Fullscreen HTML5Med Res (??MB)HTML4

Pinball The VERY BEST DIY Projects! Machine Customisable and built from scratch Comfort Indicator Measures temperature, humidity and dew point Analog Computers Part 2: modern-day examples and how they’re used in neural networks JUNE 2026 ISSN 1030-2662 06 9 771030 266001 $ 00* NZ $1590 15 INC GST INC GST SERIOUS PRECISION. SERIOUS SCALE. 16K NEW $ JUST 1849 JUPITER 2 RESIN 3D PRINTER TL4988 LIMITED STOCK. ONLINE ONLY Available for special order in-store SEE THE FULL RANGE OF 3D PRINTING ONLINE jaycar.com.au 1800 022 888 jaycar.co.nz 0800 452 922 Prices shown in $AUD, and correct at time of publication but are subject to change. Jaycar reserves the right to change prices if and when required. Contents Vol.39, No.06 June 2026 12 Analog Computers, Part 2 Analog Computers Analog computers are making a comeback because of how well-suited they are for neural network processing. We take a look at some of the modern analog computers and the features of current-day analog AI workflows. By Dr David Maddison, VK3DSM Technology feature 36 Inspection Reports for PCBs Quality inspection reports are important for ensuring that PCBs will be made to the required standard from your chosen PCB manufacturer. Supplied by PCBWay, it includes useful information on manufacturing. By Steve Mansfield-Devine for PCBWay PCB fabrication 54 Whole-Home Sound System Come and see the trials and tribulations brought about by installing a sound system that covers an entire house. It uses three amplifiers, one preamp, thirteen speakers and hundreds of metres of wiring! By Julian Edgar Home audio 66 Working with e-Paper Displays The Human Comfort Indicator (listed below) is the first project where we have incorporated a bare e-paper display into a design. This article goes into more detail on how e-paper displays work and how we are driving it. By Tim Blythman Display modules 26 Phenomenal Pinball Machine We cover how to design and build every part of your own Pinball Machine. While we provide a standard layout, you can customise it to your own tastes, with different artwork, sounds, lighting and more. Part 1 by Phil Prosser Gaming project 43 Human Comfort Indicator Displaying the temperature, humidity and dew point, this handy little device tells you if an area is comfortable to be in. It’s battery-powered with USB charging, and stores historical data for the last day, week and month. By Tim Blythman Environment measurement project 60 Simple USB Power Monitor Using under 20 parts, this project can measure voltage, current and power supplied over USB up to 36V, 3A and 108W. It displays data on a 0.96-inch OLED screen and supports USB power delivery (PD) 2.0 and 3.0. By Richard Palmer Test & measurement project 70 Micropower SSB Transmitters We wanted to see how small we could make a single-sideband transmitter, so we came up with three designs, one using just three transistors. Each of the designs only use through-hole components and no microcontrollers. By Andrew Woodfield, ZL2PD Radio communications project Cover background image: https://unsplash.com/photos/a-close-up-of-a-pinball-machine-18dGkEQ5wSM Part 2: page 12 Source: www.flickr.com/photos/jitze1942/4304353299/in/ album-72157623284316506 Page 70 Micropower SSB Transmitters 2 Editorial Viewpoint 4 Mailbag 53 Subscriptions 76 Kits 77 Circuit Notebook 80 Serviceman’s Log 86 Online Shop 88 Vintage Radio 99 Ask Silicon Chip 103 Market Centre 104 Advertising Index 104 Notes & Errata 1. iClap intelligent multi-clap switch 2. Interruptible NiMH trickle charger 3. IF signal injector for radios Sailor 66T radio by Dr Hugo Holden SILICON SILIC CHIP www.siliconchip.com.au Editorial Viewpoint Default sound settings can ruin the streaming experience Publisher/Editor Nicholas Vinen Technical Editor John Clarke – B.E.(Elec.) Technical Staff Bao Smith – B.Sc. Tim Blythman – B.E., B.Sc. Advertising Enquiries (02) 9939 3295 adverts<at>siliconchip.com.au Regular Contributors Allan Linton-Smith Dave Thompson David Maddison – B.App.Sc. (Hons 1), PhD, Grad.Dip.Entr.Innov. Geoff Graham Associate Professor Graham Parslow Dr Hugo Holden – B.H.B, MB.ChB., FRANZCO Ian Batty – M.Ed. Phil Prosser – B.Sc., B.E.(Elec.) Cartoonist Louis Decrevel loueee.com Founding Editor (retired) Leo Simpson – B.Bus., FAICD Silicon Chip is published 12 times a year by Silicon Chip Publications Pty Ltd. ACN 626 922 870. ABN 20 880 526 923. All material is copyright ©. No part of this publication may be reproduced without the written consent of the publisher. Subscription rates (Australia only) 6 issues (6 months): $77.50 12 issues (1 year): $145 24 issues (2 years): $270 Online subscription (Worldwide) 6 issues (6 months): $55 12 issues (1 year): $105 24 issues (2 years): $200 For overseas rates, see our website or email silicon<at>siliconchip.com.au * recommended & maximum price only Postal address: PO Box 194, Matraville, NSW 2036. Phone: (02) 9939 3295. ISSN: 1030-2662 Printing and Distribution: 1 Huntingwood Dr, Huntingwood NSW 2148 54 Park St, Sydney NSW 2000 2 Silicon Chip Our TV is around 10 years old now. While the picture still looks fine, its built-in “smart” features are slow and outdated. So a couple of years ago I bought an Amazon Firestick 4K, which plugs into an HDMI port and provides modern smart-TV functionality. While we use the TV’s internal speakers for everyday viewing, for movies and live concerts I have the TV’s optical output connected to my CLASSiC DAC (February-May 2013; siliconchip.au/Series/63), then into an Ultra-LD series amplifier and the Majestic loudspeakers (June 2014; siliconchip.au/Series/275). This setup provides excellent audio quality. Recently, I was watching some live concert recordings through VLC on the Firestick (streamed from a computer over WiFi) and noticed that the audio sounded very muffled. It was like someone had placed a blanket over the speakers, with a clear loss of high-frequency detail. This was odd, because I knew the concert recordings used PCM audio (ie, not digitally compressed), and when I played that same PCM audio directly through the DAC, it sounded fantastic. So it wasn’t the recording; it was something wrong with the playback chain. The audio path in this case is: File Server ▶ WiFi network ▶ Firestick ▶ TV ▶ TOSLINK ▶ DAC ▶ Amplifier ▶ Speakers. I knew the first couple of steps wouldn’t be a problem, nor anything from the TOSLINK output onward. That narrowed the problem down to either the Firestick or the TV. I went through the TV’s audio settings but couldn’t find anything suspicious. On the Firestick, navigating