Silicon ChipMailbag - May 2022 SILICON CHIP
  1. Outer Front Cover
  2. Contents
  3. Publisher's Letter: Most software is a product, not a service
  4. Mailbag
  5. Feature: All About Heat Pipes by Dr David Maddison
  6. Feature: The History of Transistors, Pt3 by Ian Batty
  7. Project: AM-FM DDS Signal Generator by Charles Kosina
  8. Project: 500W Power Amplifier, Part Two by John Clarke
  9. Feature: Air Quality Sensors by Jim Rowe
  10. Project: Slot Machine by Gianni Palotti
  11. Serviceman's Log: Where there’s a weld, there’s a way by Dave Thompson
  12. Project: Oatley LED Lighting & Driver Kits by John Clarke
  13. Subscriptions
  14. Circuit Notebook: Simple stereo microphone by Thomas Scarborough
  15. Circuit Notebook: A useful wireless charger by John Russull
  16. Circuit Notebook: Lithium-ion battery reconditioner by Dan Amos
  17. Circuit Notebook: Motion-triggered ESP32-based WiFi camera captures intruders by Bera Somnath
  18. Vintage Radio: Calstan 559M2 superhet by Fred Lever
  19. PartShop
  20. Ask Silicon Chip
  21. Market Centre
  22. Advertising Index
  23. Notes & Errata: Dual Hybrid Power Supply, March 2022; Remote Control Range Extender, January 2022; Programmable Hybrid Lab Supply with WiFi, May & June 2021
  24. Outer Back Cover

This is only a preview of the May 2022 issue of Silicon Chip.

You can view 0 of the 112 pages in the full issue.

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

Articles in this series:
  • The History of Transistors, part one (March 2022)
  • The History of Transistors, Pt2 (April 2022)
  • The History of Transistors, Pt3 (May 2022)
Items relevant to "AM-FM DDS Signal Generator":
  • AM/FM DDS Signal Generator PCB [CSE211002] (AUD $7.50)
  • ATMEGA644PA-AN programmed for the AM/FM DDS Signal Generator [CSE21100A.HEX] (Programmed Microcontroller, AUD $20.00)
  • 0.96in cyan OLED with SSD1306 controller (Component, AUD $10.00)
  • 0.96in yellow/cyan OLED with SSD1306 controller (Component, AUD $10.00)
  • Pulse-type rotary encoder with pushbutton and 18t spline shaft (Component, AUD $3.00)
  • Firmware for the AM/FM DDS Signal Generator [CSE21100A.HEX] (Software, Free)
  • AM/FM DDS Signal Generator PCB pattern (PDF download) [CSE211002] (Free)
  • Front panel label and drilling diagram for the AM/FM DDS Signal Generator (Panel Artwork, Free)
Items relevant to "500W Power Amplifier, Part Two":
  • 500W Amplifier Module PCB [01107021 RevD] (AUD $25.00)
  • Hard-to-get parts for the 500W Amplifier (Component, AUD $200.00)
  • 500W Amplifier Module PCB pattern (PDF download) [01107021] (Free)
Articles in this series:
  • Fan Controller & Loudspeaker Protector (February 2022)
  • Amplifier Clipping Indicator (March 2022)
  • 500W Power Amplifier, Part 1 (April 2022)
  • 500W Power Amplifier, Part Two (May 2022)
Articles in this series:
  • El Cheapo Modules From Asia - Part 1 (October 2016)
  • El Cheapo Modules From Asia - Part 2 (December 2016)
  • El Cheapo Modules From Asia - Part 3 (January 2017)
  • El Cheapo Modules from Asia - Part 4 (February 2017)
  • El Cheapo Modules, Part 5: LCD module with I²C (March 2017)
  • El Cheapo Modules, Part 6: Direct Digital Synthesiser (April 2017)
