listening to the beeps in the testing phase. With the PC board finished, you can cut the hole in the case for the 25-way D-socket and drill all the holes (one for your finger, one for the games port wires, two for banana sockets and two for the potentiometers). Don’t mount the PC board in the case just yet, as it This view shows the finished GSR electrodes. They are attached to two of your fingers so that your skin resistance can be monitored. October 1998  39 You poke your finger into the hole in the side of the case to take your pulse measurement and fit the electrodes to your fin­gers to take your skin resistance. is far easier to do the testing with the PC board on the bench. Connect the two wires to the 15-pin D-connector and connect them and the speaker to their respective stakes. Connect a 100kΩ resistor (the spare in the kit) across the GSR terminals and connect IRLED1 as shown so that it sits above and to the side of PD1. Testing Before you can begin any serious tests you will have to type in the software listing, unless you get it with a kit or buy it from SILICON CHIP. If you don’t know how to enter the listing, refer to the panel in this article. If you use QBasic, start at line 10. If you use GW Basic and get sick of typing, just type RUN 5 (then press Enter) from the command line to save what you have entered so far. Once the software is entered and saved, and with the cor­rect printer port being identified in line 1600, plug the 15-way connector into the joystick port and using the 25-way cable connect the male end to the printer port and the other end to the PC board. Turn the computer on and run the program. The SILICON CHIP logo and STRESSO-METER should appear along with the four coloured stress boxes. You should be prompted at the bottom of the screen to connect the GSR electrodes, insert your index finger (into the hole so your pulse rate can be checked) and press the spacebar to begin. Connect GSR leads, insert index finger then press spacebar to begin. Fig.5: this is the opening screen that appears when you load the software. The instructions at the bottom of the screen tell you what to do. 40  Silicon Chip As we haven’t connected the GSR electrodes yet, place your finger between the IRLED and photodiode then press the spacebar. The next message should tell you to wait until the beeps become regular, then press the spacebar again. You should then hear six or seven more beeps before the Pulse average reading appears below the stress levels, followed by the message ‘Now calculating Stress level’. A second or so later the stress bar should appear in one of the boxes. If you experience this sequence, then your software and hardware are OK. Software errors will be flagged and displayed to allow you to fix them. The cause of most problems will be either erratic or no beeps, which indicates that the hardware cannot detect your pulse beat. This doesn’t necessarily mean that you don’t have a pulse. If you have very thick or weathered skin, the infrared light may not be able to penetrate sufficiently. Measure the voltage at the output, pin 7, of IC1b, preferably with an analog multimeter (or oscilloscope). It should sit at 0V and move fairly rapidly to +5V then fall slowly towards ground. Reducing the 180Ω resistor feeding IRLED1 to 100Ω should give sufficient extra output to make the beeps regular. Now is obviously the time to make the GSR leads. The dia­gram of Fig.4 should give you sufficient details. Use adhesive-backed Velcro which can be obtained from your local haberdashery store. The aluminium foil came from the kitchen. The other problem you may encounter is your normal skin resistance. The Remove finger. Press spacebar to quit, R to run again. Fig.6: when you run the program, your relative stress level is displayed as a number and indicated on the colour chart. Software Listing For Stress.Bas 1 GOTO 10 5 SAVE “C:\bas\stress”,A ‘Save file on C drive in ASCII format 6 ‘Don’t enter lines 1-7 for QBasic 7 END 10 REM STRESS.BAS V1.0 R.W. 17/05/98 20 GOSUB 1030 ‘Initialise 30 GOSUB 2030 ‘Draw stress screen 40 GOSUB 2130 ‘Draw CALM block 50 GOSUB 2230 ‘Draw NORMAL block 60 GOSUB 2330 ‘Draw MEDIUM block 70 GOSUB 2430 ‘Draw HIGH block 80 K$ = INPUT$(1) 90 GOSUB 3030 ‘Read pulse rate to determine stress 100 GOSUB 4030 ‘Set D/A voltage to suitable value 110 GOSUB 5030 ‘Read GSR and show change 120 IF K< > “R” AND K < > “r” THEN 999 130 CLEAR: GOTO 20 999 CLS: SYSTEM 1000 ‘*********************** 1010 ‘Initialisation routine. 1020 ‘*********************** 1030 KEY OFF: CLS: DEFINT A-Z: DEFSTR K: DEFSNG P,T 1031 ‘A to Z integers, K is a string, P & T single precision 1040 DEF FNCENTRE$(M$) = SPACE$((79 - LEN(M$))/2) + M$ 1050 DEF FNCEOL$ = STRING$(79 - POS(Q),” “) 1060 KSP = CHR$(32) 1070 BLACK = 0: BLUEDEEP = 1: GREEN = 2: CYAN = 3: RED = 4:MAGENTA = 5 1080 BROWN = 6: WHITE = 7: GREY = 8: LTBLUE = 9: LTGREEN = 10 1090 LTCYAN = 11: LTRED = 12: LTMAGENTA = 13: YELLOW = 14: HIWHITE = 15 1100 PX = 79: PY = 140 ‘box top corner 1110 VX = 402: VY = 52 ‘box size see line 2150 1120 CX = 100: NX = 100: MX = 100: HX = 100 ‘total must equal VX - 2 1130 CST = PX + 1: CEND = CST + CX: NST = CEND + 1: NEND = NST + NX -1 1131 ‘Calm start, Calm end, Normal start, Normal end 1140 MST = NEND + 1: MEND = MST + MX - 1: HST = MEND + 1:HEND = HST + HX - 1 1141 ‘Some stress start, Some stress end, High stress start, High end 1150 ISW = 440 ‘define stress line width (10 lines) 1151 ‘Some stress start, Some stress end, High stress start, High end 1160 PORTA = &H378: PORTB = PORTA + 1 ‘LPT1, change to &H278 for LPT2 1170 MASK = 8 ‘00001000B mask all but bit 3 1180 OUT PORTA,0 ‘set D/A output to zero 1200 ‘********************* 1210 ‘Write opening screen. 1220 ‘********************* 1230 SCREEN 9: COLOR LTBLUE,LTCYAN 1240 X = 100: Y = 25: PSET (X,Y) ‘write SC to screen 1250 DRAW “u12;h12;l48;g12;d24;f12;r32;d24;l24;u12;l24;d12;f12;r48” 1260 PSET (X,Y): DRAW “l24;u12;l24;d24;r32;f12;d24;g12” 1270 PAINT (X-20,Y-5) ‘draw & fill S 1280 PSET (X+90,Y) 1290 DRAW “u12;h12;l48;g12;d60;f12;r48;e12;u12;l24;d12;l24;u60;r24;d12;r24” 1300 PAINT (X+80,Y-5) ‘draw & fill C 1310 LOCATE 3,35: PRINT “Silicon Chip”; 1320 COLOR RED,LTCYAN 1330 LOCATE 5,35: PRINT “STRESS-O-METER”; 1340 COLOR YELLOW,LTCYAN 1350 LOCATE 25,1 1360 PRINT FNCENTRE$(“Connect GSR leads, insert index finger then press spacebar to begin.”); 1399 RETURN 2000 ‘****************** 2010 ‘Draw stress meter. 2020 ‘****************** 2030 COLOR WHITE,BLACK 2040 PSET (PX,PY) 2050 DRAW”d52; r402; u52; l402;” ‘must be values VY & VX line 1430 2060 LOCATE 10,16: PRINT “Calm” 2070 LOCATE 10,26: PRINT “Normal” 2080 LOCATE 10,36: PRINT “Some tension” 2090 LOCATE 10,51: PRINT “Stressed” 2099 RETURN 2100 ‘**************** 2110 ‘Draw calm block. 2120 ‘**************** 2130 COLOR GREEN,BLACK 2140 FOR A = CST TO CEND ‘calm start to calm end 2150 LINE (A,PY+1) - (A,PY+VY-1) 2160 NEXT 2199 RETURN 2200 ‘************************ 2210 ‘Draw normal color block. 2220 ‘************************ 2230 COLOR LTBLUE,BLACK 2240 FOR A = NST TO NEND ‘normal start to end 2250 LINE (A,PY+1) - (A,PY+VY-1) 2260 NEXT 2299 RETURN 2300 ‘***************************** 2310 ‘Draw some-stress color block. 2320 ‘***************************** 2330 COLOR YELLOW,BLACK 2340 FOR A = MST TO MEND ‘medium start to end 2350 LINE (A,PY+1) - (A,PY+VY-1) 2360 NEXT 2399 RETURN 2400 ‘***************************** 2410 ‘Draw high stress color block. 2420 ‘***************************** 2430 COLOR RED,BLACK 2440 FOR A = HST TO HEND ‘high start to end 2450 LINE (A,PY+1) - (A,PY+VY-1) 2460 NEXT 2470 LOCATE 16,1 2499 RETURN 2500 ‘*************************** 2510 ‘Draw stress indicator line. 2520 ‘*************************** 2530 COLOR GREY,BLACK 2540 FOR A = ISW TO ISW + 10 ‘stress indicator line 2550 LINE (A,PY+1) - (A,PY+VY-1) 2560 NEXT 2570 OLDIS = ISW ‘save old stress indicator positiom 2580 COLOR WHITE,BLACK 2599 RETURN 2600 ‘****************************** 2610 ‘Re-draw stress indicator line. 2620 ‘****************************** 2621 ‘first re-color old line 2630 IF OLDIS > 0 THEN IF OLDIS = < NST THEN GOSUB 2130 ELSE IF OLDIS = < MST THEN GOSUB 2230 ELSE IF OLDIS = < HST THEN GOSUB 2330 ELSE GOSUB 2430 2640 GOSUB 2530 ‘then draw new line 2699 RETURN 3000 ‘********************* 3010 ‘Determine pulse rate. 3020 ‘********************* 3030 COLOR WHITE,BLACK: A = 1: K = INKEY$ 3040 LOCATE 25,1: PRINT FNCEOL$;: LOCATE 25,1 3050 PRINT FNCENTRE$(“Wait until beeps are regular, then press space-bar”); 3060 WHILE K = “”: K = INKEY$ 3070 WHILE (INP(PORTB) AND MASK) = 0 ‘wait for finger pulse 3080 PULSE(A) = TIMER ‘record start time for pulse 3090 WEND 3100 BEEP 3110 WHILE (PULSE(A) + .45) > TIMER: WEND ‘wait for 450ms 3120 WEND 3130 LOCATE 25,1: PRINT FNCEOL$;: LOCATE 25,1 3140 PRINT FNCENTRE$(“Reading your pulse rate - Please wait.”); 3150 FOR A = 0 TO 5 3160 WHILE (INP(PORTB) AND MASK) = 0 ‘wait for finger pulse 3170 PULSE(A) = TIMER ‘record start time for pulse 3180 WEND 3190 BEEP 3200 WHILE (PULSE(A) + .45) > TIMER: WEND ‘wait for 450ms 3210 NEXT 3220 IF PULSE(5) = 0 THEN 3150 3230 PAV = 300/(PULSE(5) - PULSE(0)) ‘average pulse in B.P.M. 3240 IF PAV > 110 OR PAV < 50 THEN 3150 ‘set reasonable limits 3250 LOCATE 16,1: PRINT “Pulse average - “;: PRINT USING “###.#”;PAV; 3260 PCAL = CINT(PAV) * 2 ‘use twice pulse rate as GSR count 3270 LOCATE 18,1: PRINT “Now calculating Stress level.”; 3280 K = “” ‘clear input 3299 RETURN 4000 ‘******************************** 4010 ‘Set Digital Output to match GSR. 4020 ‘******************************** 4030 COUNT = 60 ‘increment D/A output until 4040 WHILE STICK(0) > PCAL ‘STICK(0) value = count 4050 LOCATE 25,1: PRINT “Count =”;COUNT;” “;STICK(0); 4060 OUT PORTA,COUNT 4070 T = TIMER 4080 WHILE T + .05 > TIMER: WEND ’50ms delay 4090 COUNT = COUNT + 1 4100 IF COUNT > 240 THEN LOCATE 25,1: PRINT FNCEOL$;: LOCATE 25,1: PRINT FNCENTRE$(“An error has occured. Press spacebar to begin again”);: RETURN 80 4110 WEND 4199 RETURN 5000 ‘******************************* 5010 ‘Read GSR value and show change. 5020 ‘******************************* 5021 ‘take average of 10 counts in 10 seconds to determine change 5030 LOCATE 25,1 5040 PRINT FNCEOL$;: LOCATE 25,1 5050 PRINT FNCENTRE$(“Remove finger. Press spacebar to quit, R to run again.”); 5060 WHILE K = “”: K = INKEY$ 5070 FOR A = 1 TO 10 5080 T = TIMER ‘get the timer value 5090 COUNT(A) = STICK(0) ‘read the joystick value 5100 WHILE TIMER < T + 1: WEND ‘wait one second 5110 NEXT ‘get next reading 5120 FOR A = 1 TO 10: SCOUNT = SCOUNT + COUNT(A): NEXT 5130 COUNTAV = CINT(SCOUNT/10) ‘take average of ten counts 5140 WHILE COUNTAV < > COUNTOLD ‘skip rewrite if same value as last 5150 COUNTOLD = COUNTAV 5160 IF COUNTREF = 0 THEN COUNTREF = COUNTAV 5170 ISW = 1.7 * (2 * COUNTREF - COUNTAV): GOSUB 2630 ‘scale count for indicator 5180 LOCATE 18,1: PRINT FNCEOL$:LOCATE 16,1:PRINT FNCEOL$;:LOCATE 16,1 5190 PRINT “Relative stress level - “;: PRINT USING “###”;CINT(ISW); 5200 WEND 5210 SCOUNT = 0 ‘clear count 5220 WEND 5299 RETURN October 1998  41 100kΩ resistor from the 5V supply to the top GSR terminal has been chosen to cover a range of GSRs between 68kΩ and 150kΩ. If yours is outside these limits then change this resistor to be similar to your GSR. To measure your GSR, connect the leads to your fingers and read your resistance value with a multimeter. Obtain the next highest preferred value resistor and use it in place of the 100kΩ resistor on the PC board. For example, if your GSR was 163kΩ you would use the next preferred value which is 180kΩ. Once you are satisfied that the unit is operating properly, mount the PC board on the case lid using tapped spacers and M3 screws. After removing the 100kΩ resistor from across the GSR stakes, connect them to the 4mm sockets and fit the lid on the box. You should also fit rubber feet to the lid of the box to prevent it from scratching your table. Using the Stress-O-Meter Plug one end of a 25-way extender cable into the stress PC board and the other end into the parallel printer port you plan to use. Plug the 15-way connector into the games port connector. Wrap the GSR sensors around the middle and index fingers of whichever hand you prefer and connect the leads to the interface. Turn on the computer and load Fig.7: this is the full-size artwork for the PC board. Check your board carefully before installing any of the parts. the Basic you will use, then load the stress program and run it. Follow the prompts at the bottom of the screen. Once your stress level is shown, the aim is to move the bar to the left into the calmer region. Different approaches work with different people. Some find deep breathing reduces stress, others find thinking about the seashore and gentle waves lapping around their feet does the trick, while remembering some of the more pleasant episodes in your life may do the trick for you. Whatever it is, the result will be shown by the pointer and the stress value. The volume of the VCO can be turned up and the frequency adjusted to suit you, if you find this helps you to relax. Once you find the mindset that relaxes you, you can then begin prac­tising this mental exercise when things start getting you down, helping to get your stress level under control without the need for the Stress-OSC Meter. Fig.8: this is the full-size artwork for the front panel of the Stress Meter. 42  Silicon Chip CIRCUIT NOTEBOOK Interesting circuit ideas which we have checked but not built and tested. Contributions from readers are welcome and will be paid for at standard rates. Independent messages from sound recorder This circuit delivers two 8-second messages separately, on demand, from an ISD1416 sound recorder. It was built to provide voice sounds for a model train layout; eg, a station master instructing people to stand clear prior to train departure and a policeman booking a speeding motorist. The ISD1416 was fully described in the “Talking Headlight Reminder” described in the October 1994 issue of SILICON CHIP. In that article the chip is arranged to repeat the one message for a period of 30 seconds. In the circuit shown here, the address pins are arranged to enable the user to record two messages, each capable of being retrieved separately. To record the first message, switch S1 is switched to posi­tion “A”. This pulls address lines A3, A4, A5 and A6 low and initiates a record cycle from the beginning of the message space. When the record switch S3 is pushed, it pulls pin 27 low and allows the recording to be made. The address lines A3, A4, A5 and A6 actually provide for the first four seconds of the record­ing, but by keeping record switch S2 closed the device allows the recording to flow over to the next four-second segment. There­ fore, by using only four address lines you can record/retrieve eight seconds of a recording. Similarly, by switching S1 to the “B” position, ad­dress lines A3 and A5 are pulled low and by closing record switch S3 the second recording may be made. This recording begins at the end of the first recording and also has a maximum recording time of eight seconds. The record function takes precedence over all other device controls and if pin 27 is pulled low recording will begin irres­pective of the state of the other controls. If you keep the record switch closed, the device will continue to record for the full period of 16 seconds and will nullify your selective ad­dressing facility. There are two methods of retrieving messages. One is to pull pin 23 “PLAYL” low for the duration of the message and then allow it to return high. Using this method requires that a switch be held on for the duration of that message or that a timing device such as a 555 be used to keep pin 23 low for the 8-second period. The other method is to momentarily pull pin 24 “PLAYE” low and the device will play back the entire message selected by the position of switch S1 and automatically power down at the comple­tion of that message. K. Ferguson, Woy Woy, NSW. ($45) October 1998  43 Frequency doubler for a cruise control This circuit doubles the frequen­ cy output from a car speedo­ meter to make it compatible with an aftermarket cruise control system. It has a 100:1 frequency range and is capable of operation with an input frequency up to 100Hz. The speedometer’s digital signal is fed to pin 14 of a 4046 phase lock loop (PLL). This input has a 200mV sensitivity. The PLL comprises a voltage controlled oscillator (VCO), a phase comparator and a filter. The oscillation frequency is set by the .01µF capacitor between pins 6 & 7, the 10kΩ resistor at pin 11 and the DC voltage at pin 9. The oscillator signal output at pin 4 clocks a 4020 counter at pin 10 and its Q1 (pin 9) output, which is half the clock frequency, is applied to the comparator input (pin 3) of the 4046. The internal comparator compares the input signal at pin 14 with the Charger controller for an outboard motor This circuit was designed to allow an outboard motor to charge a 12V battery. The alternator output from the outboard can rise to as much as 60V at full revs. In essence, the circuit is an emitter follower boosted by two power transistors to provide high current output. Q1 & Q3 act as paralleled series pass transis­tors. Q2 has its base held at +14.2V by ZD1, D1 and D2. This maintains the emitter voltage of Q2 at 13.6V. Q2 conducts and turns on Q1 & Q3 just hard enough to always maintain 13.6V at the output. Q1 output of the 4020 counter and produces an error signal at pin 13, which is filtered via the 180kΩ resistor and 6.8µF capacitor. The resulting DC voltage is applied to the pin 9 VCO input. The voltage at pin 9 adjusts the VCO so that so that the frequency at pin 3 is equal to the input at pin 14. This forces the VCO to operate at twice the input frequency since the com­parator input is divided by two by the 4020. The two 10kΩ resistors reduce the signal level to suit the cruise control input. SILICON CHIP Q1 & Q3 are forced to share the output current equally by dint of their 0.47Ω 5W emitter resistors. Both transistors turn on just hard enough so that the voltage across the 1kΩ resistor equals the voltage across their respective 0.47Ω emitter resistor plus their base-emitter vol­tage. If one transistor tries to deliver more current, the in­creased voltage across its emitter resistor throttles it back. The 60V maximum input voltage is a major problem. With the two MJ2955 power transistors, this circuit will supply around 4A but a large heatsink will be required as the power dissipa­tion could be in excess of 180W. SILICON CHIP Fuel injector driver for added power Modified cars with engine management systems often need extra fuel flow if they are to be supercharged or turbocharged. The extra airflow means that extra fuel must be added otherwise the engine will run lean and could suffer serious damage. The solution is to add extra injectors which the engine management system will typically be able to control within its existing parameters, or if the car has been heavily modified, the system may have to be reprogrammed. Either way, the system may not actually be able to drive extra injectors and this is where this circuit comes in. It connects to an existing injector output on the engine management system and causes negligible loading. It will 44  S Chip drive anilicon injector with a coil resistance as low as 2Ω. Q1 is a Darlington transistor which inverts the injector signal from the ECU to drive Q2, a TIP power transistor. D1 protects Q2 from voltage spikes which are generated by the injec­tor coil each time the current through it is switched off. Q2 will probably need to be mounted on a small heatsink and should be kept away from hot spots near the engine’s exhaust manifold or radiator. SILICON CHIP SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SILICON CHIP If you are seeing a blank page here, it is more than likely that it contained advertising which is now out of date and the advertiser has requested that the page be removed to prevent misunderstandings. Please feel free to visit the advertiser’s website: www.jaycar.com.au SERVICEMAN'S LOG Comparing the old & the new Vintage car radios are still hanging around this month and it was quite interesting to compare their straightforward problems with the rather tricky problem I tracked down in a modern car sound system. After that, a TV set and a TV/ VCR combi­nation came almost as a relief. To kick off this month, I am reverting to last month’s story, about the three Delco car radios in vintage Cadillacs. I can now complete this story; relating how the remaining problems were solved. As