to Settings ▶ Display & Audio ▶ Audio didn’t reveal many options. However, there was one called “Surround Sound”, which was set to “Best available”. The other options were “PCM”, “Dolby Digital Plus” and “Dolby Digital”. I changed this setting to PCM and the problem immediately disappeared. Playing back the same concert video now produced crisp, clear audio, exactly as expected. What surprised me even more is that watching TV shows and movies through the Firestick now also sounds better than it did previously, particularly in terms of dialogue intelligibility. This makes me wonder why “Best available” is the default setting, when it clearly isn’t the best option for systems connected to a proper amplifier and speakers. As far as I can tell, this “best available” mode enables additional audio processing such as compression and EQ that is optimised for small, tinny speakers. While that may help on basic TV audio systems, it actively harms sound quality when used with a decent hi-fi setup. Even more oddly, the Firestick felt the need to adulterate the audio even when the source was already in PCM format. I would have expected it to simply pass the audio through unchanged, but clearly that isn’t what happens. So if you have a decent sound system connected to your TV, it’s worth checking the entire signal chain and not assuming the default settings are optimal. I’ve always found dialogue on this setup a little muffled and simply put up with it, assuming poor mastering was to blame. Now I know better. I wish I had realised this a long time ago as it would have made a lot of the media I’ve viewed a lot more enjoyable. It’s hard enough to understand what’s said in modern movies and TV shows without the audio being artificially muffled! by Nicholas Vinen Subscription Prices, effective 01/06/2026 New Prices Print Print+Online Print Print+Online Print Online 6-month $77.50 $87.50 $95 $105 $115 $55 12-month $145 $165 $180 $200 $220 $105 24-month $270 $305 $335 $370 $410 $200 New Zealand RoW Australia Prices from June 1st, 2026; all prices are listed in Australian dollars (AUD). RoW = Rest of World Australia's electronics magazine siliconchip.com.au 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”. Installing a car GPS tracker yourself Cairns has many cars stolen every year. Rather than install an engine immobiliser that I would find annoying every time I get in the car, I opted to install a GPS tracker. I selected a tracker from AliExpress that uses 4G and was labelled as “global”, using 21 frequency bands. It is called a “GPSCJ730plus 4g”, and cost $55 delivered (currently AliExpress 1005007557451623). It looks just like a standard car relay, complete with a connector plug containing five wires. Two wires are for power, with the positive lead having an inline fuse. A third wire is used for sensing if the accessory power is on, which indicates that someone has started the vehicle. The last two wires are heavier gauge and are connected to a normally closed relay contact inside the device. These can be used (if desired) to disconnect power to the fuel pump, effectively disabling the car. The device has an operating voltage of 9-95V DC, so can be used on a range of vehicles. The device needs a physical SIM card to function. I purchased a $2 SIM from Amaysim and activated it on my computer. I selected the “Pay as You Go” plan. This was a little difficult to find on the Amaysim website, but it is there. $15 was required to be paid to start this plan, which lasts for 365 days. The charges are 12¢ per minute for calls, 12¢ per SMS and 5¢ per 1MB of data. Approximately $4 of the $15 plan is used in setting up the SIM card, mostly from SMS messages sent to my mobile acknowledging the setup parameters. In the six months since then, another $3 has been used up, mostly in data charges for uploading the GPS position information whenever the car is moving. All the setup is done by sending SMS messages to the Amaysim number in the device. The device needs a “Control number” for sending SMS messages under specific conditions, so I used my mobile phone number for this. The device also needs the APN of the mobile phone carrier, which for Amaysim is “yesinternet”. There are numerous other setup options that can be activated, like automatically sending an SMS to the Control number (my mobile) if a certain speed limit is exceeded. The process I followed to set it up is: 1. Activate the SIM online 2. Power up the CJ730 and insert the SIM card 3. Wait for the CJ730 to get GPS location and mobile phone connection (two minutes) The default user & password for the device are admin & 123456 (I changed the password after it was all working OK). Two essential parameters are now required to be entered into the CJ730: 4 Silicon Chip 4. Set the Amaysim APN by sending an SMS containing “apn123456 yesinternet” (without the quotation marks). The response should be “apn ok”. 5. Set the Admin (Control) phone number in CJ730 to my mobile number. Send an SMS containing “admin123456 0419xxxxxx”. The response should be “admin ok”. After that, a few SMS commands can be sent from my mobile, to make sure it is operational: • “status”, the response includes the battery, GPRS, GSM, GPS etc levels • “where”, response = “http://maps.google. ...” with the GPS coordinates • “imei”, response = “86209206xxxxxxx” Once it is all set up, the device uploads its GPS position data to a website and you pay for this data on the Amaysim plan. I don’t believe any top-up of the original $15 will be required in 12 months. It will be necessary to pay another $15 every 365 days to keep the SIM active, though. The data uploaded consists of the date, time, latitude, longitude, speed, direction, accessories on/off and the supply voltage. The device has a motion sensor, so no data is uploaded when stationary. By default, data is uploaded about every 20 seconds while moving to www.gpscj.net, which is in China. This is a free site that stores your data for 160 days. Below is part of a screenshot of the web page. I have found this site works very well. I can access it on my computer or by using an app called “Yuntrack”. It displays details superimposed on Google Maps. You can instead use a free program called “traccar” that runs on a Raspberry Pi. I plan to go down that path in the future. You send an SMS to the device containing the URL and port number to point to the Pi. The device can only send data to one server at a time. It has an internal rechargeable battery, allowing it to send siliconchip.com.au Where Australia’s Electronics Future Comes to Life Explore