.elextor-magazine.co eMmagazine January & February 2014 USS14.00 - canada $16.00 5669824969 | |WHERE THE Crea ae THE BOARD PLS [en Maye Blain ola 0AM nom eal mA ale |-TelTelCcte Cod tal -Moal| OM oer 1Ce Tale VAC Cerrina eeo MANOA e (te MaMa aeselsxe Ke aR ae NUN) PoLeIA NMA Ue oMe Solara Cctota] Wee Mao eel creccela Cone esa Explore 100+ Technical Tutorials, Panels & Case Studies for the latest design theories, methodologies and applications Meet 150+ Exhibitors showcasing advanced design tools & solutions to your biggest SI, Pl & verification challenges NTT Cat oa aco) Le CLL a AN UL eae M CI ooo ModaaTaaLUlal forsee aT CC] ColO Mal RSMO Ma uaa ee Onl mM rearacel (oscar RS TCeR (RSM) ATLA eLoLeRSM ae Oty Aeon ool He Uo cl ena Acre] DYNO UIs an 19) NYA g ULM OT oa Co MLA NU ne Celie AEM TOC RET la asi) January 29 & 30 SPIRO Ee MM eee ACC ROP Leia wyDoes your design need class-lea code density and ADC throughput? g 32-bit MCU performance, PIC32MZ offers 330 DMIPS and 3.28 CoreMarks™/MHz performance PIC32MZ Embedded Connectivity MCU Family MB Flash, 512 KB RAM, 330 DMIPS. ‘s new PIC32MZ 32-bit MCUs achieve class leading performance, combined with ADC throughput of 28 Msps, 30% better code density and up to 2 MB dual-panel Flash with live update and 512KB RAM. The PIC32MZ Embedded Connectivity (EC) family of 32-bit MCUs introduces a breakthrough in high-end embedded control with its class-leading performance and code density in addition to new levels of on-chip memory and peripheral integration, With up to 2 MB of dual-panel Flash and 512 KB of RAM, the PIC32MZ offers 4x more on-chip memory than any other PIC® MCU, with fail-safe operation during live Flash updates. Itis also the first PIC MCU to use the enhanced MIPS microAptiv™ core which adds 159 new DSP instructions that enable the ‘execution of DSP algorithms at up to 75% fewer cycles than the PIC32MX families. Advanced connectivity is supported over Hi-Speed USB, 10/100 Ethernet and ‘two CAN 2.0b modules as well as multiple UART, SPI/PS, and °C channels. The optional on-chip crypto engine ensures secure communication with a random number generator and high-throughput data encryption/ decryption and authentication, DEVELOPMENT TOOL SUPPORT: m= Turn-key PIC32MZEC Starter Kits = Multimedia Expansion Board II = PIC32MZ2048EC Plug-in Module for Explorer 16 PICOMZEmbedded Connectivity Starter Kit (0320006 or D4520006-C with crypto engine) For more information, go to: www.microchip.com/get/eupic32mz Microcontrollers * Digital AS\ MicrocHi ignal Controllers + Analog * Memory * WirelessContents (ektor © Community 8 Elektor World + Postermania + The Elektr Studio + World Cass Tips ‘From Tracing Paper to Computer: Elektor Schematics Layout + Raspberry Pi Cookbook 10 New Look for the Elektor Store Welcome tothe revamped Hektor Store, ‘and enjoy the shopping exerience! @ DesignSpark 78 That Zero-ohm Resistor The Unbearable Lightness of Being zero ohms, 104 DesignSpark Tips & Tricks Day #7: 3D Modding and Medeling Howto create a 3D view of your project, and export it to DesignSpark Mechanical © Projects 12 ADAU1701 Universal Audio DSP Board Always wanted to start with DSPs, but afraid of the SMO? Here's the answer, 20. UItiProp Clock (2) ‘This month we conclude with the Intricate assembly ofthe propeller clock 26 Compact Tube Amplifier Hey this one works with ordinary power ‘transformers. 32 Pretty Accurate Digital Wall Clock No radio contol, yet an error of just 64 seconds per year, max! 38 RJ45 ‘Running- Cable Tester No microcontroller includee—no common ‘ground lead either /0 for FPGA Development Board Applying VHDL programming tothe ights! 4 | Januery & February 2014 | waww.elektor-magazine-com 48 54 60 70 75 Elektor FPGA dev syster 555 Class-D Audio Amplifier What do you mean Class-D aucio alvays requires complex ICs? Mixing Electronics and Mechanics with Flowcode 6 30 ie Key, The developers explain LED's Replace That Halogen Lamp A direct replacement for the standard (U4 halogen lamp, Model Train Dual Radio Control Racio takes over from infra-red Arduino on Course (6) Let's build a sounding balloon to capture ‘weather data, Active ESD Protection Some microcontrollers and other ICs can ‘do with extra protection against static discharges.Volume 40 - No. 445 & 446 January & February 2014 80 Developing with Embedded Linux “This manth we learn how te develop our ‘own applications using C+ and Eclipse 87 Simple Current Pulse Generator ‘Theory and practice or wide-range ‘current measurements on a budget. 90 Soldering LFCSP ICs by Hand Demenstrating how the ultra-tny ADSS13 DDS IC gets wires attached to its pins 92. USB Key Passport “Turn the key and have your password typed automatically 98 Nanoamps on the DMM Measure tiny currents using an ordinary digital multimeter. 100 Bird Drinking Water Heater Using @ 2N305S and a tomato paste tin © Labs 108 Arduino Yan Labs got their hands on a Yin and found ‘out ft relly bridges two works 113 What it Takes to Make an LED Blink Behold the embedded-age equivalent of a switch and a resistor, 114 Join the Fourth Industrial Revolution! [An overview of Internet of Things (IT) projects hot on our labs website 115 Where There's Smoke There's Fire ‘an atterot at bulking an electric cigarette. © Industry 116 News & New Products A selection of news items received from the electeonics industry, labs and organizations. © Regulars 120 Hexadoku ‘The Original lektorized Sudoku. 122 Retronics Eleitor AC Power supply (1984!) A useful tool if you are mineful about starting up low-voltage power supplies, Series Editor: Jan Buiting, 125 Gerard's Columns: Pro-Tronics ‘column of two from our columnist Gerard Forte, 130 Next Month in Elektor [Asneak preview of articles on the Elektor publication schedule www.elektor-magazine.com | Jnuary & February 2014 | 5eCommunity Volume 40, No. 445 a 446 January & February 2016 ISSN 1947-3753 USA /co ISSN 1757-0875 1% /Ku vvaelektercom feubh ens very Elektor international Media East Hartford, CT 06208, USA Phone? 3.860.289.0800 and Elektor international Media 78 York Street London WIM 1D, UK Phone! (+44) (0)20 7682 8348 Elektor International Media b.v. Phone: (+31) 46 4389444 Fax: (+51) 46 4370161 Escondlde, CA 92046, Phone: 800-269-6203 Evmall elektorBpespubling.com Internet: wwrw.elekter.com/members Please use London address E-mall service@elektericom Internet: wwr.clekter.com/member Email: [email protected] Johan Dif ° Phone: #32 6 15894245 small: .dik@elektor-com eraslektorcom/advertising pyight Novice py ah pte - © Elekcar International Media by. 2034 Pervean tre lsh Pelee he Mth 6 | January & February 2014 | wwwlektor 132 pages—and not on a diet used to think it was hard to prevent total dominance in Elektor magazine of articles chew- ing over projects, technologies or programming methods related to microcontroller and embedded technologies, and with our new “kitchen” website www.elektor-labs.com well up and running had reason to fear an overwhelming presence of bit crunchers and compilers. Yet, looking at the Proposals and In Progress columns on the labs website there is @ reassuring number of projects and project ideas that are sultable for microcontroller-free diets. Three areas in particular, audio, radio, and test & measurement, seem to attract designs produced without any PICs, AVRs, ARMs, or Cortex-Xs acting as black boxes. That does not mean the projects are totally free from Raspberry Pis or Arduinos though, 2s these and similar boards are now (rightly) considered “just another useful compo- nent” rather than an intricate microcontroller in a cloud of code and protocols. This extra thick edition of Elektor magazine intends to get you through the two first winter months of 2014 while attempting to give analog en non-microcontroller projects their due share of pages. To kick off, there's our DSP-without-the-math on page 12. There are two magnificent audio amplifiers, one 55S based (page 48) and fone with tubes (page 26). The amps have low cost, ease of assembly, and vintage ‘components in common—and by no means espire to reach the realms of High End ‘Audio, We could not resist though grilling them on our Audio Precision distortion analyzer. Besides the 555 another old faithful, the 2N3055, hails from the past— ‘secured to the bottom of a tomato paste tin, on page 100 it’s helping to serve @ drink of water to birds of the feathered variety during the winter months. The RJ45 Cable Tester on page 38 also excels in not having a microcontroller. Nanoamps current measurements (page 98) and current pulses (page 87) are covered In two separate articles for the T & M fens among you. The latter article is part of a series due for continuation in the March 2014 edition, Thumbs up for that AC power sup- ply from 1984 (page 122). Diets if applied should be moderate and balanced—we serve dishes with and without the opcode ingredient. Enjoy! Jan Buiting, Editor-in-Chief The Team Editor-in-Chief Jan Buiting Publisher / President: Carlo van Nistelrooy Shannon Barraclough (USA / Canada), Reoul Morreau (UK / ROW) Intemational 2dtoral Stat Harry Baggen, Jaime Gonzalez Arintero, Denis Meyer, Jens Nickel Thijs Beckers, Ton Giesberts, Wisse Hettinga, Luc Lemmens, Mart Schroijen, Clemens Valens, Jan Visser, Patrick Wielders Graphic Design & Prepress: Glel Dols Membership Managers: Laboratory Steff: Online Manager: Daniélle Mertens Managing Director: Don AkkermansPRBLEHEBELCHUCEE USA Care von Nisteny $1 ec 709-0600 Crannsteroy elt com United Kingdom Caro von Nistelrooy Ha 20 7682 8944 Cannsterop@eltr com Germany Fercnand te WoNart Ftewalvaert@elektr de France Denis Meyer oats 8035 Gimeyeraeleenr te Netherlands Fewahaart@elstor Spain ‘me Gonzalez Antero $318 1698 7465 Jgezareroagmal con Italy “392605007 mideecranairaare Brazil Jobo Motes Sr sama Jmatre@eborcom Portugal Seo Menins a1 45 0044 Jmanredeektec com India Son. Mattar ou sessi6s815 ‘S@eleeecin Russia 47 (G65) 3953336 bor [email protected] Turkey pS a8 26 [email protected] South Africa Joran Dik $31 61589 4245 J aeelktorcom china Cons Snay e521 6495 2911 Ceesbaayoonalicom @ektor EEE C circuit cellar Deters Connects You To (baal Supporting Companies verso Labcenter — = s 2 stalfonte otc a cn [Bom] "=" Designcon 2014 Mlerochi z Drm : DLP Design reo Ope, Pig en Eurocicuits a Pc E2PCB Front Panel Express TPC APEX Expo 2014 Not a supporting company yet? Contact Peter Wostre ([email protected], Phone 1978 281 7708, to reserve your avn space in Elektor Magazine, ElektorePOST or Elektorcom wwnelektor-magazine.com | January & Februsry 2014 | 7eCommunity Elektor World Compiled by Every day, every hour, every minute, at every Wisse Hettinga given moment designers and enthusiasts are thinking up, tweaking, reverse-engineering and developing new electronics. Chiefly for fun, but occasionally fun turns into serious business. Elektor World connects some of these events and activities — for fun and business. Bhs SS © GN LM vostermania These are interesting times for the publishing world. The new ‘online and content’ are giving the old, established media a run for their money, and who still believes you can make a decent profit from publishing something as old-fashioned as a book? Or take posters—in the old days Elektor often produced posters showing, for example, a bunch of connectors. You can still find them in classrooms. Is it just ancient his- tory? To our surprise the Raspberry Pi poster did rather well earlier this year. The design was downloaded over 35,000 times and as you are reading this an Arduino poster is being made available—and there's more coming. And everything is down- loadable. I know what you are thinking: “So, back to online anyway?" And yes, everything is downloadable—arguably the easiest way to deliver to your door. — The Elektor Studio ‘There are a few places in the Elektor offices that stand out. We have the permanent wire & meter mess of the Lab, the ‘dungeons’ (what secrets the walls could tell Us...) and we have our attic. For years the attic remained unused and unloved, gathering dust. We decided to bring out the brooms and convert the space into a real video & photo studio. From our attic studio we bring you our monthly webinars in cooperation with element, 14. Video will be implemented more and more in the new workings of Elektor—we are well past our first les= ssons and mistakes, and are now creating video manuals for an increasing number of products, ‘The number and form of the webinars will also be looked at—you can look forward to more even more Interesting sessions from us, Patrick Wielders will be our man behind the camere and will manage the studio. 8 | Januery & February 2014 | woww.elektor-magazine-com= 7 => World Class Tips For the past few months, Circult Cellar magazine has been posting a weekly electrical engineering tip at CircuitCellar.com, The response has been excellent. Engineers from around the globe have been reading the tips and even submitting tips of their own. Recent tips have included: ¥ Prototyping Tips for Engineers: Pro engineer and CC columnist Jeff Bachiochi presents three prototyping of options and discusses the pros and cons associated with each one, He writes: "There are two alternatives to having @ PCB house manufacture your PCBs: do-it-yourself (DIY) and routing, If you choose a DIY approach, . you'll have to work with ferric chloride (or another acid) to remove unwanted copper. You'll be able to produce “S If some PCBs quickly, but it will ikely be messy (and dangerous).” (hitp://bit.ly/1bB4a40) Mil [v.05 ve. RISC Processors: Engineer Bernard Debbash covers afew fundamental ciferences between DSP ané RISC processors. Some generic RISC processors lke the NXP LPC2138 don’t have a saturation function, he explains. “So it’s important to ensure that the input values or the size of the variable are scaled correctly to prevent overfiow, This problem can be avoided with a thorough simulation process.” (http://bit.ly/1aBFPKb) Visit the CircuitCellar.com website or follow the editorial team via Twitter (https://twitter.com/circuitcellar), Facebook (https:// wwwv.facebook.com/circuitcellar), and Google+, to get your weekly EE tips. Feel free to share you tips as well — wa From Tracing Paper to Computer: Elektor Schematics Layout IF there is one characteristic of Elektor that stands out, it must be the way schematics are drawn, From the very start In the early 1970s Elextor has had 2 very specific way of schematic drawing, and the force behind this is... Mart Schroljen, For 36 years he has taken care of the drawings; intially using a set of Rotring pens and calque tracing paper, now it’s all done with the help of com- puters and Petrick Wielders. Over the years Mart created thousends of draw- ings and also took many of the techy photographs appearing in the magazine. The biggest advantage of the Elektor schematic drawing style is instant recog- nition, Sometimes we receive project proposels covering our own publications, or we spot illegal downloads of our published articles (fun Friday afternoon times). In all cases the Elektor style is instantly recognizable. Teticious recipes a cea al 4 Se Raspberry Pi Cookbook Last year Elektor members enjoyed a series of RPi projects and arti- cles arriving Via their Elektor.POST subscription every other week. The first seven projects have been compiled into a small cookbook allowing everyone to start off right away using a Raspberry Pi micro= computer. The booklet will be available to anyone launching a project con wnw.elektorlabs.com in January. worw.elektor-magazine.com | January & February 2014 | 9(Hektor New Look for the Elektor Store Since the start of the year, the familiar address www.elektor.com takes you to a fully revamped Elektor Store. The first thing you notice on the website is the fresh new layout based on the “less is more” principle. Thanks to Responsive Web Design technology, the site looks good on all types of devices, including computer monitors, tablets and even smartphones. + However, what's even more important is that, the search function has been dramatically Improved—and we have to admit that it was not one of the major strengths of our previ- ous online presence. The main menu gives you access to the fol- lowing product categories: Books, Micro- controllers, PCBs, Kits & Modules, and even more, A new feature Is that searches now depend on criteria specific to the selected product category. For example, you can search for microcontrollers based on the product number, but you can also restrict, your search by selecting parameters such as operating voltage or package type. You can select a price range for searching In any product category, and if desired you can sort the results by price. + Each product has its own page where all the product data Is shown In an orderly manner. But that's not all: there you can also access pictures and user guides or manuals, and in many cases videos produced by the Elek- tor Labs end Editorial teams, The options for combined purchases and special offers are another attractive feature. That can save you a pretty penny! By the way, the range of payment options has been expanded and now includes MasterCard, Direct Debit (UK) and PayPal, to mention only the most important. However, we have dropped the 10 | January & February 2014 | www.elektor-magazine.com Elektor Credits option. Regular and special editions of Elektor magazine can now be purchased directly In the store without the bother of credit units. + The new Elektor Store at www.elektor.com no longer includes the well-known project pages, They have been moved to www.elek- tor-magazine.com. The project pages are the starting point for downloads for Elektor projects, such as software or PCB layouts in PDF format. You can access an individ- Lal project page quickly using the deep link www.elektor-magazine.com/xx000«, where “so000x" is the six-digit article number that is always shown at the end of each article, How= ‘ever, the old links with the format www.eleke toncom/xoocxx stil work; they are automati- cally redirected to the -magazine website. The forums are stil available to Elektor readers as usual. You can access the reader forums via the link http: //forum.elektor.com. * Finally, @ word about user accounts for the Elektor Store: If you have previously ordered something from the online store, you do not have to enter your data again. Now it’s time for us to stop talking and for you to see for yourself: simply surf to the new Elektor Store at www.elektor.com, where you are bound to find something attractive,Newsletter vewnrwscs LAE] res a LUNI-T Digital multimeter BD Sutey: car arto v Measures resistance omgoratur as well as IACIOC eure & vgs rae 49,95 + Preset emseviy- 0.95 1 Repeat accurey 05 -Cors03% aPCia2% rom € 34,95, + Upto400 1 chp: 26 mm Shag LE dep, 3909 counts + Sateys Ec 61010, aT He00v-cATI300¥ rae § 39,9. (Jreiche electronics Your competent online partner for DES ESS UNI-T- Digital memory oscilloscope with outstanding performance data and USB interface! + Scanning rate: 250 MS/s + Rew time: <'4 ne + Storage depth: 25 kpis + USB interface for expor. ofreacings PT) res UNF. UNIT: Digital memory oscilloscope + Sater CATH ‘cary 600 v + USBinetace, sofware includes 10 vessinge capsctarce, requency tenperstue ae well se GIDC eure and velape + Teenanne 050 wih 25 Miz analog bandwith + Seanang rate: 200 MB‘ + Digtsl mutter £00 courts rp r025 cL © 295,95 |e. 19) urn 135, ern Digital table multimeter + 66 it etal ply, 40,000 eau 1 pocuracy 0225 % 4 Measures resistance, eapactance, oques temperature ae well as ACIDG carers ane voiage + Diode ard sorry eck 1 Relive value measurement 4 Dataloggerturetion 1 Use gag? nlerace, stare + Satey: cATI 1000V, car PayPal tone eat cepa tants ap ds, reduc names nd ges ae prope ofeProjects ADAU1701 Universal Audio DSP Board drag ‘n drop DSP’ing — and no SMDs to solder by Ramkumar It seems that to even dip your feet in the ocean of DSP (digital signal processing) Ramaswamy (India) you need lots of SMD soldering skills and a ton of math. Seems. Let's end the anxiety: here is a Universal Audio DSP board for the DIYer, Based on the Analog Devices ADAU1701 DSP, this board has only through-hole components—except for the DSP itself, which Elektor PCB Service reflow-solder in place especially 4 U. ‘Analog Devices’ ADAU1701 DSP chip represents a wonderful way of starting out and experimenting with DSP (digital signal processing). Together with the assoclated (free!) development environment called SigmaStudio, the chip is targeted at those 12 | January & February 2014 | wiew.elektoremagazine.com who want an easy migration path from all-ana- lag audio signal processing designs and moving into the digital domain. SigmaStudio requires ho programming—Just drag and drop prebuilt blocks such as “State Variable Filter” and heyADAU1701 DSP Board presto you have a circuit in minutes which can either be implemented as a DSP platform or sim- ulated using this board and then implemented {as an analog platform! So there you have it—an ‘easy-torassemble DSP board featuring analog and digital 1/0, easily programmable with no special hardware, in an environment that allows you to go for a finished product that Is DSP-based or analog. All with no math! Out of the box, the ADAU1701 offers sampling rates of 44.1 kHz and 48 kHz, Higher sampling rates such as 96 kHz are possible but they require deeper configuration. ‘The board comes with the DSP chip (a 48-Legged SMD component!) preassembled. Even better, Elektor’s semi-kit #120232-71 contains that same board and all through-hole parts, so there's also the fun of building It yourself DSP board h/w ‘The schematic shown in Figure 1 is based on the standard “self-boot" constellation found in the ADAU1701 datasheet, with some additions. All parts except the ADAU1701 DSP chip (IC3) are standard through-hole (TH). Let’s do a quick tour, Unstabilized 5 to 12 volts DC enters the circuit via Figure 1 Schematic of the ADAUZ701 ‘Audio DSP Board, Caution: Traneietor T2 must not be not fitted, wwwcelektor-magazine.com | January & February 2014 | 13eProjects COMPONENT LIST Resistors (all xed Re 5%, 251) R1,R18 = 000 a15 = 3300. 