readers may remember, the main story was about a Delco model 7265845 using a vibrator power supply, in which the vibrator and rectifiers had failed. The faulty rectifiers were easily replaced but replacing the vibrator was the real problem. These devices no longer exist and the only solution was to substitute a solid state version. Fortunately, a suitable circuit was found and an updated version was built and fitted, with complete success. That left two other Delco car radios. These were slightly later models (7272505) from 1959 Cadillac Coupe De Villes and were hybrid types, using five valves and one transistor. The valves were also special types, designed to work with a 12V “HT” rail. The transistor was used in the power output stage, a Motor­ ola HEP231 being used in one radio and a Delco DS501 in the other. These two power transistors were rather unusual, being attached to the case via a central bolt, with their base and emitter leads on either side. Initially, I couldn’t find any data on these until I dug up an early Hong Kong transistor manual. This listed the DS501 as a PNP germanium audio power transistor with an hfe of 40 and a power rating of about 60W. The case style is a TO-36. Both these radios had the same clever features as the previous unit. The first problem I had was working out the cable connections to them, as they had a few additional features such as a footpedal switch to control the self-seek tuning and also a balance control for the front and rear speakers The set fitted with the Motorola HEP231 device in the output stage was relatively easy to diagnose and fix. This set was built on a PC board and the fault turned out to be dry joints on the terminals of the last IF coil. The other set proved to be a bit more difficult, although the fault itself wasn’t hard to find. As I quickly discovered, the DS501 audio output transistor had shorted and taken with it a 0.47Ω 5W resistor and a 1000µF 1V electrolytic capacitor. The problem was finding a replacement transistor or an equivalent. Of course, no-one had ever heard of this type; it was “used back in the Ark”. So could I substitute a modern silicon transis­ t or equivalent and work out any circuit modifications that might be needed? I decided to have another word with the author of the vibrator article – perhaps he had another article hidden away somewhere, to solve this problem! Unfortunately, he didn’t but we discussed the problem at some length. Initially, I was concerned that the different base-emitter voltages of germanium and silicon transistors (0.2V versus 0.6V, respectively) might be a problem but he felt that we should be able to solve this. I subsequently spent some time tracing out the audio output stage – see Fig.1. This shows that the output stage is transform­er coupled to the preceding 12DV8 audio amplifier valve. In other words, the transistor is completely isolated from DC signals, which meant that its bias voltages could be juggled in any way necessary to achieve the required operating conditions. So, what did we have to lose? After checking all the specifications, I eventually decided to substitute an MJE2955 TO220 transistor. This was simple to mount on the original heatsink, along with a mica insulating washer and insulating bush. On firing it up, I was surprised and happy to hear sound immediately, even though it was distorted. There was a 115Ω potentiometer in the base bias network of the transistor, so I marked its setting before adjusting it until the sound was quite clean. This turned out to be at maximum setting, which made me rather suspicious. I decided to take a closer look October 1998  53 cables and an antenna, I switched it on. There was no sound from either the tuner, the cassette player or the CD player. One problem or more? Fig.1: the transistor output stage of the Delco 7272505 hybrid car radio. Substituting an MJE2955T silicon transistor for the original germanium DS501 transistor proved to be completely successful. at the bias network itself. Apart from the 115Ω pot, this also includes an 82Ω 5W resistor and a 10Ω 0.5W feedback resistor. The latter was slightly discoloured and when I checked it, it measured over 12Ω. I replaced it with a new 5% 1W resistor and found that the sound was now the cleanest when the potentiometer was reset to its original position. The new transistor was quite cool even after it had been running for over half an hour. I checked the overall current at 4A and was satisfied that all was well before returning it to be reinstalled in the original vehicle. This by now had been converted to right­hand drive and had been re-upholstered to the tune of $3000. In fact, the whole car looked fantastically com­fortable and beautiful. I wonder what it’s like to drive? (My thanks to Erol G. Engineering, the company involved with restoring the Cadillacs, for their help and co-operation with this story). A modern car sound system It wasn’t long after fixing the old Cadillac radio that I had to tackle a modern car sound system. In this case, the unit was a 1990 Kenwood car stereo system, consisting of a digital AM/FM stereo tuner/cassette player (model KRC-810), a 10-CD auto­changer with an infrared remote control (model KDC-C300) and a separate amplifier system capable 54  Silicon Chip of driving front and rear-mounted loudspeakers (ie, four channels). It even had a muting circuit to mute the output when the car phone rang! Inevitably, it got me thinking about how far car sound technology has progressed in the last 30-odd years. How ever did we manage with a simple car radio delivering a couple of watts of AM mono sound? The complaint from the owner was that there was no sound output but he’d managed to clear the amplifier unit, apparently by feeding signals from some other source into it. As a result, the system came to me as two units: the stereo tuner/cassette player and the CD player. The tuner/cassette player was mounted inside a metal cabi­net which is normally mounted in the car. This cabinet is fitted with a detachable connector on the rear panel, to which is at­ tached an array of cables that connect to the various external units. This connector mates with a socket on the rear of the tuner/cassette player. The first thing I had to do was to withdraw the tuner/cas­sette player, then remove the plug so that I could work on the unit outside the metal cabinet. I then had to set up an external amplifier and speaker system to hear anything. The audio comes out on four RCA mono sockets and I used a bench amplifier and speaker in this role. Finally, having connected all the Initially, I felt certain that there was something common to the failure of all three units and on removing the covers, I noticed a TC4066BP analog switch near the output connections. This device is no­toriously unreliable in TV receivers and was exactly what might cause the problem. However, replacing it made no difference. One of my handiest pieces of test equipment is a small portable audio amplifier fitted with meter probes. Using this, I jabbed around inside the unit, looking for an audio signal and any other clues. I finally found that I could get sound from the tuner on test points TP2 and TP3 (more on these later) on one of the boards holding the tuner sections. In addition, I could hear the tape player at potentiometers VR1 and VR2 near the tape pream-plifier outputs. The CD offered nothing except lots of clunking noises as the autochanger worked. Unfortunately, that was about as far as I could go without the service manuals, as the rest of the system was just too com­plicated. As a result, I put the unit to one side while I waited for the manuals to arrive. They duly turned up about a month later and though beauti­fully drawn in multicolours, they were also very complex and initially rather confusing. The main problem was first identify­ing which board was which and then following the signal lines from one plug to another and through one board after another. I was beginning to think that I had got myself involved in a real can of worms. On the other hand, the most useful diagram from the read­er’s point of view was the “Block and Level Diagram” and this is shown in Fig.2 as an aid to following the fault-finding process. I was still hoping that only one fault was causing the problem, so I started at the output end; ie, at the righthand end of the block diagram. This tuner/ cassette player has four sepa­rate preamplifiers – based on dual op amps IC1 & IC2 – and these feed four RCA sockets, following the “Front Pre Out” and “Rear Pre Out” designations on the block diagram. I checked the voltages marked on the circuit for these preamplifier stages (7.5V, 3.75V and 5V) and all were correct. I couldn’t believe that all four amplifiers were dead and I was beginning to consider the possibility that it was a muting prob­lem. This unit uses a very complex muting system which shuts down the audio chain in response to a number of instructions, which I would have to trace out and examine if this was the case. I was dreading this because it is all tied up with a secur­ity pass code, an “EXT-MUTE” signal, “MUTE IN 1”, “MUTE IN 2” and “RESET-IN”, all of which looked fearfully complicated. I decided to postpone that investigation and move to the lefthand side of the diagram. In particular, I thought that I would follow the tuner audio signal from where I had discovered it at TP2 and TP3 and find where it disappeared. This meant first removing the cassette deck to give access to one side of the double-sided mother­board. TP2 is the output of the FM IF/DET discriminator (IC1), while TP3 is part of the FM MPX/NC stereo decoder (IC2). The output of the decoder goes to a synthesiser unit on the mother­ board and then to a Dolby B/C decoder (CXA 1332M) – also desig­nated IC2. This a 30-pin surface mounted IC (CXA1332M) which looked really difficult to access and replace. Despite a few difficulties, it didn’t take too long to find that the signal reached pins 1 & 30 of the Dolby decoder. Howev­er, the signal didn’t go beyond this IC, so this became my major suspect, particularly as the tape signal also went to this IC (at pins 2 and 29). Anyway, I thought I would have a go at this IC with my trusty freezer to see if this had any affect. I duly gave it a generous squirt, which also hit some of the surrounding compon­ents, and was surprised to hear sound from the tuner. The sound then disappeared again after a few minutes, as the circuit warmed up again. I repeated the freezer procedure, applying a more judicious squirt this time, but nothing happened. Feeling sure I was within an ace of cracking this, I start­ed freezing the surrounding components. Nothing happened until I reached C18 and C19, two red Elna 4.7µF 16V sub­ min­i ature electro­l ytic capacitors. These are in the audio path directly ahead of IC2. Freezing these had an immediate effect, with both channels switching on. Repeating the procedure several times confirmed that these devices were indeed heat sensitive and though they made no difference to the audio signal immediately after them, they were somehow upsetting the DC condition on IC2. And I have no doubt that it was leakage that was responsi­ble; enough leakage to apply a disabling voltage to IC2 and shut it down. I cut the two offending devices off the board and fitted new ones. And that fixed the tuner but that was all. Tape player Now what about the tape player? Reinstalling the tape mechanism and playing a tape still produced no sound until, by carefully spraying freezer underneath the deck, I found another two Elna electros – C37 and C38 – that were heat sensitive. These were marked 47µF on the circuit but were actually 4.7µF on the board. Replacing these two fixed the tape problem and so it was on to the CD player. This was an entirely new ball game because no audio was entering the front end from the external CD autochanger. I care­fully dismantled the 10-CD stacker and watched as the mechanism found the selected CD and moved it over to the player. But that is where it stayed; the motor didn’t spin the disc and the dis­play was left showing the “LOAD” designation. Now some CD players require the laser to focus first before letting the disc spin, while others do it the other way around. Similarly, for the October 1998  55 starting position of the sled motor. The problem here was how to check the laser focusing, as access was rather difficult. The only thing to do was to remove all the covers until the underneath of the CD could be seen. I was in the course of doing all this when the improbable happened – the CD started to work properly and I couldn’t make it fail. I can only put it down to either poor connections in the ribbon cable connectors or a dirty/noisy limit switch (S902) on the sled. Anyway, I cleaned the laser lens, checked its alignment and put it aside to soak test. Fig.2: the block diagram for the Kenwood KRC-810 car sound sys­tem. Princess or frog? 56  Silicon Chip My next job belonged to Mrs Payne and involved a Princess TV set. Unfortunately, it was behaving more like a frog than a princess and no amount of kissing would alter its state! More precisely, it was a Princess model 14CT9 TV set, which is really a Goldstar PC04X chassis in disguise. It came in origi­nally for a faulty on/off power switch but as these sets are getting a bit long in the tooth, I quoted to also replace all the troublesome electros (C801, C802, C810, C817 & C818) in the power supply which can cause very similar symptoms. The job was quite routine, although I did have to unsolder and remove the heatsink for the chopper transistor to allow access to the electros. Anyway, when all was done the set per­formed well and was allowed to soak test. The lady ultimately collected it and everyone was happy. That is until two weeks later when a less than happy customer returned with her TV set which was no longer producing a picture – just a bright (yellowish) raster. She gave the strong impression that the new symptoms must be entirely due to the work I had completed earlier. Such com­plaints go over my head but I listened sympathetically and prom­ised I would give my most earnest and immediate attention to her Princess. I have written before (April 1997) on this very problem and felt that, given my time and experience with this model, I should crack it quickly. By changing modules, I quickly eliminated the chroma module and its sometimes troublesome IC501 (TA3562A), eventually isolating the problem to the CRT base board. Measuring the voltages on this board didn’t reveal anything specific, other than that they were all incorrect. So I started to tediously replace the components one by one. When I reached the three BF421 transistors (Q902, Q904 and Q906), I checked them with a multimeter and they all measured OK. However, as I had just completed the new Automatic Semiconductor Analyser kit (SILICON CHIP, July 1998), I thought I would give it a go on these transistors. The display correctly identified them as being PNP types and the hfe was around 125 for two of them. But one – Q906 for the blue gun – had a gain of only 55. I re­placed it and switched it on to be rewarded with a perfect pic­ture. But how or why did this happen? The transistor had only changed its gain and everything else was normal. My only theory is that perhaps a flashover within the tube had somehow created this freak change in hfe value. Anyway, to give SILICON CHIP a plug, I don’t think I would have picked this fault without the Automatic Semiconductor Analyser – buy yourself one now! Mrs Payne was still a pain when she picked the set up but I charged her again anyway. Accidents happen The next customer brought in a 1993 Teac MV1490 Televideo – a combination 34cm TV set and VCR – and it had had a hard life. The owner was an advertising agent and used it to present and describe the products he was advertising. It was rusty on all the metal work and the top of the case had been warped in the sun – presumably because it had been left in the back of his station wagon on a hot day. Mr Anderson’s complaint was that it was chewing up tapes, particularly, he perceived, on the righthand (takeup reel) side. I was dreading repairing this as access is normally appalling on this sort of equipment so, when I managed to get it out, I was pleasantly surprised at how easy it had been. The only major problem was how to service it with any degree of access because the leads are far too short. In the end, I unplugged the half dozen leads that plug into the back of the VCR, turned the deck around and plugged them back in again. Even so, it was still pretty tight and as I had a me­chanical fault, I needed to get underneath the PC board on top as well as underneath the deck. Finally, I managed to balance the VCR on its side and examine what was happening. The device was made in Korea which whittled the manufactur­ er down to Samsung, Goldstar or Dae­ woo. My guess is that it was a Daewoo mechanism, as it is very similar to an NEC VCR. The prob­lem was insufficient torque on the take-up spool and there was a fair amount of noise coming from the reel idler gear assembly. As all these gears are plastic, I decided to quote to re­ place all of them. In the course of trying to identify precisely where the problem lay underneath the deck, the whole VCR moved too close to the TV chassis until suddenly there was a burning noise (how do you define that?) and a smell. The main PC board of the VCR had made contact with the CRT socket board. “Oh bother” I said – or words to that effect – and added “copy of one circuit diagram” to the order (the Service Manual costs $80). When the gears arrived from Melbourne, I eventually pin­pointed the VCR problem as a broken plastic retainer on a bracket assembly. This was allowing the end of a leaf spring to hit and jam the idler gears. All was now OK with the tape deck. As to the TV set, my carelessness had resulted in no pic­ture or sound from the tuner. I tried to pinpoint exactly where the two boards had made contact and I worked this out to within few centimetres. I then checked the voltages but there were no clues to be had here – they were all as one would have expected. The most likely suspect was the LA7555 IF processor (IC701), as nothing was coming out of it. Replacing this 24-pin high density IC restored the picture but the sound was low com­pared with the tape playback. Of course, I only discovered this when the covers were all back on. Anyway, I delved back in and followed the sound out from IC701. Well, there was sound aplenty on pins 1 and 3 but not enough, I surmised, on pin 5. I followed this track to R417, C418 and pin 18 of IC401 (BA7790LS). The sound was OK on playback on pin 16, so what had I missed? The answer was a line from pin 18 going down to Q481 and Q480, which appear to be some sort of muting circuit. Unsoldering the collector of Q481 restored the full volume but replacing it only made things worse. Replacing Q480 as well finally fixed the problem. Believe it or not, I managed to get it all back together again without any further accidents and it is now on soak test waiting for the customer. However, I can’t help wondering how much wire costs the manufacturer. Is it so expensive that they could­n’t spare another 7cm or so, so that one can easily can work on the deck with sufficient room at the rear? Alternatively, perhaps they could cover the track side of the CRT socket board with plastic as they used to do in the SC early Rank Arenas. October 1998  57 PRODUCT SHOWCASE 25-turn industrial trimpots Jaycar Electronics now has available a range of 25-turn trim­p ots in values of 100Ω, 500Ω, 1kΩ, 2kΩ, 5kΩ, 10kΩ, 20kΩ, 50kΩ, 100kΩ and 1MΩ. The adjustment for the trim­ pot is located on the top which makes it easier to change any settings after equipment has been installed. Selling for $2.95 each, the pots have a power rating of 0.5W at 70°C. The pins are aligned in a row and are spaced at 2.5mm intervals. Details from Jaycar Electronics, 8-10 Leeds St, Rhodes, NSW 2138. Ph: (02) 9743 5222; Fax: (02) 9743 2066 AUDIO MODULES broadcast quality Manufactured in Australia Harbuch Electronics Pty Ltd 9/40 Leighton Pl. HORNSBY 2077 Ph (02) 9476-5854 Fx (02) 9476-3231 58  Silicon Chip Rockby Catalog Along with a very comprehensive range of everyday-type components the 188-page Rockby Electronics catalog also features a wide variety of parts, bits & pieces and hardware suitable for the electronics service industry. Of particular note are the 40 or so pages of components specifically for TV/monitor and VCR servicing. All