breakthrough technologies, design solutions and manufacturing advances for electronic products. SMCBA CONFERENCE Engage with Australia’s industry leaders and experts at the SMCBA Electronics Design and Manufacture Conference. Details at www.smcba.asn.au In Association with Supporting Publication Organised by an SMS if the primary power is disconnected. Disconnection of primary power is considered an urgent situation, so as well as sending an SMS, the device also rings my mobile number to get my attention. This was inadvertently tested when I replaced my vehicle battery. I can send an SMS of “555” to the device to operate its internal relay, cutting off the fuel pump if it is wired up that way. The device has smarts that cause it to intermittently operate the relay if the car is travelling above a set speed and fully operate the relay when the speed has reduced. Another SMS text of “666” will drop out the relay, allowing the motor to be started again. The GPS receiver in the device works with three satellite systems: GPS, BeiDou and GLONASS. It reports the number of satellites for each system that it can currently see. I have the device permanently powered on in my 2005 Mazda 2. Its power consumption is 40mA, or approximately 1Ah per day. That isn’t much higher than the normal parasitic drain, so the battery should last about four weeks without driving the car. More commands can be found by searching the web for “cj730 commands” or “cj720 commands”. I have yet to experiment with commands that limit the power consumption when the vehicle is idle. One command will set the device to only send data when the accessory power is on. Other commands cause the device to go to sleep and wake up when its sensor detects vehicle movement. Sid Lonsdale, Cairns, Qld. Transistors should be matched on Vbe, not hfe After reading the article for your new Calliope Amplifier (April 2026; siliconchip.au/Article/20084), I have doubts about the value of matching the input transistors for hfe. Keeping the input transistors matched and minimising the input offset voltage does provide benefits. Not only is the output offset voltage minimised, but mismatch in a longtailed pair increases distortion. Whether one deems the extra effort worthwhile is another matter. The BC558 and BC556B are from the same family of devices, with the 556B selected for higher gain. Selecting the 556B alone will reduce the offset voltage without the added complication of matching by reducing the base current through the base resistors, so swapping the input devices between these two devices (556 and 556B) is not a fair measure of the advantage of matching. As best as I can tell, matching the hfe of the current mirror transistors will have virtually no effect on the input offset voltage. The current balance in the mirror is primarily fixed by the requirements of the input stage and the amplifier after global feedback is applied, and the bases are tied together, so any mismatch will have an insignificant effect on the balance. The parameters of most importance to input offset are the Vbes of the input pair and the mirror resistors, and as far as I know, there is no significant relationship between hfe matching and Vbe matching. The greater impact on input offset from hfe in the input transistors is simply the combined effects of the transistors’ hfe (and hence base current) and the resistances seen at their bases. So, to minimise the impact on the offset, the sensible approach would be to keep those resistances as low as practical and to maximise the current gain of the input pair. 6 Silicon Chip The factors that do significantly impact input offset are the Vbe matches of the input pair (Q7 & Q8) and the current mirror (Q15 & Q16). I did some research and circuit analysis. Data from Philips’ 1977 “Low frequency transistors” data book (SC2 11-77) shows a BC556’s hfe varies from 220 to 480, which would result in a worst-case offset voltage (cause by mismatched bias currents flowing through input bias resistors) of approximately 81mV. The Vbe of a BC556 at 2.5mA varies between 600mV and 750mV, which implies a maximum Vbe mismatch (ie, input offset voltage) of 150mV. The same figures for the BC549 are 580mV and 700mV, so a maximum Vbe mismatch would be about 120mV. An amplifying factor of 1.47 must be applied due to the emitter resistors in the input pair and the current mirror, so the maximum offset error caused by a mismatch in the current mirror Vbe could be over 170mV. These are the absolute worst-case contributions to input offset voltage due to mismatched current mirror and input transistors. These figures are blurred by the fact that the resolution of the graphs I have had to use is not great. Another complicating factor that I cannot account for is the usual distribution of parameters within their published ranges. I don’t claim they are particularly accurate, but they do give a clear indication of the greatest causes of input offsets. Still, the implications are clear: the overall input offset voltage of the complete amplifier could depend as much as four times more on Vbe matching than hfe matching. My advice is to match input transistors for Vbe. Problems caused by mismatched hfe are of minor significance by comparison, and the bad effects can be overcome by simply selecting for higher hfe (eg, using BC556Bs instead of BC558s). Matching only the hfe might appear to provide benefit, but it is quite likely that it coincidentally produced a better match in Vbe as well. Phil Denniss, Darlington, NSW. Comments: we prefer using higher-gain BJTs where possible since, as you point out, it reduces their base current, which can bring certain benefits. While difficult to find now (apparently due to a lack of demand), some Asian fabs still bin BC556Cs and related devices, which should (in theory) give even more benefit than the BC556B. You’re right that Vbe matching is more valuable, and it’s easy to do with a basic DC power supply, a few resistors, a breadboard and a multimeter. All that’s really needed is to apply the same voltage to the base-emitter junction via the same resistance for both transistors (say, 5V & 1kW), measure the voltage across the junction, and select two transistors from a small batch purchased together with almost identical readings. We strongly recommend using transistors from the same batch for pairs in amplifiers (input, current mirror etc). The Vbe variation within a batch is usually on the order of 10-20mV, not the 100mV+ suggested by the data sheets. Feedback on the Internet Radio project After constructing the