  • El Cheapo Modules, Part 7: LED Matrix displays (June 2017)
  • El Cheapo Modules: Li-ion & LiPo Chargers (August 2017)
  • El Cheapo modules Part 9: AD9850 DDS module (September 2017)
  • El Cheapo Modules Part 10: GPS receivers (October 2017)
  • El Cheapo Modules 11: Pressure/Temperature Sensors (December 2017)
  • El Cheapo Modules 12: 2.4GHz Wireless Data Modules (January 2018)
  • El Cheapo Modules 13: sensing motion and moisture (February 2018)
  • El Cheapo Modules 14: Logarithmic RF Detector (March 2018)
  • El Cheapo Modules 16: 35-4400MHz frequency generator (May 2018)
  • El Cheapo Modules 17: 4GHz digital attenuator (June 2018)
  • El Cheapo: 500MHz frequency counter and preamp (July 2018)
  • El Cheapo modules Part 19 – Arduino NFC Shield (September 2018)
  • El cheapo modules, part 20: two tiny compass modules (November 2018)
  • El cheapo modules, part 21: stamp-sized audio player (December 2018)
  • El Cheapo Modules 22: Stepper Motor Drivers (February 2019)
  • El Cheapo Modules 23: Galvanic Skin Response (March 2019)
  • El Cheapo Modules: Class D amplifier modules (May 2019)
  • El Cheapo Modules: Long Range (LoRa) Transceivers (June 2019)
  • El Cheapo Modules: AD584 Precision Voltage References (July 2019)
  • Three I-O Expanders to give you more control! (November 2019)
  • El Cheapo modules: “Intelligent” 8x8 RGB LED Matrix (January 2020)
  • El Cheapo modules: 8-channel USB Logic Analyser (February 2020)
  • New w-i-d-e-b-a-n-d RTL-SDR modules (May 2020)
  • New w-i-d-e-b-a-n-d RTL-SDR modules, Part 2 (June 2020)
  • El Cheapo Modules: Mini Digital Volt/Amp Panel Meters (December 2020)
  • El Cheapo Modules: Mini Digital AC Panel Meters (January 2021)
  • El Cheapo Modules: LCR-T4 Digital Multi-Tester (February 2021)
  • El Cheapo Modules: USB-PD chargers (July 2021)
  • El Cheapo Modules: USB-PD Triggers (August 2021)
  • El Cheapo Modules: 3.8GHz Digital Attenuator (October 2021)
  • El Cheapo Modules: 6GHz Digital Attenuator (November 2021)
  • El Cheapo Modules: 35MHz-4.4GHz Signal Generator (December 2021)
  • El Cheapo Modules: LTDZ Spectrum Analyser (January 2022)
  • Low-noise HF-UHF Amplifiers (February 2022)
  • A Gesture Recognition Module (March 2022)
  • Air Quality Sensors (May 2022)
Items relevant to "Slot Machine":
  • Slot Machine PCB [08105221] (AUD $5.00)
  • DFPlayer Mini audio player module (Component, AUD $5.00)
  • Micromite Plus LCD BackPack kit without touchscreen (Component, AUD $45.00)
  • Laser-cut acrylic pieces for the Slot Machine (Component, AUD $10.00)
  • Firmware, sounds and images for the Slot Machine [SlotMachine V10] (Software, Free)
  • Slot Machine PCB pattern (PDF download) [08105221] (Free)
  • Cutting diagrams for the Slot Machine (Panel Artwork, Free)
Videos relevant to "Slot Machine":
  • Slot Machine
Items relevant to "Lithium-ion battery reconditioner":
  • Firmware for the Lithium-ion battery reconditioner (Software, Free)
Items relevant to "Motion-triggered ESP32-based WiFi camera captures intruders":
  • Firmware for the Motion-triggered ESP32-based WiFi camera (Software, Free)

Purchase a printed copy of this issue for $8.50.