items are well indexed and illustrated, either by line drawing or photograph, with pricing details included. Rockby Electronics, based in the Melbourne suburb of Huntingdale, claim that most items in the catalog are ex stock and will express post, overnight air freight or local courier (Melbourne only) with just a $10 minimum order. They also have a web site which often has products not featured in the catalog (www. rockby.com.au). For a copy of the 1998 Rockby electronics component catalog, contact Rockby on (03) 9562 8559, fax (03) 9562 8772 or email sales<at> rockby.com.au Panel meters for power analysis The new Elcontrol Power and harmonics analyser allows accurate monitoring of the quality of three‑phase supply, generation within an installation and the importing and exporting of power. The VIP96 PLUS is a panel‑ mou nting instrument providing visual readout as well as digital output via a multi‑drop communication. The instruments measure the value for line voltage and current power factor, analysis for voltage and current to the 24th order including percentage values as well as the individual displacement phase angles and total harmonic distortion (THD) expressed as a percentage. The VIP96 PLUS measures power (kW, kVAr) and energy with programmable integration times ranging from 1‑60 minutes. The instrument is fully programmable for any CT or PT combination. Networking software is provided by the VIPLINK/VIPLOAD and VIPVIEW packages, permitting up to 247 instruments to be linked to a central PC. For further information, contact Nilsen Technologies 150 Oxford St, Collingwood, Vic 3066. Phone (03) 9419 9999; Fax (03) 9416 1312. Dick Smith Electronics stores have available an adhesive remover which not only removes labels without damaging either the label or the surface it was on but allows the adhesive to regain its stickyness so the label can be used again. “Un-Du” is claimed to remove most types of sticky labels and tapes from paper, glass, plastic, walls, most clothing, carpet, furniture, metals and leather without leaving any adhesive residue. It comes with a tool which assists in removal. A 30ml bottle retails for $9.95 (Cat. N-1204). It is available from all Dick Smith Electronics stores and dealers. For fur- ther information, contact Dick Smith Electronics, Lane Cove & Waterloo Rds, North Ryde NSW 2113. Tel 9937 3200; Fax 9888 1507. CD pickup meter for jitter measurement Leader Instruments have released a CD Pickup meter designed for real time measurement of jitter in EFM signal output from a CD player. Stable measurements can be performed even when track jumping occurs or when measuring stained or scratched discs. The instrument automatically tracks an input signal over a range from 30mV to 3V. The measurement is compared with a preset value and displayed in GO/ NO GO format. A DC voltage proportional to the meter indication can also be obtained. For further information, contact Stantron Australia Pty Ltd, Suite 1, Unit 27/7 Anella Avenue, Castle Hill, NSW 2154. Phone (02) 9894 2377; fax (02) 9894 2386. Wideband receiver for EMC testing The new AFJ EMI modular receivers cover the frequency range from 9kHz to 1GHz and are suitable for full EMC compliance and operate in spectrum analyser mode. They are equipped with quasi‑peak detectors and can have up to 10 fixed and tuned preselector filters, providing more than 40dB attenuation for intermediate frequency, image frequency and intermodulation effects. Optional features include short and long click counting, as well as continuous interference observation via tuning to CISPR‑required frequencies. The ER55 receivers are PC‑based and are controlled by Windows software. This enables the operator to set up parameters as specified by CISPR 16 or according to individual requirements. The program allows the setting of frequency range, frequency step, selection of detectors (peak, quasi‑peak and average) and antenna correction factors. Customised software for communication with controllers of slide‑bars, turntables and antenna masts is also available. A wide range of accessories includes LISNs, bi‑tonic, log‑periodic and broadband antennas, loop antennas, near‑ field probes, passive probes and absorbing clamps. For further in­-formation, contact Westek Industrial Products Pty Ltd, Unit 2, 6‑10 Maria Street, Laverton North, Vic 3026. Phone (03) 9369 8802; Fax (03) 9369 8006. ELECTRONIC COMPONENTS & ACCESSORIES RANGE OF ICs, RESIS•  LARGE TORS, CAPACITORS & OTHER COMPONENTS •  MAIL ORDERS WELCOME! CROYDON STORE ONLY ELECTRONIC DISPOSALS CLEARANCE! FRAME 240V INDUCTION •  OPEN MOTORS 600 WATT AND 900 WATT. 600 WATT - $15 EACH OR 10 FOR $100 900 WATT - $18 EACH OR 10 FOR $120 VARIETY OF DISPOS•  LARGE ALS TRANSFORMERS AT GIVEAWAY PRICES! Croydon Ph (03) 9723 3860 Fax (03) 9725 9443 MilduraPh (03) 5023 8138 Fax (03) 5023 8511 M W OR A EL D IL C ER O M E "Un-Du" adhesive remover Truscott’s ELECTRONIC WORLD Pty Ltd ACN 069 935 397 30 Lacey St Croydon Vic 3136 24 Langtree Ave Mildura Vic 3500 P.C.B. Makers ! If you need: •  P.C.B. High Speed Drill •  P.C.B. Guillotine •  P.C.B. Material – Negative or Positive acting •  Light Box – Single or Double Sided – Large or Small •  Etch Tank – Bubble or Circulating – Large or Small •  U.V. Sensitive film for Negatives •  Electronic Components and •  •  Equipment for TAFEs, Colleges and Schools FREE ADVICE ON ANY OF OUR PRODUCTS FROM DEDICATED PEOPLE WITH HANDS-ON EXPERIENCE Prompt and Economical Delivery KALEX 40 Wallis Ave E. Ivanhoe 3079 Ph (03) 9497 3422 FAX (03) 9499 2381 •  ALL MAJOR CREDIT CARDS ACCEPTED October 1998  59 ! k c a t t A h s Fla . . . adding an external battery pack to a flashgun Build yourself a high-power battery pack for your photographic flashgun and spend almost no time waiting for the “ready” light to come on! By JULIAN EDGAR For years, I used a series of old 35mm Ricoh and Pentax camera bodies for all my photographic requirements. While long in the tooth, these cameras worked well enough to produce high-quality photographs that were published in many magazines. But although I could get away 60  Silicon Chip with relatively cheap cameras, one thing became apparent earlier in my photography career – I needed a good-quality flashgun. When using a flashgun, it’s normal to shoot at medium or open apertures like f4 or f5.6. In these conditions, most garden-variety flashguns work well. However, the depth of field (the zone of the picture in focus) is quite shallow and this is unsuitable for a lot of photographic work. For good depth of field, you need to shoot at a stopped-down aperture like f11 or f16. And if you shoot at this aperture, you need a very powerful flash. For this reason, I lashed out and bought a secondhand Metz 60CT-1. This flash pumps out a lot of light – so much so that if it were powered from AA cells, they would be flattened quite quickly. To avoid this problem, the Metz 60CT-1 flash uses a sealed lead acid (SLA) battery pack that hangs from a shoulder strap. At 1.6kg, this battery pack isn’t exactly light but it gives the Fig.1: the battery is connected to the power output socket through one pole of the DPDT switch. The other pole turns on the flashing LED, which provides power on/off and on-charge indication. flashgun two major advantages. First, the Metz SLA battery lasts for a long, long time. In my case, I charged it after every 10 or so 36-exposure films and it never let me down. Second, the SLA battery pack provides a very short recycle time for the flash. That’s because the large battery is able to deliver sufficient current to quickly recharge the high-voltage capacitor inside the flash unit. In an environment where the flash is being used at full power and you need to make a couple of quick exposures, one immediately after the other, this is invaluable. So is that the end of the story? Was Julian a “happy chappy” with the Metz? Not quite. As my demands grew, the shortcomings of the Ricoh and Pentax bodies gradually became more apparent. Their lack of auto-focus and other features meant that I could see better results coming from a new camera. The Metz unit also couldn’t perform fill-in flash functions (where shadows are subtly illuminated) and as for a motor drive, what was that? The battery pack has only an on/off switch, a LED and the power socket. The large red LED flashes whenever the power pack is switched on or the battery is being recharged. A new camera There was nothing for it but to bite the bullet and so it was off to the store to buy my second new camera in 15 years. After a great deal of deliberation, I settled on a Nikon F5 with a Nikon SB28 flash and a few lenses. It’s great gear and works very well indeed but I soon realised that the SB28 flash had a massive appetite for AA cells! With four brand new AA cells fitted, I found that shooting with full power flash would flatten the batteries in just one roll of film! What’s more, the recharge time was inordinately slow, even when there was still plenty of juice left in the batteries. I started taking the “flat” Inside, the battery pack appears to be . . . well, all battery! If you look closely, you can also see the 15A blade fuse (top). Note how part of the central aluminium boss at the top has been ground away to allow clearance for the positive battery lead. batteries out of the flash and putting them aside for my torch but when the box marked “torch” had no less than 32 cells in it, I decided that a new approach was required. I examined what was available commercially and found that there are a few different external battery packs available for the Nikon flash – no surprises there. But the cost! October 1998  61 Nikon has a pack that carries external dry batteries and it will set you back $200. Another company, Quantum Instruments, produces a series of external battery packs using SLA batteries. They feature battery level indicators plus a nice case and that’s about it. They are widely used by professional photographers and cost about $1000. Gulp! OK, so why not make one for myself? Well, I did and it cost me less than $130. The parts Fig.2(a): the flashgun normally uses four AA cells, installed in this manner. Fig.2(b): only two cells are installed when the external battery pack supplies the flashgun – one with a lead soldered to its positive terminal and the other with a lead soldered to its negative terminal. Note that the supply current does NOT flow through the cells! 62  Silicon Chip The battery pack was made to suit my specific requirements and some parts and techniques may not be applicable to you. So feel free to mix and match to come up with what you need. One of my requirements was that all parts for the battery pack had to be available “off the shelf”. My other requirements concerned durability and ease of use in the field. I started off with a standard diecast alloy case that measures 171 x 121 x 55mm. It’s available from Jaycar (Cat. HB-5046) and costs $27.95. The reason that I picked the alloy case is because it’s extremely strong while still being lightweight. In the environments where I am often taking flash photographs (drag racing, automotive workshops, etc), the battery pack is quite likely to hit concrete floors and knock against tripod legs, so it needs to be capable of standing up to a certain amount of abuse. For more normal household use, a sealed ABS-plastic enclosure would probably suffice – at a lot lower cost. Oh yes, another thing. Be warned that the chosen SLA battery is a very tight fit inside the specified alloy box. In fact, you will have to grind away some of the cast bosses within the case and part of the inner sealing channel of the lid if you are to fit the battery into it. The advantage? – it keeps the battery pack more compact and no battery clamp is needed. If you want a slighter larger alloy box, Dick Smith Electronics sell one that it is just a little longer than the Jaycar unit. 10 amp-hours. Using this battery, I measured an initial current draw of no less than 13A for the Nikon SB28 flash but this quickly drops away as the capacitor charges! More importantly, the flash results so far have been excellent. With the battery and box selected, you could go ahead and build the pack. But there are a few other requirements if it is to be reliable and effective in the field. An on/ off switch is useful and the $3 Dick Smith P-7720 rocker switch was the shot. I also selected a large (10mm) flashing red LED to serve as a power indicator and this operates whenever the power pack is turned on. The LED used came from Jaycar (Cat. ZD-1965) and cost $1.50. Next up was a plug and socket. Easy, huh? Well, no it wasn’t. Initially, I selected a standard low-voltage plug and chassis mount socket similar to that used in cassette recorders and other equipment. That proved to be a bad move. The plug was too flimsy and the socket actually fell apart while it was being screwed to the box! Scratch that one. Obviously, I needed something that was much more rugged and Dick Smith Electronics eventually provided the plugs and sockets that I ended up using (1 x P-1820 and 2 x P-1832 at $4 each). These polarised plugs and sockets are actually designed for microphones but they have Battery choices Talking about batteries, which one was used? I needed a 6V battery (1.5V x 4 = 6V) and so I selected the Jaycar SB-2497 which cost $29.95. In the catalog it is rated at 12 amp-hours but the rating on the battery itself is This close-up shows how the switch, power socket and indicator LED are mounted at one end of the case. Note also that the solder lugs on the back of the switch have been shortened slightly, so that they clear the battery. get away with just a single-pole switch but if you did that, there would be no way of isolating the flashing LED from the battery on the one hand or from the battery charger on the other. So the 2-pole switch is needed. The single socket is used for both charging and flash supply. When the battery needs to be charged, the flash cord is unplugged and the charger plugged in instead. The switch then needs to be turned on before the charge current can flow and the LED flashes during charging. I deliberately chose to have the LED flashing as a clear indicator that the pack was on, as I had sometimes previously forgotten about the battery pack being on charge. A blade fuseholder was also squeez­ ed into the case and is positioned adjacent to the positive terminal of the battery. This was fitted with a 15A fuse. If you are using the same alloy case as shown here, the wiring will be very tight. The fuseholder sits between the battery and the wall of the box and is held in place with double-sided tape. Use reasonably heavy-duty hook-up wiring for the supply current wiring to the switch, socket and fuse. The brackets for the carry strap were made from scrap aluminium and riveted to the side of the metal case. several important features that make them suitable for this application: (1) they’re made of metal; (2) they have large contact pins; (3) they have good cable anchoring devices; and (4) they have a screw-on collar that securely attaches the plug to the socket. And lest you think that I am overemphasising the importance of rugged socket connections, think what it would be like to be halfway through a photo shoot when someone points out that the flash has stopped working and you look down to see the cord dangling free! For the same reason, the cable connecting the pack to the flash was carefully selected. Jaycar provided 1.5 metres of Response Professional Microphone Cable (WB1530) which is flexible, has heavy conductors and isn’t too thick. It costs $2.20 a metre. Charging the battery Finally, how do you charge the battery? There are plenty of kits around for SLA chargers but that didn’t fit very well into my “off the shelf” criterion. In the end, Jaycar supplied their MB-3516 plugpack style SLA charger for $29.95. It automatically switches to trickle charge when the battery is fully charged and can deliver 0.5A. However, a new plug had to be fitted to this charger to match the battery pack. Connecting the flash unit The electrical connections from the external power pack to the flashgun were made by soldering the two leads to discarded AA cells, which are then installed at either end of the flashgun’s 4-way battery holder. Note that the current must NOT flow through the batteries, which means that the two centre batteries are not installed! My next problem was how to connect the battery pack to the flash unit, without making any modifications to the flashgun itself. Initially, I made elaborate plugs that fitted into the standard battery holder. These replicated AA cells in size but had metal contacts at one end, which were connected to the power supply This view shows how the two cells are inserted into the 4-cell holder. The black lead runs to the external battery pack. Putting it all together I wired the battery pack as shown in Fig.1. The battery is connected to the socket through one pole of the DPST switch, while the other pole turns on the flashing power indicator LED. Some readers may think that we could October 1998  63 A small U-shaped opening was filed in the flashgun’s battery compartment door, to provide an exit for the power supply cable. An SLA plugpack charger easily charges the external battery pack. Note that the spade terminals on the charger have been swapped for a plug that matches the battery pack’s socket. cable. That way, the positive and negative leads could be connected to the correct battery contacts. A small U-shaped opening was carefully filed in the battery compartment door to allow the leads to escape. However, I wasn’t happy with this arrangement. No matter how carefully I made the dummy AA cell connectors, they didn’t work very well. In the end, I decided to use real cells instead of dummies. I selected two AA cells from my enormous collection of flat batteries, tinned the positive terminal of one cell and then carefully and quickly soldered a wire to it. I then inserted this cell at one end of the battery compartment, as shown in Fig.2(b). I then did the same for the negative terminal of the second cell and inserted this at the other end of the compartment. The other (centre) two cells must be left out of the battery compartment, so that there is no complete circuit through the batteries. In other words, the two dummy cells are there only to terminate the leads from the external battery pack and to make the connections to the positive and negative terminals at either end of the flashgun’s battery compartment. Note that the two cells must be inserted as shown in Fig.2(b), so that the current from the external battery pack does NOT flow through the AA cells! Note also that this approach could NOT be used with a flashgun that used only two cells, since that would form a complete circuit (nor can you buy 3V SLA batteries, for that matter). Refinements & performance To finish off the pack, I made some aluminium brackets and attached a tripod carry strap that I happened to have handy. If you wanted the pack to look even more professional, you could have it powder-coated black but I didn’t bother. With the pack connected to the flash, the recharge rate when firing the flash on full power is only 2.8 seconds, about a quarter of the time taken with fresh AA cells. When operating at less than full power, the flash recharges almost immediately. At the time of writing, I am yet to flatten the battery pack sufficiently to require charging. In fact, the measured battery voltage has only dropped from 6.5V to 6.2V after 144 full power flash SC exposures! 