Internet Radio from the February & March 2026 issues (siliconchip.au/Series/458), I would like to share some of the deviations made from the build article. Initially, I had no audio from the Pi 4. The article suggests using a 3.5mm stereo jack plug, but the Pi 4B has a Australia's electronics magazine siliconchip.com.au Introducing ATEM Mini The compact television studio that lets you create presentation videos and live streams! Blackmagic Design is a leader in video for the television industry, and now you can create your own streaming videos with ATEM Mini. Simply connect HDMI cameras, computers or even microphones. Then push the buttons on the panel to switch video sources just like a professional broadcaster! You can even add titles, picture in picture overlays and mix audio! Then live stream to Zoom, Teams or YouTube! Live Stream Training and Conferences Create Training and Educational Videos Monitor all Video Inputs! ATEM Mini’s includes everything you need. All the buttons are positioned on With so many cameras, computers and effects, things can get busy fast! The All models have built in hardware streaming engine for live streaming via its ethernet connection. This means you can live stream to YouTube, Facebook and Teams in much better quality and with perfectly smooth motion. You can even connect a hard disk or flash storage to the USB connection and record your stream for upload later! the front panel so it’s very easy to learn. There are 4 HDMI video inputs for ATEM Mini features a “multi-view” that lets you see all cameras, titles and program, connecting cameras and computers, plus a USB output that looks like a webcam plus streaming and recording status all on a single TV or monitor. There are even so you can connect to Zoom or Skype. ATEM Software Control for Mac and PC tally indicators to show when a camera is on air! Only ATEM Mini is a true is also included, which allows access to more advanced “broadcast” features! professional television studio in a small compact design! Use Professional Video Effects ATEM Mini is really a professional broadcast switcher used by television stations. This means it has professional effects such as a DVE for picture in picture effects commonly used for commentating over a computer slide show. There are titles for presenter names, wipe effects for transitioning between sources and a green screen keyer for replacing backgrounds with graphics. www.blackmagicdesign.com/au ATEM Mini Pro..........$469 ATEM Software Control..........FREE Learn More! four-pole 3.5mm audio jack socket, as the audio jack also has the composite video signal. Swapping the 3.5mm stereo jack plug to a TRRS type fixed the sound. Since the 3D-printed Internet Radio case has ample room, I added a cooling fan on the Pi 4, although I am not sure if it is necessary. The rest of the construction and software installation went smoothly. After creating some radio station text files and saving them to the Pi desktop as per the instructions, I decided it would be convenient if the Internet Radio automatically opened my favourite radio station on start-up. The radio station text file on my desktop that I want to run at startup is named “rock.m3u”. Raspberry Pi software has multiple different approaches to do this. After some research, I decided to create a .desktop file and run it in the autostart folder. The first job is to create a .config/autostart folder. To do this, I opened a Terminal window and typed: mkdir -p ~/.config/autostart “~” expands to your home directory, which in my case is /home/pi. Be sure to add the full stop in front of config. I then created a VLC autostart file that will be in the .config/autostart folder. In the same Terminal window, I typed: nano ~/.config/autostart/vlc_start.desktop This opens the nano text editor. In the text editor, type: [Desktop Entry] Type=Application Exec=vlc /home/pi/Desktop/rock.m3u Replace “pi” if your username is different and replace rock.m3u with the radio station text file of your choice. Press Ctrl+x then y to save the file and then press Enter to close the text editor. By default, the Pi File Manager doesn’t show the .config folder (the “.” at the start means ‘hidden’). To view the folder, go to the “View” menu and click Show Hidden in the File Manager. To change which radio station autostarts, use the File Manager, navigate to .config/autostart and edit the vlc_ start.desktop file by right-clicking on the file and selecting “Open with text editor”. With autostart configured, I then tinkered with VLC. The default VLC window is quite small. To permanently resize and save it as the default size, you need to close down VLC via the menu “Media” and “Quit”. Once you’re happy with VLC window size, you can close VLC with the X or just shut down the Pi to retain your new default window size. VLC has some audio visualisations that are accessed via the Audio → Visualization menu option. Unfortunately, these are not saved as a preference, so they need to be selected every time VLC is opened. To permanently set an audio visualisation, use the Tools → Preferences → Audio menu item. Under Effects Visualization, select either 3D OpenGL Spectrum Visualization or Visualizer Filter. Those are the only two that can be set as the default. Mathew Prentis, Port Augusta, SA. Phil Prosser comments: It is great to see people doing their thing with the design. I love that Mathew has extended the installation and I will be updating the radio in the shed SC with his advice right away. 8 Silicon Chip Australia's electronics magazine siliconchip.com.au $20,000 INSTANT TAX BREAK WOW! IT 'S FOR THE BUSINESS! YEAH, SURE..... UP TO u30% OFF BEGINS MAY 11, ENDS JUNE 30 HOLE SAW SET - HSS • 11 piece set • M42 Bi-Metal high speed steel • 19, 22, 25, 32, 35, 38, 44, 51, 57, 64, 76mm • Includes 3/8” & 1/2” arbor shank & pilot drill METRIC CARBIDE TIPPED HOLE SAW SET • 8 piece set • 16, 20, 22, 25, 28, 30, 32, 35mm • Carbide teeth Ø10mm shank END MILL SET • Manufactured from micro fine solid carbide • 4-flute • Plain shank SLOT DRILL & END MILL SETS • Manufactured from HSS M2 • Plain shank • Includes 5 end mills & 5 slot drills 10 PIECE METRIC 6, 10, 12, 16, 20mm INDUSTRIAL DRILL SETS • Precision ground flutes • M35 Grade HSS including 5% Cobalt • 135º split point centre 25 PIECE METRIC 1-13mm (x0.5mm) $85.00 (M3301) $85.00 (D1272) 10 PIECE IMPERIAL 1/4", 3/8", 1/2", 5/8", 3/4" 29 PIECE IMPERIAL 1/16"- 1/2" (x1/64") $85.00 (M3321) $95.00 (D1282) • Precision ground flutes • HSS M2 bright finish $109.00 (D102) $105.00 (D108) 4 PIECE METRIC 6, 8, 10, 12mm 51 PIECE