4,R10,R11 212,813, 21,R22,R23,R24 RD = 1ko RB.RS,RI4 = 2.240. 225,R25 = 10k. 16,R19,R20 R17'= ATK G)R7,RB,RO = IMO. °1,P2,03,P4 = 10k. potentiometer, near law 5 = 4700 preset (trimmer) 4709 oy Capacitors (c1,c2 = 22pF Ga'= 3 anF (€24,26,28,c30 5 = SenF (€3,C8,9,C10,C13,c12,C14,C15,C21,C22,C32,C34 C36 = L00NF (€5,C7,C13,C16,C33,C35 = 20,F 26, radial, 2.5mm pie (€23,C31 = 47pF 25V, racial, 2.5mm pitch (€17,€18,15,C20 = 1oopF 18v, 2.5mm nF Figure 2. ‘Semiconductors :,02,03 = 1NSB17 1D4,05,06,07 = 82X79-C3V3 3.3V zener diode Ici = 24L0256-V/P,DILB case 3C2 = Mc76004, DILI4 case 1C3 = ADAUL7OL3577 (LOFP4B case) 1e4 = T529400z-3.3 Bcs27 ‘72 = bo nor Fir Miscellaneous K3K4,KS,K5,<7,KB = RCAJcinch socket, k= DC paver adaptor connector ‘Si = pushoutton w. tactile feedback Xa = 12.288HH2 quartz crystal 3P1 = 2-pin pinheader, 0.1" piteh ‘4-way pin header, 0.1" piteh S-way pin header, 0.1" pitch L4-nay pin header, 0.1” pitch ‘Teway pin header, 0.1" pith Dumper PCB, Eletor Store # 130232-1, comes with ADAU- 47019STZ presoldered Semi-kit, Eletor Store # 130232-72, inchdes PCS 130232-1 and all trough hole parts Creu board design for the L_J J ‘ADAUL701 Universal Audio DSP Board, Conveniently the board comes withthe DSP chip soldered in place. Do rot ft 2, 14 | January & February 2014 | www.elektoremagazine.comsocket K1. Schottky diode D1 provides protection against reverse polarity of the input voltage. All other circuitry on the board operates off a 3.3 V voltage supplied by regulator 1¢4. ‘The stereo analog input sound is applied to DSP ‘audio inputs ADCO and ADC1 via K3 and K4. The digitally processed audio from DACO, DAC1, DAC2 ‘and DAC3 is available on cinch connectors KS-K8. Pots PL, P2, P3 and P4 are analog inputs for the DSP, They deliver a voltage from 0 V to a max- imum level defined by preset PS. The pot wiper voltages are buffered by rail-to-rail opamps IC2.A through 1C2.D. Pushbutton St is used to reset the DSP via its RESET pin (5). Jumper 1P1 on the DSP’s CLATCH line must be installed to program the on-board AT24cP EEPROM (IC1). Note that the DSP will not boot with JPI fitted. After programming remove JP1 ‘and press S1 to allow the DSP to boot. K10 is for programming the EEPROM using special programmer. K12 allows using the Elek- ADAU1701 DSP Board tor FTDI-BOB module # 110553 to program the EEPROM using the Elektor application S. Studio to EEPROM Converter. This is rather slow though. KG Is for programming the EEPROM using cus- tom hardware like an Arduino board. Its pins accept 5-V swing and are protected by 3.3 V zener diodes D6 and D7. K11 finally gives access to all GPIO pins of the DSP. Four lines (2, 3, 8, 9) are also used by the potentiometers P1-P4, Note however that these four lines can only be used to access the outputs of the opamp 1C2, and not as inputs. DSP board assembly Elektor Labs have designed @ compact PCB for the circuit—the component layout is shown in Fig ure 2. This board Is supplied with the DSP chip pre-assembled, so the rest of the construction ‘should be easy. The introductory photograph and the one in Figure 3 show the prototype assem- bled by Elektor Labs. Note that transistor T2 must not be fitted. It was required during an Figure 3 View of the completed board ‘and enclosure bottom hal, www-elektor-magazine.com | January & February 2014 | 15eProjects Figure 4 Example of the ADAUI701 SMD chip fitted on an SMO= to-DIP adapter board—i this case aVot type PC3-06 (LOFP48), For convenience, ‘the board pinout anc -ADAUI701 DSP pin index is Pictured also. early phase of the project but causes maifunc tion in the final setup. ‘Those of you keen to have the ADAU1701 DSP chip only on a separate board for dropping into their own project (stomp box, audio processor, ham radio TRX) will be delighted to know that LQFP48-to-DIP adapter boards are available from various sources like Voti [1], Dipmicro [2], Proto Advantage [3], Schmartboard [4] and... EBay! ‘The Schmertboard is our favorite for solderabilty, kind staff and the pinneaders included. Figure 4 shows the Voti adapter board assembled by Labs. Our DSP board accepts both 0.6 inch and Figure 5. 0.7 inch wide adapter boards—either can go in Creating a project in PCB position MOD1, but only if the DSP chip is Sigmastudio, not fitted on the DSP main board, This provision 16 | January & February 2014 | www.elektoremagazine.com should suit those of you having persisted in etch- ing and drilling their own circuit board from the artwork provided [5]. DSP’ing the graphic way DsP sadly but falsely remains associated with mathematics which tends to put people off. Happlly, the Analog Devices’ SigmaStudio envi- ronment follows an almost entirely graphical approach to configuring that complex DSP chip. With SigmaStudio it’s drag and drop mostly Instead of mathematics. Let’s see how it’s done In a Tutorial, ‘The prerequisites are: + Windows 7 x86/x64, Viste, XP Professional or Home Edition with SP2; + at least one COM port or a USB-to-serial adapter; + snet framework 4 + 3.5; + SigmaStudio 3.9 (download at [7], free but account needed); + Elektor SigmaStudio to EEPROM Converter utllty [5], an Arduino Uno or Mega board, or 2 12°C EEPROM programmer supporting the AT24CP, Step 1, Create a SigmaStudie project Start SigmaStudio. Select File + New Project and wait for the project to be created. You will see elements in the toolbox on the left; drag and drop “ADAUL701","USBi" and “E2Prom” blocks to the main window named Hardware Configu= ration, Now link one of the blue dots of the blockUSBi to the green dot of the ADAU1703, link the other blue dot of USBi to the green dot of the E2Prom, The result should look like in Figure 5. ‘Save the project (¢.g. as “Tutorial”), in a folder of the same name. Now Click on the "Schematic" tab. Notice the change of toolbox at the left + Under 10 ~ Input, drag and drop an “Input” element. + Under [0 ~ Output, drag and drop an “Out- put” element twice + Under GPIO = Input, drag and drop the “auxiliary ADC Input” element. + Under Volume Controls ~ Adjustable Gain ~ Ext Control ~ Clickless SW Slew, drag and drop the “Single slew ext vol”, + Right click on the “SW vol 1” block, select Grow Algorithm — 1. Ext vol (SW slew) + 1 Block "SW vol 1” should now have 3 inputs and 2 outputs. + From Filters ~ Second order ~ Single pre- clsion 2 Ch drag and drop "Medium Size Eq." ‘Right click on the “Mid EQ 1” block, select Grow Algorithm — 1. 2 Channel ~ Single Precision + 3. Block "Mid EQ 1” should now have four sliders. + Connect the blocks as pictured in Figure 6. Now we need to configure GPIO9 as an auxiliary ADC input. + Return to the “Hardware Configuration” tab and then click “ICL -170x\140x Register Control” tab at the bottom of the window. + Select MP9 In the GPIO block and change “Input GPIO Debounce” to “ADCO, then return to “Config” tab as illustrated in Figure 7. Return to the Schematic and change the values of the slider frequencies in the equalizer and set the gain to your taste. The project is ready, save It. Press F7 or select from the menu Action -> Link Compile Download, You will see a Comms error because duh you do not have the offical program- mer. At some point during the creation of the ‘schematic you will see a message mentioning USB problems, You can safely ignore this message. Return to the Hardware Configuration Tab, right lick on IC1 and select “Write Latest Compilation ADAU1701 DSP Board Sees to E2PROM". This will create an executable file that we can program into the EEPROM, ‘Step 2, Programming the EEPROM ‘Three possibilities exist—each is discussed below, a) Programming over the serial port using the Elektor Sigma Studio Serial 12C EEPROM Pro- grammer utility (t's rather slow). See Figure 8, Figure 6. ‘These four sliders are the Virtual equivalents of the pots an the DSP board, Here we're creating an equalizer function. Look, no maths! Figure 7. Lets do some GPIO configuring in SigmaStudio Figure 8. Elektor' sigma Studio Serial PC EEPROM Programmer utility in action. «Eom wz —~ [Premio] wwwelektoremagazine.com | January & February 2014 | 17eProjects foie tiny ML et ete OD ‘One of the first things you will need to get hold of is the somewhat hard-to-find method to capture the EEPROM text flle generated by ‘SigmaStudio and convert that to something you cen burn. I have distilled the procedure from various posts on the SigmaDSP forums and itis presented here for ready reference, ‘Step 1, Open the project file (.dspprog) in SigmaStudio, and navigate to the schematic page. ‘Step 2. Click Hardware Configuration tab. Press F7 or menu item Action = Link Compile Download. (Af the Analog Devices Eval Board Is not connected, you will get @ ‘message saying “Communication Failure” that may be respectfully ignored.) ‘Step 3. Right click on the ADAU1701 box and choose “Write latest ‘compilation to E2PROM.” In the project folder (which contains the B 22 | January & February 2024 | wwww.elektoremagazine-com‘activated and should run at full speed; the volt- ‘age for the transformer is also present, Using an oscilloscope and temporary 1-10 k@ resistors connected between pins 1-2 and 2-3 on J4, you will be able to observe the 50 kHz square-wave drive signal with an amplitude of 9 V. Let’s now test the propeller independently ofits base. Power it via a 9 V battery on 11 (the polar- ity doesn’t matter), The two LEDs closest to the center ought to light at once, If you now pass Qt or Q2 in front of the infrared diode DSS on the base unit, a third LED lights indicating correct reception of the positioning signal. If you obtain. this result, your propeller is ready for use. Bring the central part of the propeller closer to the base, directing U7 towards the ring formed by D54-D62. The propeller circult confirms the reception of the data sequences by lighting in turn the fifty diodes on the two blades. You can’t find U7? Look under the propeller, Were these tests satisfactory? Well let's move on to the assembly then! Mechanical The biggest step in the mechanical construction ‘consists in mocifying an 80 mm fan, common in computers, keeping only the motor itself, and winding the transformer onto it. To avold hav- ing to go into the details of this operation here, I've prepared @ document [1] that J invite you to read: it explains everything step by step in detail with pictures, Follow these instructions to the letter and you will obtain without any trouble (but with some patience) the vital transformer. ‘The cross-sectional view of the top (Figure 12) ‘and the photo (Figure 13) give an idea of the assembly of this spinning team. The primary, ‘Wound on the outside, Is fixed to the base and so doesn’t move. The motor stator is also glued to the base. ‘The secondary, with a slightly smaller diameter, is positioned in the center of the primary and Figure 13, ‘lose ups ofthe transformer after winding (2), then ‘separated (b) and lastly assembled (¢). In (2) we can make out the strip of card which was used during winding to maintain 2 space between the primary and secondary windings, as well asin (c) the two green disk spacers that space the propeller away from the mator to Which itis ued, Ultiprop Clock Tie phic eee wae wwwcelektor-magazine.com | January & February 2014 | 23eProjects Figure 14 Using the Apple remote-control long glued around the fan motor, So it turns with the propeller, which is itself glued to the motor hub. To keep the turns in place while winding the transformer, you'll need to insert between the layers of copper wire a good (!) thin, double-sided adhesive tape, then add instant glue as you go along to keep the turns finally in place, ‘The two crescent-shaped pleces are the feet, to be fitted by soldering. Fitted perpendicular to the base unit PCB, they will support your clock if it Is placed on a flat surface. Don’t solder them til everything else is finished If you prefer, the hole at the top of the base unit will et you hang your clock on a small nail on the wall or any other appropriate support. You can thread the power cable through the hook- shaped notch mace for this purpose at the foot of the base unit. This cable mustn’t get tangled up with the propeller! We also draw your attention to the mechanical fragility of certain components, in particular at the ends of the blades (QU and 2); when the propeller turns at full speed, you ‘must avoid these components hitting an obstacle The clock’s not slow ‘There's ventilation in the air! If you have checked the operation of your two boards, in a few sec fonds and in a gentle waft of air you're going to discover the magic of your UltiProp Clock, All that Is missing is @ power supply between 10-14 V DC, capable of supplying at least 250 mA at cruising speed. When power Is applied, the variables and peripherals are initialized, then the motor and transformer start up. On the propeller side, D1 ‘and D26 light constantly to confirm the presence of the supply. A few moments later, LED D65 is lit in turn, to bring the display to life. The time is then shown in the top half, and a welcome message scrolls. Once started, the clock can be controlled by either the remote-control (Figure 14) [3] or the rotary encoder, Press this button to put the system into stand-by, or to wake it up, Turn it to the right or left to scroll at will trough several tens of display modes, with various combinations of colors and representation of the clock-face and the hands, Resistors SMD 0805, 5% RIMR5,R23,R30 = 47402 R6,R7,R16,R22, 26,827, R31,R32,R36,837 = ana RE-RI2,R24 = 22k0 8,c16,c20 = 10nF Sov a7’ anF Sov 38-41 = 33pF 10V 13,14 R15,R17,R19,R21,229 = 1002 R5,R34 = 1.542 R28 = 3300. R33,R35 = 10k0. Rae'= Mo SMD 1206, 5% Capacitors Ceramic, SMD 0805, 20% Ceramic, SMD 1205, 20% (63,C4,C7,C8,10,C17 = 22uF 10V hi,ct8,ci9,c43 = 1yF Sov C14,cis = 10pF 16V Miscellaneous (C31,C32 = 100UF Lev, Vishay 5930 (G33 = 47yF 25V, Vishay 5930 Gi (9682703) (61 €2,€5,06,C12,C13,20,C21,C24,C25,C27 (€22,023,C26,28,029,C35 = 1yF 10V [IF 5,5V, Panasonic EECFSRSUI04 ductors LLL = 68H 0.844, Bourns SDROGO4-SBOKL (428011) L2 = 220,H 0,38 A, Bourns SOROGO4-221KL (1828016) LA = im 0,120, Bourns SDROSO4-102KL (1828020) US = 334H 3A, warth 744771133 (2082608) Semiconductors D1-D50 = LED, bicolor, PLECA, choose your color, e.9. Vishay VLMKE3400-GS08 (1328370) ')51,052,033,063,056,068,06 MBRS140T3G, ON Semiconductor (9557237) 1p54-D82 = 5F+421-7, infrared LED, Osram (1228348) 24 | January & February 2014 | www elektoremagazine-comwith or without the seconds displayed, as well as the date and ambient temperature, and the choice between analog or digital display. Once you have made your choice, don’t touch anything for ten seconds and this mode will be saved into non-volatile memory so it can be restored next time the unit is turned on. Press the encoder for more than two seconds to enter the con- figuration menu and set the time, date, choice of language, rotational speed in day and night modes, and the brightness, etc, You can navi- gate through this menu by short presses to select the required parameter and then by rotating to adjust the value. Choosing zero speed in night mode allows the Unit to go into stand-by and wake up again auto matically according to the ambient light in the room, ideal if the project is fitted in a bedroom. ‘This option is handy, but it can catch you out. If by chance your clock refuses to start up, this may be because there’s not enough light and it thinks you are sleeping. So before you start looking for 2 fault, do make sure the surroundings are bright ‘enough, Next time you adjust the parameters, remember to raise the stand-by threshold a little This 1s where our electronic clock adventure together ends. Over to you now to get stuck in. T hope this project will give you as much plea- sure as I have derived from developing it, and that lots of you will come and share your expe~ riences and enthusiasm in the Microcontrollers & Embedded area on the Elektor forum. 730386) 64 = BATS4 12 = 0532315, Maxim (1593292) TRI = transformer with two concentric wind- 55 = VSLYSBSO, infrared LED, Vishay ULB = TS0P6238, 38kHz infrared receiver, Ve _ings [1] (1870807) shay (4913220) Enameled wire, 0.5mm diam. (23 AWG), ap~ 57 = LED, orange, 1070, D71 = SMB)48A (1899472) 1, 2 = TEMT1020, phototransistor,Vishey Miscellaneous {see construetion manual [1]) 470165) YLY2 = 20Miz quartz crystal, HC PCB nes, 120732-1 through -7 3, 24, Q7 = Bc847B 78-20.000MAA)-T (1841988) Q5, Q6, QB = IRIMLO060, transistor, Interna 3 ~ PEC21-4215F-S0024 rotary encoder, ‘lonal Rectner (1783927) Bourns (1653380) La Ut = MAXSSS7AAX, Maxim [Digikey # eI = VT935G, photoresstor, Excelias Tech MAX6SS7AAX4T-ND] (1652538) US, UB » LN22674M:-5.0 (1679666) 22, 13 = 6-pn pinheader (2x3), 0.2” piteh U5, U0 = Aimega326P-AU (1715486) 14'= 5-pin pinhender, 0.1” pitch U7'= SF112400FA-Z, photeciode, Osram 15 = minijack socket, 2.1mm, Clif OCIOAS (1226482) (1888308) ug = L2570S0-AD) (1286848) 436'= ventilator, 8omm, 2000-3000 rpm cee UL4, USL = TC4427E0A, Microchip (9762647) prox. 25m (75 f.) Ultiprop Clock Internet Links [1] 120732 - UtiProp Clock Assembly. pct [2] Downloadable software, including source code: www elektoremagazine.com/120732 [3] Remote control for Apple Universal Dock http://store.apple.com/en/product/MC746ZA/A/ apple-universal-dock Instant glue and double sided adhesive tape Neb ry aretha Farallon weet wwwcelektor-magazine.com | January & February 2014 | 25eProjects by Michiel Ter Burg (Netherlands) Compact Tube Amplifier Using ordinary power transformers To obtain @ wide hi-fi bandwidth and an output power of 10 watts or more, you usually need big transformers for the power supply and driving the loudspeakers. However, in practice it doesn’t take a lot of power to achieve a good sound level In an average living room with a pair of reason- ably efficient loudspeakers. With regard to the frequency range, most music has virtually noth Ing the low bass range (below 50 Hz) or the high treble range (above 12 kHz), $0 you can achieve savings there as well This means that if you are willing to relax your requirements a bit, itis possible to build a small tube amplifier with a few watts of output power per channel using ordinary PCB-mount power transformers as the output transformers. The other components are also readily available, for 26 | January & February 2014 | wwww.elektoremagazine-com Many audiophiles have occasionally thought about building a tube amplifier, but they are deterred by the high cost, large enclosure, hefty transformers and complicated wiring harnesses. For all the people in this group, the author has developed a compact, low-cost option using readily available compo- nents, so that everyone can enjoy the warm sound of a tube am- plifier in their living room or den example from mail order companies, and not expensive, The tube used here Is a PCL86 (near est US equivalent: 14GW8), a reasonably modern audio tube (dating from 1961) that combines 2 preamp triode rated at 0.5 W with a power pen- tode rated at 9 W. The design uses two of these tubes, with the triodes wired as a phase splitter and the pentodes operating in a conventional push-pull configuration. ‘The lower limit of the bandwidth Is really deter- mined by the primary inductances of the trans- formers, while the upper limit is determined by the coupling factor of the output transformer. ‘The coupling factor K is very low with this sort of transformer because the primary and second- ary windings are in separate sections to meet an Insulation voltage spec of 5 kV.