64  Silicon Chip Stop signal overload with this versatile Guitar Limiter Add some interesting effects to your guitar with this versatile Guitar Limiter. It can be used to restrict any loud signals to a fixed level or adjusted to provide sustain or au­tomatic level control. By JOHN CLARKE This versatile Guitar Limiter not only provides signal limiting but provides some other really useful effects as well. They are all easily selected and adjusted using a variety of rotary controls lined up on the front panel. These controls, from left to right, are: Gain, Output, Limit, Decay and Attack. A power switch and an indicator LED are also provided on the front panel of the unit. The rear panel carries two 6.35mm jack sockets (for the input and output signals), a bypass 66  Silicon Chip switch and a DC power socket. The main effect is the limiter and this can be used in any number of ways. First, you can set the Limit control to a high level (ie, fully clockwise) so that normal signals are unaffect­ed. At the same time, high level signals from the guitar will be limited to prevent signal overload in your amplifier. Conversely, reducing the limit control will produce compression on loud guitar passages (ie, the unit functions just like an automatic volume control). The limiter action can be controlled so that it operates quickly (ie, with a fast attack setting) or relatively slowly when the attack setting is at minimum. A fast attack prevents any overshoot of the signal above the limit level, while a slow attack allows some degree of signal increase before limiting occurs. In other words, a fast attack setting reduces the dynamic range of the music while a slow attack permits a wider dynamic range. The Decay control sets the rate at which limiting ceases after the signal drops below the limit level. If the decay rate is set for a fast response, there will be little or no limiting of low signal levels when this occurs. On the other hand, a slow decay rate means that some limiting of low level signals will briefly take place after the signal drops below the limit level. The Attack and Decay controls also affect the distortion of the signal when it reaches the limit setting. Main Features •  Low noise and distortion •  Constant level over a 52dB input range •  Adjustable limit level •  Adjustable attack and decay •  •  •  •  times Output level control Adjustable input gain 12AC plugpack powered Compact size Slow attack and decay settings provide the best distortion figures. Sustain & ALC Above the limit setting, the signal is compressed so that the output level remains constant even if the input level varies. This can be used to maintain the output signal at a constant level (volume) as the input signal dies away. This effect is called “sustain” and it is achieved by setting the Limit control to a low level. Fairly obviously, you can vary the amount of sustain by varying the setting of the Limit control. Combining a low limit setting with fast attack and decay settings makes the unit operate as an automatic level control (ALC). In other words, the Guitar Limiter will maintain a fixed output level for a wide range of input levels. Fig.1 shows how the Guitar Limiter Fig.1: these oscilloscope traces show how the Guitar Limiter typically responds when presented with 20dB signal bursts. The top trace is the input signal, while the lower trace shows the output from the limiter. typically responds when presented with 20dB signal bursts. The top oscilloscope trace is the input signal, while the lower trace shows the signal output from the limiter. These waveforms were obtained with maximum attack and decay rates. Note the slight overshoot at the begin­ning of the input burst and the 150ms recovery time for the waveform to return to normal after the burst signal has ended. Block diagram The general arrangement of the Guitar Limiter circuit is shown in Fig.2. The input signal is first amplified by IC1a and then fed to a voltage controlled amplifier (VCA). The gain of the first stage can be adjusted from about 3-46 using poten­tiometer VR6. The VCA stage is based on an Analog Devices SSM2018 chip. This device features a 117dB dynamic range, .006% THD at 1kHz and unity gain, and a 140dB gain control range. In addition, the output amplifier can be set to operate in either class-A or class-AB mode. We used the class-A mode since this provides excellent distortion characteristics (the classAB mode improves noise perfor- Fig.2: the block diagram for the Guitar Limiter. The signal is limited by using a voltage controlled amplifier (VCA). October 1998  67 68  Silicon Chip Fig.3: the complete circuit uses just three ICs and a few other parts. The incoming signal is first amplified by IC1a and then fed to IC2 which is the voltage controlled amplifier (VCA). Its gain is set by a control signal derived from a precision rectifier based on IC3a, IC3b, IC3c and diodes D1 & D2. mance by 3dB but the distortion is 10 times higher). Changing from one mode to the other simply involves changing a resistor value. Put simply, the VCA changes its gain (ie, amount of am­plification) according to a control voltage applied to one of its inputs. This control voltage is derived as follows. Following the VCA, the signal is applied to both level control VR5 and to a precision full-wave rectifier. The latter produces a DC output voltage whose level is directly related to the signal level at the output of the VCA. This “control” voltage is then smoothed using a variable RC filter network and applied to the control input of the VCA. As a result, the gain of the VCA is automatically adjusted so that it produces a constant signal level at its output. The response of the variable filter determines how quickly or slowly the gain of the VCA is controlled. The gain limiter block (IC3d, VR4 & D4) prevents the control input of the VCA from going below a certain preset voltage. This limits the overall gain of the VCA by ensuring that limiting is initiated at a certain preset mini­mum signal. Returning now to level control VR5, its output is applied to amplifier stage IC1b. This IC functions as a unity gain buffer stage and its output signal is fed to the Effects switch (S2). This switch then selects either the processed signal from IC1b’s output or the unprocessed signal at the output of amplifier IC1a. Circuit details So much for the circuit basics. To find out how it’s all achieved take a look now at the full circuit diagram shown in Fig.3. There are just three ICs and all the major circuit parts can be directly related to the block diagram (Fig.2). The guitar signal comes in via the 3.5mm input socket and is fed to pin 5 of op amp IC1a via a 10Ω resistor and a 10µF capaci­tor. The associated 22kΩ resistor to ground provides the 0V reference for the signal path while the 10Ω resistor and 10pF capacitor filter out RF signals which could otherwise cause distortion or overload in the op amp. IC1a is a low-noise low-distortion op amp, part of a dual LM833 package. Its gain is set by the 100kΩ feedback Performance Frequency response: -3dB at 30Hz and 15kHz (measured below compression limit) Signal-to-noise ratio (with respect to 1V): 80dB with 22Hz to 22kHz filter and 82dB A-weighted at 100mV signal limiting and minimum gain; 75dB with 22Hz to 22kHz filter and 80dB A-weighted at 8mV limit­ing Harmonic distortion at slowest attack and decay settings: < 0.07% at 1kHz & 10kHz for 18mV to 1V input Limiting range: 2.5mV to 2.5V Attack time: 1ms to 150ms Decay rate: 50dB/second to 8dB/second Maximum input signal before clipping: 2.5V RMS at minimum gain; 150mV RMS at maximum gain Output level: 0-1V RMS resistor (between pins 7 & 6) and the total resistance between pin 6 and ground. This latter resistance consists of a 2.2kΩ resistor and a series 50kΩ pot (VR6) which is used to vary the gain between three and 46 times. The maximum gain of 46 occurs when VR6 is at its minimum (ie, gain = 1 + 100,000/2200 = 46), while the minimum gain occurs when the pot is at its maximum value. The 68pF capacitor across the 100kΩ feedback resistor rolls off the frequency response for signals above about 23kHz and is included to prevent high-fre­ quency oscillation in the op amp. The amplified signal from IC1a is coupled to pins 6 & 4 of IC2 which is the Voltage Controlled Amplifier (VCA) stage. As discussed previously, its gain is set by the voltage applied to its control input at pin 11. The 18kΩ input resistor and the 18kΩ resistor between pins 3 & 14 set the VCA gain to 1 when pin 11 is at ground. The 680pF capacitor between pins 5 and 8 is included to compensate the amplifier and prevent instability. Similarly, the 390pF capacitor across the 18kΩ resistor at pins 3 & 14 provides high frequency rolloff. Precision rectifier stage The output from IC2 appears at pin 14 and is fed to the full-wave precision rectifier stage via a 3.3µF bipolar capacitor and a 20kΩ resistor. This stage comprises op amps IC3a, IC3b & IC3c (all part of a quad TL074 IC package), plus diodes D1, D2 and associated resistors. When the signal on pin 6 of IC3a swings positive, the signal at the pin 7 output swings negative and diode D2 is for­ward biased. The gain of this op amp stage is thus set to -1 by the 20kΩ input and feedback resistors. Because D2 is forward biased, the signal on pin 7 of IC3a also appears at D2’s anode and is coupled to the inverting input (pin 2) of IC3b via a 10kΩ resistor. IC3b’s gain is set by this 10kΩ input resistor and a 180kΩ feedback resistor to -18. This means that the overall gain for the signal from pin 6 of IC3a to pin 1 of IC3b is -1 x -18 = +18. However, IC3b doesn’t only derive its input signal from IC3a. The signal from IC2 is also fed to pin 2 of IC3b via a separate 20kΩ resistor (ie, pin 2 of IC3b is fed via two differ­ent signal paths). In the latter case, IC3b operates with a gain of -9 and adding the two gains (for the two paths) gives us an overall gain of +9. For negative input signals from IC2, IC3a’s output is clamped because D1 is now forward biased. The input signal can now only pass via the 20kΩ resistor that connects to pin 2 of IC3b. As before, IC3b functions with a gain of -9 and so the resulting signal on pin 1 is again positive. In summary, the circuit operates with an overall gain of +9 for positive input signals and -9 for negative input signals. The output from pin 1 of IC3b is thus always positive and so the circuit functions as a full-wave rectifier. The resulting full-wave rectified October 1998  69 Fig.4: follow this wiring diagram when mounting the parts on the PC board. Take care with component orientation and note that IC2 faces in the opposite direction to IC1 and IC3. signal from pin 1 of IC3b is filtered using D3, VR1, VR2 and its series 68kΩ resistor, and two series-connected 100µF capacitors (C1 & C2). D3 allows the 100µF capacitors to charge via VR1 (Attack) but at the same time ensures that they can only discharge via VR2 (Decay) and its series 68kΩ resistor. This effectively provides us with separate control for the VCA attack and decay times. The filtered DC voltage is applied to pin 11 of IC2 (the voltage controlled amplifier) to control its gain, as described previously. VR3 sets the no-signal control voltage applied to the VCA. This trimpot can be adjusted to produce a voltage on pin 10 of IC3c anywhere between ±12V. It is normally set at about 0V 70  Silicon Chip and this is buffered by voltage follower IC3c which sets the bias on pins 5 & 3 of IC3a & IC3b respectively. Essentially, VR3 sets the no-signal offset voltage at the output of IC3b. In practice, it is adjusted so that the signal at pin 14 of IC2 is at 1VAC under high input-signal conditions. Gain limit adjustment Trimpot VR4 is used to set the maximum gain of the VCA (ie, the gain limit). The voltage on its wiper can be set anywhere in the range from -1.2V to +1.2V and this is then buffered by IC3b and fed to diode D4. In operation, this sets the minimum voltage that can be applied to IC2’s control input (pin 11). It does this by clamping the voltage on pin 11 so that it cannot go below the voltage set by VR4. If the filtered control voltage from the precision recti­fier does fall below this level, D4 conducts and applies the clamp. Conversely, the voltage on pin 11 can rise above this mini­mum clamp level. That’s because if the control voltage from the precision rectifier rises, D4 will be reverse biased and so the voltage on pin 11 of IC2 is free to rise to limit the gain of the VCA. Output stage The pin 14 output of IC2 is AC-coupled to level control VR5 via a 1µF capacitor. From there, the signal on the wiper is coupled via another 1µF capacitor to pin 3 of op amp stage IC1b Table 1: Resistor Colour Codes  No.   1   3   1   1   4   3   2   1   1   2   1 Value 180kΩ 100kΩ 68kΩ 24kΩ 22kΩ 20kΩ 18kΩ 10kΩ 4.7kΩ 100Ω 10Ω IC1b functions as a non-inverting unity gain buffer stage. Its output appears at pin 1, after which the processed signal is fed to one terminal of switch S2 (Effects/Bypass). S2 selects between the processed (ie, limited) signal coming from the VCA, or it can bypass the effects circuit by selecting the signal at the output of IC1a (via a 100Ω resistor and 1µF capacitor). The 100Ω resistor in series with the output of IC1b prev­ents this stage from oscillating when long leads are connected to it. Power supply Power for the circuit is derived from a 12V AC plugpack via on/ off switch S1. Diodes D5 and D6 provide half-wave rectifica­ tion to give ±16V (nominal) rails which are then filtered using 470µF capacitors. The resulting DC is then applied to 3-terminal regulators REG1 & REG2 which provide the ±12V supply rails. The 10µF capacitors at the regulator outputs are included to ensure stability, while LED 1 and its associated 4.7kΩ current limiting resistor function as a power on/off indicator. 4-Band Code (1%) brown grey yellow brown brown black yellow brown blue grey orange brown red yellow orange brown red red orange brown red black orange brown brown grey orange brown brown black orange brown yellow violet red brown brown black brown brown brown black black brown done, install the eight PC stakes at the external wiring positions (six for the connections to switch S2, plus one each for S1 and the power socket). Next, install the ICs, diodes, resistors and links. Take care with the orientation of the diodes and ICs and note that IC2 faces in the opposite direction to the other two. Table 1 shows the resistor colour codes but we suggest that you also use a digital multimeter to check the values as the colours on some brands can be difficult to distinguish. The capacitors can go in next (watch the polarity of the electrolytics), followed by the two 3-terminal regulators. Be careful not to get the regulators mixed up and make sure that their metal tabs go towards the adjacent 10µF capacitors. You must use the 7812 device for REG1, while REG2 is the 7912 device. The PC board assembly can now be substantially completed by installing the potentiometers, the trimpot and the 3.5mm jack sockets. Note that several different pot values are used, so 5-Band Code (1%) brown grey black orange brown brown black black orange brown blue grey black red brown red yellow black red brown red red black red brown red black black red brown brown grey black red brown brown black black red brown yellow violet black brown brown brown black black black brown brown black black gold brown Table 2: Capacitor Codes  Value  0.47µF  680pF  390pF  220pF  10pF IEC 470n 680p 390p 220p 10p EIA 474 681 391 221 100 be sure to choose the correct value for each position. Push each pot down onto the PC board as far as it will go before soldering its leads. The same goes for the 3.5mm jack sockets. Finally, the power indicator LED can be fitted to the board. This should be installed at full lead length so that it can later be bent over at right angles and pushed into a bezel on the front panel. Final assembly The assembled PC board fits neatly inside a low-profile plastic Construction The Guitar Limiter is really easy to build since virtually all the parts, including the pots and 3.5mm jack sockets, mount on a single PC board. This board is coded 01308981 and measures 117 x 102mm. Fig.4 shows the parts layout on the PC board. Before mount­ing any of the parts, check your etched board for any defects by carefully comparing it with the published pattern. This The rear panel carries the two 6.5mm jack sockets, the bypass switch and the power connector October 1998  71 Above: use shielded cable for the bypass switch (S2) connections and note that the pot bodies are all earthed. Fig.5 (left) shows the full-size etching pattern for the PC board. instrument case measuring 140 x 111 x 35mm. As shown in the photos, two self-adhesive labels are affixed to the front and rear panels to give a professional finish. Affix these labels to their respective panels, then drill out the holes for the potentiometers, power switch and 3mm LED on the front panel, followed by holes for the 6.35mm jack sockets, the Effects/Bypass switch and power socket on the rear. The best way to go about this is to first drill small pilot holes at each location and then carefully enlarge each hole to size using a tapered reamer. This done, secure all the various hardware items to the panels. Note that the potentiometers are not directly fastened to the front panel, as such. Instead, the front panel is 72  Silicon Chip simply slipped over the pot shafts and the knobs attached. The PC board can now be positioned inside the case, along with the front panel which slides into a retaining slot at the front. Secure the board to the integral standoffs on the bottom of the case using the self-tapping screws provided. All that remains now is to install the internal wiring – see Fig.4. Be sure to use short lengths of shielded cable for the connections to S2, as shown. The remaining connections are run using light-duty hookup wire. Note that a length of tinned copper wire is soldered to all the pot bodies, which are then earthed. You will need to file away a small amount of the cadmium plating on each pot body to allow the solder to “take”. Testing Now for the smoke test but first check your work carefully to ensure that all parts are correct and the wiring has been completed. This done, apply power and use your multimeter to check for +12V on pin 8 of IC1, pin 2 of IC2 and pin 4 of IC3. Similarly, check that -12V is present on pin 4 of IC1, pins 16 & 10 of IC2 and pin 11 of IC3. If the LED doesn’t light, it’s probably connected the wrong way around. The output preset adjustment, VR3, can only be set by applying a signal to the input. You can use a signal generator set to about 500mV output and 1kHz, or you can use a guitar signal. To make the adjustment, first set the Limit control (VR4) to minimum and the Output control (VR5) to maximum. Now connect a multimeter Parts List – Guitar Limiter 1 PC board, code 01308981, 117 x 102mm 1 plastic case, 140 x 111 x 35mm 1 front panel label, 131 x 28mm 1 rear panel label, 131 x 28mm 2 SPDT toggle switches (S1,S2) 1 16mm 50kΩ lin pot (VR1) 1 16mm 1MΩ lin pot (VR2) 1 16mm 22kΩ lin pot (VR4) 1 16mm 10kΩ log pot (VR5) 1 16mm 50kΩ log pot (VR6) 1 5kΩ vertical multi-turn trimpot (Bourns 3296) (VR3) 5 15mm knobs 2 PC-mount 6.35mm jack sockets 1 insulated panel mount DC socket 2 M3 screws and nuts to mount DC socket 4 self-tapping screws for mounting PC board 1 250mm length of light duty hookup wire 1 100mm length of green hookup wire 1 150mm length of single shielded cable 1 150mm length of 0.8mm tinned copper wire 8 PC stakes 1 12VAC 300mA plugpack 1 SSM2018P VCA (IC2) 1 LF347 quad op amp (IC3) 1 7812 positive regulator (REG1) 1 7912 negative regulator (REG2) 4 1N4148, 1N914 signal diodes (D1-D4) 2 1N4004 1A diodes (D5,D6) 1 3mm red LED (LED1) Semiconductors 1 LM833 dual op amp (IC1) Miscellaneous Cable ties, solder, etc. Capacitors 2 470µF 16VW PC electrolytic 2 100µF 16VW PC electrolytic 4 10µF 25VW PC electrolytic 1 3.3µF bipolar electrolytic 1 2.2µF 16VW PC electrolytic 4 1µF 16VW PC electrolytic 1 0.47µF MKT polyester 1 680pF ceramic or MKT polyester 1 390pF ceramic 1 220pF ceramic 2 10pF ceramic Resistors (0.25W, 1%) 1 180kΩ 2 18kΩ 3 100kΩ 1 10kΩ 1 68kΩ 1 4.7kΩ 1 24kΩ 2 100Ω 4 22kΩ 1 10Ω 3 20kΩ set to read AC volts to pin 1 of IC1b and adjust VR3 for a reading of 1VAC. The Guitar Limiter is now ready for use. Connect it between your guitar and the amplifier and check that it gives a constant volume output when the Limit control is at its minimum setting. Adjust the Output control to set the required volume with the Gain control at about mid-setting. Finally, you can experiment with the Attack and Decay con­ t rols to verify their effect. You can also try varying the Limit control, to change the level at which limiting occurs. The Gain control sets the overall gain of the unit and can be adjusted to cope with input signals ranging from below 150mV to about 2.5V. A few minutes spent twiddling the controls will soon demonstrate the capabilities Fig.6: you can use these two full-size artworks as drilling templates for the front and rear of this versatile unit. SC panels. Drill small pilot holes first, then carefully ream them to full size. October 1998  73 This trickle charger is intended for permanent “float” charging of a leadacid battery which could be used for a garage door opener, security system or an emergency lighting system. A 12V trickle charger for “float” conditions This 12V battery charger will charge at up to 2A and then switch to “float” and keep the battery