METRIC 1-6mm (x0.1mm) $149.00 (M337) $99.00 (D1285) HUGE DRILLING RANGE COUNTERBORE DRILL SET • 6 piece set • M35 Grade HSS including 5% Cobalt • TiAlN coating • Plain shank • Range: M3, M4, M5, M6, M8, M10 $115.00 (M3381) RADIUS CUTTER SET • 6 piece set • M35 Grade HSS including 5% Cobalt • Radius sizes: 4, 5, 6, 8mm • Shank sizes: 12, 16, 16, 20mm • HSS M35 5% Cobalt bright finish • Plain shank • 2 flute • Shank sizes: 6, 10, 10, 12, 16mm 5 PIECE METRIC 6, 8, 10, 12, 16mm $115.00 (M3391) METRIC PRECISION 2MT MORSE TAPER DRILL SET • M35 Grade HSS including 5% Cobalt • Precision ground flutes • 135º point for tougher materials • 10pc in sizes - Ø14.5, 15, 16, 17, 18, 19, 20, 21, 22, 23mm • 2MT shank with tang • Bronze finish $389.00 (D122) BALL NOSE CUTTER SET $85.00 (M3355) INDUSTRIAL COUNTERSINK SETS • 6 piece • 4 piece • M35 grade HSS including 5% Cobalt • Cross hole type • Range: Ø2-5, 5-10, 10-15, 15-20mm • Angle 90° • M35 grade HSS including 5% Cobalt • 3 flute type • Ranges: Ø1.5-6.3, 6.3-8.3, 8.3-10.4, 10.4-12.4, 12.4-16.5, 16.5-20.5mm $95.00 (D1051) $105.00 (D1061) REDUCED SHANK DRILL SETS • M35 Grade HSS including 5% Cobalt • 135º cutting point METRIC RANGE - 10PC 14.5-23mm 2MT (Morse Taper) IMPERIAL RANGE - 8PC 9/16"- 1" Shank Ø1/2" $239.00 (D1181) $195.00 (D1171) View and purchase these items online: www.machineryhouse.com.au/SIC2605 OVER 4,000 PRODUCTS ON SALE! Y Y F F O E EO NEW RELEASE IP-54 RATED PROTECTION AGAINST INGRESS OF DUST & WATER Q170 RECHARGEABLE DIGITAL CALIPERS • Magnetic charging in under 3 hours • Long battery life with low battery alert Q177 • Metric, Imperial and Fraction measuring modes • 4-Way measuring • Auto-off function 3 sizes available from $75 E L SA IP-67 RATED WATERPROOF UP TO 1M FOR 30 MINUTES 3 sizes available from $109 DIGITAL OUTSIDE MICROMETER 10-1245 METRIC OUTSIDE MICROMETERS 6 PIECE SET - 20-118 • 25 - 50mm range • Resolution 0.01mm • 0.004mm accuracy • Ratchet stop for exact repetitive readings • 150-300mm Range • Easy adjustment for recalibrating with ratchet spanner • Carbide tipped anvils, Flatness 0.0008mm • Micro-fine clear graduations on satin chrome finish $50.00 (Q105) $737.00 (Q118) METRIC OUTSIDE MICROMETER 10-105 • 0-25mm/0-1” range • ±0.001mm accuracy • Large LCD for easy reading • Carbide measuring faces PROTECTED TO IP65 SPECIFICATIONS $187.00 (Q1245) HUGE MEASURING RANGE DIGITAL INDICATOR 34-2205 • 12.5mm/0.5" range • Zero setting at any position • Metric/Imperial system • 55mm dial face • Data output interface $99.00 (Q2205) • BSS:939 Grade "B" • Hardened spring steel blade • True right angle inside and outside 23.00 (Q644) 32.00 (Q645) $ 150mm 36.00 (Q646) $ 225mm 55.00 (Q647) 100mm $ $ INDICATOR - LONG RANGE DIAL - 34-510 INSIDE MICROMETERS ROD TYPE 23-148 • 0-50mm capacity • Steel Body with black & white steel bezel • Jeweled movement • Steel back with lug • Lifting screw and revolution counter $177.00 (Q510) • 25-50mm capacity • The anvil and spindle are hardened & ground • Interchangeable rod is marked with the range SHOCK PROOF MOVEMENT $199.00 (Q148) MEASURING BOX SET 70-605 IP-65 DIGITAL PROTRACTOR 35-2041 DIAL THICKNESS GAUGE 34-506 $79.00 (Q605) $139.00 (Q2041) $99.00 (Q506) • CNC machined for high accuracy • Ground measuring face • Black anodized coating for a protective anti rust coating • Precision laser engraved markings • Full 360° range (90° x 4) • 0-10mm • 0.05° resolution • Standard steel contact point & anvil • ±0.15° accuracy Ø10mm diameter • IP-65 waterproof rating LARGE BACK LIGHT LCD • Compact design • Magnetic base SCREEN • Ideally suited for quick inspection Specifications & Prices are subject to change without notification. All prices include GST and valid until 30-06-26 COMPETITIVE FREIGHT RATES DELIVERED TO YOUR DOOR MORE THAN 4000 PRODUCTS ON SALE ONLINE, INSTORE OR CALL SYDNEY MELBOURNE BRISBANE PERTH ADELAIDE (02) 9890 9111 (03) 9212 4422 (07) 3715 2200 (08) 9373 9999 (08) 9373 9969 04_SIC_250526 75mm • Large measuring range 1.6mm - +1.6mm • Measurement in both directions steel ball end • Gradient of 12 degrees on dial face • Mono enclosure for air tight seal $168.00 (Q218) PRECISION STEEL SQUARES Size DIAL TEST INDICATOR METRIC - 34-218 Part 2 by Dr David Maddison, VK3DSM Analog Computers A recreation of Charles Babbage’s Difference Engine at the Computer History Museum in Mountain View, California. Source: Jitze Couperus – www.flickr.com/photos/jitze1942/4304353299/in/album-72157623284316506 Analog computers are making a comeback because they are well-suited to neural network processing. We’ll cover some of the theory behind that, then look at some of the new analog computers that are being developed. T he extremely high speed of modern digital computers and the relative ease of programming compared to analog computers accelerated the decline of the latter in the 1970s. However, analog computing is experiencing a major resurgence, albeit in a somewhat different form from traditional analog computers. While modern analog computers still rely on analog signals (voltages, currents, resistances or even light) to perform calculations, their design bears little resemblance to the analog computers of the mid-20th century. 12 Silicon Chip Instead of racks of discrete op amps wired via patch panels and potentiometers, today’s implementations are built on silicon chips using memristors, floating-gate transistors, switched-capacitor arrays or photonic waveguides. They are often digitally programmable. They are also far smaller than the analog computers of yore, and more precise, being targeted at specific tasks like AI matrix-vector multiplications or AI inferencing (using a pre-trained AI model to produce an output). That makes it possible to use them in Australia's electronics magazine smartphones or embedded devices, alongside existing digital processors. The main benefit of modern AI analog computing is high energy efficiency and high processing speed for specific AI tasks, like inferencing and matrix-vector multiplication. These functions are implemented in AI accelerator chips. One common feature of AI accelerators (hardware optimised for AI tasks) is their massively parallel nature. By using many parallel units operating at a high speed, they can perform thousands of calculations simultaneously siliconchip.com.au and complete billions per second. They are designed for linear alegbra and the tensor mathematics used in AI applications. Digital AI accelerator chips include GPUs (graphics processing units, widely used for training), TPUs (Google’s tensor processing units for neural networks), NPUs (neural processing units for on-device AI), ASICs (application-specific integrated circuits for specific AI functions) or FPGAs (field programmable gate arrays, reconfigurable chips for various tasks). On the other hand, modern analog accelerator AI chips generally fall into the following categories: • Resistive/electrical AIMC (analog in-memory computing). • Neuromorphic (analog, digital, or mixed; when analog, they often overlap with AIMC). • Photonic analog AIMC is ideal for matrix-vector multiplications directly in memory arrays using resistive devices like PCM (phase change memory), RRAM (resistive random-access memory) or flash memory. Examples include IBM’s PCM-based chips, Mythic’s M1076 (flash analog), EnCharge AI’s capacitors and Peking University’s RRAM prototype. These focus on efficient deep learning inference. We will describe in-memory computing later. Neuromorphic chips emphasise brain-inspired designs, often with spiking neural networks (SNNs), event-driven/asynchronous processing and sparsity. These can be analog, digital or hybrid. Sparse models and SNNs will be described shortly. Photonic chips use light-based processing, like Lightmatter and Microsoft’s AI chip. They have the potential to use even less power than the other types of analog AI chips. Features of modern analog AI Modern analog AI computing has the following characteristics. that have been trained using digital AI. Currently analog AI computers are mostly used only for energy-­efficient inferencing, not training. However, research is underway to develop analog AI training models, and it has been experimentally demonstrated. Analog computing suits specific AI workloads For specific AI workloads, especially inferencing, in-memory matrix operations and ‘sparse’ models such as ‘mixture of experts’ (MoE) or spiking networks, analog computing offers dramatic efficiency gains. In sparse models, a significant portion (often the majority) of connections (weights) or activations are intentionally set to zero or left inactive to boost efficiency in terms of memory, computation and energy without drastically harming performance. Dense models, by contrast, maintain near-­ complete connectivity, which can capture more complex patterns but at a higher resource cost. Sparsity is especially beneficial in contexts like analog AI, where it aligns with the nature of the analog hardware. Analog AI has been suggested as being highly suitable for MoE models. These are neural network architectures in which a large model is split into many smaller sub-networks. An ‘expert’ is a small specialised part of the overall neural network model. A lightweight routing network dynamically decides, for each input, which expert (or experts) to activate for a particular problem. So, instead of running the entire massive model for every input, only a small subset of experts is used. This makes the system much more efficient than traditional dense models (where everything runs every time). In analog implementations of MoEs, the unused experts can be completely powered down, leading to even greater Analog AI is not a complete replacement for digital AI Analog (or analog-inspired neuromorphic/mixed-signal) computing cannot yet fully replace a digital GPU cluster like xAI’s Colossus because it is not a drop-in replacement for general large-scale AI training. Due to present limitations of analog AI, for inferencing, analog AI uses models Fig.37: the model of experts (MoE) concept. siliconchip.com.au Australia's electronics magazine power savings. These analog MoE systems can be fully analog or hybrid (for example, with digital routing and analog experts). In reference to Fig.37, a gating network uses weights to adjust each expert’s contribution to the final answer. The gating network learns from experience and decides which is the most appropriate expert to send the data to. By sending the data to the ‘best’ expert, the processing is more effective. This is more effective than just using a single expert. Analog MoE models have been implemented by IBM using phasechange memory (PCM) crossbars for the expert model weighting; a small analog router selects 2-4 experts per input for vision tasks. Mythic’s analog matrix processors use flash memory to carry the expert weighting. Lightmatter is using optical routing to switch light paths to different refractive expert layers. A spiking neural network (SNN), or spiking network, is a type of neural network that mimics how biological brains work more closely than standard artificial neural networks. With analog AI, the reported possible savings in power consumption are 10× to 100× (or even as much as 1000× for analog-optimised tasks) for inference applications compared to digital AI. Edge computing Because analog AI chips consume so little power, they allow advanced AI models like large language models (LLMs) to mimic the human brain’s efficiency and to run on small ‘edge’ devices like smartphones, autonomous robots, UAVs, wearables etc without the need for a ‘cloud’ connection. Energy efficiency Currently, xAI’s Colossus AI training supercluster in Memphis, Tennessee is June 2026  13 recognised as the world’s most powerful AI compute facility with around 200,000 NVIDIA H100/H200 GPU equivalents. It consumes 280-300MW peak power. Meta’s clusters and Microsoft/OpenAI facilities are in the 100-200MW range. US data centre power consumption today, a large portion of which is due to AI, is around 40GW and is projected to reach 78-123GW by 2035. Even with the low cost of electricity in the USA compared to Australia, with such high power consumption, the running costs are significant. In a hypothetical 300MW data centre with an inference-heavy workload, partial analog adoption could reduce power consumption by 50-90% (eg, 30-150MW). For full training, that is not as feasible today, but perhaps a 20-50% reduction is possible with hybrid computers. Matrix operations A fundamental operation in neural networks is matrix-vector multiplication (often matrix multiplication in layered neural networks). Analog circuits can perform this extremely efficiently and in parallel by exploiting natural physical laws such as Ohm’s law (V = IR) and Kirchhoff’s current law (the total current entering a junction must equal the total current leaving), which are intrinsic to electronic circuits. Inputs are applied as voltages across a resistive ‘crossbar array’ (typically memristors, resistive RAM [RRAM] or resistor grids), and the resulting currents at each column naturally sum to produce the dot-product outputs, which represent the matrix