MAILBAG your feedback Letters and emails should contain complete name, address and daytime phone number. Letters to the Editor are submitted on the condition that Silicon Chip Publications Pty Ltd has the right to edit, reproduce in electronic form, and communicate these letters. This also applies to submissions to “Ask Silicon Chip”, “Circuit Notebook” and “Serviceman’s Log”. Vale Brendan J. Akhurst I have been a reader of Silicon Chip (and Electronics Australia) for many decades now, but this is the first time I’ve been motivated to write (and it isn’t even about electronics...) I have just read Nicholas Vinen’s “Editorial Viewpoint” in the February 2022 issue and was very sad to learn of the passing of Brendan Akhurst, your long time contributing cartoonist. I first met “BJ” when I was just a lad and his father was my teacher at school. That was way back in 1969 – it was clear that he was a very talented artist even then. I have followed his work ever since. His distinctive style and gentle wit were always a highlight of your magazine for me – particularly when he added illustrations to your Serviceman’s Log section. I was surprised to learn from your editorial that he had also been a Police Officer, for I too had a career in Policing, but we never managed to cross paths again. Thanks to your editorial, I have passed on the news to Brendan’s former workmates in the NSW Police Force, and many of them have remembered him fondly. Please offer my sincerest condolences to his family and thank them for sharing him with us – your readers – for so many years. Bob Stephenson, Canberra, ACT. A collection of Brendan’s cartoons It is sad to note the passing of your cartoonist, B. J. Akhurst. Like you, I found his work highly amusing, especially the visual puns. His work will be missed. It was funny how he highlighted the fact that Dave Thompson is a New Zealander, with Kiwi birds and volcanoes in the distance. It would be nice if you could publish a “Best of Brendan” type of publication with some of his best efforts, although I would understand if you considered this is outside your brief as an electronics magazine. Ray Chapman, Pakenham, Vic. Comment: that is a good idea. We will certainly consider it. More on possible clone CP210x chips I read the letter on CP210x chips in Mailbag (April 2022) with some amusement; my brother gave me a USB/serial cable because it wouldn’t work on his Windows box, yet it worked just fine on my MacBook, FreeBSD server, and Linux laptop. He has another cable that works for him, so I assume it has the “correct” chip. There was a discussion on a private technical mailing list of which I am a member. The consensus was that 4 Silicon Chip Microsoft did indeed hobble the driver, given that a previous version is known to work with these devices. If it were a “driver bug”, as you postulate, surely it would have been fixed by now in one of their tedious “Patch Tuesday” updates. The story goes that the genuine chip has an undocumented feature (documentation available only under NDA) that makes it work faster to give it a competitive edge over clones. Dave Horsfall, North Gosford, NSW. Comment: this is just something to consider, but Microsoft has left serious bugs (including security holes) unpatched in Windows for long periods, see siliconchip.au/link/abdg Colour Maximite wanted Do any readers have a fully working Colour Maximite 2 (V1) they are willing to sell? The fully-assembled CMM2 I can purchase online is too expensive because they want to be paid in US dollars. Ric Mabury, Melville, WA. Twisted rather than separate mains supply lines Energex, the Queensland government corporation that owns the electricity network in SE Queensland, replaced the poles and wires in the eastern end of my street. They installed new poles several weeks ago and then, on Thursday, they replaced the wires. The contractors replaced the supply wires for more than twenty houses in just over four hours. The new poles and wires are totally different from the old system with cross-bars having four individual conductors supported on insulators. Instead, there is a single cable of four twisted wires, and the house supplies are connected via tapping blocks of some sort. I have never seen this system and when I looked around the neighbourhood, my street plus a side street are the only ones using it. I do not know if it is common in other areas. Perhaps electricity distribution methods might be another subject for an article considering that this method appears to be new technology. George Ramsay, Holland Park. Qld. Comment: we have seen that style of mains supply wires before (also called aerial bundled conductors). Presumably, it is quicker and easier to run because there is only one wire bundle, and it has better aesthetics. They must be confident that the inter-wire insulation