topped up, without any intervention on your part. It will run the lead-acid battery in a garage door opener, gate opener, security system or any other permanently powered system. By RICK WALTERS The major problem with most cheap “off the shelf” 12V chargers is that they cannot be permanently connected to a bat­tery. They must be disconnected after they have given the battery a boost or else they will inevitably overcharge it. This causes excessive gassing which splashes acid all over the top of the battery and does nothing 74  Silicon Chip for the battery life. Don’t get us wrong, the standard 12V auto battery charger, available from auto spare parts and hardware stores all over the country, has its place. It will get you out of trouble if your car battery goes flat and it’s handy for giving the boat or caravan battery a charge before its next day out. But these same 12V chargers are no good for “float” use. This is where the charger is connected permanently to the bat­tery, as is the case in burglar alarms and security systems. Here the battery must be kept up to the right voltage at all times but it must not be overcharged. This project started out to be a trickle charger for the battery which was used in the Garage Door Opener published in the March & April 1998 issues of SILICON CHIP. The intermittent use and (normally) long time period between each use means that a high current charger is not required. This charger, with the transformer specified, will put up to 2A into a “flat” battery. As the battery voltage rises, the charging current naturally reduces. But it does not reduce enough Fig.1: the circuit of the battery charger uses power transistor Q4 as the current control element. Q2 monitors the current through resistor Rs and removes base drive to Q3 if the average current exceeds 2A. Q1 monitors the battery voltage and gradually shuts down the circuit as the voltage rises to around 14V. to avoid overcharging the battery. So when the battery is fully charged, the control circuit switches to trickle mode, in which the battery can be left connected without damage or worry. The trickle mode of operation is ideal for topping up a car battery with an overnight charge, and is more than capable of keeping the battery fully charged when operating the garage door or any other permanently connected application. Charger circuit Let’s now have a look at the circuit of Fig.1. This is really just a conventional charger circuit with the current and voltage limiting circuitry grafted on. This part of the circuit has three separate phases. First, it limits the charging current to 2A. Second, the battery voltage rises and the charge current reduces. Third, it switches to trickle mode. The power transformer T1 and bridge rectifier B1 comprise the conventional charger circuit. This provides a full wave rectified pulsating DC voltage which is fed directly to the positive battery terminal. The charging current then flows through the 12V battery, transistor Q4 and the 0.47Ω 5W resistor Rs, then back to the negative terminal of the bridge rectifier. Q3 and Q4 are connected as a Darlington (compound) transis­tor which has a current gain which is the product of their indi­vidual gains. Q1’s gain is a minimum of 20 and Q2’s is a minimum of 100, so the combined gain is a minimum of 2000. This means that a base current of 1mA into Q3 can control a minimum of 2A collector current in Q4. The base of Q3 is connected to the positive supply through a 2.2kΩ resistor. This results in a base current of around 6.5mA which turns Q3 and Q4 hard on, and would allow Q4 to pass peak currents of 13A or much more. This is well outside the current rating of our transformer and would result in a blown fuse and possibly a damaged transformer. The saving grace is the resistor Rs. The voltage dropped across this resistor is applied to the base of Q2 via a 2.2kΩ resistor. As soon as this voltage reaches around 0.55V the tran­sistor begins to turn on, taking some of Q3’s base current and tending to turn Q3 and thus Q4 off. If they did turn off there would be no current flowing through Rs and therefore Q2 would not be turning on. The circuit does a neat balancing act, which lets just enough current flow through Rs to just turn Q2 on which steals just enough current from Q3 to just let enough current flow through Q4 to just let enough current flow through Rs and so on. By selecting a suitable value for Rs we can set the maximum battery charging current. The only drawback to this scheme is that when the current is limited like this, Q4 is not saturated, and as it has some voltage across it and current flowing through it, it must dissipate this power as heat. Once the battery voltage rises to a level where the current is less than the limit, Q3 and Q4 will saturate (be turned hard on) again, and the circuit will stay in this state until the battery voltage reaches around 14V. Q1 monitors the battery voltage through the two resistors connected to its base. When the battery voltage is high enough, this transistor starts to turn on pulling the base of Q2 high through its 2.2kΩ collector resistor. Once this occurs we have the same situation as before, except this time we don’t have current limiting but voltage limiting. The waveform of Fig.2 shows the voltage across Rs. As in any conventional battery charger, the charge current takes the form of pulses at a rate of 100Hz; ie, twice the 50Hz mains frequency because of full-wave rectification. As the battery voltage rises, the current pulses become smaller until October 1998  75 Parts List 1 PC board, code 14110981, 59 x 40mm 1 metal case, 185 x 70 x 160mm, DSE Cat H-2744 or equivalent 1 power transformer, 15V 2A, DSE Cat M-2156 or equivalent 1 TO-3 finned heatsink, DSE Cat. H-3400 or equivalent 1 small finned heatsink (for bridge rectifier) 1 3AG safety fuseholder, DSE Cat. P-7916 or equivalent 1 3AG 0.5A slow-blow fuse 1 250VAC 3-core mains cord and moulded 3-pin plug 1 cordgrip grommet to suit mains cord 2 2-way insulated terminal block 2 solder lugs 1 red battery clip 1 black battery clip 1m polarised (red/black) figure-8 cable (for battery leads) 2 3mm x 10mm threaded spacers 2 20mm x 4mm screws 8 4mm nuts 4 15mm x 3mm screws 6 3mm x 6mm screws 8 3mm nuts 5 3mm flat washers 4 3mm star washers Semiconductors 1 BC557 PNP transistor (Q1) 1 BC639 NPN 1A transistor (Q2) 1 BD139 NPN power transistor (Q3) 1 2N3055 NPN power transistor (Q4) 1 400V 6A bridge rectifier (BR1) Capacitors 1 22µF 16VW electrolytic 1 0.22µF MKT polyester 1 .047µF MKT polyester Resistors (0.25W, 1%) 1 39kΩ 1 100Ω 3 2.2kΩ 1 0.47Ω 5W 1 1.8kΩ Miscellaneous Hookup wire, heatshrink sleeving, solder 76  Silicon Chip Fig.2: this the pulse current waveform fed to the battery. It is measured across resistor Rs. finally they are tiny blips. There are three capacitors in the circuit and they are there to ensure AC stability. The .047µF capacitor between col­lector and emitter of Q4 and the 0.22µF across the 2.2kΩ collec­ tor resistor for Q2 effectively stops the Darlington pair (Q3 & Q4) from breaking into supersonic oscillation. The 22µF capacitor at the base of Q2 filters the voltage appearing across Rs. Without the 22µF capacitor, Q2 tends to turn on hard as soon as the voltage at its base exceeds 0.6V. That immediately kills the current through Rs which removes the base voltage to Q2 which turns off again. The result can be a “squegging” effect whereby Q2 switches on and off extremely rapidly at 100kHz or more. This radiates interference right across the AM broadcast band which is undesir­able to say the least. With the 22µF capacitor in place though, the circuit is as docile as a lamb. Grounded collector There is another wrinkle to the circuit that may not be readily apparent. Unusually, the negative output of the charger is earthed to the case and back to the 50Hz AC mains supply. The reason this has been done is to simplify the mounting of the 2N3055 power transistor, Q4. Instead of using mica washers and insulating bushes, the transistor’s case is bolted directly to its finned heatsink and to the chassis. While this method of transistor mounting is easier than using a mica washer and so on, it could have drawbacks if the charger is to be used to charge a battery where the load circuit is earthed in some other way. This could be the case in a burglar alarm or security system, for example. If that is the case, then the 2N3055 must be mounted on its heatsink with a mica washer and insulating bushes to ensure that the charger output is fully floating. Finally, we should note that if you want to duplicate the oscilloscope measurement shown in Fig.2, you will need a scope with differential inputs or you will need to temporarily discon­nect the case earth. Assembly Most of the circuitry is accommodated on a small PC board measuring 59 x 40mm and coded 14110981. The component layout diagram is shown in Fig.3. The board assembly is quite straight­forward and should not take long. Mount the PC stakes for all the external connections first, followed by the resistors, the capacitors and then the transis­tors. Note that Q1 and Q2 have different pinouts, as shown on the circuit of Fig.1. With the board complete, put it Fig.3: follow this diagram when installing the wiring in the case. Note that transistor Q4 is mounted upside down under its finned heatsink so that the base and emitter connections can be easily soldered. October 1998  77 Silicon Chip Binders REAL VALUE AT $12.95 PLUS P&P Fig.4: actual size artwork for the PC board. These beautifully-made binders will protect your copies of SILICON CHIP. They feature heavy-board covers and are made from a dis­ tinctive 2-tone green vinyl that will look great on your bookshelf.   High quality.   Hold up to 14 issues (12 issues plus catalogs)   80mm internal width.  SILICON CHIP logo printed in gold-coloured lettering on the spine & cover. Yes! Please send me ________ SILICON CHIP binder(s) at $A17.95 each (includes postage). Australia only – not available elsewhere. Enclosed is my cheque/money order for $­__________ or please debit my  Bankcard    Visa Card    Master Card Card No. Signature­­­­­­­­­­­­_________________________  Card expiry date______/______ Name _____________________________________________________ Street _____________________________________________________ SILICON CHIP PUBLICATIONS PO Box 139, Collaroy Beach, NSW 2097, Australia. Phone (02) 9979 5644  Fax: (02) 9979 6503. 78  Silicon Chip ✂ Suburb/town __________________________ Postcode______________ aside so that work can be done on the case. You will need to drill holes to mount the transformer, bridge rectifier and transistor heatsink, insulated terminal blocks and an Earth solder lug. As well, you will need holes for the output lead grommet, the mains fuseholder and cordgrip grommet for the mains cord. Drill all the required holes in the case and make sure that they are all de-burred and cleaned of drilling swarf. Mount the transformer and bridge rectifier in the case in the positions shown in the chassis diagram of Fig.4 and in the photograph. The 2N3055 power transistor (Q4) is mounted on the heatsink upside down, as shown in the photos. This has been done to make it easier to solder on the base and emitter leads, after the transistor has been bolted into place. Note that you should smear some heatsink compound onto the transistor’s mounting surface before bolting it down. The wiring connections between the components can be seen in Fig.3. The 240VAC mains cord should be anchored in a cordgrip grommet and the Earth wire connected to an adjacent solder lug on the case. Make sure that any paint around the solder lug mounting hole is scraped clean of paint and use a star washer under the nut to ensure a good connection. The transformer has lugs for the mains connections and these should be sleeved with heatshrink as best you can. We mounted the transformer so that the mains terminals are placed at the rear, well away from where you will be measuring. If you strip your mains lead back far enough you could run the blue lead direct to the transformer lug without using an insulated terminal block. The terminal block for the battery leads should be used as it acts as a strain This view inside the chassis shows the general wiring. Note that the 2N3055 power transistor is mounted upside down, underneath its heatsink, to make it easy to connect the base and emitter leads. relief and also prevents the leads from twisting inside the case. When the wiring is complete, check it all carefully against the circuit of Fig.1 and the wiring diagram of Fig.3. Testing To test the charger, first make sure you have fitted the mains fuse. It’s amazing how many projects don’t work first time because of this. It’s no use flicking the leads together to see if you get a spark as there is no output capacitor to discharge, and the current limit also comes into effect. You can measure the output voltage using a digital multi­meter although the result is no guide to the eventual output voltage when a battery is connected. It should read around 10V (±10%). If this is the case, switch to the 10A current range and measure the output current with the meter leads connected directly to the battery leads. You should read around 2A or so. Don’t measure the current unless the output voltage is within range as you will blow the meter fuse and perhaps even the meter itself. If you have problems, check that you haven’t interchanged the two transistors on the PC board. Next, check the wiring and lastly that all the parts are in the right location on the PC board and that each value is correct. If you wish to check what the actual fully charged battery voltage will be, place a 2200µF 25V electrolytic capacitor across the battery leads (positive to positive) and measure the voltage again. Our unit measured 14.25V. Anywhere between 13.9V and 14.3V SC is acceptable. Resistor Colour Codes       No. 1 3 1 1 1 Value 39kΩ 2.2kΩ 1.8kΩ 100Ω 0.47Ω 4-Band Code (1%) orange white orange brown red red red brown brown grey red brown brown black brown brown not applicable 5-Band Code (1%) orange white black red brown red red black brown brown brown grey black brown brown brown black black black brown not applicable October 1998  79 Hifi Review Dual CS505-4 belt-driven turntable The German manufacturer Dual has long been a manufactur­er of record players and turntables and they’re still going strong. Here we review their model CS505-4, a belt-driven semi-automatic turntable fitted with an Ortofon magnetic cartridge. By LEO SIMPSON What’s this? A review of ancient turntable technology for vinyl records in this digital age. Well, surprising though it may seem, now that it is 16 years since the compact disc was intro­ duced, there are still lots of turntables being sold every year in Australia. The reason is that many people still have big collections of vinyl records and they still wish to play them. Most music systems these days are not sold with a turntable so there are quite a few buyers in the market for a player. And even if you have an older turntable, the chances are that its styling looks quite out of place with modern audio gear which is predominantly black in finish. OK, so we have established that there is a market for turntables which is why we are reviewing this model from Dual. For anyone who hasn’t looked at turntables recently, the Dual CS505-4 is up-to-date in its styling by virtue of its all black vinyl finish on the timber plinth and the use of large circular feet with anodised aluminium dress rings. First used on upmarket CD players years ago, this styling feature is now widely used on all types of domestic audio equipment. 80  Silicon Chip In other respects, the Dual turntable is traditional, with a smoked Perspex lid fitted with spring-loaded hinges, an alumin­ium platter and a balanced tonearm. In case you’ve forgotten, turntables are large and this one has overall measurements of 440mm wide, 372mm deep and 149mm high with the lid closed. With the lid fully open, it is 410mm high, so you need plenty of clearance if it is mounted on a shelf. Tonearm & headshell The low mass tonearm is straight, not curved, and it has an angled headshell to orient the magnetic cartridge correctly with respect to the record grooves. The tonearm is dynamically bal­anced with a rotatable counterweight behind the pivot but the tracking force is applied by a spring dial. Also provided is anti-skating compensation, via a small spring dial near the arm pivot. There are two anti-skating scales, one for cartridges with a spherical stylus and one for those with an elliptical stylus. The headshell is removable and is secured with a locking collar although this is somewhat smaller than the common EIA collar found on turntables made around 20 years ago. It is fitted with an Ortofon OMB10 magnetic cartridge which has a removeable (elliptical) stylus assembly. It has a listed tracking force of 15 milliNewtons. Now that is a change from the unit of grams which used to be used for cartridge tracking weights. The conversion factor from kilograms (weight) to Newtons is 9.8 metres/second 2 and so 15mN is actually close to 1.5 grams. Funnily enough, the tracking force dial on the tonearm is cal­ ibrated in grams not milli­Newtons. The frequency range of the cartridge is quoted as 10Hz to 25kHz but without limits and there is no figure for separation between channels. Semi-automatic operation Dual’s CS505-4 is a semi-automatic belt drive turntable. The “semi-automatic” approach is for those buyers who want the convenience of automatic shut-off at the end of a record but don’t fancy the extra mechanism attached to the tonearm to give fully automatic operation. Personally, I would rather pay the extra money for the convenience of fully automatic operation because the tonearm is effectively decoupled from any drive mechanism during normal play operation. The semi-automatic “modus operandi” of the CS-505-4 is as follows: As you move the tonearm away from its rest position, the platter starts revolving and the lift/lower lever is in the UP position. You then position the stylus over the beginning of the track you wish to play and flick the knock on the timber plinth without disturbing the tracking of the stylus. Performance testing The Dual CS505-4 is a belt-driven semi-automatic turntable fitted with an Ortofon magnetic cartridge. lift/lower lever to its DOWN position. The stylus then gently lowers into the groove and the music starts to play. At the end of the record, the tonearm is lifted off the surface and the platter stops rotating. You then manually move the tonearm back to its rest and the lift/lower lever flicks back to the UP position. If you want to stop play before the end of the record, you raise the tonearm with the cueing lever and then move the arm back to its rest to stop the motor. It’s simple and it works well. Belt-drive system The belt-drive system is interesting too and is different to that used on belt-drive turntables made around 20 years ago. These turntables usually had a flat belt running around an inner rim of the platter which was typically around 220mm in diameter. This entailed an expensive casting which usually required at least some machining to ensure it was balanced and so on. By contrast, the Dual’s platter is an aluminium pressing with a polished dress rim and additional mass inside the rim to provide extra inertia. It weighs 1.2kg. The platter sits on a plastic drum about 90mm in diameter and this is driven via a flat belt from the 12-pole synchronous motor. Two speeds are provided, 33.3 and 45 rpm, and a pitch control with a total range of 6% is included. This knob is coupled via a toothed belt which actually expands or reduces the diameter of the main drive pulley. A printed strobe disc is supplied to let you set the speed precisely using light from the 50Hz mains supply. Actually, the 50Hz mains supply is reputedly not very precise but when we measured the speed using a 3kHz test track on a disc, the speed was found to be within .05% which is close enough, even for those with absolute pitch. To isolate the whole record playing system from floor-borne and external vibrations, the turntable’s pressed steel base is sprung and this works quite well, to the extent that you can Testing the performance of the CS505-4 turntable and car­tridge combination proved to be something of a hurdle as far as we were concerned, as our much-vaunted Audio Precision equipment, in conjunction with somewhat worn test records, proved to be not up to the job. The main problem was