‘The input transformer used here (originally intended for a light organ) results in a fairly low input imped lance of approximately 1 k®, which Is necessary to obtain a reasonable open-loop bandwidth. This should not cause a problem ifthe input signal comes from a good preamp or the headghone output of a ‘computer, MP3 player or similar source, Schematic Figure 1 shows the schematic diagram of the ‘amplifier. The audio input signal on connector K1 is fed to input transformer TR2. The secondary of this transformer drives the two triode sections of the tubes in opposite phase via resistors R8 and R12. The anodes of the pentode sections of tubes VI and V2 are connected to the primary winding of transformer TR1 (an encapsulated PCB-mount ower transformer), which is recast as an output transformer here. The output signals from the tri- ‘odes are fed to the control grids of the pentodes vvia networks C3/R3 and C5/R16. The two second- ary windings of TR1 drive the loudspeaker. You ‘can fit jumpers on JPL to connect these windings In series or in parallel, JP2 and R26 provide the negative feedback path from the output to the Compact Tube Amplifier input. You can experiment with different values for R26, or you can disable negative feedback entirely by removing the jumper on JP2. A jumper can be fitted on JP3 to connect the input ground to the loudspeaker output. ‘Trimpot P2 sets the quiescent current of the tubes. It adjusts the negative bias on the con- trol grids of the pentodes, which is derived from the negative filament voltage V,. Pi adjusts the balance between the two tubes. Header K3 allows the quiescent currents of the ‘two tubes to be measured individually using a voltmeter. The scale ratio Is 10 volts per ampére, so with a current of 25 mA (the recommended value) the reading is 250 mV. You should let the tubes warm up properly before setting the qui- escent current, All the filaments are connected in series, and the filament current should be 300 mA. To prevent excessive filament current at switch-on (inrush current) due to the low cold resistance of the filaments, a simple current source built around an LM337 is included on the circuit board. A sim- ple negative voltage supply with an output of at least 31 Vis all you need for the filament current. Figure 1. This simple tube amplifier is built around a pair of PCL86 (1468) tubes. wwwcelektor-magazine.com | January & February 2014 | 27eProjects Figure 2. Tre PCB layout fora single amplifier charnal; two of these must be assembled for a stereo amplifier ‘The allowable range of the anode supply voltage Vp is 160 Voc to 200 Voe, A suitable power supply for the amplifier is described below. PCB design Elektor Labs designed a double-sided PCB layout for the tube amplifier, The PCB design is posted at 1]. The board has a large ground plane on one side to shield the mostly high-impedance com- ponents on the other side, with extra-wide iso- lation on account of the high anode voltage. The input transformer is located as far away from the output transformer as possible to minimize the magnetic coupling between the two transformers Component List Amplifier Resistors (@%, SW, 350V): RI RS R6RE,RG.RII,RI2.RI5.R1G,R19;ROS R10 = 2200. R20,R21 = 47k0. R22 2.7K PL = 108 crimp, 0.15, horizontal mounting P2 = 2ka trimpot,0.15W, herzontal mounting Capacitors C1,C3,C5 = LOOnF 250V, 5%, MKP, 5, 7.5, 10, 15 0° 225mm pitch 2,68 = 100nF 100V, 10%, MKT, 5 or 7.5mm ptch U7nF 250V, 10%, MR, 5, 7.5 af 10 mm 10UF 100V, radial, 2.5mm piteh, 6.3mm diam. (C7 = 10pF 250V, radial, Sm plch, 10 mm dlam, Semiconductors Di = 1Ne007 ica = uaa Tubes Viwv2 = PCL86 or 14cwe Miscellaneous ‘TR1 = power transformer, Block type FL 34/5, 2x15 primaries; 2x6V secondaries; 14 VA TR2 = 1:5 aucio transformer eg, [TEI19/KD-0703 (Conrad Electronics # 515701-89) 1,K5,1P2.°3 = 2-pinpinheader, 0.1" pitch Kaka = 2-way PCB screw terminal bleck, Sm pitch 3 = Z-way PCB screw terminal block, 7.5mm pitch JPL = 4-pin pinheader, 0.1" piten 3 or 4 jumpers for J>1, 992, 393 Vi,v2 = tube socket, ceramic, pin Noval, PC Heatsink for IC1, 20K/W (e.. Fischer Elektronik SK 125432) PB # 130385-1, see (1) Power Supply Resistors RL = 27042, 5, 350V 2 = 10k2, {25 250V Capacitors C1-C4 = 4,7HF 400V (€2-€8 = 47nF 100V ceramic 9 = 100,F 350. C1o,cit = 1000pF Sov Semiconductors D1-D4 = 2N4007 D5-Ds = 1NS@29 18,010 = LED, low-current cellaneous Fi = fuse, 200mAVT (115VAC: 400mA/T) with holder F2 = fuse, 750mAVT with holder Ki)K2 = Zoway PCB screw terminal Bock, 7.5mm pitch KS = 2-way PCB serew terminal block, Simm pitch ‘TRI = power transformer, Block type FL 30/12, 2x115V primaries; 2*12V second- ares, 30VA ‘TR2 = power transformer, Block type FL 18/12, 2e115V primaries; 2x12V second aries, 18 VA. 28 | January & February 2024 | www.elektoremagazine-comTo keep the PCB as compact as possible, several tracks are routed underneath the input trans- former. As a consequence, this amplifier must never be powered directly from the rectified AC Tine voltage ‘The photos clearly show how everything should be mounted. All of the components are normal leaded types. PCB screw terminal blocks are used {or the supply and loudspeaker connections. Due to the high anode voltage, a terminal block with 2 lead spacing of 7.5 mm Is used for K3. A heat sink is fitted on the LM337 regulator. The size depends on the level of the supply voltage; the temperature rise should preferably be kept below 40 degrees. After assembling the circuit board, you should Immediately fit several jumpers before connect ing the supply voltages. First fit a jumper on JP2 (negative feedback) and JP3 (ground connection between input and output). You can use pinheader JP3 to select either series or parallel connection of the two secondary windings. Fit two jumpers to connect them in parallel, or fit one jumper In middle of 3P1 for to connect them In series, Power supply ‘There are various ways to Implement the power supply for the tube amplifier. For the prototype, the author built an external switch-mode supply with a transformer for galvanic isolation from the AC line, The guys at Elektor Labs opted for ‘another approach with 2 power supply using two PCB-mount power transformers with their sec ‘ondary windings connected back to back. This results in good galvanic isolation from the AC line without the need for a transformer that may be hard to obtain, It also allows the filament volt- ‘age to be derived from the secondary voltage of the first transformer by means of bridge rectifier D5-DB, filter capacitors C10/C11 and associated ‘components. This is why the first transformer is ‘2 30-watt type and the other, an 18-watt type. If you are on a 115 V grid fit jumpers JP1 and JP3, and a 400 mA slow-blow fuse for FL, The high voltage at the output of TR2 will not be the targeted 230 volts, but instead lower because the data sheet says that the no-load voltage of TR2 is 1.22 times higher than the nominal voltage. The actual voltage is therefore 188.5 V (230/1.22). This will drop by an additional factor of 1.22 under load, so we end up with a plate voltage In the vicinity of 154 V, As PCB transformers of this sort are not designed to be used backwards, Compact Tube Amplifier large losses will occur when the filter capacl~ tors are being charged by short current spikes. In our test setup, the measured voltage with no input signal was slightly more than 188 Voc, dropping by ten volts or so at maximum output power. Pay attention to the voltage ratings of the various capacitors in the amplifier circuit. With 2 higher supply voltage, it is advisable to use 350-V types instead, ehgurston ter te For both supply voltages there is an LED indica- Srraiien Here to power tor that shows whether the voltage is present. transformers are connected ‘The LED for the high voltage supply also forms 2 “back to back” Figure 3. ‘suitable power supply swwwelektor-magazine.com | January & February 2014 | 29eProjects Ee gtd Plot A shows the total harmonic distortion plus noise versus frequency at an output power of 1 watt into 8 2 at 1 kHz. The two secondary windings are connected in parallel, and feedback jumpers JP2 and JP3 are fitted. This clearly shows the effect of the inferior characteristics of a power transformer in comparison to a real output transformer. Plot B shows the frequency characteristic of the amplifier at 1 watt into 8 ohms. The two secondary windings are connected in parallel. The curve with the larger bandwidth was measured with negative feedback (input level 850 mV); the other curve was measured without negative feedback (input level 235 mV) For comparison, the two curves were normalized at 1 kHz Plot € shows the total harmonic distortion plus noise versus output power with an 8-9 load, The two secondary windings are connected in parallel. The green curve is with negative feedback, and the blue curve is without. Plot D shows the Fourler spectrum of a 1 kiiz signal with 1 W into 8 @ and the secondary windings connected in parallel. The total harmonic distortion plus noise 1s 0.4%, Along with the two harmonics of the 1 kHz signal, a broad noise spectrum can be seen. It is caused by the ripple on the supply voltage. (ewtescmeconwinneeie tena; 1.3 n= hot) Inge impedance wie Lot ka cnet ere a eee baw ae appro 1 mex 12 4 98 0. emeretart ert] aera aghe iWeb 1/8 oleomacin wth negara etc; > 64.48 7208) THD nose at 1 W/8.2 Gate epacien wth eave edit ote aa * Damping factor at 1 W/ 89. uate enmecien th eave ede 25 (tes recon th ng ein: as 33 ma = 3 Ve=81Y) 30 | January & February 2024 | www.eleKtoremagazine-comload for the supply, to prevent the voltage from rising too high right after switch-on when the tubes are not yet drawing any current. Ifyou wish, you can experiment with a somewhat higher supply voltage and ditferent settings for the tubes, The author managed to achieve an ‘output power of about 10 watts (at 10% THD) with the prototype, using a beefier power sup- ply with an anode voltage of 200 V and a higher quiescent current Putting it all together After you have assembled the amplifier board and the power supply, you can adjust the quiescent current and the balance after letting the tubes warm up for a while, As already mentioned, you ‘can start with 2 quiescent current of 25 mA per tube (250 mV on KS) and adjust the balance so that the currents through the two tubes are exactly the same. Compact Tube Amplifier ‘Then you're ready to start listening to the ampli- fier, The author obtained the best results with the two output windings connected in parallel and a 4.7-k9 resistor in the feedback path, as Indicated on the schematic. With various speakers (either box enclosure or panel type, and regardless of the impedance) it sounds like what you would expect from a tube amplifier: good detailing at low volume and smooth overdriving at high signal levels. Although the measured bandwidth at full power may be somewhat disappointing, in practical listening situations you don’t miss anything because the bandwidth is fairly large at lower output levels, extending to above 10 kHz, 303854 Internet Link [2] www .elektorsmagazine.com/13: Elektor RF & Microwave Toolbox @ for Android “The RF& Microwave Toolbox contains 55 ‘aleulation and conversion tools for RF, microwave and electronics in genera Whetheryou are an RF profession, rao-amateur, astronomer or hobbyist. thisapp puts some ofthe most impor tanttaoksright your fingertips. erally Highlights: + Amplifier cascade (NF, Gain, P15, 1P2, 13) + Feld intensity and power density converter (Wjm2, Vim. Ai, Tesla, Gauss, W) + Pw Trace calculator (impedance/dimensions) + PlandT attenuator + tenn temperature (Kelvin) + EMC (EIRP. ERP, dBVimn) + Fter Design (Butterworth, Chebyshev, prototype) + And much more Penne www.elektor.com rf-app MaxBotix High Performance Ultrasonic Rangefinders foe ere Email: sales @maxbo! Outdoor Sensor Indoor Sensor O°) & & HRWRCLine HRWRLne HRLVLine LV Line WRCLine — WRLine & Ps} USB Line ‘Save 10% NOW Enter Code: elektor daeaiannucad XL Line Order today at: www.maxbotix.com/elektor ‘Common Applications: + Industrial ‘Applications + Liquid, Water, ‘Snow Level Measurement + Proximity & People Detection Store Displays, Booths & Kiosks ‘yewyeetektersonagazine.com | January & February 2014 | 31By Niras C.V. (nei) Pretty Accurate Digital Wall Clock -000202943 % slow or fast max. The crucial component in this project is a Maxim IC type DS3231, qualified by its manufacturer as an “extremely accurate I2C real time clock (RTC) with inte- grated temperature compensated crystal oscillator (TCXO) and crystal.” Maxim also says that the integration of the crystal resonator enhances the long-term accuracy of the device, guaranteeing a maximum error of less than 64 seconds over a year, and over a temperature range 0 to 40 °C (32 to 104 °F), The de- vice incorporates a battery input which maintains running of the device in the absence of external power. Circuit description Hyper-accurate as it may be, the DS3231 is unable to do anything useful without the help of a few other components. Referring to the sche= matic in Figure 1, A PICL6F876A microcontrot bate a oe PICIGF876A controlled Maxim DS3231SN RTC chip 1.5 inch red 7-segment LED displays Max. error 64 seconds per year Optional IR remote control 9V @ 500 mA max. power supply (DC adapter) Free C source code Free DesignSpark PCB files 32 | January & February 2024 | wwww.elektoremagazine-com ler ([C4) Is used to interface to the RTC (IC3) as well as drive a bunch of 7-segment LED displays (discussed further on). The microcontroller has an °C port which makes for easy Interfacing to the D$3231 RTC, ‘The microcontroller clock operates at 20 MHz due to quartz crystal X1 and its two load capacitors C1 and C2, not forgetting the appropriate con- twol word in the microcontroller “config settings”, Port lines RAG=RA3, RAS and RCI of the PIC micro switch Individual LED displays on and off through drivers/level changers T10-T21, Note the use of an npn/pnp transistor pair on each line to handle (1) the level conversion from 5-V swing (PIC side) to 9-V swing (display side), and (2) feeding current from the 9-V rail to each dis~ play via Its CA (common anode) terminal under multiplex controlWall Clock |. a ona © ca The supply voltage for the PIC micro is regu: aprototype of the clock, a current consumption of — igure lated with a 5-V linear regulator (IC1), and the 320 mA was measured on account of displays are connected to 2 higher (9-V) unreg- section operating in multiplex mode. ulated supply. This mekes the design capable The design is for a simple 24/12 hour clock which of driving bigger displays with a larger voltage _ incorporates six 7-segment LED displays and two drop per segment—such as 6.8 V—due to more input switches, The display section schematic is, LEDs in series, PORTB pins RB1-RB7, and RCO, shown in Figure 2, One of the attractions of this activate the individual LED segments through an design is it 1.5" bright red 7-segment LED dis- array of switching transistors T1-T8. plays, so the clock can be read easily from con- A9-V, 500-mA unregulated power supply (power siderable distances. The displays are multiplexed adapter) is sufficient for the circuit to provide at 1 kHz but provide sufficient everage currents good brightness, and a CR2030 lithium battery for high brightness. The individual segments (a-g) (BT1) is used as a backup supply for the RTC. On of displays LD1-LDS and the decimal point (dp) 3 display Schematic of the ‘microcontroller section. KZ connects to the common anodes ofthe 7-segm displays, K2to the bussed segments www-elektor-magazine.com | January & February 2014 | 33see) Figure 2 Schematic of the display section, \D3 and LDS are mounted upside down with adapted control in software, to create colons between hours and minutes, and minutes and seconde, (cb | ict BEET] wore of the displays LD?-LDS are connected and brought out to pins on Ké. Displays LD3 and LDS are mounted upside down on the board (note the dots!) to create an oblique colon (:) between hours and minutes, and minutes and seconds. The PCB design takes care of this, The LED dis~ plays are multiplexed through their common-an- ‘ode (CA) terminals, brought out to pins on KS connecting to K3 on the microcontroller board. Pushbuttons S1 and S2 for controlling the clock functions are connected to the microcontroller also by way of K4 on the display board connect= Ing straight to K2 on the microcontroller board, ‘The display brightness can be adjusted as a user setting stored in the PIC micro. This function may be implemented using the flag ‘DisplayOn’ in the code, Finally, on the display board, a bicolor LED (D2) Is used to Indicate AM (green) or PM (red if the clock operates in 12 hour mode. 34 | January & February 2014 | www elektoremagazine-com Infrared remote control: software ramifications The TSOP31238 in position 1C2 on the dis play board Is an infra-red receiver IC responding to 38 kHz signals from an RCS (or compatible) IR remote control. LED D1 and its driver circult form an. infrared transmitter. An Infrared ‘send! function, was originally not implemented, hence if you do not need it, D1 and associated parts may be omitted. As the Wall Clock project progressed, changes were made to the source code, as follows: + Xtal changed to 20 MHz enabling the PIC to execute RCS protocol section without any hiccups in the original code. ‘+ RCS protocol section Implemented In theinterrupt routine, microcontroller now scans port pin RC7 status; checks received signal against RCS protocol. + In the main section: addresses and com- mands get extracted from the received RCS data. + If the command received is 16 or 17 (address and all other commands except 16 & 17 ignored) the program takes it as but- tons $1 or S2 being pushed, respectively. + Implemented Watchdog functionality in-or- der to avoid any uncertainty in the program With IR reception implemented you can control the clock from a distance wielding your (RCS compatible) remote, Don’t tell the kids! Real Time Clock (RTC) type DS3231 ‘The DS3231 Is 2 serial RTC driven by a tem- perature compensated 32-kHz crystal oscillator (TCXO), and provides @ stable and accurate ref erence clock, The temperature sensor, oscillator, ‘and control logic form the TCXO. The control logic reads the output of the on-chip temperature sen- sor and uses a lookup table to deter mine the capaci- tance required, adds the aging correction in the AGE reg- ister and then sets the capaci- tance selection registers. The AGE function is not Used in this project—use of the aging register Is. not needed to achieve the given accuracy, With the clock source, the RTC provides seconds, min- utes, hours, day, date, month, and year infor- mation, all accessible through the I2C bus. The device monitors its VCC level to detect power failures and to automatically switch to the backup supply when necessary. On first power up the PIC microcontroller initial~ izes the RTC to generate a 1-Hz square wave at the INT/SQW pin by writing 0x00 to the RTC's control register. This is connected to the external interrupt (INT) of the microcontroller, effectively setting the PIC’s INTF flag on each High-to-Low transition at RBO/INT. This is used to initiate a reading of the RTC's time registers, A 1-Hz (ie. 1 -second) flashing colon can also be derived from. polling the status of RBO/INT. Construction In terms of hardware the clock is divided into two sub-circuits: microcontroller/driver and display. Each is built on its own circuit board of which. the component overlays are shown in Figure 3, Excepting the RTC, the entire design Is imple- mented in old skool through-hole (TH) parts so should be easy to build if you apply care and pre- cision in reading and soldering. Here are just two mistakes Elektor tech staff in their 35,_, years of publishing on electron ics have heard about 28 people make: (2) mixing up npn (like BC548) and pnp (like BC327) nsistors ("man, they look Identical don’t they"); (2) wrong values for the 22 pF caps (man, T used the nearest equivalent”) k2 on the microcontroller board is linked to K4 on. the display board through 2 14-way SIL pinheader and receptacle, The same for K3 on the micro- controller board and KS on the display board ‘These connections allow the display board to be stacked on top of the microcontroller board. Operation ‘The circuit has only two pushbuttons to perform user control and adjustments, Press SI for 1 sec- ond to take the circult into time adjusting mode, Blinking digits means they're open to having the Wall Clock www-elektor-magazine.com | January & February 2014 | 35a Pate ig as a COMPONENT LIST . Combined for microcontroller board, display board Resistors (5%, 0.25W) Ri = 820 RO-R7)RIS = 