is not going to degrade. Real-time clock option on Pico BackPack I have built two of your Pico BackPacks (March 2022; siliconchip.com.au/Article/15236) successfully according to the article. One useful minor amendment to add Australia's electronics magazine siliconchip.com.au to the list of commands in the third column on page 40 are these commands to configure the optional DS3231 real-time clock IC: RTC SETTIME 2022, 03, 10, 09, 00, 00 (change numbers to the current date & time) OPTION RTC AUTO ENABLE Mike Sunners, Nairne, SA. Why use a half-wave rectifier? In the Driveway Gate Remote Control article by Dr Hugo Holden (siliconchip.com.au/Article/15197), the circuit diagram on pages 80 & 81 shows the main rectifier for the motor as half-wave. This means that the transformer is subjected to a net DC through the windings, which will cause heavy saturation and exceptionally poor transformer utilisation. This is easily demonstrated, and I cannot imagine why a half-wave rectifier was specified. A single chassis-mount bridge would improve the DC output and greatly reduce the stress on the transformer and filter capacitor (which is also subjected to much greater ripple current). I realise that it’s only used intermittently, but half-wave rectification is the worst possible option. Transformers can tolerate severe overloads for a short time, but given the modest cost of a ‘proper’ bridge rectifier, there’s no reason not to have used one. Rod Elliott, Thornleigh, NSW. Response: If this was a clean-sheet design using all-new components, you are correct that it would be sensible to use a transformer only just large enough to power the gate motor with a bridge rectifier. That would be more cost-­efficient and power-efficient, although that’s hardly critical given its very intermittent usage. But consider that this project is designed to be retrofitted to a pre-existing gate controller. In many cases, the existing designs use a large power transformer with half-wave rectification. Then there’s no real advantage to changing that to full-wave rectification, as detailed in Dr Hugo Holden’s reply to your comments: Typical gate motors run at low average powers in the region of 10-30W. The large power transformers used in most gate controller boxes look good for at least 100W. It doesn’t excessively stress a transformer to draw current on half-cycles if the power drawn is well below the transformer’s rated power. And the gate is for intermittent use; it is not moving 24/7. When the gate is in motion, at a constant speed, the current requirement is in the region of an amp or so because the motor is only overcoming the friction of the moving parts, not accelerating a mass or lifting a weight against gravity. If we assume that the motor supply is around 22V (which it would be with the filter capacitor), motor power is in the region of 20W, but only for the time the gate is travelling. The whole thing has a very low duty in terms of motor on-time. Many low-power appliances use half-wave rectification. Car battery chargers with similar proportions are often based on half-wave rectifiers. There is no risk to the transformer if the power is below its maximum rating. It would work with a full-wave rectifier, but since there is no PWM controller to reduce the motor energy (as in 6 Silicon Chip some commercial controllers), the gate speed would be too high, at least for my gate. The controller I built for my gate doesn’t have the filter capacitor, and the speed is about right with half-wave power from the pre-existing transformer. It has been working like that for over 15 years. (End of Dr Holden’s reply) We agree that it would make sense for anyone who has to purchase a transformer for the Remote Gate Controller to add a bridge rectifier between the transformer secondary and CON4. In this case, for a 24V rated motor, a transformer with an 18V output under load plus a bridge rectifier might give about the right power to the motor (depending on the filter capacitor value). A large 24V transformer will probably deliver too much power to the motor if a bridge rectifier is used. Software obsolescence is a problem I read with interest your editorial in the February edition about devices that cease to work due to phone or PC operating system upgrades. I have had to throw out perfectly good (and expensive) equipment because it would no longer work with my PC or laptop computer. This includes a perfectly good laser printer, an expensive video capture card and an even more expensive WinRadio receiver, all due to the drivers not being updated to work with the latest versions of Windows. Also, I can no longer get data from my solar panels because the inverter manufacturer has gone away. I am now very wary of buying any equipment that requires a PC, laptop, phone or an internet connection to the manufacturer