that the interruptions to the track signals as each frequency is announced on the record played merry hell with the relay switching circuitry inside the Audio Precision test gear. To overcome this problem, we had to resort to older analog instrumentation (to wit, our AC Millivoltmeter, described in the July & August 1988 issues of SILICON CHIP). The upshot was that the CS505-4 performed well. The Ortofon cartridge appears to be a particularly good tracker, handling the heavily recorded tracks with aplomb. And its wave­shape on sinew­aves in the critical region from 3kHz to 10kHz was also commen­ dably free of the “sawtooth” effects you can see with many car­tridges. It is clearly up in the top echelon of magnetic car­tridges. As you would expect, the Dual CS505-4 performs quietly and with no fuss at all times. It just plays records, pure and sim­ple. The price And what about the price? For those who have become used to cheap and cheerful consumer goods, the price might be something of a shock. However, while a CD player might presently cost $300 to $500 and typically last five years after which it is not worth repairing if it fails, turntables generally last for decades. There is no reason why this won’t apply to Dual turntables which have always had a good reputation. Recommended retail price of the Dual CS505-4 turntable, complete with the Ortofon OMB-10 cartridge, is $799.00. Availability For further information about the range of Dual turntables, contact the Australian distributor, Scan Audio Pty Ltd, 52 Crown Street, Richmond, Vic 3121. Phone (03) 9429 2199; fax SC (03) 9429 9309. October 1998  81 RADIO CONTROL BY BOB YOUNG The art of the F3B glider This month, we will look at some aspects in the design of F3B gliders. These are the most exotic of all radio-controlled sailplanes. They are fast, difficult to fly and they are big, with a 3-metre wing span. In the course of our discussions on gliders over the past few months, I have often referred to the most difficult and exotic class of glider or more correctly, sailplane, the F3B model. So what is an F3B model and why is it so exotic? The F3B class of sailplane model is an internationally recognised competition class, which is governed by the FAI (Federation Aeronautique Inter­ nationale), the body which oversees all international aviation activities, both model and full size. The rules for this class are too complex to present in an article of this nature but they may be found in the MAAA Official Rules and Instructions Handbook, General Regulations and Special Rules (page 94). Australia has been well represented over the years in international F3B events and our ranking is quite high as a result. Briefly, the specifications of the model are laid down as maximum surface area (150dm 2), maximum flying mass (5kg), sur­face loading between 12-75g/dm2, minimum radius of the fuse­lage nose (7.5mm). There is one major difference to the 2-metre rules and that is a defined maximum surface area. This includes the tailplane area and it is this rule that dictates the trend to smaller tailplanes that I have often mentioned in recent articles. In the F3B design, the tailplane area subtracts from the wing area and so the F3B designer cannot afford the luxury of large tailplanes, unless he has very definite views in that direction. In the 2-metre design there is no limitation on wing area, only span, so we can afford to be conservative with tailplane size. The necessity for the defined nose radius becomes imme­diately obvious upon witnessing your first launch This photo shows a typical F3B glider with V tail. Note the very smooth nose which contributes to the long landing distance of these gliders. By the way, one wing appears to be longer than the other because of the wideangle lens used to take the photograph. 82  Silicon Chip and especially, the first bad launch. At the speed these things operate at, they would bore a hole through a brick wall. Landing problems Although safety is one aspect of the nose radius specifica­tion, there is probably a more mundane reason behind this radius and it may be found in the specification governing the underside of the model. No fixed or retractable arresting device may be fitted to the underside of the model (no bolt or sawtooth like protuberance). The only fittings allowed underneath are the tow-hook and control linkages and these are defined in size. The reasoning behind these rules is linked to the scoring method applied to spot landings. The distance from the spot is measured from the nose of the model, after it has come to rest. The problem is that these clean models, when they enter “ground ef­fect”, will float on and on forever. And when they do finally touch down they will slide for some distance, particularly on some types of grasses or wet grass. This makes judging the landing a very uncertain activity indeed and losing points on spot landings is a serious business in competitions, where placings revolve around one or two points. Now this gives rise to the one aspect of glider competitions that I do not like and that is the landings. Because of the foregoing rules there are a large number of landings which are nothing more than driving the model into the ground like a javelin. Thus, the rounded nose will not penetrate the ground as well as a sharp point. If the radius were not defined then you can absolutely guarantee some modellers would be flying javelin-style fuselage noses, in spite of the hazard they would present to people on the field. And so we come to the first of the reasons why F3B models are so complex and exotic: the need for some form of drag brake to slow the model down for landing. However, if we continue our reading of the rules, the main reason for the exotic nature of the F3B machine becomes imme­diately apparent when we examine the definitions. The F3B contest is defined as a multi-task event consisting of a duration task, a distance task and a speed task. Fig.1: the plan-form of the Stingray-3M designed to F3B specifications. The main features are the 3-metre wingspan and the full span controllable trailing edges on the wing and tail­plane. Each control segment is fitted with a separate servo to allow it to be operated independently. These three tasks call for three entirely different air­frames with conflicting aerodynamic parameters and yet the F3B competitor is required to combine all three into a single model; a formidable task. Thus, we have the foundations upon which the variable geometry airframe arose, controlled by an increasingly complex radio control system. It was largely the needs of the F3B flier that drove the development of the computer radio with dozens (recently nearly a hundred) of model memories. In order to achieve maximum performance, each task requires the model to be configured aerodynam- ically to a different speci­fication. In many cases, these configurations are stored in sepa­rate memories and these memories are switched in and out during flight. We will look at some of these configurations shortly. Some models use up to six or more memories to store all the configurations required for the three tasks. Thus a 40-memory transmitter can really only store sufficient information for about six models and a serious competition modeller will have at least three or four models on the go at any one time. I often hear modellers talking about computer radios and they are often seen scratching their heads during October 1998  83 Fig.2: the F3B glider has a number of wing section configurations required for the three tasks. As well as the speed, distance and duration configur­ ations, there are settings for launch and CROW (flaps lowered, ailerons raised). these discus­sions about why anyone would want 99 model memories. This is one of the reasons. But it is also true that the average sport flier does not need a radio of anything like this complexity. Unfortunately, the marketing methods now used in all forms of merchandising con­stantly stress that it is impossible to live without the latest gadgets and the average sports flier now burdens himself with an over-complex and overly expensive R/C system just to keep up with the Joneses, or at least the international F3B flier. Fortunately the tide is starting to turn on this trend and I have noticed some interesting non-computer ra84  Silicon Chip dios starting to show up in overseas magazines, designed with the sports flier in mind. Fig.1 shows the plan-form of the Stingray-3M that is de­signed to F3B specifications. The main features are the 3-metre wingspan and the full span controllable trailing edges on the wing and tailplane. These full span controllable trailing edges are the heart of the F3B model. Each control segment is fitted with a separate servo that allows each segment to be operated independently from the transmitter. Thus, the manner in which these independent control surfaces are coupled together via mixers in the transmit­ ter will determine the aerodynamic characteristics of the model. If we consider each of the tasks separately, we will find that they call for completely different aerodynamic parameters. Therefore we must begin our design so that the degree of control we have over the geometry of the control surfaces allows us to compensate for the shortcomings introduced by the most difficult task. In this regard the speed task is the odd one out. Here we are looking at an airframe that has a thin wing section and a smaller wing. Opposing that is the duration task which calls for a very efficient air­ frame, a thicker wing section and a high aspect ratio wing. In the middle is the distance task, calling for a well-balanced airframe combining elements of speed and duration. It would be possible to write books about the design of an F3B model, especially about the choice of wing sections, so I will keep the discussion as simple as possible. Basically, the lift of any one wing section can be in­creased by drooping the trailing edge of the wing and the drag can be reduced by raising (or reflexing) the trailing edge. Therefore to configure the model for a speed run, we need to raise the trailing edge of the entire wing, both flaps and ailer­ons. To configure for a duration task we need to droop the trailing edge. Here we must be careful because the wing tips will tend to stall first, so it is usual to lower the flaps more than the ailerons. The distance task calls for a very carefully selected bal­ance between the two. It is immediately obvious that here is a flying competition that will very quickly separate the men from the boys (and the husbands from their wives), for there is enormous scope for sub­ tlety and understanding of the F3B machine at a holistic level. It is absolutely no accident that the same handful of fliers dominate F3B competitions. Ready-to-fly airframes If you do not have a sound grasp of the interactions that take place between engineering and aerodynamics then you are doomed to failure in this competition. For this reason, many serious competition fliers use almost ready-to-fly airframes and these models are very expensive, costing around $1500 apiece. The wings come complete, ready to mount servos and control surfaces. These wings are masterpieces of materials engineering, being formed from various foams and covered with fibreglass, Kevlar or carbon fibre. Reinforced with carbon-fibre spars, they are immensely strong and yet at one recent competition held on a very windy day, something like 16 models crashed, many suffering wing failure during launch. We have discussed the relationship between wing strength, spar depth, wing sections and aspect ratio in the Stingray articles during the past two months. So even starting with a correctly designed commercial model it is obvious that there is still plenty to occupy the minds of serious F3B competitors and they will travel to the ends of the Earth to obtain that competitive advantage; thus the drive behind the computer radio. It is the computer radio or at least the smart transmitter that has made the modern F3B machine possible. Having said that, I still believe that much of what takes place in modern F3B models is overkill and a good natural flier who practises regularly and pays attention to detail will still outperform the gadget man with all his technology. Setting up the F3B model Fig.2 shows the various wing section configurations re­quired for the three main tasks. These include the speed configura­tion, a possible distance setting and a duration setting. Fig.2 also shows the normal, launch and CROW configurations. The description that follows is a guide to fundamental principles only and the methods used for programming each trans­mitter will vary with brand and model. The ailerons are set up with one aileron servo plugged into the aileron channel on the receiver. The second servo is slaved via a mixer, either inverting or non-inverting, depending on which way the servos are mounted in the wing. The slave servo is usually plugged into channels 6, 7 or 8. A non-inverting mixer results in the two servos moving in the same direction of travel. An inverting mixer reverses the direc­tion of travel on the slaved servo. The correct result in this setup is for one aileron to go up when the other goes down and vice-versa. The throttle stick is often used as the flap master con­trol. One servo is plugged into the receiver throttle channel and the second flap servo is slaved via a mixer to give a result whereby both flaps move in the same direction with the throttle stick movement. This is the basic wing setup. Now comes the tricky part. It may be deemed desirable to couple some droop into the ailerons when the flaps are lowered, in order to increase the lift over the entire wing. This would normally be the case in the duration task, for example, and also to increase lift during launch to maintain the line tension at as high a level as possible. In this case, a small amount of flap setting is mixed into the aileron channel, moving both servos in the same direction. Here an inverted and a non-inverted mixing component is required. One must be careful during launch not to stall a wingtip. For this reason the ailerons are lowered to a smaller degree than the flaps. SILICON CHIP This advertisment is out of date and has been removed to prevent confusion. October 1998  85 move in various degrees. Oddly enough, when full rudder and elevator are applied in one particular direction, only one surface moves to full exten­sion whilst the other remains at neutral. Which way does it turn? Another typical F3B model, again showing the V tail and rounded nose. Many serious competition fliers use almost ready-to-fly airframes and these models are very expensive, costing around $1500 apiece. It is the requirement for mixing ailerons into flaps and flaps into ailerons that dictates the use of two servos, one on each control surface. Likewise it may be seen as desirable to couple some aileron into the flaps to obtain improved manoeuvr­ability. Such a scenario could arise in the speed run where the pylon turns at each end of the course require powerful ailerons. Once again, an inverting and a non-inverting mixing component is required. A variation on this theme is the snap-flap configuration. In this application, some elevator input is mixed into the flaps in an opposing direction. Thus, when the elevators are pulled UP the flaps both move DOWN. This increases the lift on the wings and tightens the pylon turns dramatically. Let me tell you an F3B pilot is as busy as a little beaver switching all of these configurations in and out during a 100km/h speed or distance run. Crow configuration An unusual configuration commonly used for landing is what is termed CROW or Butterfly. In this case the flaps are lowered to their maximum extension and the ailerons are both raised. The raising of the ailerons reduces the overall lift on the wing and prevents tip stall at slow speed, while the lowered flaps in­ crease the lift over the centre section 86  Silicon Chip and also increase drag. The net result is a slower flying, highly controllable model that is much easier to put on the spot. Some care is required when setting the aileron “UP” move­ment, as too much UP will reduce the effectiveness of the ailer­ons. The throttle stick is usually only used for landings to activate the full CROW configuration. The camber-changing config­urations are usually introduced by switches located on the trans­mitter face and call for much smaller control increments. From the foregoing it is easy to see why there has been a need for smart transmitters. Some of these mixing arrangements are very complex and the electronics required calls for some smart design. Coming back now to Fig.1, if we move on to the tailplane, once again we are presented with a mixing problem on the “V” tail. In this case each control surface must share the rudder and elevator functions. We covered this in the 2-metre articles but briefly, in this setup, rudder is mixed into elevator and elevator is mixed into rudder in an essentially cross-coupled mixing arrangement. Each surface has its own servo and when elevator is applied both control surfaces go up or down and when rudder is applied, one goes up and the other down. When both rudder and elevator are applied, the controls For those dumbos like myself who have difficulty figuring out which way the model will turn when setting up the controls in a “V” tail model, here is the rule. The model will turn in the apparent natural direction of the controls; ie, the model will turn right if the trailing edges move to the right of the model (viewed from the direction of travel). I have been flying since 1955 and had never owned a “V” tail model until the Stingray 2M and I was surprised to find that I was uncertain as to which way it would turn. When I presented the model at the field I was surprised to find that all of those present, including people who were flying “V” tail models, were also uncertain. We finally settled the issue by turning on one of the “V” tail models and looking at how it was set up. This may be a simple thing but it is quite confusing, even to so-called ex­perts. One final mixing function is CAR or “coupled aileron/rud­der”. This is sometimes necessary on sailplanes be-cause of an effect known as “aileron reversal”. In this case, the down-going aileron will increase the drag at the wingtip on the outside of the turn, whilst the up-going aileron on the tip at the inside of the turn is reducing drag at that tip. The resulting unbalanced drag forces drag the outside tip back and induce a yaw opposite to the desired direction of turn. This is especially the case in flat-bottom wing sections (Clark-Y, for example) and high aspect ratio wings with under-cambered sections. The answer is to mix some aileron input into the rudder channel, thereby increasing the yaw component in the desired direction. So there you have a very brief overview of an extremely complex subject. I hope it gives you some insight into the art of F3B flying. SC Bob Young is principal of Silvertone Electronics. Phone (02) 9533 3517. Their web site is at: www.silvertone.com.au VINTAGE RADIO By RODNEY CHAMPNESS, VK3UG Behind the lines, Pt.2: the Type 3 MkII (B2) & Type A MkIII spy radios The Type 3 MkII & Type A MkIII were among the best of the spy radios from World War II. In this article, we take a closer look at these two transceivers. Following World War 2, the Type 3 MkII was an extremely popular high-frequency (HF) transceiver with amateur radio opera­tors. It was considered a good Morse code (CW) transmitter and many ingenious amateur operators modified it for amplitude modu­lation (AM) as well. It covered three of the most popular amateur radio bands (3.5-3.8MHz, 7-7.15MHz & 14-14.35MHz) with no modifi­ cations and the receiver was stable and sensitive. Amateur operators extensively modified it to suit their perceived needs when these sets came on the surplus market after the war ended. With such an enthusiastic endorsement of its value by amateur operators, it must have been a good set for its time. And so it was. It wasn’t the smallest or the lightest set around but it was arguably the best for long-range communications over 800km. The smaller sets with more limited frequency ranges and lower power were inadequate for reliable communications over these longer distances. The Type 3 MkII sets were supplied in either a suitcase or in two water- The receiver section of the Type 3 Mk.II consisted of two sub-assemblies, with the RF/converter in one and the IF and audio stages