272. RE-R19,R16,R22,R25,R28,R31,R34,R37 5802 RI7,R18,R28,R39,340,R4: 20)R23,R25,R29,232,R3: R21,R24,R27,R30,R33,R3) Ra8'= 470 Ra-RAS = 1k. 10Ks. 100K 1509 Capacitors C,C2 = 229F ceramic ©3-c7,c10,c11 = 1000F €8,c9 = 470pF 16V redial Semiconductors LED, red, Sram 2 = LED, sleolr, e-9. Kingbright Leaweew. ich = L7BO5CT 1C2 = 15031238 (Vishay Semiconductor) {C3 = DSB231SN (Maxim Integrated Products) Ic4 = PICLSFS76A-1/52, programmed, Elek tor Store 9 110167-41 LDI-LD6 = SA15-1152WA 7-segment le- play, CA ¢Kingbright) TEs = ecsae nd receptacle), Dats] uanaL Figure 3, Separate printed circuit boards were designed for the microcontroller display sections, both using DesignSpark PCB. Boards shown at XX per cent of actual size. The display board ‘accommodates Kingbright 15-inch (38 mm) SALS-11SRWA 7-segment displays. K2/K4, K3KS = 14 pin SIL connector par (pinheader 716-721 = 8C327 51/52 = pushbutton, PCB mount, tactile feedback, €.9. Alps SKHHALADIO. Mise Xi = 20Mi2 quartz crystal BBI1 = 3V Lithium button cell with holder Poe # 110167-1 (mierocontreller board). KI = Z-pin PCB screw terminal black, 0.2” pitch. displayed value changed by you, the user. Digits can be selected individually by a short press of 'S1—from seconds, through minutes, hours, 24/12 hour selection, to exit. Pressing $2 Increments the selected digit to its highest value then rolls over, except for the ‘seconds’ digits, these will be changec to zero. Also, if the seconds exceed 30, the minutes’ digits will increment by one. Press ing S2 (‘up’) with the clock not in time adjust ment mode shows the temperature in °C with @ minus sign for really cold rooms, The tempera~ ture sensor has an “accuracy” of +3 °C and a resolution of 0.1 °C, Conclusion The circuit provided here is a basic wall clock with sizeable (1.5") displays. Features such as alarm or synchronization between PC via infrared link= Ing are optionally possible. The schematic, PCB 36 | January & February 2014 | wwww.elektoremagazine-com Poe # 110167-2 (dsplay board) design and PCB Gerber files produced by Elek- tor Labs India Dept. using DesignSpark PCB are available for free downloading at [1]. The same applies to the PIC source code written In © ‘The project is expressly pitched at everyone wish- ing to extend it with their own functionality by adding software of their awn creation, lke display brightness adjustment mentioned above. Lett designer and the community at wwrw.elektor-labs. com know how you are getting along. alektor PCB Service Professional Quality Trusted Service Secure Ordering 4 Targeted pooling services and 1 non-pooling service Free online PCB data verification service Online price calculator available No minimum order value No film charges or start-up chargeseProjects By Flemming Jensen (@enmark) RJ45 ‘Running-Lights’ Cable Tester Remarkably, it has no microcontroller (hooray!) and no common ground lead (huh?). This cheap & cheerful circuit is indispensable if you are suspicious about 2 defective RI45 cable In the patch cabinet or anywhere else for that matter. Connect the Master (sender) to one end of the cable under test (C.U‘T.) and the Sleve (receiver) to the other end. If the LEDs light up In succession then the cable is okay. If not, flaunt the cable to the IT Manager, drop it in His trash bin (with a fanfare) and ask a raise as a reward for keeping the company in business. Let's look at this clever money making gadget then, Remarkably, it has no microcontroller (hoo- ray!) and no commen ground lead (huh?). Each of the 4017's counter outputs CTO-CT7 has its own LED and wire in the cable under test but no wire is required for the ground return cur rent. When one of the 4017 counter’s outputs Js logic High due to the clock pulses from IC1 2 current flows through the corresponding LED ‘and normally retums to ground (GND) via diodes D1-D8 or the reverse connected LEDs, by way 38 | January & February 2014 | wwww.elektoremagazine-com of the Low outputs of the 4017. You either fit eight dual colour LEDs in positions D11-D18, or eight diodes in positions D1-D8 and eight high efficiency LEDs in positions D11-D18, No buffers are needed between the 4017 and the LED array at the other end of the C.U.T. consid= ering the IC is capable of sourcing the necessary current. Current limiting resistors R4-R11 are required though in view of the 9-volt supply volt- age. The speed of the running lights is adjustable on preset P1 within a certain range. ‘The LEDs tells a thing or two on the cable under test, as follows. + RU45 cable all right and straight through: LEDs light in succession like 2 small running lights. ‘+ One or several wires are shorted: one or several LEDs light all the time. ‘+ One or several wires broken: one or several LEDs off permanently, ‘= One or several wires connected to a wrong pin: running lights erratic (jumping around),Cable Tester ‘The Master and Slave circuits are built as sep- Slave board you fit either the dual LEDs or the arate units pictured here. The two boards are single color LEDs and diodes D1-D8. The photo- separated by sawing along the dashed line on graph shows the first option, the component overlay. Note again that on the aneso2) COMPONENT LIST corre Resistors Ri = Sk Ro = 33K Ra = 4.7k9 RAR = 10. PI = 100K preset, vertical Capacitors SUF 16V radial Semiconductors (Cl = C4011 or HEFAOT ; Di-pe = 1Na4a* K1,K2 = RI45 CATSE socket, PCB mount, e.g 1D9-D16 = LED, dualcolourlow-current, or ‘Farnell # 2060738 * either/or (cee text) single colour LED* weww-elektor-magazine.com | January & February 2014 | 39Projects By Andreas MokroB, Dominik Riepl, Christian Winkler and Professor Thomas Fuhrmann (Germany) Figure 2. Tre board showing the welcome screen, Programming in VHDL rs In the December 2013 edition we described the FPGA development board hard- ware. With this board it’s a simple job to interface the FPGA to real world devices and events, The extensive range of peripherals include sensors and an LC display, ‘The next step is to program the FPGA and get them talking to one another. ‘The Elektor FPGA board was originally featured in December 2012, It has the part number 120099 91 and can be ordered directly from the Elektor Store, This FPGA board alone does not contain any peripheral chips. To make it more useful ‘as an educational ald an expansion board was developed by the students at the Ostbayerischen ‘Technischen Hochschule in Regensburg, Germany. ‘The expansion board has a number of peripheral sensors and an LCD which interface to the FPGA 40 | January & February 2014 | www.elektoremagazine-com board. They can be easily configured and con- trolled using VHDL and put to use in all sorts of applications. The complete development environ- ‘ment with both boards is shown in Figure 1, In part 1 [1] of this project the development board hardware was described, in this second part we describe how these peripherals can be controlled using VHDL. This will give you a good grounding In the technology so that you can go on to use it for your own applications,‘The complete project has been developed using the XILINX ISE 14.5 Design Suite which is freely available to download from the Internet [2]. Ahierarchical approach For large projects in VHDL It Is sensible to ‘approach the design in the way you would a soft- were project by breaking down the program into small manageable parts. Unlike software VHDL doesn’t use functions or classes, instead we use descriptions of hardware, the so-called Module. Each module is described by its own file and has its own defined interface to the outside world. ‘A module functions as a self-contained unit and can be simulated with the help of @ Test-Bench. It can be seen as @ functional black box which Integrates into the complete project. The use of modules greatly simplifies the process of sys- tem debugging Based on the sensors, display, control elements ‘and their control, the VHDL project is divided into the following areas (see Figure 2) + menu_controt: The central module; this is where all the data comes together and the menu functions are taken care of. + taster: Interfaces to the pushbuttons with debounce logic. + Led: Drives the LCD and loads the display data, + gps_controt: Controls the GPS module and receives Global position information and time of day. + ADc_contrel: Controls the A/D converter. At the top level the individual modules are com- bined and no logic process is described. The con nections are made in the sub-modules defined in the top_Level description using the conponent key word. The compiler is thus informed which ‘components are used and what inputs and out- puts are available, Connections to the module are defined in the port map, where the input ‘and output of the sub modules are linked to the signals defined in the top level While VHDL is 2 Hardware Description Lan- guage, they get translated directly into gates ‘and because the modules work in parallel itis not necessary to consider the order of the modules. All modules have a connection to the 8 MHz clock so that every process Is synchronous with the FPGA Expansion Board | Pushbutton 9] puaten | | | ona Main cone I ! Leo Leo conta! *7] clock. Apart from this the module’s data output has an enable signal which goes to a logic ‘1’ state for one clock period when all the process- ing in the module is completed. It indicates that output date is stable and can be used elsewhere, Package: self_defined_types ‘This Package contains the global definitions for the complete project and frequently used data types and conversions to make them more readable, It defines the following data types: + BYTE: An 8-bit wide array of type std_togic. + BYTE_ARRAY: An arbitrary width array of type eye In addition the function HEXtexsciz is defined which converts 2 4-bit wide ste_togie_vector with a hexadecimal value into a displayable ASCIL character. ‘The packages are linked into this and any other library using the commang Library works use work.self_defined_types.all; The ADC Modul ‘The first module described here is used to con- trol the A/D converter and read the output data ‘The module uses a 3-bit long std_togic-input vector in_channel to select the analog channel and an 8-bit long sté_Logic-output vector out_ sdvet to output the data, The other inputs and Figure 2 ‘The projects block diagram, wwwcelektor-magazine.com | January & February 2014 | 41eProjects insar=t (wansmission starts) outputs are necessary to communicate with the IC (in sar, ino, out_clk, outdi, out_not_cs, out_not_se), Figure 3 State diagram of the ADC read processes. The Clock Process ‘According to the data sheet of the A/D converter, [3] it should be provided with a clock signal in the range from 10 kHz to 400 kHz. For this appli- cation we use 100 kHz, The clock Is generated a dedicated process, see Listing 1. The counter cnt_clock is incremented on each rising edge of the 8 MHz input clock in_ clk and the internal clock aee_ctk at (const_ éivider ~ 1) / 2 and (const_divider ~ 3) Listing 1 constant const_divider: integer := $0; signal ent_clock: integer range @ to const_divider - 1 signal rising clk signel falling clk: sté_togic; signal ade_clk: std_logic := 10"; std_logic; process(in_clk) begin if rising edge(in_clk) then if ent_clock = const_divider - 1 chen adc clk <= 11"; Fising.clk <= 11); cent_clock elsif ent_clock = (const_divider ~ 1)/2 then adc_elk < "0; FalLing_ctk ent_clock else rising.clk <= FalLing_clk ‘ent_clock end 36; ene if; end process; ron 10; = ent_elock + 15 42 | January & February 2014 | wwww.elektoremagazine-com which is inverted, so that the time that the timer runs from 0 to resat corresponds exactly to clock period of the generated clock. ‘The constant const_divider is the ratio of the clock frequency of the external oscillator and the required converter clock frequency: oscillator clack frequency const _divide converter clack frequency Apart from this the Enable signal rising_ctk and falling_clk are generated for state machine communication with the converter, The reading process: implementing a State Machine For the sake of clarity control of the FPGA func- tions are Implemented in modules using state machines. The following implementation will be for the A/D converter. The state transition dia~ gram of the machine is given in Figure 3, Each state is represented by a circle and transition to another state is indicated by an arrow labeled with the transition condition. ‘Two states are required to read data from the ADC: In wait_state the state machine waits until data is available indicated by the SARS out- ut from the ADC going to a logic ‘1’, The state machine then jumps to the send_state which reads the 8-bits from the convert ‘To describe the state diagram in Figure 3 using VHDL a separate data type state_type_read is defined with all the possible states and then 2 signal that Is this type. The compiler con~ verts this construct Into a timer, For 2 devel- ‘oper the approach used here is more easily understandable. Itis written as a process using a case statement, in which all the possible states of the machine are described, see Listing 2. ‘The two separate states of the machine can now be described: ‘+ wait_state: When the A/D converter out- puts a new word the SAR status output is set toa logic ‘1’, This makes the machine state change to reac_state, + read_state: This Is where the ADC serialoutput value is read bit by bit and starting with the MSB, written into a shift register. When 8-bits have been read the Value-En- able-Bit int_vat_en is set to logic ‘1’ and returns to the wait_state. After the change this will be reset to logic ‘0’ again ‘The ADC outputs 2 new bit on every falling clock ‘edge of the 100 kHz clock. Each bit is read on the rising edge of the clock with help from the Enable signal from the clock process ‘This implementation is used in the same way by all the other state machines so only the individ- ual states and transitions will be described and not the principle itself, The sending process A/D conversion is initiated by sending a telegram to the A/D converter. This Is performed using fa separate process. According to the ADC data sheet (data sheet [3], Figure 20) it reads con- trol signals from the FPGA on rising clock edges. ‘The A/D converter outputs a new bit at rising clock edges. To ensure that the data is stable the implemented state machine reads the value of these bits on the falling clock edge. ‘The state diagram showing the sending pro- cess structure is given in Figure 4, The process is implemented as a state machine with three state + wait_state: Waits for the Channel_Enable signal to start the conversion. When this signal is logic ‘1’ the chip select signal is set to logic ‘0" and the first bit of the telegram is sent, The state machine changes to the send_state state and decodes the remaining bits in the telegram, + send_state: In this state bits are sent one after another on the data line to the A/D. converter. The number of sent bits are counted until all the bits have been sent then the state changes to woit_for_rec_ready. + wait_for_rec_ready: In this state the walt for the conversion process of the analog voltage is finished, This is done with help of the int_vat_en signal. When reading out is finished communication with the A/D con- verter is ended by switching the chip select signal to logic ‘1’. The machine returns to the output state wait_state and the next request can be processed. FPGA Expansion Board (data to channel int vale (Goished reading converted data) Figure 4 ‘State diagram of the ADC send processes. Listing 2 type state_type_read is (wait_state, read_state); signal read_state_machine : state_type_read := wait_state; process(in_etk) begin if rising_edge(in_clk) then if rising_clk = "1! chen case read_state machine 4s when wait_state => read bit_count int_valen| if in_sar = "1" then read_state_nachine end if reaé_state; when read_state => if read_bit_count /= @ then ‘int_adval (7 downto 1) <= int_adval (6 downto 8); ‘int_edval (8) < read bit_count else indo; read bit_count + 2; read_state_machine wait states Gnt_valen <= 1"; end ifs end case; end ifs end if fend process; www-elektor-magazine.com | January & February 2014 | 43eProjects The pushbutton input module ‘The process which reacts to pushbutton activity {s contained in a separate VHDL module, Its main function Is to perform contact debouncing (see article on the board hardware [4]). To achieve debounce a counter is started when the signal level from the pushbutton input changes state, The counter is used to produce a delay so that the signal level is only valid once this counter has finished counting. Along press would result in the counter timing out several times and regis- tering several presses. To avoic this situation the active pushbutton is polled at every rising clock ledge to check if the press has already been reg- istered. The pushbutton input signal state is com- pared with its state stored when the counter last out_enabte else end ifs out_enable else end Sf} else enc if; Listing 3 generic (WALT_40MS: integer := 320190; WAIT_4_1NS: integer := 32900; WAIT_2_S2¥S: integer := 12200; WATT_19@US: integer := 900 WAIT_38US: integer := 490; WAIT_45@NS: integer := 10); Listing 4 when sené_éata => if wait_counter = WAIT_450NS - 1 then state < wait states wait_counter woiteounter <= when watt_state if wait_counter if prev_state state < init state < write data; waitcounter <= 95 waitcounter <= wait counter + 1; aa <8 waiteounter + 15 ‘e wait_tine = 1 then 10! then 4 | January & February 2014 | wwww.elektoremagazine-com elapsed. When the two states are the same (long press detected) the counter is reset otherwise it runs until the debounce delay time finishes Any contact bounce will be finished before the counter reaches the end of its counting period. ‘The signal level is now stored to temporary mem- ory and a short pulse is output. The LCD Module ‘The LCD has a parallel interface so all control and data information Is sent in the form of parallel words. After each word it is necessary to intro- duce a wait period to allow the LCD board con- troller to process the information, Itis therefore necessary to generate some wait periods. This is achieved with the genere command where con- stants valid in the module are placed, These are defined in the Entity declaration, as in Listing 3. ‘The wait period is defined by the integer value Which defines the maximum value of the counter clocked at 8 MHz. The LCD state machine A state machine with six states Is implemented to control the LCD. Figure 5 shows a simplified state diagram for the LCD. The state machine starts with start_up. Firstly there Is a 40 ms delay introduced to allow for the LCD to power up. Next is the init state to initialize the LCD. After initialization It automatically jumps to the wait_for_date state and stays here until in_ ata_en (an external input) is logic ‘1. This indi= cates that display data Is avallable to be written to the display. Next it jumps to the write_data state, where all the data Is written to the display. For every character written the LCD interface requires a '1' of at least 450 ns on its Enable input. This is taken care of after each character is sent out in the sand_cate state. In here the out_enable is set followed by a 450 ns wait. ‘The display requires 2 processing time of 38 us after each character Is sent to the display. This is generated by using a wait_state before the next character is sent. After each wait_state elapses it retums to write cata state until there are no more characters left to send to the display. Listing 4 shows the relevant section of VHDL code which handles sené_data and Its wait_stateops ‘The VHDL description of the GPS module is made Up of four individual modules, The top module for GPS control is gps_controt which contains three sub-modules eps_serva\_parallel, gps_check- sun and gps_parser GPS control (gps_control) In the top module gps_control the other sub-mod- Ules referenced above are declared as components. ‘and linked to the corresponding ports. The pro- ‘cess to turn the GPS module off and on is in this module. In addition a short process flashes an LED each time a valid GGA-type sentence is read. ‘The Conversion Process (gps_serial_paraltel) In the gps_sersal_paratiet module the serial UART protocol data from the GPS module Is con verted into a one byte wide parallel signal. Using ‘a previously calculated divider constant the com- munication speed with the GPS module (here we Use 4,800 bit/s) can be adapted as necessary. It Is Important that the sampling points of the received GPS data stream are synchronized to the data rate. To achieve this, the falling edge of the start bit at the beginning of every byte is detected, the GPS data gps_data input signal is shifted into the vector data_shift using the internal 