to operate. Mike Hammer, Mordialloc, Vic. Editorial on apps and obsolescence I read with a mixture of amusement and horror your editorial comments about the problem of devices that need apps. I have three internet radios that are now e-waste because Qualcomm pulled the plug on their so-called Aggregate website (Reciva, https://radios.reciva.com/index). All internet radios need such a site for listing and accessing the internet radio stations. I can now only receive the very few stations I had stored on buttons in the radios. I think this is, sadly, a case where it is difficult to follow your understandable advice that “all hardware devices should be able to be used in a standalone mode”. In the meantime, I have had to purchase yet another internet radio from a different company that uses a different aggregate page. Let’s hope that will be usable for a few more years. Christopher Ross, Tuebingen, Germany. Response: one of the aspects that bothers us the most is that it isn’t all that time-consuming or expensive to continue supporting many of these products. Website hosting doesn’t cost much, and the amount of labour required to maintain the site is minimal. It shows a lack of respect for the customers who purchased their products. Misleading battery capacity ratings My son-in-law recently asked me to look at a compact battery pack with USB outputs. He had bought it believing Australia's electronics magazine siliconchip.com.au Development tools in one location Thousands of tools from hundreds of trusted manufacturers Choose from our extensive selection au.mouser.com/dev-tools australia<at>mouser.com its 26,800mAh advertised capacity but found it could not recharge his Macintosh laptop more than once. I recharged it, and it took around 24,000mAh before the charge rate dropped. I then easily got more than 20,000mAh into a LED lamp. The reason for his problem then dawned on me. The ‘battery’ must be a single lithium-ion cell, nominally 3.7V. In converting that to the 5V USB output, he would not get more than 19,800mAh, even if the process was 100% efficient. Then there is the laptop battery, which is described as 10.8V and around 7,000mAh. To recharge that would require around 12.8V, so even at 100% efficiency, no more than 6,700mAh would be delivered. It seems to me that a better solution would be to buy a 12V battery booster with a USB output. So this was a cautionary tale in understanding battery capacity in the real world. Graham P. Jackman, Melbourne, Vic. Comment: giving battery mAh ratings is misleading as it is not a unit of energy. Units of energy are either joules (J), milliwatt-hours (mWh) or watt-hours (Wh). Milliamp-hours or amp-hours is only a useful metric if you want to know how much current a battery can deliver over a certain amount of time. It is not helpful in comparing the capacities of batteries unless you know they have the same voltages. The Wh figure for a battery can be easily estimated by multiplying the Ah capacity by the nominal battery voltage. But note that this will only ever be an estimate because the battery voltage won’t necessarily change linearly as it discharges. My adventures into SMDs When ordering the SMD Trainer kit from Silicon Chip, the gentleman asked if there was anything else that I wanted. When I said yes, he said, let me guess – the SMD Tweezers. No! I was after a Micromite LCD BackPack. After thinking this over for half an hour or so, I thought I was being silly, so I rang back and put the Tweezers on the same order. Over the years, I have built many projects. Apart from crystal sets, the first (in the late 1950s) was a five-band, five-valve radio. When the first stereo record came to town, I added a reel-to-reel tape recorder and turned it into a stereo system. That’s the sort of thing I’m used to building, so these new projects would be a game-changer for me due to the small parts. When the package arrived, I had to wait for some flux gel to arrive. As I usually work in a dusty old shed with a lathe, mill, welders, drills etc, I thought it wise to find a cleaner spot, so I cleared an area in my home office. I then had to make many trips back-and-forth for forgotten things (part of getting old). Was the soldering iron too big? A jeweller’s loupe OK? I was pretty keen to get started, but disaster struck as I opened the packet. Removing one part to try to determine what it was, it flicked out of my fingers and disappeared forever! I wonder what it was. The next trick was figuring out how to open the small packages. After fiddling for a long time and reading and asking Mr Google, I emailed Silicon Chip. The reply was to peel off the plastic layer, which I really hadn’t seen. I 8 Silicon Chip Australia's electronics magazine siliconchip.com.au was then able to extract one chip, but what was it? It had eight pins, so it had