in the other. proof steel boxes. Neither assembly was small or light and concealment of the equipment would not have been easy. The suitcase version, for example, measured 47 x 34 x 15cm and weighed around 15kg. The waterproof steel-box version measured 33 x 25 x 15cm and 34 x 29 x 15cm and had a combined weight of over 25kg – ideal for someone doing weightlifting! Although packed into a case larger than those used for some other sets of the era, this was made up for by the amount of equipment that was included. That said, the space efficiency didn’t compare with the later Type A MkIII. Physically, the Type 3 consisted of three separate assem­blies, the power supply being the heaviest. This comprised an AC power supply suitable for 110V or 230V mains plus a 6V vibrator supply – all using the same transformer. In standard form it produced 500V and 230V DC for the transmitter and receiver, as well as 6V for the heaters (not a power supply to be treated carelessly when switched on). Although the supply used a completely isolated primary winding (so it isn’t a “hot chassis” set), a nasty bite could be experienced when the mains socket was attached to the set one way. There was no 3-core mains lead on this set. The bite was due to capacitor C28 (in the power supply) being wired from one side of the mains to chassis. This capacitor was part of the vibrator “hash” filtering network. A common technique was to mark the socket and plug so that it was only plugged in one way, so that the chassis remained cold. Even so, an unearthed chassis like this must still be regarded as potentially dangerous. October 1998  87 Correction Our circuit for the Paraset on page 77 in last month’s issue omitted the keying circuit for the transmitter. The corrected circuit section is shown above. The Type 3 Mk.II’s transmitter covered from 3-16 MHz, using two valves to give an output of 20 watts CW. Another little nasty with the transmitter involved the Morse key. Due to the fact that the key plug could be inserted either way around in the socket, the frame of the key could be live at 200V (depending on the polarity of the plug). As with the AC power lead, the key plug and the socket were sometimes marked with matching dobs of paint to indicate the polarity. This made it much easier to ensure that the key frame was at earth poten­tial. The receiver consisted of two sub-assemblies, with the RF/converter in one and the IF and audio stages in the other. It had three wavebands from 3.1-15.5MHz, two IF stages and one audio stage. The transmitter covered from 3-16MHz, using two valves to give an output of 20 watts CW. The design of the Type 3 MkII Specifications For Type 3 MkII Spy Radio Transmitter Frequency Coverage: 3-16MHz in eight bands Power Supply: 500V <at> 60mA, 230V <at> 18mA & 6.3V <at> 1.1A Circuit: crystal oscillator working on fundamentals or harmonics (EL32), feeding a class C power output stage (6L6). Power Output: 15-20W, depending on whether mains or battery supply was used and whether operating on a harmonic of the crystal frequency. Receiver Frequency Coverage: 3.1-15.5MHz in three bands. Power Supply: 230V <at> 28mA, 6.3V <at> 1.2A & -12.5V to -14V bias Circuit: 4-valve superheterodyne, essentially designed for CW reception. Intermediate Frequency: 470kHz, BFO 470kHz ±3kHz Sensitivity: 1-3µV for 10mW output at 1kHz (CW input and BFO on) Selectivity: 1kHz bandwidth for 3dB down from peak; 9kHz band­width for 20dB down from peak Maximum Output: 50mW into 120Ω headphones Power Supply Mains Supply: 97-140V AC and 190-250V AC, 40-60 Hz, using a combined AC/ vibrator multiple wound transformer. Power Consumption: 57W transmit; 25W receive Battery Supply: large 6V automotive battery. A single vibrator was used, with a spare carried in the power supply case. 88  Silicon Chip commenced in 1942 but from the information I have it would appear that the design was not finalised until 1943. How many of these sets were produced is un­ known, although it must have been several thousand. As well as being used in Europe, a tropicalised version was also used in South-East Asia during the latter part of the war. They were particularly valuable in Asia because of their good performance over long distances of the order of 2600km. A comprehensive list of transmitter, receiver and power supply specifications were given with the set, plus the circuit, some servicing and installation information. Whilst the circuit and servicing information was probably good to have, one wonders where friendly servicing facilities existed in occupied coun­tries. Certainly wandering down to the local Gestapo radio man for parts and facilities to service the set was not an option. The accompanying panel lists the abbreviated specifications of the unit. In view of its intended role, it was quite a good performer for that era. A explained last month, the Gestapo would remove power from a block of buildings when close to a clandestine radio station and observe whether the agent’s radio transmissions ceased immediately. If they did, they had him or her bottled up in a small area and would soon find whoever it was. For this reason, the ability to change over from mains to battery operation within a second or two whilst oper- ating was important if the operator wasn’t going to be caught. The Type 3 could be changed over quite quickly, by switching the power off, extracting a plug from a socket on the power supply front panel, turning it through 180 degrees and re-inserting it into the socket and turning the power on. In all, that took around 2-3 seconds and was supposedly fast enough to prevent a Gestapo radio detection group from detecting the momentary gap in transmission. Personally, I be­ lieve that the time to make the changeover was too great and the detection group would have been suspicious. In fact, an automatic change-over when mains power was lost could have been achieved with only a slightly more complicated power supply. In fact, changeover was effected much more quickly in the Type A MkIII which became available the following year. So let’s now take a look at the Type A MkIII. The Type A MkIII I first came to know the Type A MkIII as a young lad in­volved with communications for the Emergency Fire Services of South Australia. This particular set had been considerably modi­fied, in that the power supply had been removed and a plate and screen modulator had been installed in its place, converting the transmitter from continuous wave (Morse code) transmission to voice (AM) transmission. It worked quite satisfactorily putting out a magnificent 3 watts into a 2.4-metre loaded whip antenna on a frequency of 5790kHz. It was great fun being able to talk “over the air” and it started me on my way to obtaining an amateur radio licence. The Type A MkIII is an interesting little transceiver and I do mean little. If you take a look at the average domestic receiver of the early 1940s you will appreciate the compactness of the Type A MkIII. The unit consisted a 2-band 4-valve superheterodyne receiv­er capable of CW and AM reception, a 2-valve CW transmitter capable of 5W output, and a power supply that could operate on either 110V or 240V AC power. This was all packed into a case that measured just 21.6 x 19 x 8.3cm and weighed 3.75kg (or 7.7kg in its suitcase packaging). A separate adaptor box containing a vibrator supply allowed Type A MkIII transceiver used a 2-band 4-valve superheterodyne receiv­er capable of CW and AM reception, plus a 2-valve CW transmitter capable of 5W output. It is shown here without its power supply. the set to operate off 6V DC. It was quite a remarkable achievement considering that there were few miniature parts in those days and the valves are full size loctals. Being so compact, the set generated a lot of heat and so there were lots of perforations in the case to pro­vide ventilation and aid cooling. The set had a normal range of around 800km. If greater ranges were required, the Type 3 MkII was used. As with the Type 3, the Type A WARNING! The two transceivers featured in this article do not meet modern electrical mains safety standards and could be quite dan­ger­ous to operate. If you come across these transceivers or any similar equipment (eg, any of the equipment mentioned last month in Pt.1), we recommend that you leave them strictly alone. Do not be tempted to restore, modify or operate them in any way unless you are a qualified person who has the necessary expertise to ensure safe operation (and you have an amateur radio operator’s licence). transceivers were supplied in suitcase packaging or in two fully waterproof steel boxes. Just how many of these units were actually made is difficult to say. The set in the photograph carries a serial number of “32441” which is just visible at the top right. Only a few hundred agents were equipped with radios, so these high serial numbers appear to be a ploy to make the enemy think that there were many more agents in the field than was actually the case. A comprehensive list of transmitter, receiver and power supply specifications were given with the set, plus the circuit and information on how to set the equipment up and use it. An extract of these specifications is shown in the accompanying panel. Performance comparisons So how does this set compare with the 4/5 valve domestic radios of the same era? Actually, the receiver compares very favourably. A typical domestic set had a sensitivity of 5µV on the broadcast band and 15µV on shortwave. The selectivity was about the same but the image rejection of the Type A would have been quite a bit better because it used a 1215kHz IF compared to the 455kHz IF used in most domestic sets. On the other hand, the audio output is well down, the Type A MkIII being designed to work into headphones October 1998  89 This view inside the Type A Mk.III transceiver shows just how tightly the parts were packed together. only. After all, they didn’t want to entertain the local Gestapo officers! There was also one very nasty design inadequacy in the receiver. It wasn’t nice to be listening on the headphones if a burst of static occurred. The intensity of the noise was so high that it nearly lifted the operator’s head off. This problem could have been easily solved by wiring back-to-back diode strings across the headphone socket. With this scheme, the audio quality is unaffected but when a static crash (or some other form of electrical interference occurs), the diodes conduct and quench the interference spikes. In fact, some military sets from World War II did include “crash limit­ers” and these used copper oxide diodes. One clever aspect of the design Specifications For Type A Mk III Spy Radio Transmitter Frequency Coverage: 3.2-9MHz in two bands. Power Supply: Built in supply or 6V battery pack. 270V <at> 50mA and 6.3V <at> 0.75A (1.65A with receiver heaters). Transmitting Circuit: Pierce oscillator (7H7) impedance coupled to class-C beam tetrode (7C5). Power Output: 5W on fundamental; 4W on second harmonic. Receiver Frequency Coverage: 3.2-8.55MHz in two bands. Power Supply: Built in supply or 6V battery pack. 250V <at> 35mA and 6.3V <at> 0.9A (1.65A with transmitter heaters). Receiver Circuit: 3-valve superheterodyne with 1215kHz IF (the 7H7 audio output stage also served as the oscillator valve for the transmitter). Sensitivity: 2-4µV CW for 1mW output into 800Ω. Selectivity: 10kHz bandwidth for 20dB down at critical reaction and 30% modulation. Maximum Output: 100mW. AC Power Supply Mains Supply: 100-130V AC and 200-250V AC 40-60Hz Power Consumption: Transmit - 30W key down and 20W key up; Re­ceive - 25W Battery Supply: 6V accumulator (large capacity) Current Consumption: Transmit - 5.5A key down, 3.8A key up; Receive - 4.3A 90  Silicon Chip was that the audio output valve for the receiver was switched to become the crystal oscil­lator for the transmitter. This saved a valve and required just one switch section to effect the changeover. In operation, the Type A was mostly used on the mains but with the DC vibrator power supply all connected up and ready to use in case of a mains interruption. If the AC power was cut off, the agent immediately pulled a ring (shown on the centre left of the photograph) which changed the set over to DC operation. In all, this took no more than about 1-2 seconds. This meant that the Gestapo couldn’t be sure if the radio transmitter was actually in that block of buildings, as the transmissions continued on. At least that was the theory. In practice, it was probably only effective until the Gestapo captured one of the sets and discovered that they had dual power supplies. Warnings Electrically, the Type A MkIII (and many of the other spy radios of the World War 2 era) could be quite dangerous. This is best summar­ised by a warning in the handbook: “WARNING - The HT, -ve line & parts of the chassis & metal components are connected directly to the AC mains and if touched, may give a dangerous shock”. And that really was quite an understatement. The Morse key could also be at mains potential. For this reason, the key supplied with the set was fully insulated to prevent the operator from coming into contact with any live parts. In fact, the only way to safely service such sets was to use an isolation transformer and even then, you had to be care­ful. Finally, if anyone would like to find out more about the activities of the Resistance and the radios they used, I recom­ mend that you try to obtain a copy of “Secret Warfare” by Pierre Lorain, translated/adapted by David Kahn and published by Orbis Publishing Limited, London. Another book well worth reading is “S.O.E. The Special Operations Executive” by M.R.D. Foot, published by Mandarin. I’m not sure whether or not these books are still in print and I understand that very few copies of “Secret SC Warfare” came into Australia. ASK SILICON CHIP Got a technical problem? Can’t understand a piece of jargon or some technical principle? Drop us a line and we’ll answer your question. Write to: Ask Silicon Chip, PO Box 139, Collaroy Beach, NSW 2097. High voltages in insulation tester On completing the Insulation Tester kit from your May 1996 issue, I found the voltages at the cathode of D3 are slightly higher than indicated in the article. I measured 121V, 273V, 549V, 633V and 1048V. The voltage at TP2 and ground is 1.98V. I have rewound T1 and checked the resistors around the error ampli­fier but have come up with the same results. Are the voltages within tolerance and I need not worry about them or do I have a problem? The circuit otherwise is working fine. (G. M., Seven Hills, NSW). •  While the 100V test voltage is 21% high and the others are also high, this will not cause any problems with the functioning of the Tester. Use it without worrying. More on decent sound from phono preamps I can sympathise with P. D. of Mt Colah (Ask SILICON CHIP, July 1998) in his attempt to wring decent sound out of an IC-based phono preamp. I too built the Series 5000 system but did not bother to check the phono input until some years later. I used a Shure N97HE stylus in a Rega turntable and initially thought that it lacked bass. I checked the phono input with a signal generator, audio SSVM and CRO and can confirm that RIAA equalisation was spot-on and that the output was an accurate copy of the input. However, as I listened it became obvious that the sound was overly bright, harsh and metallic in the mid to high frequencies. However, the problem was more than just a boosted mid-range; it was also distorted. In just the same way that some power amps of this era could easily pump out several hundred watts of sinewave energy into a resistive load but still produced a distorted sound on a musical input (finally identified as transient intermodulation distor­ tion), I assume a similar problem afflicts these IC-based pre­ a mps. SILICON CHIP Modification for Sun tracker I am building a sun tracker unit for solar panels, as fea­tured in the January 1995 issue of SILICON CHIP. At the end of the article, the editor noted that a toggle switch could be inserted in series with diode D1, which would stop any erratic movement of the panel on cloudy days. My solar panel array will be approximately 7 metres above the ground; ie, 4 metres to the roof and 3 metres above the roof to raise the panels somewhat above reflected heat. My question is, is there a small electrically-operated on/off switch available that could be mounted on either the PC board or close to it, that could be turned on or off by an elec­trical impulse through light Certainly no commercial hifi preamp used ICs at this time at any point in their circuitry and it is probably still the same today. I did not persevere with tracking down the problem since I only had to reconnect the turntable into a cheap NAD amplifier to have superb sound again. I also cannot comment on the SILICON CHIP design since I have not built it but I am not surprised at P. D.’s findings. I wonder whether some designs are actually tested in the real world? While Shure state in their literature that their cartridges “like to look into” a 47kΩ resistive load in parallel with 200-300pF of capacitance, I do not think this should be the prob­ lem, since their cartridges work just fine in commercial amplifi­ers, even cheap combination units, without any wiring to possibly a momentary toggle switch, which would be on the main control board in the building. I don’t want a radio-controlled switch (if such is avail­able) but of course would go this way if it was the only option. (I. M., Miles, Qld). •  One possible solution to a remote switch for the output of comparator IC1b is to use a transistor to short out resistor R6 instead. The accompanying circuit should do the trick. This advertisment is out of date and has been removed to prevent confusion. October 1998  91 modification. I cannot recall whether I did actually check the Series 5000 preamp in this regard – maybe these ICs do have an internal capacitance which upsets the Shure cartridge?. (Note that the cartridge loading also includes tonearm wiring). I also built the metal detector kit in the May 1998 issue but the unit will only oscillate after switching off and on again by “kick starting” it, either through touching the resistors near the trimpot or by adjusting or touching the trimpot itself. I was going to mount an external trimpot on the case but thought that you might have a better suggestion. (T. G., Henty, NSW). •  As far as the ETI 5000 preamplifier is concerned, if it is distorting then that indicates an overload somewhere in the system. Our LM833 design published in the April 1994 issue has very good bass as it is very close to the RIAA equalization characteristic. In fact, one of our staff recently spent several hours listening to our remote controlled preamplifier (it uses 5534 op amps with the same feedback components) and the sound was superb. Many IC preamplifiers are not designed with sufficient input overload margin. Our RIAA preamp design has an input signal capacity of 300mV (at 1kHz) which is more than adequate for any magnetic cartridge. As far as the Metal Locator is concerned, hand capacitance has quite a large effect as you are adjusting the trimpot, so to get it working correctly, VR1 has to be adjusted a number of times. Plaudits for class-A amplifier module My hearty thanks for the new class-A module, coming as it does in the depths of winter when such thermal generators are most appreciated. Joking aside, this article has done much to rekindle my interest in audio amplifiers, particularly as you have gone to some lengths to overcome the development hurdles to present it. With the considerations you have drawn attention to in your article, I plan to adapt the module as an AB amplifier with appropriately higher rail voltages (and of course, uprated driv­ ers, etc) to produce about 40 watts. If nothing else, this should give a bit of useful headroom over the 15W design. However, I am reluctant to use such expensive (however excellent) devices as the MJL21193/4. Noting that the linearity of these devices con­ tributes to the lower distortion of this design, I would not like to substitute inappropriately. Consequently, assuming that by transistor linearity you mean Hfe versus collector current (or is it base impedance?), would you consider publishing the relevant linearity and SOA graphs, possibly serialised, for a few commonly available power transistors? May I suggest, for example, the venerable Motorola types 2N3055/ MJ2955, MJ802/4502, MJ15003/4, MJE3055/2955 and the C versions of Texas