8 MHz clock and compared with the bit sequence ‘1110’, Once the falling clock edge is detected the fol- lowing GPS data will be sampled one half of a bit width later i.e. mid-bit, and then shifted into the int_data shift register until a complete byte has been received. The valid data Is now stored in snt_data and written to the parallel data output our_date and the Enable-signal set. Checksum calculation (gps_checksun) ‘The gps_checksun module calculates the check- sum on al the transmitted data bits and compares it with the checksum value sent from the GPS module, This ensures that there are no errors in the receives sentence. When an error is detected the corrupted sentence is discarded Astate machine with fve states is used to read-in, calculate, validate and them output the result + reset: All of the signals used for these cal- culations are first reset to zero, When valid FPGA Expansion Board end count ata array, data val —> —— (| data from the serial/paraliel converter (out. Figures data_enable = 1) is available the state of State cagram ofthe zeichen.in changes, as soon as a$' symbol display Is detected in the data. This symbol is the GPS sentence start character. + zeichen_in: This detects where the check- sum begins In the received sentence and changes to the checksum_in_1 state, An ‘if condition is used to detect an asterisk which marks the end of the sentence data. The two characters following the asterisk are the two-byte sentence checksum * checksun_in_t: This reads in the first check- sum character, This is XOR'ed with the sum of the input characters, When it is not valid the signal int_checksum_err will be assigned logic 1" + checksun_in_2: The second checksum char- acter is read in here, Otherwise identical to checksun_in_t, output: When the checksum is valid then the signal int_checksum_ok Is given the value logic ‘1’ otherwise it has the value ‘0, Reading GPS data (z9s_parser) ‘The gps_parser module filters out the relevant information from the received GPS data stream and prepares it for further processing, The GPS. module sends al its data sentences sequentially weww-elektor-magazine.com | January & February 2014 | 45eProjects Figure 6 Tre menu options. according to the NMEA protocol. It is necessary to identify the sentence of interest (for our pur- poses the GGA sentence containing positional fix Information) and recover it from the data stream. A'S’ character identifies the start of every new data sentence. A state machine checks when this ‘occurs. Following this character is the 'GPGGA’ sequence which is the preamble to the GPS data of interest to us, There is a state for each data sentence of interest in which the data is read in Each data field in the sentence Is separated by a comma and this is used to change the state of the machine. One after another all the data is read in and sent to the corresponding output. itt, nh 46 | January & February 2014 | www-elektor-magazine.com Menu control in VHDL. ‘A menu has been implemented on the display to allow user control of the GPS module and A/D converter. Pushbuttons under the display allow intuitive interaction with displayed menu options. Figure 6 shows the menu layout. After tthe start screen there is an option to select sub menus ‘GPS! or ‘ADC’. ‘The GPS sub-menu firstly gives you the option to turn it off or on. Other pages give you the option to view additional information such as your current longitude and latitude. The com- mand ‘up’ returns you to the next level up in the menu structure. Selecting ‘ADC’ from the menu allows you to select a channel of the A/D converter, The mea sured values are displayed on a page in the sub menu. Pressing ‘ref’ (refresh) causes the ADC to make another measurement of the displayed channel and update the display with the new value, The GPS menu ‘The menu control s also built with a state machine and can be easily restructured (by changing the state transition diagram) or expanded (add new states). The menu structure and associated state machines are described using the GPS menu as an example. Each page in the menu has a corresponding state of the state machine for control of the menu, The current state Is stored in the signal state, The machine starts in the state_init state and then changes to the state_welcone state. Now the welcome screen is shown. A press of the pushbutton on the right changes to the state_gos state. This state builds the highest level of the GPS menu In the lower line of the display are arrows pointing to the left and right. Pressing the button beneath the arrow changes the state to state_ade and now the A/D converter menu options are displayed. Staying in the GPS menu you can press OK to get to the first GPS sub-menu, here you have the option to switch the GPS module on and off (state_sPs_toggle). When powering down the GPS module it is important to observe the cor rect power-down sequence to avoid any possible Internal memory data corruption. On one level with the state to switch on and off there are dis- play options for GPS data such as longitude and latitude which can be selected using the left and right pushbuttons,‘As an exemple we can show how the longitude is displayed ‘on the LCD. In the state_tong ‘tude state it will (automat- ically when valid data is available) assign to the vector ele ments of out_icd_tine the display data elements. In the VHDL description, for example, the data element in_ton_pre Is assigned to the vector element out_led_Linel(5). Atthe fifth position on the first display line is the value of sn_ton_ pre which in this case will be either the letter ’E" or ‘W’ ie ‘east or west of the central meridian, In accordance to this principle each position of the display will be assigned the character to be displayed. After this process refresh_tcd refreshes the display and displays the characters. The principles of state changing and display of date described above also operate in the same way for the other menu pages. To sum up ‘As an example project we have demonstrated @ menu driven control of the FPGA expansion board. All the Modules consist of systematically implemented state machines. This project used up 40 % of the gates and look up tables in the FPGA. ‘There is still enough in reserve for additional applications, It would be fairly easy, for example to add @ function to convert the A/D output values into a temperature reading or a voltage level, (130380) Internet Links [1] Multi 1/0 for FPGA Development Board, Part 2 www elektor-magazine.com/130148 [2 Xtinx 1S: www xilinx.com/support/download/index.htmY/content/ xilinx/en/dovinloedNav/design-tools/v2012_4=r-14_5. html [3] Data sheet for the A/D converter wwrw.ti;com/lit/ds/snas531b/snas531b.pdf [4] wwwelektor.comy130380 EiPololu Poa are ‘ 3pi Robot ey ITEM 4975 $9gQe5 nee + User programmable in C/Ce* (sarple programs avalabe) + Designed exe! at maze solving ane tne faloning * Ships fly asverblewith motor, LCD, burr, ane five rtlectance oe Micro Metal ‘A4990 Dual Mote Gearmotors Driver Carrier 15 tes «+ bonensofmotarsnnding ard» 6Vto32V opening range gest rlio combinations stocked + 07 continuous per chanel + Dualshatt versions avalablefor + current conta bts peak mounting encedert cument 199A per motor Mini Maestro 12-Channel_ Power HD High-Torque USB Servo Controller Serve 1501MG Trem m352 ITEM #1057, $2995 $1995 + Us8,seralandinteralserpting + Speceatev: contol SG ezin + 5,18 and zechannel versions + 014 seclso" he walle + Mata gens and ball bexngs Addressable RGB 30-LED Strip, ESS: 5 V, 1m (WS28128) Temes Es > $1695 =aaa Wacerroo,aruevaly adressble LED sip thatrans on SV. Can be chained o form longer spe oeutr shorter sections Other lengths and LED densities svalble 12 V Step-Up Voltage Regulator u3vsoF12 ITEM 92568 $1395 ] + Accepts inputvltages alow 3729 + SAswiteh allows forinput cents upto A + Fed! 56,9, 12 and 24V versions az well a adjustable versione avaible Priori net ented ee aor ne a eee) Ca scge nie Le a cd in: Pololu has the unique products — from actuators to wireless modules — that can help you take your design eres Find out more at: www.pololu.com wwwcelektor-magazine.com | January & February 2014 | 47eProjects By Frederik Crevits @elgium) 555 Class-D Audio Amplifier A novel use for an old-timer The celebrated ‘555’ IC was originally developed as a timer device, but over the years this golden oldie has been used in all sorts of other applications, including some never imagined by its inventors. This article describes a small class-D stereo audio amplifier built around Sa the oscillator and modulation sections of a 555, The simple design makes it easy to build the circuit yourself. Thousands of different applications can be imple- mented using the popular 555 timer IC, Here this IC is used as the basis for a simple class-D audio amplifier that operates without overall negative feedback. Only standard components are used in the circult, In addition to 48 | January & February 2014 | wwww.elektoremagazine-com Sf & making the circuit easy to build yourself, this makes it a good learning project. Fur- thermore, this project shows that class-D does not have to be exotic or difficult. ‘The key component of this circuit is the NESS timer IC, Originally developed in 1971 (ancient times In the Integrated Circult world), It Is still enormously popular due to its simplicity, rellabil- ity and versatility. A pair of MOSFETs (type IRF530) driven by a voltage level shifter enable the circuit to deliver sufficient output power, To avoid making the cir- cuit unnecessarily complicated, we opted for @ single supply voltage, which means that an out put capacitor Is required,555 Audio Amp EE Measured Performance @ Elektor Audio labs Figure 1, which shows a complete stereo ampli- ‘* Input sensitivity: 580 mV (THD+N = 1%) fier. To avoid having to mention a whole raft 830 mV (THO=N = 10%) of component numbers in the following circuit | » tnput impedence: ka description, we limit ourselves tothe components | « Continvaus output power: -—_~6.3. W (THD+N = 1%) in the upper channel. 10.8 W (THN = 10%) ‘+ Power bandwidth: 11 Hz ~ 37.5 kHz (-3 dB) Ina class-D audio amplifier, the analog eudio ils Heat wis (las) signal is converted into a pulsewidth modulated signal that drives the output transistors. Here this is implemented by employing the 555 as an + Signal to noise ratio: 69 dB (1W/ 89; B = 22 He 22 kHz) ++ Total harmonic distortion + noise: 0.23% (1 kHz} 1 W/ 8 2) ‘stable multivibrator and using the analog audio | * Channel separation: 42 dB (200 Hz, Prin) signal to modulate the voltage for the charge/ 54 dB (1 KH2, Pras) discharge capacitor (C2). In standard 555 cir- 60 dB (20 KHz, Prax) cuits this capacitor is usually charged by | _+ Current consumption: OBA a2x6.3W/82 Figure 1. Fall stereo version of the class-D amplifier based on way | the S55IC. A simple switeh- fon delay circuit i included wwwcelektor-magazine.com | January & February 2014 | 49eProjects Waa eines ‘Although the arrangement with T3 and T4 dramatically Improves the linearity, It |s somewhat less than perfect. The current. through the capacitor Is the sum of the current determined by the base-emitter voltage over R6 and the base current of T3 oF T4, Since the collector current of T3 or T4 depends on the current through R7, the base-emitter voltage (and therefore the base current) also depends on this current. This relationship Is not linear, and another factor is that the two transistors, re not perfectly complementary, so they have slightly different gain curves. (Diz = 151/0.5W zener dlode HSI = heat nk for MOSFET, aluminum eombonenacaag (13 = 1N4148 plate 130 x 50 mm (5.1 x 2 inch), 1-2mm Tirsmiomiams ~ ecsarc thie Taran = 8557 4 soiting wars fr 0-220 case (e Tia © Boi Bergquse IL-PAD K0, 006", 10-220) Tors © B10 4 coining bush, 3mm Iainea16/809 = ko Tas = 2N2222 Heatsink fo 13, 30K/W (@, Fischer Elek nee Tengmi7i6 = 1RFS30 tronk S« 12 SA 32) nea rehic2 = Tucssscp ssnrteemede power spl, sec, 24V, 2.58 721 R29 = 2.242 ie'=7as nin Rojas = eka pcb # 1303441 1) Wizcllneous 13,42 ~ pin pnneader 0.3" pitch K3(L51,L52 = 2-way PCB screw terminal ‘block, 0.2" pitch REL = Relay, 24V, 12000, 8, DPDT-CO (es, Finder 40,52,7.024,0000) ass fuse, 2A), slow-blow, with PCR. f P1,P2 = 2k0 mUltur preset (Vishay Sfern- mount holder’and cover ice T93¥B202KT20) 11,192 = 1 pin of pinheacer Capacitors C1,C15 = 2.2uF Sov, Smm o17-S.mm pitch F (eg. Panasonic EcQV2H2259L) (€2,C16 = 330pF 1%, polystyrene, 7,18mm pitch (e.g. LCR Components EXES/HR 330PF 21%) (€3-€6,¢8,C10,c17~c20,©22,¢24,C30,C32 = 00nF, X7R, 0.2" pitch (€7,€9,C21,C23 = nF MKT, Sm piteh €14,C25,C28 = 1000pF 35 v, radial 12.5mm dia, Sm piteh (e.g, Rubycon 35¢xF1000MEFCT2 5x25) (€12,C26 = s80nF polypropylene, 15mm pitch (€'g. Panasonic ECWF26E1)AQ) (€13,€27 = 220nF metalized fim (HAT), mm piteh (14,28, = 2200) 35v, radial, 18mm ‘diam. Sram or 7.5mm pitch (€.g. Panasonic EeutP1v222) (C31 = 10pF 100V, radial, 6.3mm iam, 2.5mm pitch (©33 = 47pF 35V, radial, 8.5mm max. dlam., 2.5mm pitch Inductors LiL2 = 47H, 21mn/8.5A, pot core type (Murata Power Solutions 1447385c) ‘L004, 35m5/SA rng core tyoe (Warth Flektronik 7447070) u Semiconductors Ds-D8 = 1NS#19 9,010 = LED, rec, 2x5mm rectangular Dil = LED, red, 3mm Figure 2. The PCB layout is uncrowded, with room for an aluminum plate heat sink for the [MOSFET:—although the dissipation is low with the lass-D architecture, 50 | January & February 2014 | www.elektoremagazine-com@ constant voltage. This leads to nonlinearity because the charge and discharge curves are ‘always exponential. That's not an especially good basis for building an amplifier, To eliminate this problem, here we charge the capacitor with a constant current instead of a constant voltage. This is handled by a current source built around 72 and a voltage to current converter built around 13 and T4, This results in a fairly linear triangular voltage waveform on C3, ‘and the ratio between the PWM output signal of the 555 and the input signal is close to linear. The input signal on connector Ki affects the charge/ discharge time of the capacitor via T1, and in this way modulates the output signal. The switching frequency is approximately 250 kHz. Abuffer stage consisting of a BD139 and 2 BD140 ‘complementary pair (TS and T6) at the output of IC1 prevents excessive loading by the down- stream circuitry. ‘The resulting PWM signal drives a push-pull out- put stage with two MOSFETs (78 and 79), which are able to deliver enough current to drive a 4-9 or 8-9 loudspeaker. Tt is essential to ensure that T8 and T9 never conduct at the same time, since that would short out the supply voltage. However, the dead time (the time when neither of the MOSFETs is conducting) must be kept as short as possible in order to minimize harmonic distortion. This creates a dilemma. With the 15 V supply voltage for the 555, It is not possible to deliver very much power to @ 4-9 loudspeaker, We solve this problem by connecting the MOSFETs to a higher supply voltage—in this case 24 V. Since the high-side MOSFET (T8) always sees the low-side MOSFET (T9) as a load, the voltage Ucs from the driver stage willl never be high enough to drive TS fully on. As a result, the output volt- age will never rise ebove 15 V and the rest of the power will be dissipated by the MOSFETs as heat. This Is not how class-O Is supposed to work. The circuitry around 77, which bootstraps the gate of TS, remedies this situation. When the ‘output of the T5/TS buffer stage is high (15 V), 7 is driven into conduction and cuts off T8. Capacitor C9 is then charged through D2 and T9 (which is conducting because Its gate is connected directly to the buffer stage) to a level close to 24 V, When the output of the T5/T6 buffer stage goes low (0 V), T7 and T9 are cut off. This puts the 24 V supply voltage in series with the voltage on C9, resulting in a level of approximately 45 V relative to ground. This voltage is sufficient to drive T8 fully on, so the circuit works the way it should. At the output there is a PWM signal with an amplitude of 24 V, while T8 and T9 remain nice and cool. ‘The network D1/R12 is included to control the dead time. It causes the turn-on and turn-off times of T9 to be different because the gate capacitance is charged through R11 but dis~ charged through D1, which is much faster. ‘An LC filter is placed at the output to suppress the 250 kHz square-wave signal, The output from the filter is a clean audio signal that can be feed to the loudspeaker through 2 cable. The filter is dimensioned to have its 3-dB corner frequency at approximately 37 kHz. The network R13/C13 prevents undesirable oscillation when no loud- speaker is connected. ‘The power supply section is very simple, A 7815 provides 2 regulated 15 V supply voltage for the 555 stage. Due to the simple design, the amplifier does not have any real protection circuit, but the circultry around 719 provides a switch-on delay to prevent audible switch-on pops. ‘A good choice for powering the amplifier is 2 low-cost switch-mode power supply with 2 reg Ulated 24 V output voltage. Such power supplies are available from electronics distributors at rea- sonable prices. RFI choke L3 is included in the power supply section to block any noise from the switch-mode power supply, so that it does not reach the amplifier stages. Construction Figure 2 shows the circuit board layout designed for this amplifier. As already mentioned, this is a stereo version that only requires the connection of an external power source, All components are leaded (through-hole) types to make assembly easy. There is room in the mid~ dle to fit a small aluminum plate that provides extra cooling for the output transistors. This is not absolutely necessary, since the transistors remain fairly cool. A few detalls deserve mention, You should use the best possible components for the frequen 555 Audio Amp wewwcelektor-magazine.com | January & February 2014 | 51eProjects EM eer suo Due to the simple design and the absence of overall negative feedback, you shouldn't count on especially low distortion figures—but all things considered, the results are much better than expected. During the measurement session, it was interesting to see that the distortion components from the different stages of the amplifier partially cancel each other at some output amplitude levels. Plot A shows this effect (THD versus input level, 1 kHz, THD-+N With 8 = 22 kH2), The biue curve shows the distortion at the output of the 555 (pin 3) after filtering out the oscillator frequency. The red curve shows the distortion at the output of the amplifier. At input levels between 7 mV and 40 mV, the distortion at the amplifier output is lower than the distortion at the 555 output. This may be due to the dead time in the output stage (similar to the distortion of 2 standard class-B output stage), or it may be due to the output filter. Plot B shows the frequency characteristic of the amplifier at 1 W into 8 £2 The lower corner frequency Is 11 Hz, and at the other end the -3 dB pont is at 38 kHz. There is a small bump (0.56 dB) in the vicinity of 19 kHz, The slight overshoot can be reduced by decreasing the value of filter capacitors C12 and C26 to 390 nF, but this also reduces the attenuation of the modulation frequency by some 4 8. Plot C shows the total harmonic distortion plus noise versus frequency with an 8-2 load, The red curve shows the distortion ahead of the output fiter of the amplifier, while the blue curve shows the distortion at the output (after the filter). It is evident that the filter does its job properly and effectively suppresses the intermodulation products in the higher frequency region. Both measurements were made using the class-D measurement filter described In Elektor in 2011 [3], with the bandwidth of the analyzer limited to 80 kHz. Plot D shows the Fourier spectrum of a 1 kHz signal (1 W 78 2). The five harmonics responsible for the majority of the distortion (THD+N = 0.23%) are at a level of ~60 dB. Intermodulation products are also visible with a spacing of 326 Hz. They result from the difference between the clock frequencies of the two modulators, However, these products are under -85 dB. They disappear if one of the channels is switched off, To avoid these intermodulation products, the clock frequencies would have to be at least 40 kHz apart. 