to be the 7555. I studied this for a long time to determine where the number one pin was. There was no dot in the corner, and no corner cut off. I woke in the middle of the night with the answer; the whole of one side may have a chamfer. This proved to be correct, but I found it was very subtle and hard to see. The next day, I decided to explore some more of the mysterious mini packages and had no idea what they were, so I decided to build the tweezers. Those worked first time, so I think I scored my first Brownie point. The next problem was to work out what size parts were in which packet, so I opened them and soldered them in the appropriate places. It wasn’t long before I had all the parts above the line in, including the 1μF, which is the one I lost and Silicon Chip kindly replaced. And it worked! I was starting to find that one and a half to two hours of concentration was enough, and as I had other projects on the go, it was a few days before I tackled the next bit. I next tackled the resistors and was quite pleased with my progress until I got to the M1005/0402. I managed to solder that one, but it was crooked. Still, I figured that it was too small for anyone to notice. The M0603/0201 was harder, though. I was still trying to get the plastic off and found it had dropped onto the desk, and then it stuck to the iron, but eventually, it made it into the board. I checked it with the tweezers and it looked good! Another Brownie point! After telling my friend how clever I was, I picked up the Trainer board and brushed my fingers over it, and the M0603 part fell off. 90 minutes later, I had managed to lose all the other M0603 resistors, either by having them stick to the iron and frying or losing them. Loss of all Brownie points! After spending some more hours, I managed to get a few LEDs flashing, so maybe I scored one back. The problem with the LEDs is that the green line is hard to see, especially when darkened on one end with heat from the iron. My recommendations from all this are, if you are serious, look at plenty of YouTube videos, buy two or three trainer kits or boards. Be prepared to destroy one, and buy plenty of spare components to replace the ones that get lost and destroyed. I found that Altronics have a list in their catalog. My conclusion is that I need a better iron, better lighting and a better magnifier (preferably a microscope with a screen). Better eyesight and steadier hands would be good too. I am certainly more confident with SMD parts; M2012/0805 may be my limit for reliable results. What a great project; thank you, Silicon Chip. I can now confidently tackle the Micromite BackPack. David Lloyd, Clare, SA. Comments: here at Silicon Chip, we were all initially hesitant to tackle SMD components as we feared working with them would be difficult. We mostly found that all but the smallest are manageable with a bit of practice. It helps to start when you are younger and still have reasonable eyesight. It also helps to work with smaller parts regularly to maintain that important close-up vision. You are correct that pin 1 can sometimes be hard to determine. We prefer it when there is a pin 1 dot or divot, but that isn’t always the case. And we too find that working siliconchip.com.au Helping to put you in Control LabJack T7 Data Acquisition Module A USB/Ethernet based multifunction data acquisition and control device. It features high data acquisition rates with a high resolution ADC of 4 ksamples/s at 18 bits to 50 ksamples/s at 16 bits. SKU: LAJ-045 Price: $902.00 ea + GST Temperature probe 5m Teflon Cable RTD probe with magnet fixing for surface temperature measurement. -50 to 200 ºC range and 5meter teflon cable. SKU: CMS-007T Price: $153.95 ea + GST J Thermocouple Temperature probe with magnet fixing This J type Thermocouple sensor has magnet fixing for surface temperature measurement. The 2 wire sensor has a silicone cable which is 3m long. Temperature range is -50 to 200 ºC. Class B. SKU: CMS-017J Price: $142.95 ea + GST 400W ACM Brushless AC Servo Motor Leadshine ACM604V60-T-2500 400W brushless AC servo motor with 2500 line encoder suitable to work with the ACS806 brush-less drive. SKU: MOT-450 Price: $347.60 ea + GST ACS806 Brushless Servo Motor Drive Brushless servo motor driver for 50 to 400 W, AC brushless motors with encoders. SKU: SMC-410 Price: $319.00 ea + GST LCD Temperature and Humidity Sensor The Pronem Midi from Emko Elektronik are microprocessor based instruments that incorporate high accurate and stable sensors that convert ambient temperature and humidity to linear 4 to 20 mA. Dimensions are only 60x 126 x 35mm. SKU: EES-020A Price: $241.95 ea + GST TxIsoloop-1 Single Loop Isolator Loop isolators provide signal protection by electrically isolating the 4-20mA input signal from the 4-20mA output. SKU: SIG-201 Price: $168.19 ea + GST For Wholesale prices Contact Ocean Controls Ph: (03) 9708 2390 oceancontrols.com.au Prices are subjected to change without notice. Australia's electronics magazine May 2022  9 with small parts can be fatiguing, and it’s essential to take breaks to avoid eye strain etc. Yes, M0603/0201 components are very difficult to hand-solder, but it can be done with some effort. It’s much better to destroy or lose parts in the Trainer kit than something more expensive. Some SMD micros can cost upwards of $20 each! We prefer to test the LEDs with a multimeter set on diode test mode rather than try to identify the anode and cathode visually. When they light up in that test, the red probe is on the anode and the black probe on the cathode. Clever SMD holding tool I see that you are including an increasing number of surface-mount devices in your projects. I am also finding that most of my personal projects and those for my small design business are forced to use more SMD components because they are all that is available. Recently, the component shortage has forced my hand even more into the SMD “corner”. With this came the realisation that SMD components not only are smaller but they can be mounted on either side of a PCB without interfering with the other side. This allows a doubling of component density and consequential reduction in overall size. My current project has over 120 SMDs. While I have the usual through-hole tools, including a magnifying lamp, I have added a hot air soldering station. Initially, dealing with tiny components was challenging enough without them getting blown away or flipped over by the airflow! Holding SMD transistors, diodes, resistors and capacitors accurately in place could be a problem. As a result, I developed this simple “holding tool”. It is made of 0.5mm steel plate with one edge folded by 90° for strength and springiness, and it articulates across the full extent of a 100 x 80mm board. It is temporarily secured to the PCB with an M3 threaded spacer with Nylon washers to allow movement while minimising slippage. Please refer to the attached photographs. I am in my mid-60s and can say that SMD is not just the domain of the youngsters! I even use QFN (quad, flat, no-leads) packages, although I extend the pads beyond the normal footprint of the chip to allow easier soldering and debugging. A hot-air station is essential for those chips. Peter Gee, Inglewood, WA. Comment: what an ingenious device, and simple too! We especially like how you use one of the existing mounting 10 Silicon Chip holes to attach it to the board. Some of the advantages of SMDs that you’ve noted, like being able to mount them on both sides of the board, are part of the reason they are so popular with manufacturers. We don’t always put parts on both sides just because it complicates assembly; we have some designs coming up where we do, though. Driving loudspeakers with a current source It is my assertion that, for voice coil speakers and ribbon tweeters, the use of voltage feedback is incorrect as this is affected by changes in impedance with frequency. The best result can be obtained by feeding back a voltage produced by sampling the current through the speaker and feeding back voltage proportional to the current. Simply a series resistor. It is also necessary to feed back the DC level to prevent DC from being fed through the speakers. You can find a simple design for this at www.inja.com. au/wp/ or www.inja.com.au/diy.php It can be proven that the magnetic force that applies in these speakers is proportional to current, not voltage, as follows. Wind two coils on two small diameter ceramic formers, one using Nichrome wire spaced to prevent shorted turns, the other with copper with the same number of turns and the same spacing (Jaycar WW4040 and WW4013). You will need an old compass. Connect the coils in series across a low-voltage supply so that they repel opposite ends of the compass. Now place the coils on the same side of the compass such that each coil is attempting to repel the opposite ends of the compass. You will notice that the forces are the same because the same current passes through identical coils. This shows that current, not voltage, determines the magnetic force in a coil as almost all the voltage is across the Nichrome coil with microvolts across the copper coil. If voltage caused the force, the end facing the Nichrome would have produced a much greater repelling force. John Cornwall, INJA Comment: We are publishing this letter in the spirit of open debate; we do not necessarily endorse its content. Note that we have never argued against the notion that magnetic field strength is related to the current through a coil. That is basic theory, and we are pretty sure that SC most speaker designers would agree. Australia's electronics magazine siliconchip.com.au