Instruments’ TIP31/2, TIP41/42, TIP33/4 and TIP35/36. The TOP3/TO218/SOT93 packages which are compatible with T0264 pin spacings are preferred, where optional. If this is not possible, perhaps you could make relative assessments and suggestions on the above types. Also, in your comments on the design of the output stage, you stated that the current feedback pair was more linear that the usual Darlington. Can you enlarge on this? Whilst you Which power module to use? What’s in it for you? Inverters have matured very quickly in the last two years. Technology was forced forward to meet the demands of an aware market (people). The largest global manufacturer of inverters and chargers have released the 1998 models. NEW! Really, NEW technology, features and benefits! Call us now for your nearest Dealer. Australia wide. AU STR AL IA Simply Brilliant Technology BAINBRIDGE TECHNOLOGIES PTY LTD 77 Shore Street, Cleveland Brisbane Qld 4163 PH: (07) 3821 3333 Fax: (07) 3821 3977 Email: baintech<at>powerup.com.au Internet: www.statpower.com 92  Silicon Chip I am somewhat confused as to which power amplifier to use in my stereo system. Currently I have an electronic crossover (3-way) which has electronic delay compensation for the tweeter and gradual boost above 7Hz to 35kHz. The bass is likewise boosted below 50Hz to compensate for hearing loss at those frequencies and this in turn drives six power amplifiers which currently are the old Sanken hybrid 50W jobs (2 x 20W for the tweeters actual­ly). After reading your magazine for several months there ap­pears to be several choices to upgrade my power amplifiers. First, there is a Mosfet 50W module, then there is the National 50W IC modules and now, as published in the July & August issues, there is a third option of a 15W class-A module. Could you advise me as to which of the above modules (or any others I’m not aware of) would give the best sound for a home system. By the way, the speakers are KEF units but I’m upgrading the midrange and tweeters to Dynaudio units in the near future. (S. P., Auckland, NZ). •  In terms of all-round value for money it is hard to go past the National 50W IC modules, especially if you need quite a few of them for a multi-channel system. The 15W class-A system is superb but would be a very expensive proposition for your system and its power level may not be adequate. Accelerometers for experimentation I am a retired physicist who would like to experiment with accelerometer ICs. I know they are expensive but I don’t know where to get them. Analog Devices have them but only with a 50G range, far too insensitive for my application. Exar manufacture them with various ranges down to 2G but I don’t know their ad­dress and they would not sell one or two to individuals anyway. I would be grateful for any information you may be able to provide. (G. F., Vermont, Vic). •  We know of no source of accelerometers which have a range of operation below several G. This is de­ scribed how it is understood to operate as an amplifier, I could not tell from the discourse where relative performances were implied. (I. F., Urunga, NSW). •  If you want to substitute transistors, the MJ15003/4 are the closest equivalents, with the MJ802/4502 being less preferable. We would not recommend any of the others for a low distortion design. Linearity in a transistor context is beta versus collec­tor current. The current feedback pair is more linear than a Darlington emitter-follower because of its 100% current feedback. This means that the distortion before overall negative feedback is applied is much lower. We are not at all confident that similar results to the class A design would be obtainable by upgrading the design and operating in class AB mode. If it was that easy, our previous class AB amplifiers should have produced because most accelerometers have become available due to their use in automotive airbag electronics and they need to operate at fairly high G. The Analog Devices ADXL76 can be used for applications well below 1G but only by adding an external op amp to increase the output level. The data sheet on this device shows how to do this and sensitivities of 400mV/G can be obtained. Higher gains than that shown on the data sheet can be obtained providing you choose a low drift op amp. Analog Devices components can be obtained from Insight Electronics Pty Ltd in Melbourne. Phone (03) 9760 4277. better results. They were good but the class A module is far better. Wanted: plastic tank tracks for large models Help! Do any SILICON CHIP readers know where to get plastic or rubber caterpillar-type "tank tracks" as used on robots, etc? I'm after some fairly large ones for a project. (Ross Tester ross<at>silchip.com.au) Notes & Errata Motor Speed Controller, June 1997: the text on page 30 states that “you should be able to measure about +12V at pin 16 .. of IC1”. Pin 16 is grounded. It should refer to pin 12. Positive Earth HEI, November 1997: the circuit on page 90 shows a BC337 for Q3. It should be a BC327. The .01µF capacitor associat­ed with D1 should be marked C2, not C1. High Energy Ignition, June 1998: the 0.1µF capacitor shown on the overlay diagram for the points version (near diode D1) should be .01µF, as shown on the circuit. On-Board Mixer for R/C Receivers, July 1997: the circuit on page 79 shows diode D3 reverse-connected. Its cathode should connect to pin 3 of IC2a. Opus One Loudspeaker System, August 1998: we have been advised by Altronics that the crossover network circuit on page 5 should show the tweeter reversed in phase, to agree with the pre-assembled crossover SC networks. WARNING! SILICON CHIP magazine regularly describes projects which employ a mains power supply or produce high voltage. All such projects should be considered dangerous or even lethal if not used safely. Readers are warned that high voltage wiring should be carried out according to the instructions in the articles. When working on these projects use extreme care to ensure that you do not accidentally come into contact with mains AC voltages or high voltage DC. If you are not confident about working with projects employing mains voltages or other high voltages, you are advised not to attempt work on them. Silicon Chip Publications Pty Ltd disclaims any liability for damages should anyone be killed or injured while working on a project or circuit described in any issue of SILICON CHIP magazine. Devices or circuits described in SILICON CHIP may be covered by patents. SILICON CHIP disclaims any liability for the infringement of such patents by the manufacturing or selling of any such equipment. SILICON CHIP also disclaims any liability for projects which are used in such a way as to infringe relevant government regulations and by-laws. Advertisers are warned that they are responsible for the content of all advertisements and that they must conform to the Trade Practices Act 1974 or as subsequently amended and to any governmental regulations which are applicable. October 1998  93 MARKET CENTRE Cash in your surplus gear. Advertise it here in Silicon Chip. FOR SALE SPEAKERWORKS: specialist in speaker repairs and parts. DIY refoam kits: 31/2", 4", 5", 6", 7", 8", 9", 10", 11", 12" and 15" $39.95. Includes shims, dustcaps and adhesive. Largest inventory of cones, surrounds, gaskets, spiders, dustcaps, grilles, foam and cloth and 4,700 custom voice coils. Phone 02 9420 8121, Fax 9420 8131. ELECTRONIC ENGINEERING SERVICES: digital & analog, embedded & Windows/PC based designs, complete solutions or design advice/assistance. Phone 03 9807 9886. Email caddy<at>netspace.net.au PIC84/12 PROGRAMMERS: Many models available. Also other PIC-driven devices. EST (02) 9789 3616 or www. internetezy.com/au/~sesame PCBS MADE, ONE OR MANY. Low prices, hobbyists welcome. Sesame Electronics (02) 9554 9760 CLASSIFIED ADVERTISING RATES Advertising rates for this page: Classified ads: $10.00 for up to 12 words plus 50 cents for each additional word. Display ads (casual rate): $25 per column centimetre (Max. 10cm). Closing date: five weeks prior to month of sale. To run your classified ad, print it clearly on a separate sheet of paper, fill out the form below & send it with your cheque or credit card details to: Silicon Chip Classifieds, PO Box 139, Collaroy, NSW 2097. Or fax the details to (02) 9979 6503. ____________ ____________ ___________ ___________ ___________ ____________ ____________ ___________ ___________ ___________ ____________ ____________ ___________ ___________ ___________ ____________ ____________ ___________ ___________ ___________ ____________ ____________ ___________ ___________ ___________ Enclosed is my cheque/money order for $­__________ or please debit my ❏ Bankcard   ❏ Visa Card   ❏ Master Card Card No. Signature­­­­­­­­­­­­__________________________  Card expiry date______/______ Name ______________________________________________________ Street ______________________________________________________ Suburb/town ___________________________ Postcode______________ 94  Silicon Chip sesame<at>internetezy.com.au http:// members.tripod.com/~sesame_elec C COMPILERS: everything you need to develop C and ASM software for 68HC08, 6809, 68HC11, 68HC12, 68HC16, 8051/52, 8080/85, 8086 or 8096: $145.00 each. Macro Cross Assemblers and Disassemblers for above CPUs + 6800/01/03/05, 6502 and 68HC12 now combined at the new low price of $75. Debug monitors: $75 for 6 CPUs. All compilers, XASMs and monitors: $480. 8051/52 Simulator (fast, now incl. 80C320): $75. Try the C-FLEA Virtual Machine for small CPUs, build a “C-Stamp”. Demo desk: FREE. All prices + $5 p&p. Atmel Flash CPU Programmer: Handles the 89Cx051, the 89C5x and 89Sxx series, and the new AVRs in both DIP and PLCC44. Also does most 8-pin EEPROMs. Includes socket for serial ISP cable. $189, $35 tax, $10 p&p. 20-pin SOIC adaptor only $70. Credit cards accepted. GRAN­TRONICS PTY LTD, PO Box 275, Wentworthville 2145. Ph (02) 9896 7150 or Internet: http://www.grantronics.com.au RAIN BRAIN AND DIGI-TEMP KITS. Also 60 channel Moni-temp with alarms and PC Data logging. Mantis Micro Products, 38 Garnet Street, Niddrie, 3042. (03) 9331 4786. Fax (03) 9331 4782 http://www.home.aone.net.au/mantismp WEATHER STATIONS: Windspeed & direction, inside temperature, outside temperature & windchill. Records highs & lows with time and date as they occur. $399.00 complete plus sales tax if appli­cable. Optional rainfall and PC interface. Used by Government Departments, farmers, pilots, and weather enthusiasts. Other models with barometric pressure, humidity, dew point, solar radiation, UV, leaf wetness, etc., etc. Just phone, fax or write for our FREE catalogue and price list. Solar Flair/Ecowatch ph: (03) 5968 4863 fax: (03) 5968 5810, PO Box 18, Emerald, Vic., 3782. HOMEBUILT DYNAMO, engineering dreams into reality. “An absolutely marvellous book for the true ex­ perimentalist!” Elektor Electronics. (www.onekw.co.nz) TELEPHONE EXCHANGE SIMULATOR, SC February 1998. Test all sorts of equipment without the cost of extra telephone lines. Melbourne 9806 0110. RTN Australia Parallax distributor: Basic Stamps, SXKey develop­ ment tools and SX chips. Wireless RF modules, serial LCD modules, Basic Stamp Bug, etc, etc. FerretTronics >R/C servo control chips. NEW: HandyScope 2 from Europe, 2 channel/12 bit portable measur­ i ng WATER RESISTANT DMM Rugged construction Drop proof to 10' Water resistant Push-button range selection Min/max memory Ideal for field service Computronics Corporation Ltd 6 Sarich Way, Technology Park, Bentley, WA, 6102 Ph. 08 9470 1177 Fax 08 9470 2844 Specifications at www.computronics.com.au Need prototype PC boards? We have the solutions – we print electronics! Four-day turnaround, less if urgent; Artwork from your own positive or file; Through hole plating; Prompt postal service; 29 years technical experience; Inexpensive; Superb quality. Printed Electronics, 12A Aristoc Rd, Glen Waverley, Vic 3150. Phone: (03) 9545 3722; Fax: (03) 9545 3561 Call Mike Lynch and check us out! We are the best for low cost, small runs. KITS-R-US PO Box 314 Blackwood S.A. Ph/fax 08 8270 3175 FMTX2A Universal Stereo Coder $49 FMTX2B 30mW Xtal Locked 100MHz Transmitter $49 FMTX1 1-3 Watt Free Running Transmitter $49 FMX1 200mW Full Broadcast Transmitter, built & tested $499 FM220 10-18 Watt FM BGY133 Philips Linear $499 FM1525 25 Watt Discrete Linear FM Band $499 FM2100 110 Watt Discrete Linear FM Band $699 FM3000 300 Watt Discrete Linear FM Band $1499 Philips 828E/A VHF Receiver Boards (6 metres) $9 AWA 721 VHF Receiver Boards (2 metres) $9 AWA 721 VHF transmitter boards 1 watt (2 metres) $19 Philips 323 UHF transmitter boards 500mW (70cm) $19 AEM 35 Watt Little Brick Audio Power Amp $15 Digi-125 200W RMS Audio Power Amp $39 CA Clipper Compiler, new in box $49 6dBd Gain Colinear FM Band Antenna $999 Roll Smart-1 FM Station Audio Processor $999 Free catalog on disk of discounted surplus components Same day shipping, credit cards OK, circuits supplied. SPECIAL STEAM BOAT KITS $14 !!!CAR ALARMS $99!!! VIDEO SURVEILLANCE & CCTV CAMERAS & EQUIP­MENT. SPECIALS: 380 + LINE x 0.2 Lux SILICON MODULE only $69! DOME HOUSINGS only $10! 50 LED DIY Infra-Red Illuminators only $19! MODULES: AWFUL-CMOS only $49! PREMIUM 400 + Line x 0.05 Lux SONY H.A.D. CCD & CHIPSET from $91. CAMERAS: Mini 36 x 36 from $88. Dome from $91. DIGITAL COLOUR CAMERAS & MODULES: 400 + Line from $180! DOME from $189! 600 + Line from $346! ACCESSORIES: 30 + Lenses, Infra-Red Illuminator Kits, IR LEDs, Polarising, Co­lour, Infra-Red, Temperature Conversion, Cut & Pass Filters for Image Enhancement, Exposure, Colour Correction, Focus & Glare Control. ANCILLARY EQUIPMENT: Quads 4 pix 1 screen from $280. SWITCHERS 4 & 8 Ch from $126. MULTIPLEXERS FULL-SCREEN FULL-RESOLUTION VCR Recording/Playback from $826. ALSO: Monitors, Outdoor Housings, Brackets, Dummy Cams, CCTV-TV/ VCR I/F Modules, Motorised Pan Units etc. CCTV-TV/VCR Modulator/Mixer/Amplifier Modules from $14. PACKAGED SETS! QUAD + 4 CAMERAS + Power Sup­ plies from $689. 400 + Page CCTV Technical Reference Manual $95 or FREE! 2 Year WARRANTY available for most items! DISCOUNTS: based on ORDER VALUE, BUYING HISTORY, for CASH/CHEQUE & NEW ZEALAND BUYERS! BEFORE you BUY Ask for our Illustrated Catalogue/Price List with Application Notes. Allthings Sales & Services 08 9349 9413 Fax 08 9344 5905. Positions At Jaycar We are often looking for enthusiastic staff for positions in our retail stores and head office at Rhodes in Sydney. A genuine interest in electronics is a necessity. Phone 02 9743 5222 for current vacancies. instrument, it’s a voltmeter, digital storage CRO, transient recorder and spectrum analyser. All in a very small box powered off a parallel port. DOS and Windows software provided. Ph/ Fax (03) 9338-3306. email: nollet<at>mail.enternet.com.au http://people.enternet.com.au/~nollet LOGIC ANALYSER 100Ms/s 32-Channel Kit $1275. Stand alone, not a plug in PC Card. Requires a VGA or EGA monitor - user supplied. Edge and Level Triggering. Multiple Triggering Modes such as, Trigger on pulse width too long or too short, Clock Stop, User Defined Storing, 2 Level Sequencer. Request brochure from: Peter Baxter, Tantau Australia, PO Box 1232, Lane Cove 1595, Sydney. Ph: 02 9878 4715 Fax: 02 9888 7679 Email: peter.baxter<at>tantau.com.au. All manuals on the website: www.tantau.com.au. Revised, no prototype area, “8051 Proto-Board” EA Feb 93. $30. AMATEUR, CB RADIO & other Consumer Electronics Trading Centre can be found at www.mackay.net.au/~ajl ELECTRONIC ENGINEERING SOLUTIONS: No matter what problem what industry we will find you a solution that meets your needs. Specialising in schematic & PCB design, custom Windows based software, embedded control, Windows/PC based test equipment, turnkey solutions. Fast turn around with competitive rates. DAMUE PTY LTD, 46 Whitby Road, Kings Langley NSW 2147. Phone (02) 9624 2802. Fax (02) 9624 2651 or E-mail alovell<at>ibm.net A NEW address for Acetronics http://www.acetronics.com.au On-line PCB quotes, free software, DIY PCB supplies plus many other items & services. 02 9743 9235. KIT ASSEMBLY ANY KITS ASSEMBLED: professional, speedy service. Phone Neville Walker (07) 3857 2752. KITS ASSEMBLED: $20/h, max. fee $60. Phone Russell Grif­fiths (03) 5486 5410. RMB 3170 Rochester Vic. 3561. October 1998  95 14 Model Railway Projects Advertising Index Bainbridge Technologies..............92 Computronics..............................95 Shop soiled but HA LF PRICE! Our stocks of this book are now limited. All we have left are newsagents’ returns which means that they may be slightly shop soiled or have minor cover blemishes. Otherwise, they're undamaged and in good condition. SPECIAL CLEARANCE PRICE: $3.95 + $3 P&P (Aust. & NZ) This book will not be reprinted Yes! Please send me _____ copies of 14 Model Railway Projects at the special price of $A3.95 + $A3 p&p (p&p outside Aust. & NZ $A6). Enclosed is my cheque/money order for $­A__________ or please debit my Dick Smith Electronics..................... ................................ IFC,OBC,12-15 EMC Technologies.......................17 Harbuch Electronics....................58 Instant PCBs................................95 Jaycar .............................. 45-52,95 Kalex............................................59 Kits-R-Us.....................................95 Microgram Computers...................3 Printed Electronics.......................95 Procon Technology......................95 Quest Electronics........................93 Scan Audio.............................85,91 Silicon Chip Bookshop.................11 Silicon Chip Subscriptions...........21  Bankcard     Visa Card    MasterCard Silicon Chip Wallchart..................65 Card No. Solis.............................................96 Signature­­­­­­­­­­­­___________________________  Card expiry date______/______ Taig Machinery............................56 Name ______________________________________________________ PLEASE PRINT Street ______________________________________________________ Suburb/town_________________________________ Postcode_________ Send your order to: SILICON CHIP, PO Box 139, Collaroy, NSW 2097; or fax your order to (02) 9979 6503; or ring (02) 9979 5644 and quote your credit card number (Bankcard, Visa Card or MasterCard). HELP SAVE THE NIGHT SKY! We are losing our heritage of starry night skies. Poor, inefficient outdoor lighting is causing glare and “light pollution”. This wastes energy and increases greenhouse gas emissions. You can help by joining SYDNEY OUTDOOR LIGHTING IMPROVEMENT SOCIETY (SOLIS). SOLIS aims to educate and inform about quality outdoor lighting and its benefits. We also lobby councils, government and other bodies to promote good lighting practice. SOLIS meetings are held third Monday night of each month at Sydney Observatory. Individual membership is $20 pa. Donations are also welcome. Cheques payable to “SOLIS c/- NSAS”, PO Box 214, West Ryde 2114. Email: tpeters<at>pip.elm.mq.edu.au 96  Silicon Chip Truscott’s Electronic World...........59 Valve Electronics.........................58 Zoom EFI Special......................IBC _____________________________ PC Boards Printed circuit boards for SILICON CHIP projects are made by: •  RCS Radio Pty Ltd, 651 Forest Rd, Bexley, NSW 2207. Phone (02) 9587 3491. •  Marday Services, PO Box 19-189, Avondale, Auckland, NZ. Phone (09) 828 5730. Circuit Ideas Wanted If you have a good circuit idea, sketch it out, write a brief description & send it to us for publication in Circuit Notebook. We pay up to $60 for a good circuit but don’t make it too big please. MORE FROM YOUR EFI CAR! Own an EFI car? Want to get the best from it? You’ll find all you need to know in this publication EFI TECH SPECIAL Here it is: a valuable collection of the best EFI features from ZOOM magazine, with all the tricks of the trade – and tricks the trade doesn’t know! Plus loads of do-it-yourself information to save you real $$$$ as well . . . HERE ARE JUST SOME OF THE CONTENTS . . . n Making Your EFI Car Go Harder n Building A Mixture Meter n D-I-Y Head Jobs n Fault Finding EFI Systems n $70 Boost Control For 23% More Grunt n All About Engine Management n Modifying Engine Management Systems n Water/Air Intercooling n How To Use A Multimeter n Wiring An Engine Transplant n And Much More including some Awesome Engines! AVAILABLE DIRECT FROM SILICON CHIP PUBLICATIONS PO BOX 139, COLLAROY NSW 2097 - $8.95 Inc GST & P&P To order your copy, call (02) 9979 5644 9-5 Mon-Fri with your credit card details! FROM THE PUBLISHERS OF “SILICON CHIP”