52 | January & February 2014 | www.elektoremagazine-comcy-determining capacitors C2 and C16. Polypro- pylene and silver-mica are both good choices, To keep the temperature coefficients of the current sources T2 and T11 as low as possible, the (rect angular) LEDs D8 and D10 should be fitted in close contact with the corresponding transistors (12 and T11), They are therefore located close together on the PCB, Murata pot cores and poly- propylene capacitors are recommended for the ‘output filters (L1/C12 and L2/C26). In the prototype we used ceramic insulators for the MOSFETs, but you can also use insulators made from other materials (mica, Kapton, etc.) because the power dissipation is low. First mount the MOSFETs firmly in the right positions on the ‘aluminum plate, and then bend the leads so they fit precisely in the corresponding holes in the PCB without exerting any lateral force on the MOSFETs. The aluminum plate can be mounted on the PCB using two small brackets. This also ensures that the plate is connected to circuit ‘ground, Mount the heat sink on the board before soldering the leads of the MOSFETS to the board. Fit a small heat sink on the 15 V voltage regula- tor (IC3). A Fischer type SK 12 SA 32 (30 K/W) is suitable, but you can also use a small plece of aluminum. Make sure that the heat sink is not jammed against C33. Adjustment Short the input to ground and then adjust mul- titurn potentiometer P1 to obtain the best pos- sible symmetry of the output signal. The voltage measured at the output (behind L1) should then be exactly half the supply voltage, ie. 12 V. That's it—now the amplifier is ready to use, Con ect it to a pair of loudspeakers and an audio sig nal source, sit back and enjoy the pleasant sound of this amplifier, which in many ways resembles the sound of a tube amplifier. Internet Links 1 vowweelektormagazine.com/130144 2] yw.elektor-labs.com/130144 3] wwwwelektor-magazine.com/100540, 555 Audio Amp Figure 3. This side view clearly shows the postion ofthe aluminum plate Figure 4 Inthe prototype the MOSFETs were mounted on the aluminum plate heat srk Using strews with 3-mm insulating washers and ceramic insu wwwcelektor-magazine.com | January & February 2014 | 53eProjects By Paul Newill (plications Engineer, Matrix Multimedia Lt, UX) Mixing Electronics and Mechanics with FLOWCODE G&G Electronic system design software as explained by the developer Flowcode 6 is the latest version of the flowcharting software from Matrix Ltd. As well as improved functionality over the previous versions, v6 includes many new features, This article provides an overview of one of the key new features, 3D modeling and electro- mechanical system simulation. Drag and drop programming Flowcode is @ graphical based programming lan- guage, where flowcharts are the method of cre- ating the program structure rather than the use of 2 text based programming language. Not only does this provide a method of programming which Is free of syntax errors, it also allows for a simple drag and drop of Icons. Flowchart functions are configured by adjusting their parameter proper- tues, For example, within the ‘properties panel’ of 2 ‘while loop’ the user can choose the loop con ditions, such as the number of loop executions or the break condition of the loop, such as when a variable reaches a certain value. Within Flow- code v6 we have also developed a large library of standard components, utilized through component macros, such as LEDs, switches, LCD screens, various communications modules and more, Com= ponent macros vastly simplify the programming ‘54 | January & February 2014 | wwww.elektoremagazine-com of otherwise complex systems. For example, by dragging a single component macro into a flow- chart the user can initialize an LCD display— typically a whole section of C or assembler code would be required to achieve this, The first ‘stepper’ steps One of the included components available with Flowcode v6 is the stepper motor. A stepper motor is typically use¢ in applications such as an XY plotter or a 3D printer due to its ability to achieve very precise movements, with the rota~ tional step angle of the motor being typically around 1.3 to 10 degrees. In this article we will proceed to progress through examples of increas ing complexity, but observe the simplicity with which this can be achieved. In the first example I will demonstrate the sim= plicity with which the stepper motor can be addedto the design, In Flowcode there are two pan- ‘els; the Dashboard and the System Panel. The Dashboard provides a 2D view, so objects sulted to this are placed here. Objects such as LEDs, switches and keypads are typically placed on the Dashboard. The System Panel offers a 3D view, and is typically more suited to objects such ‘as motors, servos or solencid valves. Once the stepper motor was placed on the system panel | ‘added a primitive shape to the panel, The shape of the primitive was then modified to represent ‘a long bar which was placed on the shaft of the motor. ‘The flowchart used to rotate the bar can be seen in Figure 1, where it can be seen that only four functions are required. The first component macro enables the stepper motor. The flowchart then proceeds to the next command—a loop which is configured to execute 50 times. The stepper motor was configured to have a total of 100 steps per revolution. Therefore, after 50 iterations the motor will have rotated through 180°, Within the loop there are the final two commands. The first Is to Increment the step of the motor, while the ‘second introduces a small delay, This delay is Used to slow the simulation down and allow the user to see the animation. If the program was to be downloaded to 2 chip, this delay would be essential to stop the motor stalling and vibrating as it tried to spin too fast. Figure 2a shows the 3D component for the step er motor. Figures 2b-d show the bar attached to the shaft and the simulation at the start, mid= point and end where it can be seen that the motor rotated 180° as desired Linear movements The second example will demonstrate how Flow+ code can be used to convert a rotational move- ment from the stepper motor into a linear move- ment. As previously mentioned, stepper motors are used in applications such as XY plotters or 3D printers. In these applications, the rotational Flowcode 6 MECH Seserewery Ss) movement from a stepper motor must be con- verted from rotational to linear in order to move the XYZ axes. This can be achieved by attach- Ing a threaded bar to the motor shaft, and a threaded nut is fixed to the axis. As the threaded bar spins, the axis platform will move accordingly. In this example, the stepper motor properties were adjusted to simulate a linear movement ‘each time the Increment Step component macro was called from the main program. This linear movement was then assigned to the unique han- dle of @ shape within the 3D design environment. ‘The program is too large to include @ screenshot within this article, but the file is included within the Flowcode example files to allow you to look over the flowchart yourself. Figure 3 shows the 3D model constructed to demonstrate two axes of linear movement. As the stepper motors rotate, the gold objects move along the rails in a linear motion. This example works by utilizing a flag within Flowcode to determine the direction of travel. If the flag |s set to 0, the motor will step and rotate the shaft anti-clockwise, If the flag is sset to 1, the direction of travel will be the oppo- site. The difficulty encountered in this program was determining when the moving object had reached the limit of the axes and the simulation Figure 1. Exemple 1: flowchart for stepper metor simulation Figure 2 Example 1: stepper Motor 3D model and simulation screenshots (2-4). www-elektor-magazine.com | January & February 2014 | 5SeProjects Figure 3 Example 2: simulating linear ovement from steaper rotors within Flowcode 5 Figure 4 Example 2: collision detection macro. ‘must subsequently reverse the direction of travel. Since these examples are simulation only, we Implemented commands known in Flowcode as “SIM Command”. A SIM command is one that will only execute within Flowcode and if the design was compiled to a microcontroller the SIM com= mands would not be sent. In this example, @ push-to-make switch component could be con- figured within Flowcode to operate in the same way, however, SIM commands were used here to demonstrate their simplicity and flexibility as there will be times when you only require SIM commands. There are many more SIM commands, In Figure 3b, it can be seen that the top yellow marker has changed to red, This has occurred due to Flowcode detecting a collision between two objects, determined by the section of code shown In Figure 4. When a collision is detected, the ‘yes’ branch of the decision statement executes. In this branch, the first icon shows that a macro named ‘Trig_High’ executes. Within this macro the author has created a flowchart that ternpo~ ‘56 | January & February 2014 | wwww.elektoremagazine-com rarily changes the color of the yellow marker to red, to notify the user that the collision has occurred. The second command within this branch then changes the value of the direction flag to 1, reversing the direction of travel. 3D models ‘The final example provided within this article demonstrates a progressive step from those pre- viously detailed. Here an XY plotter was simulated and also built from actual parts, to demonstrate the complexity of systems which can be created within Flowcode v6. In this example the 30 model of the plotter was first created in SolldWorks— Figure 5, The model was then separated into five sections and each was exported from Solid Works as an STL file. These five files can then be simply imported into the Flowcode environ ment by drag and drop onto the system panel once imported, each of the five 3D models were configured as required, ‘The specification of the plotter was that the pro- gram was required to read in a set of XYZ coor dinates from an array, and draw a shape deter mined by the coordinates. The given coordinates saved in the array were in the unit of millimeters, therefore In order to reach the coordinate posi= tion the distance between the current point and the new coordinates must be calculated, This value was then scaled by the thread pitch of the threaded bar, anc the step size of the motor. In our plotter the stepper motors have 200 steps per 360° revolution, The pitch of the threaded bar was 1.25 mm, and therefore each motor ‘must through rotate 160 steps to achieve Immof linear movement. For diagonal movement 2 function was also created to allow both motors to rapidly alternate movement. The gradient of tthe diagonal line was dependent on the value of fa scaling factor, calculated by the Bresenham's line algorithm. The final coordinate to consider was the Z axis. Despite the plotter being only 2, the Z axis was required to lift and lower the pen in order to draw the required shapes, This was performed with a flag; when the flag was set to 1 the pen would lower and allow drawing and when the coordi nate value was 2 0 the pen would raise up. A solenoid valve was used to raise and lower the pen as required. Use ‘GetCoords’ in the real world For the simulation the XY plotter requires sev- ‘eral of the 3D models to be combined using the ‘group function. This is particularly useful for sys- tems where individual parts move, but are linked, like in the x-axis where both the pen and pen holder must move together as the x-axis moves. However, they must also be able to move inde pendently to allow the solenoid valve to lift and drop the pen. Simulation commands are used to provide the ability to simulate, ang therefore mir- ror, the movement of the pen around the drawing ‘area in the XY plotter machine, This movement of axes is performed by linking the 3D objects within the System Panel to each stepper motor, Flowcode 6 MECH and achieved by configuring the properties of each stepper motor. However, this still does not provide the ability to draw a shape within the simulation which mirrors that which the hard- ware will draw. For this we need to implement additional SIM commands. Here, the command GetCoords is used to regularly check the XYZ Position of the 3D object representing the pen. By knowing the position of the pen, a second SIM command can be used to draw a circle point at tthe exact XY location, Doing this for each incre~ ‘mental step of the motor produces a pseudo line, constructed from a series of dots. Since this is a complex system, the design utilizes more than 20 flowcharts, as well as the use of hardware and software interrupts and many simulation com- mands which allow the same flowcharts to be used for both simulation and programming to the microcontroller. From software to hardware ‘The hardware of the XY plotter requires a microcontroller board, here the ECIO40 which Is attached to the EB061 to provide break out ports A-E, PORTA is connected to a motor board, from which the subsequent stepper motor is con nected to the X-axis to provide horizontal move- ment. The second motor board for the Y axis is connected to PORTD. PORTS is connected to a numerical keypad, which provides @ user inter face, This also allows manual movement of the X Figures. Example 3; XY Platter in SolidWorks. wwwcelektor-magazine.com | January & February 2014 | $7eProjects Figure 6. Example 3: 3D model of XY Plotter within Floweode v. Figure 7. Example 3: XY Plater model and Y axes if and when required. PORTC is con nected to a terminal Eblock board, which allows the axis limit switches to be connected, These are connected to the microcontroller as hardware Interrupts, which when switched will immediately halt the plotter. These are required if the coordi- nates provided exceed the drawing area of the plotter, which could cause damage to hardware. ‘The final Eblock is @ power board used to pro- vide the required 7.5 volts for the solenoid valve, which Is used to lift and drop the pen. To conclude ‘The following figures show both the simulation of the XY plotter within Flowcode, and the con- structed hardware. In Figure 6 we can see that the simulation of the XY plotter is capable of drawing a shape, constructed purely from XY coordinates. To conclude, Figure 7 pictures the electro-mechanical hardware of the XY plotter. Additional Information Elektor International Media is official distributor of Flowcade 6 and E-blocks products. More articles and sales information can be found at www.elektor.com/flowcode Follow Matrix on Twitter: @MatrixFlowcode ‘Check out the Matrix blog: www.matrixltd.com/blog ‘58 | January & February 2024 | www.elektoremagazine-comHe should have used a LabSal GPS simulator - Record once, replay endlessly Recreate real world conitions Try LabSat FREE for 14 days now www.labsat.co.uk/triall4 tate Ue Ea eae) CONFERENCE & EXHIBITION > March 25-27, 2014 MANDAL ie ireeh MeL 5 LAS VEGAS, NEVADA ane aan FOR NEW HORIZONS - BUOY eee ee eee eee + EXPLORE emerging technologies andthe atest innovators thot ere eect reas OO) OU Nate Oe ead crea ret etn one etd bee ee BN MINOR egy erred ea rg aS www.l PCAPEXEXPO.orG pao eee Bee Nee Teel oN ee Ra woeeProjects By Fons Janssen (Maxim Integrates, Netherlands) LED’s Replace That Halogen Lamp Economical alternative for the GU4 There are numerous LED alternatives available these days for most types of in= candescent and halogen lamps. But as an electronics hobbyist you can of course design and build something like this yourself, as this article illustrates. The LED lamp presented here can replace a GU4 model 10-watt halogen lamp, uses only one-tenth of the power and is barely bigger than the original LED lighting has experienced a great ascent, par- ticularly since several government commissions have banned the good old incandescent lamp. Since 100% compatible LED retrofit lamps are very complex, because of all the optical, ther~ ‘mal, electrical and commercial requirements, the design of such a lamp is normally reserved for large development teams. This however, does not have to hold us back from designing such an LED amp ourselves! ‘The real Elektor reader does not shy away from a few compromises to nevertheless build a working lamp which can also save a lot of energy. LED alternative for a 12-V capsule lamp ‘The LED lamp proposed here is an alternative for 2 so-called halogen capsule lamp, also known as the GU4 (the 4 refers to the distance of 4 mm between the pins). The headline photograph 60 | January & February 2024 | wwww.elektoremagazine-com shows both lamps side by side. In practice the light output of the LED lamp at 1 watt is com- parable to the light output of a 10-watt halogen capsule lamp. This is a saving of 90%! By selecting 12 V as the power supply voltage, we don’t have to worry much about electrical safety, That is because the 12 VAC transformer provides sufficient isolation from the powerline so that it is safe to touch the circuit without having to place trust in an earth-leakage circuit breaker, As can be seen from the photo, the dimensions of the LED variant are somewhat larger than the original, with the consequence that itis not pos sible to replace every GU4 in any light fixture. ‘There are, however, also applications where the additional space requirement is not a problem. A rice example is the so-called ‘starry sky’, where a matrix of lamps is spread across a celling. Our alternative fits exactly in the socket and, at most,will stick out a little further than the original. Because the lamp is not enclosed itis very easy for it to dissipate its heat. Because an LED is a point source it is not pleasant to look at it directly. A special lens solves this problem and provides a nice beam of light. Construction Which challenges remain? The design of the lamp must be simple, robust and frugal. The individual ‘components have to be readily available and the whole assembly cannot be too expensive. As can be seen from the photo, this lamp shines from simplicity. There is no need for a sepa~ rate enclosure; two circuit boards are connected together using standard right-angle header pins ‘The LED board Is clamped firmly to a TO-220 heatsink using two M2 bolts. The small lens is attached with @ couple of drops of glue to the LED board. The main connection pins use the same header pins as those used to interconnect the boards, ‘The LED that has been used here is the so-called Rebel from Philips Lumileds. This is readily obtain« able and available In various colors and color temperatures. By limiting the power to 1 watt we kill two birds with one stone. Firstly, it is possible to use any type of Rebel-LED with that a similar power rating. The second advantage is that the heatsink can remain quite small and finally the LED will operate more efficiently since It doesn’t become so hot. That is because LEDs do not like heat and operate more e'ficiently at lower temperatures. Electronics schema Figure 1 shows the schematic for the circuit for the drive electronics, Diodes Dia through Did form a bridge rectifier for the incoming 12 VAC voltage. The rectified voltage is filtered a little by C1. A tantalum capacitor is used because a ceramic type was found to produce an annoying 1100 Hz (120 Hz) audible hum. D3 was added at a later stage of the project, after a few proto- types without this transient over-voltage prox tection went up in smoke, The LED driver IC, 2 MAX16820 van Maxim (IC1), has a ‘buck’ archi tecture and uses an external switching FET (1), ‘The IC attempts to regulate the voltage across shunt resistor R1 to 200 mV, so that the current through this resistor, the coil and the LED equals 200 mV/620 m® s 320 mA. The forward voltage drop across the LED Is about 3 V, so thet the LED-Halogen Replacement i. power in the LED amounts to about 1 W (3 Vx 0.32.A). The control itself is quite simple. When the voltage across the shunt resistor Is smaller than 190 mV FET T1 will be turned on. The cur- rent will increase linearly at a rate of AL / dt (Vig ~ Yeo) /L As soon as the voltage has reached a value of 210 mV the FET is turned off again, with the result that the current will decrease at a rate of DE dt = Veo /L noting that the current is returned to the input via D2. As soon as the bottom threshold of 190 mV Is reached again, the FET is again turned on and the cycle repeats. The result is a sawtooth shaped current as shown in Figure 2, The switching frequency depends on the input voltage. Since the input voltage is not constant, the switching Figure 1. Schematic diagram for the control electronics. Figure 2 ‘The sawtooth shaped current through the LED. wwwcelektor-magazine.com | January & February 2014 | 61eProjects Component Resistors i = 0.6200 (sMD1205) 2 = 14a (SMD0503) Capacitors ch = 2.2pF 25 (sMp3528) (2 = 1)F 16 (s¥D0603) Inductors Li = 200pH (e., L?55030-104ML8) Semiconductors a,b,c.d,02 = PMEG4020CE] (S0D323F) 3 = SMAI24A (SMA) LED1 = Luxeon Rebel LED Th = IRLML2030TRPBF (SOT23) UL = MAX16820ATT+ (6 TOFN-EP) PLL = right angled pinheader (4 per lamp) Lens: RS Components # 697-4288 Figure 3. The printed circuit board design comprises two parts, Because of space conetraints Dla through D1d are indicated with a