multiplication result instantaneously (limited only by circuit settling time, typically nanoseconds to microseconds). A dot product is a simple mathematical operation involving multiplying two vectors to form a single vector. A basic 3×3 crossbar array is shown in Fig.38. The inputs are voltages V1, V2 & V3 on the rows representing the input vector (eg, pixel brightness values from an image being analysed). The matrix weights (obtained via prior training representing the importance of an association) are the conductances (Iij = 1 / Rij) at each crosspoint of the matrix elements (higher conductance = more weight). The outputs are currents I1, I2, I 3 down the columns, which are 14 Silicon Chip Fig.38: a 3×3 resistive crossbar array, the core of many modern analog AI accelerators (eg, the memristor or resistor grids in Mythic or IBM chips, respectively). It performs vectormatrix multiplication instantly via physics. automatically the matrix-vector product with no mathematical operations used; just Ohm’s and Kirchhoff’s laws doing the work in parallel. By Ohm’s law, the current through each crosspoint Iij = Vi × Iij. By Kirchhoff’s current law, the output current Ik = ∑(Vi × Iij) for j = k. Thus, Ik is the dot product of the input vector V with column k of conductance matrix G. Matrix multiplication is the mathematical powerhouse behind nearly every modern artificial intelligence system, especially deep neural networks (the foundation of models like GPT, BERT, Stable Diffusion and computer vision). Matrix operations are how neural networks ‘think’. A neural network layer takes an input vector, such as numbers representing pixel values in an image or words and produces an output vector. This transformation is almost always a matrix-vector multiplication: output = weights_matrix × input_vector + bias. The weights matrix contains millions (or billions) of learned parameters that encode what the network has ‘learned’ during training. Each layer performs this operation, stacking many layers to create complex representations (eg, recognising a cat from pixels). The bias is an extra adjustable parameter added to each neuron (or node) in a layer, alongside the weighted sum of inputs. It provides a ‘starting point’ for any opinion the AI might formulate, and this is the basis Australia's electronics magazine for any political or other bias that AI engines are deemed to have. In simplified terms, a neural network layer can be thought of as a team of experts (neurons). Each expert gives an opinion weighted by their expertise (matrix weighting values) on every piece of input data. Matrix operations are the bottleneck for speed and energy use in AI training and inference. Around 90-99% of computation time and energy in deep learning is spent on matrix multiplications. That’s why in digital AI, GPUs (with thousands of cores for parallel maths) and specialised chips (eg, TPUs, NVIDIA H100s) are designed to accelerate matrix operations. Analog AI computing is ideal for matrix multiplication because it utilises physical laws, as explained above, allowing it to perform billions of multiplies/adds in parallel instantly and with little power consumption. After the matrix operation, the resulting current (or voltage after conversion) can then be fed to an analog-­ to-digital converter (ADC) for further processing, in the case of a hybrid computer, or it can pass directly to the next layer of the neural network. Such analog circuits are ideal for real-time audio or video processing since they lack the delays inherent in digital conversion and processing. In-memory computing means the computation occurs directly within memory cells, using phase-change memory or RRAM, eliminating the need to shuffle data out of memory for processing. The challenges of analog AI processing This is not without its challenges, which include: Precision and noise Analog signals are susceptible to noise and non-repeatability due to the variable values within the tolerance of electronic components. Thus, a calculation won’t necessarily give the same result every time, although it will be close enough for some purposes. Digital computers in contrast are precise and repeatable. Programmability Software development tools for CUDA (NVIDIA’s programming model) have been in development for decades. siliconchip.com.au Whole new software suites have to be developed for analog computing; companies like IBM have open-sourced toolkits such as aihwkit (https://github. com/IBM/aihwkit) to ease the transition. close to a large source of power, and the electrical grid doesn’t have to be extended to accommodate it. Circuits for modern analog AI • Optical components in photonic chips; microLED arrays for light sources to represent the input vector or neural network activations; spatial light modulators to store neural network weights and perform multiplication with incoming light; photo-­ detector arrays to convert light signals back into electrical signals for further processing; and photonic waveguides as conduits for light that steer and manipulate it to perform mathematical operations. • Phase change memory – described in the IBM entry later. • Resistive RAM (RRAM) – the practical implementation of a technology that uses memristors to store information. • Switched-capacitor arrays – described in the EnCharge entry later. We will now look at some modern experimental or commercial chips analog computing chips and systems. Key electronic components in modern analog AI chips are: Hybrid necessity • Analog-to-digital converters AI systems involve two main (ADCs) and digital-to-analog convertphases: training and inferencing. ers (DACs) to interface between the Training is the computationally digital and analog worlds. expensive phase where a model (gen• Calibration circuits, to mitigate erative or discriminative) learns pat- the natural variability in analog comterns from massive data, often requir- ponents, noise and thermal drift. ing hyperscale data centres. Infer• Field programmable analog arrays encing is the computationally lighter (FPAAs) – more on these later. phase, where the trained model is used • Ferroelectric devices (emerging) – on new inputs to produce outputs, these use ferroelectric materials where eg, answering questions, generating the polarisation state can be switched images, or classifying objects. to modulate conductance or capacThe two major model categories are itance, enabling non-volatile analog generative models, which create new weight storage. content, and discriminative models, • Floating-gate transistors – stanwhich make