single ‘D1 Starry Night set: ¢g. Conrad Electronics #'s 570590— 89, 570591-89, 570592-89 er 570593-89, frequency will vary with the input voltage. For a detailed explanation of this LED driver refer to the article ‘Get a Grip on LED Drivers’ in Elektor April 2009 [1]. The design of the circuit boards ‘The printed circuit boards were designed using the well-known CAD program DesignSpark. The only through-hole components are the header pins—all the other parts are SMD types, which allows the boards to be as compact as pos- sible. The design is shown in Figure 3. The DesignSpark design files are available as a free download [2] To make the cooling of the LED as good as is possible, there are a large number of thermal vias around the LED. The purpose of these is to Figure 4, Osciloscope printout showing the input voltage ((op) and the LED current (Gottom, 62 | January & February 2014 | wwww.elektoremagazine-com conduct the hest from the LED to the other side of the board, which in turn is in contact with the heatsink. Use the LED board as a template to locate the holes in the heatsink. The standard hole in a TO-220 heatsink corresponds with one of the LED terminals. The other three holes can be drilled using a 2-mm drill The solder mask on the bottom of the LED-board covers all the electrical connections to prevent short circuits, The header pins therefore have to be soldered only on the top side of the circult board; make sure that the header pins do not short to the heatsink. Tt is therefore very important that the holes are exactly in the right place. On both sides of the LED are two more small holes, which are for the aligament pins of the lens. The mounting of most SMD components using tweezers and a fine soldering iron should not be a problem if you have some experience with soldering. The IC and the LED are a little more troublesome and require a hot-air workstation or a reflow oven. Measurements Figure 4 shows the LED current together with the input voltage. As can be seen, the current is regulated nicely at 320 mA for most of the time. However, the driver circuit cannot oper- ate when the input voltage drops below 4.5 V, The consequence of this is that the LED current Is Interrupted around the zero-crossings of the line voltage, This is however not visible becauseLED-Halogen Replacement the human eye cannot sense this frequency of 100 Hz (120 Hz). As 2 result the filter capacit ‘ean remain small. The LED Is on for about 80% of the time, so that the effective power consumption is 0.8 x 320 mA x 3 V = 0.77 W. The efficiency of the driver circult, Including the losses in the bridge rectifier, amounts to about 77%, so that the lamp has a total power consumption of about 1 W. By changing the value of R1 it is possible to increase or decrease the LED current to your requirements, Just make sure that the tempera- ture of the LED does not became too high (20n5) Electronic transformers Figure 5. Here you can see the various stages of building the lamp: the individual circuit boards (but with LED mounted), the holes inthe bottom of the heatsink the assembled boards with mounted heatsink, the boards soldered together and finally 2 detail ofthe lens in front of the LED. With thanks to Phils Lumileds {for help with the mounting of the LEDS 0° the circuit Boards forthe prototypes. Internet Links [1] www.elektor-magazine.com/080888 [2] wwwelektor-magazine.nl/130315 A source of headaches with 12-V LED lighting Is the compatibility with electronic halogen transformers, These transformers come in many variants, but have as a common requirement that they need a substantial load connected to their output. If that is not so {end this is the case with LED lamps), they will protest. The consequence is that the LED amp will not turn on at all or starts to flash annoyingly. The solution to this requires quite a complex circuit that is outside the scope of this application. This lamp can therefore only be powered from a standard 50-Hz (60-H2) line transformer. By the way, an old laptop power supply is also a good alternative. In principle the diodes Dia through Did could be omitted, because such a power supply has 2 DC output. However we recommend that you leave them as they will then operate ‘as reverse-polarity protection, for itis easy to make the mistake of reversing the lamp in its socket. Keep in mind that the LED power when operating a DC input will increase by about 25% because there are no zero-crossings and the LED is on continuously. wewwcelektor-magazine.com | January & February 2014 | 63eee) c) By Walter Trojan (Germany) Model Train Dual Radio Control Using low-power radio devices Here's the problem—two grandchildren each with a remote controlled train set but both IR remote controllers use the same channel! Short of banishing one to play in another room the author came to the rescue with the help of radio modules, ‘The Royal Express [1] is a cheap model train set manufactured in China and traded from Hong Kong by Golden Bright Manufacturer Ltd It comes complete with steam locomotive, ten- der and wagons with trees, station, houses and rocks—in fact everything you need for a simple layout. Train control is via an infra-red remote controller which allows you to make the train go forward or backward, make noise, smoke and lighting. Above all the set runs from batteries so there is no risk from dangerous voltages. The set Is advertised as sultable for children of ages 4 to 8, Above all its relatively inexpensive so if anything goes wrong during the hacking process it will not be too much of a loss although there may be tears shed. From IR to RF For such a low cost kit t's probably a bit unfair to expect anything too sophisticated in the equip 64 | January & February 2014 | www.elektoremagazine-com ‘ment design. The main problem for the author is that the IR remote controller units use the same “channel”. With two layouts in the same room (or two trains on the same layout) signals from the remote controllers interfere with one another so the trains cannot be independently controlled. First thoughts were to check out the remote con- troller to see if was possible to switch channels. With the covers off I couldn't identify any obvi- us wire links or pads on the PCB that could be bridged to change the channel. All of the train sets must work on just the one channel. Without a circult diagram it was going to be difficult if not Impossible to make the necessary modifications to the transmitter and receiver. It just so happened that I had been experiment: ing with some wireless transceiver modules type RFM12, The planned home control system can be put on hold; harmony in the household gets higher priority, First it was Important to find outif there was room for additional circuitry, Under the roof of the engine tender there was space ‘enough for 2 small PCB, more than enough for the radio receiver in fact. So the solution was clear: an RF data link could be used to transfer control information. At the receiver end we just need to translate the received commands into the signals that would be produced by the standard IR receiver that came with the set, These will then be passed on to the train controller electronics. For the wireless link, RFM12 modules can be used which operate in the 433 MHz ISM radio band, In countries where 433 MHz is not allocated to license-exempt SRD (short-range radio), 868 MHz or 915 MHz mod- Ules should be used instead. ‘The first task was to work out the IR transmitter signals corresponding to the individual commands sent by the remote controller. Hooking up a logic analyzer to the IR receiver output it was possi- ble to capture the control pulse patterns. The control message consists of a sequence of five pulses. The commands are coded by varying the pulsewidth and space ratios. The message pulse sequences are given in Table 1 In place of the IR messages the new RF link will send command bytes that the receiver micro- controller decodes and translates into the pulse sequences recognized by the train’s built-in. controller. A secure RF messaging protocol had already been developed by the author (see the RFM12 Library article in this magazine) and is outlined in Table 2. Some characters (such as OxAA) are RFM12 con- trol characters and can’t be used in the mes- sage strings of version 1 of the protocol (in the mean time release 2,0 of the protocol has been developed which gives better transmission speed thanks to the use of Hamming-Code (2]). A trans- mission rate of 4800 Baud is more than adequate for this application. Transmitter and receive almost the same hardware The transmitter schematic in Figure 1 is a combi- nation of 2 20-pin Microchip microcontroller type PICL8F14K22 (IC2) and a small transceiver mod- ule type RFM12B-433-D (MOD1) from Hope RF [3]. The module supplies a crystal-derived clock to the microcontroller pin 2 (RAS/Osc1). Data transfer between the chips is taken care of by Figure 1, The transmitter circu Table 1, Functions, IR messages and RF comi Function IR Message RF Command Slow forward TLIPIPIPTE|Oxsi Fast forward TPELIPIP® F | oxs2 ‘Stop TPiPIPILI Ee Oxs9 Slow reverse TLILIPIPIE|Oxss Fast reverse TPIPILIPIE|Oxs4 ‘Sound Horn TLIPILIPIE|OxSS ‘Sound? Bell TPELILIP TE | oxs6 ‘Sounds Bridge TLELILIP TOS? Key: I= Pulee, active Low, 660 ps duration = short pause, High, 1130 ys duration |= long pause, High, 2270 ys duration E= End, High, 5 ms Irmo nee Byte Function Value ° Target device address 0x22 1 Transmitter address oxi 2 Message length 0x08 3 Command 0x51,..0%59 4.6 Reserve bytes 0x55, 7 RCI high xX 2s RCI ow xXx weww-elektor-magazine.com | January & February 2014 | 655 | DESIG! INS RK PCB fl = Beware of confusion: The receiver cicuit i almost identical tothe transmitter’, SCK of the RFM12 board. Generation of the SPI signals is taken care of by firmware. Signal nSEL/ C6 activates the transmitter chip and the signal IRQ/RCO indicates when the transmitter buffer is emply. Further tests on the PIC SPI interface indi- cated that the RFM12 board can handle a serial clock frequency up to 2.5 MHz. A quarter-wave length of wire is used for the antenna. Pull-up resistors Rd, RS, R6 ensure the correct quiescent level on the Active-Low signals nSel and nlRQ. ‘Command Fast reverse 23% 43 ‘Slow reverse 44 to 82 ‘Stop 83 to 104 ‘Slow forward >104 Fast forward 66 | January & February 2014 | www.elektoremagazine.com The controller is slightly modified from the orig inal unit, A pot is now used in place of the five pushbuttons to provide slow/fast forwards and slow/fast reverse plus stop. The stop position is the control knob at mid-travel. Turning clock- wise is slow forward and fully clockwise is fast forward. Counter clockwise from the central posi- tion achieves the same speeds in the reverse direction. Compared to the original controller it Is more intuitive and the young train enthusiasts actually prefer to use it. From the circult function the analog Input RAZ sees an input voltage varying in the range from zero to half supply voltage. The measured value is converted into an 8-bit digital value in the range of 0 to 128 by the A/D converter, This translates to 5 speed ranges (this is not an analog control function) given in Table 3. Also on the controller are pushbuttons for the sound generator: Bell (S2), Horn (S3) and "Bridge’ (S4) (sounds like the train is passing over a bridge). These signals connect to inputs RBS, RB7 and RAG (with pull-up resistors) on the microcontroller, The LED connected to RCL flashes each time a command message is sent. ‘The double-sided transmitter PCB has been designed in the Elektor Lab using the DesignSpark [4] software. Component fitting should be quite easy; no SMD components are used in the design. The RF module is available with either ‘SMD mounting pads or a DIP version with two rows of pin headers (at 2 mm grid spacing). This version can be soldered directly to the PCS. Note that the antenna, made up of a 17 cm (6.5-inch) length of copper wire (for 433 MHz), is soldered fn the topside of the board by the SMDs (see Figure 1). For 868 MHz, the antenna length is 8.5 cms (3.3 inches) A9V battery connected at K1 provides power to the circuit and low-drop regulator LP2950CZ-3.3 provides the on-board 3.3 V, ensuring that every last drop of juice gets used up from the battery, Diode D2 in the supply protects the circuit from accidental reverse polarity connection, The switch type for $1 is not important and depends on the type of housing you use for the transmitter. The transmitter circuit's current drain on the 9 V bat- tery pack is 4.5 mA quiescent, rising to 11.8 mA when transmitting. The PIC controller can be ordered (just like the PCB) ready programmed from [5]. It Is also pos sible to program it yourself if you have a pro~Model Train Remote Control ‘grammer. Free software downloads for this project from the Elektor site include the ready-assembled firmware and Pascal source fil, Firmware for this project can be compiled using versions 5.60 or 6,01 of Pascal Pro from MikroElektronika [6]. The fully functional compiler is free to use for pro- {grams less than 2 KB in size. The microcontroller connector K2 hooks up to a Microchip PICKit3 programmer [7] for in-circuit programming ‘The receiver circuit (Figure 2) corresponds quite losely to the transmitter circuit, Not only is there the same PICL8F14K22, RFM12B~433 combine~ tion, they are also identically connected. The microcontroller built into the train is controlled by active-low pulses from the receiver unit via a two stage transistor driver consisting of TL and 72. The output control pulse sequences are given in Table 1. Power is supplied by the battery pack (2 x 3 AA cells) in the train. The battery supply attaches to connector K1 through protection diode D1 and voltage regulator IC2. The 3.3 V output powers the circuit, LED D2 is a status indicator, it gives 2 short flash every 2 s while no data is received and extends to double the length when 2 message is received. The PIC is programmed in exactly the COMPONENT LIST, Receiver Resistors RI-R6,R7,RE = 10K R6RO = 1k a5 = 3309 Capacitors ci,c3-cs 2's a7pF 16v Semiconductors 1,03 = BATA D2'= LED, 3mm, red Ta = Bcs47 ICi = PICI8r2422-1/, programmed, Elektor Store # 130160-42, [5] Miscellaneous MODI = RFY126-433-D, 3.3 version (Hope RF}, oF 1858 Miz 3.3V version (country-specific) Ki = 4-ph pinheader 2 = 5-pin pinheader Pcp # 130160-2 (5) wwwcelektor-magazine.com | January & February 2014 | 67same way as the transmitter unit using PICKit3 connected to K2. ‘The recelver circuit's current drain on the 4.5 V battery pack Is 19.5 mA quiescent, rising to 21.2 mA during message reception mware ‘The transmitter firmware continually reeds the value of the speed/direction pot and pushbuttons, When a change is detected the corresponding command is sent to the train receiver. After the PIC microcontroller and the RFM12-433 module have been initialized the firmware enters an end- less loop where it reads the input voltage level from the speed control pot anc generates a com- mand, Note that 868-MHz modules may require different initializing codes. When this command is identical to the previous one nothing is sent out. It will only be sent ifit is not the same. The microcontroller also takes into consideration the current status of the train: suppose you were to spin the pot quickly from say maximum forward to maximum reverse, The controller first sends ‘out a Stop command and then a Slow Reverse command followed by Maximum Reverse. This puts less strain on the mechanism and pauses between the commands help to give @ more realistic motion. In the second part of the end- less loop the sound generator pushbuttons are polled. When the train is running the correspond- ing sound command Is sent to the train. Similarly the receiver firmware executes an end- less loop after everything has been initialised In the loop It checks if a new message has been received, When an error-free message is received the command byte is interpreted to select the corresponding message defining the pulse/pause sequence sent to the train’s built-in controller, For security the process is repeated three times. Now with the new controller working, peace has broken out in the playroom. The only source of conflict now is whose turn it is to use grandpa’s controller. 04 Internet Links 11] Manufacturer: www.golden-bright.com/ (2) eee: 3] www.hoperf.com/tf/fsk_module/RFM12B.hom 4] www.designspark.com/eng/page/ designspark-peb-home-pas wikipedia org/wiki/Hamming_code 'S]_www-elektormagazine.com/130160 6] wiw.mikroe.com 7] wine-microchip.com/piekit3 COMPONENT LIST, Transmitter Resistors 1,R4,R5,R6 = 10k Ro'= 3300. R324 7ka, PL = Ska trimpot Capacitors i= 10yF 25v c= 4a7yF ov 63-€6 = 1000F Semiconductors DL,02 = BATA3 D3'= LED, 2mm, ree ICE = PIC1B1422-1/, programmes, Elektor Store # 130360-41, (5), 1c2 = LP295002-3.3/NOPB Miscellaneous. MODI = RFN128-433, 3,3V version (Hope RF), 925 ‘Miz oF 868 MHz 3.3V version depencing on area, Ki = 2-pin pinneacer K2 = S:pin pinheacer Si = 2-pin pinheader andr sie switch, 1 make ‘contact '52,53,54 = pushbutton PB #'130160-1 (5] 68 | January & February 2014 | www.elektoremagazine-comThrow off the 8-bit ball and chain! c Sor Scale: 1:1 CFA10036 ARM9 + Linux SOM “ast 454NAZ ARM. = /SD:4GB10 6468 Linux maine kere {cep 12025MB. USBUARTSPURC + KADCIEXPRAUCAN ‘ide: 01/126 GPIO + debupiitus OLED +7401 to $40@0500 Get your project o marks fast: lay down a standard SODIMM connector, Shap ins CPAt00%6 Syetem-On Module and you slaty have scooas fo tone of GPIO and the power of Lnux: Leave te &-0t dark ages behing CFA920-TS Ths ty ever enaied Sone te Cratos P00 cor TET resis buch een SEWER ESSN [1084.25 -hannel aol gy"o ‘ s14ocia1 io $105220500 Crystalfontz www.crystalfontz.com +1 509 892 1200 USB Add USB to your next project. It's easier than you might think! DLP-USB1232H: USB 2.0 UART/FIFO HIGH-SPEED Rise, Osa, 480Mb/s ‘© Royalty-free, robust USB drivers ‘© Muttipurpose: 7 interfaces © No in-depth knowledge of USB required ‘© Standard 18-pin DIP interface; 0.6x1.26-inch footprint DLP-I08-G DLP-IOR4 8.Channel Data Acquisition “Channel Relay Cable DLP-TH1b ony TerBiHumity Cable $29.95! A USB-to-Xiinx FPGA Module DLP Design www.dipdesign.com © 8 1/0s: Digital YO ‘Analog In Temperature © USB Port Powered Single-Byte Commands CUM a La ea Ly 25% Discount on new E PTS an Le UMC eee Te LCLTE 90-day-launch discount on new Elektor PCBs!eProjects By David Cuartielles, Clara Leivas, and Tien Pham (Arduino Verkstad) Arduino on Course (6) For the mobile weather station, we wanted stu- dents to be able to actively gather data from different environments—in the city, through the parks, and along the riverbanks. Each environ- ment might provide a different set of data for the students to use for comparison. We decided to use tethered balloons so that stu= dents could control the placement and gather data at higher altitudes. This also gave us the oppor tunity to include a time-lapse camera enabling the kids to correlate and visualize their data to the different environments. Decisions, decisions We set about gathering the required electron- ics—a temperature/humicity sensor, a micro- phone, a lightweight spy camera, and a CO, sen- sor. To our knowledge, there were no ready-bullt wind speed sensors that would fit our require- ments, Most anemometers require a stable sur~ face or object to attach to In order to get thelr readings, but because ours would hang from a 70 | January & February 2014 | wwww.elektoremagazine-com Report on the Mobile Weather Station experiment Arduino Verkstad (“Workshop”) was tasked to create a set of education- al experiments aimed at introducing Highschool students to electronics. The primary goal was to make it affordable, yet engaging and challenging enough to trigger their curiosity. Included in the set of experiments is The Mobile Weather Station, a cross-curricular experiment combining natural sciences, physics and technology. balloon, we had to figure out a different method. ‘That, along with a few other unexpected issues, complicated what we thought would be a rela~ tively simple project. In order to keep costs low, we decided to use a cluster of balloons rather than a single large Weather balloon. We purchased a consumer-size helium tank along with regular party balloons. Searching online, we were able to estimate that it would take about 25 or so balloons to lift the Arduino, a shield, the sensors and a lightweight frame or container. Next up was the anemometer. To measure wind speed accurately wasn’t our goal so much as to be able to compare the speeds. Deciding to only measure rotations per second, we thought to use a light dependent resistor (LDR) and an LED to create @ sensor. The idea was to place the LDR opposite the LED with a paper disk in between the two, The disk would have @ slot cut out and as the blades would turn and intermittently allow light to pass through to trigger the LOR.Problem solving ‘As part of our first attempt we tried placing the LED very close to the LDR hoping the magnitude of the readings would obviate the need for shielding. As suspected, this wasn’t the case, so we decided to find a dark box to house the electronics. ‘This presented another set of problems regard- ing mechanic assembly, and so we set out minds ‘on using an infrared (IR) sensor and a black and white paper disk instead, This looked more hope- ful as it required just one electronic component, ‘meaning less room for mechanical error. We tested the assembly and got It to work. The next part was a bit trickier—how to actually catch the wind in order to spin the disk. Because our ‘aim was to build a lightweight weather station, we couldn't exactly use its welght to stabilize it so that an anemometer could spin while having fa stationary point as reference. ‘Again, to make the assembly reproducible any- where by students, we decided to try papercraft We made a simple turbine blade out of paper and tape (Figure 1). We mounted it horizontally so that regardless of wind direction, It would spin. In order to keep the balloons stable, we clustered tthe balloons at the four corners hoping that the weather station wouldn't spin. Blowing in the wind Our paper mockup proved that it might work. In any case, we got the IR sensor and paper disk to work so we were quite happy. Time was now upon us to test the proof of concept—we only had the temperature/humidity sensor, IR sensor, land spy camera We first tested the temperature/humidity sensor ‘and camera indoors. We walked over to the Uni- versity bullding nearby, inflated the balloons and were quite happy to realize we didn’t have to use that many balloons (Figure 2). Up it went—we did our high-fives as we loaded the results from the SD card. Now it was time to test the make- shift anemometer, ‘The West Harbor in Malm, Sweden, is very windy and we were hoping that It would yleld good results, So, out we went with the balloons, As soon as the sliding doors opened the balloons {got sucked out and the tether pulled taut. The balloons started to spin wildly and tangled their strings. Not only that, but the winds were so strong that the balloons and weather station blew ‘almost horizontally, seemingly close to the point of breaking the tether string. Figure 1. A paper-and-tape crafted modal ofa turbine for se as an anemometer. Hopefully! Figure 2 About a dozen balloons appeared enough to Tit the Arduino Mobile Weather Station assembly, rather than 25 as initially calculated, ww.elektoremagazine.com | January & February 2014 | 71eProjects was to mount the turbine vertically, The problem was getting it to align to the wind so we tried adding a tail similar to that of weather vanes. That shifted the weight of the unit and simply was too unpredictable as it oscillated in the wind. rectional turbine Finally, we decided to design an¢ 3D-print a tur- bine that would blow in both directions. This was achieved by twisting the shape of the blades along its spinning axis. Since the test school we were working with already had a 3D printer, they could at least print the turbine out at 2 reason- able cost. It was a compromise, but we couldn't think of anything better at the time (and time was pressing) ‘The person in charge of the mechanics (that would be me, David) didn’t have much experi- ence with 3D printing—in fact, it would be his Figure 3 Initially impressed It was only then that we realized what should first time. After 2 few tries, a rough, workable with the Rokkak kite hhave been obvious. The design of the anemom- prototype was printed eter wouldn’t work because the balloons would We tested it again and were glad to find that, never be upright if tethered, At best, it would be yes, the anemometer worked at almost every ata slight angle. Perhaps it was because of the angle, At this point, we decided to also laser-cut traditional design of free-floating weather ba a frame out of MDF (plywood) so that it could Figure 4. Connection diagram intended for Honeshout ide wostng loons that we didn’t consider this, but having a be mounted property, and construction would be tithe "connect-nbesctoe tether changed the dynamics completely. consistent. This would help to enable the sensors, wires’ level Defeated, we tried several designs, one of which especially the IR sensor, to work well. (In terms of cost, it would be relatively inexpensive for us to use the communal laser cutter to produce lightweight kits that we could send off. Again, @ compromise, but still affordable.) Cur joy was short lived as we again went to test it outside (no balloons, just outcoors). Somehow, the IR sensors were receiving a lot of interfer- ence being outside, The only solution we could find was to spray-paint the plastic casing black. Eventually, we would have to print the casing Using only black plastic since other colors allowed Interference. The IR sensor proved to be just as temperamental as the LDR. Moving the sensor a few millimeters closer or farther from the disk would prevent the sensor from reading values correctly. In fact, it seemed that the IR sensor we were using had a small working range toler~ ance, but we finally managed to get it working. Goodbye balloons We found that 30 printers, while being great for prototyping, weren't necessarily the most accu- rate (perhaps it was simply a matter of inexpe- Fence) or smooth, Small points of friction where the print was bumpy meant it now required more 72 | January & February 2014 | www.elektoremagazine-com‘wind, thus raising the speed threshold. Another compromise, but one we again decided to take. We thought we were nearing the finish line but then remembered the Issue regarding the bal- loons. The idea of a kite came up and we figured it would be best to try it that way. So, it was off to research kites. Our first foray into kite making was a box kite. Someone mentioned an anecdote about a box kite capable of producing enough pull to lift human, so we decided that would be our first model. After building a rather sad looking kite Using dowels anc trash bags, we managed to get it flying... for about ten seconds before it crashed back down. Adding wings as stabilizers helped, but the area we were in had unpredictable winds and the kite came crashing down—hours of work gone with a few broken dowels. Eventually, we happened upon a kite design called ROKKAKU~a six-sided kite originating from Japan that was said to be the most stable single-string kite available. We built a mockup with bamboo ‘and plastic sheeting, and got it up and flying (Fig ture 3). We were quite happy with the result, it being the first kite that any of us had ever built We were set to bring the whole thing together fora test run . . . when it was then decided that all of these elements would be a bit too unpre- dictable for our purposes. We needed something 2 bit more foolproof and so we compromised yet again by deciding to hang the mobile weather station from a long stick with a curved hook. This actually turned out be @ good and interesting decision because it felt a bit more interactive. Students could direct the weather station into the trees to see if greenery would affect readings, or hover it over a bridge directly above the water. Satisfied, we went back to working with the elec tronics, We added the sound sensor and it worked perfectly. The CO, sensor was another matter. While it did work, there were complications. It requires warming up and demands quite a lot of care on the source code, and therefore it gets hard to explain to the kids participating in the ‘workshop how the sensor works, Figure 4 gives {an Idea of how everything gets connected up to the Arduino, While we aren’t 100% completely satisfied with the way the experiment turned out, we are sat- isfied with the learning curve, We messed up and failed many times, but eventually, we got there, Sometimes, there is success In failure. ‘The completed Mobile Weather Station assem~ Arduino on Course (7) bly is pictured in Figure 5 as an exploded view. More of these drawings showing the electronic side of things are available for free download- ing from (1). Running the show We tried the weather station hanging fram bal- loons, as well as from a kite. Both attempts at getting meaningful pictures out of the experiment turned out bad. Therefore we decided to hang the station from a long pole with @ staff exten- sion (Figure 6). ‘The first time this weather station was put into practice was at the “Etopia Kids” Tech-Camp in Zaragoza (Spain) that took place In June and July 2013. During this event, kids get exposed to dif- ferent interactive technologies during 2 period of 5 days. They learn how to count in binary, how to program Arduino boards, get robots to move, and how to read environmental data and show it on a display. ‘The kids first assemble the station and then go for a hike to capture data at different locations. ‘Temperature, humidity, sound level and wind speed are captured at a constant pace together with an Image of the location from above, The participants then return to the computer lab, and using the ‘Processing’ software application, they map the information with the images and proceed to create a small presentation for the other groups explaining what they learned about the different locations. The Processing "sketch” developed for the project is also available at [1] Figure 5 Final assembly of the Mobile Weather Station hardware, Temperature, humidity, altitude and C0; level are measured, processed by Arduino and written to an 59 card. The system als takes pictures fram above, Note the 3-D printed turbine incorporated in the Pc, wwwcelektor-magazine.com | January & February 2014 | 73eProjects Figure 6. After ‘ruiless Lessons learned attempts with balloons and As the time of writing we have been running the [1] Arduino software Processing sketch (.2/p le); kites, it was decided to hang experiment with kids for three days. It will be done electrical and assembly drawings (.zip file) the Arduino Weather Staton total of 15 times over the following three weeks wov.elekto~magezine.com/120043 (free {rom a tong poe eres and 150 kids will gt to try the weather stations download). eae eee One thing we have seen is that there are fr too fowskehidbult she many cables on the design, Fora successful con- turbine parts were produced tinuation of this or even the expansion to more indecent clr. sites, the current Arduino shleld design needs to have the components soldered to a board, Also, the camera we hacked for taking the pictures tumed to be as weak as one could expect for the price we paid. We had to resolder the wires that “"hack" the button on it several times. We'll definitely develop this experiment further and if you have any feedback, ideas on how to Improve It, or know of similar successful exper iments that we can refer to, please feel free to let us know at [email protected]. The illustrated story n ‘The set of tech drawings presented to students fMalelugrales Uren cutie le) to assist in the assembly of the Arduino Mobile | arduino Verkstad (“ workshop" in Swedish) Weather Station Is available for free downloading | is a newly formed team—each of us from the Elektor website [1]. One set of draw- | specializes in different areas, some of which ings illustrates the connection of the external | overlap, electronics to the Arduino main board (‘Balloon- | Not everyone was directly involved in the Experiment-xx.gng), the other, the assembly of | process, but nearly everyone was able to the complete unit, including the pole it is hung | contribute a bit of advice and guidance to from (MDF_WeatherStation_Instructionsxx.png). | the main developers, Tien and Clara, 74 | January & February 2014 | wwww.elektoremagazine-comActive ESD Protection For microcontrollers and more This circuit overshoots and undershoots. Normally when dealing with @ microcontroller or other digital circuit the connections on the device are protected against electrostatic dis- charge. Nevertheless engineers are ever taking special precautions when handling such devices to avoid the risks of ESD: the lab will have an anti-static covering on the floor, and nylon clothes and shoes with soles made of insulating material are avoided, And, in case that is not enough, it is normal to wear an anti-static wrist band when moving devices from their anti-static bags to the anti-static bench surface. But what exactly do we mean when we talk about ESD? The ‘human body model’ and others ‘The first model for static discharge, mentioned as ‘early as the nineteenth century, was the ‘human body model’ (HBM). This takes as its starting point 2 voltage of up to 40 kV, a body capaci tance of a few hundred picofarads and a (skin) resistance of 1.5 k2. We find that even with a static voltage of only 10 kV, as might easily be acquired by walking across an artifical fiber car- pet in shoes with synthetic soles, it is possible to discharge through a fingertip at peak cur- rents of up to 20 amps! The discharge also hap- pens in a very short period, perhaps measured in nanoseconds. ‘The HBM was adopted in the electronics industry in the 1970s with the introduction of sensitive JFET devices in space applications. The compo- nents were tested using a simple RC circuit like the one shown in Figure 1, The discharge cur- rent depends only on the resistance in the cir- cult, and the damped discharge curve is lergely free of oscillation and is accurately reproducible, designed to be used when an application requires a greater degree of electrostatic discharge (ESD) protection than that provided by an IC on its 1/0 pins, Although arrays of ESD clamping devices are available, they rarely offer precise limiting of voltage ‘There are also other models that deal with dis- charge through a sensitive component, for exam ple when a low-resistance electrical connection is made between two devices (the ‘machine moder’, or MM), or when a static charge present on the device itself is discharged (the ‘charged device model’, or COM). Good introductions to this sub- ject can be found at [1] and [2]. ESD clamp circuits Figure 2 shows the typical protection circuitry provided on a microcontroller’s 1/0 port. This example is from an ATmega; other microcontrol- lers and logic devices use similar arrangements. ‘Two bipolar protection diodes conduct discharge currents that could cause undershoots or over~ ESD Protection by Peter Kriiger (Germany) Figure 1. ‘Standard test crcuit and current yraveform for the human body model weww-elektor-magazine.com | January & February 2014 | 75eProjects IE ns | Figure 2. Typical E50 protection I circuit, a found in an Atmel ‘ rmieroeortvole, Seance t SSE Figure 3. ‘Current waveform under the eats systenvdevel model Be Se ate 2 200m 76 | January & February 2014 | wwww.elektoremagazine-com shoots to one of the supply rails, either Voc or ground, However, the diodes take about 6 ns before they conduct fully, Since ESD transients can sometimes be considerably shorter than this it is possible that the CMOS circuit structures will be damaged long before the diodes spring into action, The parasitic capacitance of the pin is around 6 pF, and this is quickly charged up by the energy in the electrostatic discharge, Unfortu- nately we cannot increase this capacitance with ut increasing the Impedance of the pin, which is not desirable. Standard ESD protection circuits Ike this one are designed to meet the particular requirements set by the ESD association [3]. However, it is becoming apparent that the traditional models are not appropriate for modern applications, Recent efforts have been directed toward developing a new ‘system level model’ (SLM) which takes into account the different aspects of the older models. This model employs two stored charges that are discharged in different ways, creating @ high-amplitude current pulse that decays very quickly plus a low-amplitude pulse that dies away more slowly. The energy transferred in a dis~ charge under the SLM can be very much higher than that in the traditional models (Figure 3), Itis readily apparent that the conventional 1/0 pin clrcultry on the IC Is not sufficient to provide ESD protection under this model. Also, the con tinuing industry pressure to make smaller and ‘more complex structures makes it very difficult for design engineers even to maintain current levels of ESD protection, let alone improve on them, In other words: the silicon area needed to provide ESD protection in accordance with the SLM Is simply not available! For this reason external ESD clamp circuits (see [4], for example) are becoming more rele vant. If 2 component provides only a low level of ESD protection (or even none at all) itis possible to add such a circult at the points most at risk The clamp circuits usually use so-called transient suppression diodes (transils or tranzorbs) which, like Zener diodes, start to conduct at a specified Figure 4. Thie protection circuit clampe voltage transients jutside defined upper and lower thresholds.threshold voltage, However, unlike Zener diodes, they react quickly and can. withstand much higher current transients, There are many variations on the circuit design, but none has exceptional performance and none offers precise ‘lamping of voltage undershoots and overshoots. State of the art ESD clamping If we are in the lucky position of not having to worry about the last cent of materials cost or the last square millimeter of board area we can easily cre~ ate a ‘state of the art’ active ESD protection circuit from discrete components (Figure 4) The transistor circuit forms a kind of regulated voltage divider. The current through the two resistors R2 and R3 Is such that the voltages across them {are just enough that transistors T1 and T4 start to conduct and T2 and 73 are |just short of saturation, So we have one base-emitter voltage (2bout 600 mV) ‘across each of these two resistors, which means in turn that the emitters of 2 and T3 are 600 mV below VCC and above ground respectively, The circult as shown Is suitable for a 5 V supply; Ri can be changed to suit supplies of 3.3 Vor 2.7 V if needed. ‘What is the point of this complexity? If the 1/0 pin is high (at +5 V) the upper :1N4148 switching diode will conduct fully as its cathode is at only 4.4 V. If a positive voltage transient should occur it will be conducted by the 1N4148, without switching delay, to the positive rail by 1NS817 Schottky diode D2, which acts quickly and has a low forward voltage, The same thing happens with polarities reversed when a negative voltage transient (below ground) occurs. Hence the digital inputs and outputs are protected against voltage excursions outside the range of the supply rails. In addition, voltage peaks are limited by the use of suppression inductors, The Murata BLM series induc- tor [5] presents a relatively high impedance to signals in the 100 MHz range and so can significantly reduce the level of transients. Although the approach we have described works well with digital levels, it is Not suitable for use with signals destined for the analog-to-digital converter (ADC) on a microcontroller. In this case 2 reverse-biased diode between the signal and each supply rail is required to clamp overshoots and uncershoots, with a pair of 10 k® series resistors to limit the transient current, ‘The series-connected capacitors C2 and C3 present a low-impedance path for transients between VCC and ground, and hence spikes on the supply rails will aso be conducted away. 30201) Internet Links [1] vrwws.teseq.de/de/de/service_support/technical_information/01_Transient_ immunity_testing_e.pdf [2] wrwwrtl com/lt/sg/sszb130b/sszb130b.pdF [3] http://wwwnesda.org [4] wuw.serntech.com/circultsprotection/esd-protection/ [5] www. murate.com/products/eme/basic/feature/bl_Intro.html COM~101 $99.95 BOMHz SCOPE $DS5032E $299 ‘soe 2-cn scope warn soowsa/s rate ‘EnugetoMsa memo {8° color TFT-LCD & FREE Carry casel spse062 $349 10OMHz SCOPE Higrrenc oveds 2h 168a/s bencnscope wen wsa memory ‘and USB port + FREE ‘scope carryoase. Super bw pict DSi02E $399 1OOMHz SCOPE soot 2-cn scope with 18/6 sempie ‘ate ana 8" color 7 Lcd, Huge amounts ‘OF memory + FREE scope carry case. sps7102 $429 1OOMHz MSO chooses ‘scope + B-cniogic analyzer. USB 20 and ‘4 sarngies etorage per channet wth ‘20venced triggering & math functions. 53284 $1359 20MHz HANDHELD Fast & accurate nananets BoM: Fen osclloscope. =100.M/S samplerate *3851n. color TFF-cO ~Bnour patteryare FREE rugged, moact resistant caset HDS1021IM $269.95 ‘wwwelektoremagazine.com | January & February 2014 | 77