decisions or classifica- dard transistors with a ‘floating’ gate tions. that traps a variable amount of charge, While analog computing excels at storing information. This charge modACCEL siliconchip.au/link/aca7 edge AI inferencing (eg, on smart- ulates the transistor’s conductance, The All-analog Chip Combinphones or IoT devices) and lower-­ allowing precise analog storage of neu- ing Electronic and Light computpower tasks, it is not yet suitable for ral network weights. They are used in ing (ACCEL) is an experimental hyperscale training. flash memory, such as by Mythic. photonic-­ electronic chip from ChiA full switch to 100% analog is • Gain cells (emerging) – a type of na’s Tsinghua University, announced speculative; a hybrid digital/analog memory with two or three transistors in 2023. It is claimed to classify highapproach is a possible short-term path. and possibly a capacitor that can store resolution images over 3000 times In a hybrid system, a digital processor a variable amount of charge represent- faster and with up to four million times handles logic and control, while ana- ing the stored information. less energy than state-of-the-art GPUs log accelerators do the maths. • Memristors – resistors with a like NVIDIA’s A100. memory. Their resistance depends on An input image is processed in the The AI scaling crisis the amount of charge that has flowed optical domain for feature extraction; This is the realisation that simply through them in the past, so the resis- the resulting light field strikes a phoadding more data, computing power tance can be adjusted to the desired todiode array, converting it to phoand electrical energy to AI models value. tocurrents that feed directly into an will hit physical, economic and practical limits. AI-driven data centres use Ternary computing so much electricity that the availabilTraditional digital computers use ‘binary’, with memory cells and logic lines ity of electricity in specific regions (‘bits’) being in one of two states (0 or 1). In contrast, analog computers operis the limiting factor, not chips. As ate with a continuum of values. a result, companies like Amazon are An intermediate concept is ternary logic, which uses three possible states purchasing nuclear-powered data for a bit or ‘trit’, typically -1, 0, +1. They could be encoded electrically as (for centres. example) 0V, half supply and full supply, or even active low, high-impedance There is also the problem of ‘plaand active high. teauing intelligence’. Simply increasOne of the earliest examples was Thomas Fowler’s mechanical ternary caling the AI model size does not result culator in 1840. The first electronic ternary computer, Setun, was built in the in much increase in reasoning ability. Soviet Union in 1958 by Nikolay Brusentsov. Fifty units were produced until Also, the AI industry is running out of 1965. It was a remarkably balanced and efficient design that unfortunately high-quality data to train models on. lost out to the mass production of binary systems. Analog AI computing offers a possible Interest in ternary computing largely faded in the West due to the domisolution to these problems. nance of binary hardware, but like analog computing, it is attracting renewed We’ve already discussed how research interest today for much the same reasons as analog, primarily for analog AI techniques overcome the its theoretical energy efficiency (fewer state transitions per information unit) power consumption problem. Another and potential advantages in certain algorithms. advantage of this is that decentralisaHowever, practical AI applications remain experimental and far less develtion becomes possible; it will no lonoped than analog or neuromorphic approaches. ger necessary for data centres to be siliconchip.com.au Australia's electronics magazine June 2026  15 Fig.39: an example of emulating convolutional neural network (CNN) layers in the optical domain using diffractive layers. Original source: www.mdpi. com/1424-8220/23/12/5749 electronic analog computing unit for final classification, all without the need for an ADC. The front-end uses diffractive optical analog computing. An input optical image shines through a series of engineered diffractive layers. Each layer causes light waves to interfere and diffract in a way that naturally performs linear transformations (dot products & convolutions). The final light pattern at the output plane encodes the result, with no active electronics involved. It is completely passive (like a lens), ultra-fast (limited only by the speed of light), and consumes almost zero energy in the optical part. The resulting light field hits a photodiode array, converting it to analog electrical currents that feed into the electronic analog computing unit for final classification, keeping the entire pipeline analog end-to-end. Its main advantage is that the diffractive part handles the bulk of the matrix multiplications passively at light speed. The ACCEL chip mimics convolutional neural network (CNN) operations of a digital or modern analog computer in its optical front-end. This diffractive optical neural network (DONN) performs feature extraction equivalent to convolutional layers in a CNN, while the electronic analog backend handles non-linear classification – see Fig.39. ACCEL is currently specialised for vision tasks (static images), but could be scaled to other linear operations. Anabrid https://anabrid.com Anabrid has several analog computing projects. In 2024, it produced their 16 Silicon Chip ABRGPX1 Anabrid General Purpose Experimental Test Chip analog multiplier (see Fig.40). It is a test chip for the upcoming commercial offerings, a hybrid digital/analog chip that allows analog capabilities to be integrated with digital systems. Anabrid are also the makers of The Analog Thing, the open source analog computer mentioned last month. lucidac is a reconfigurable analog/ digital hybrid computer intended for early adopters and educational purposes. It has suggested uses in robotics, speech recognition and automotive electronics – see Figs.41 & 42. redac (Fig.43) is said to be the world’s first reconfigurable analog supercomputer. It has six modular clusters with 432 multipliers, 864 integrators, 1728 summation lanes, 124,416 switching elements and 3456 scaling elements. It has digital interfaces for programming and control, using the Python and Jupyter languages. Its purpose is to solve complex ordinary and partial differential equations, solve optimisation problems and act as a testbed for unconventional computing. It is claimed to be significantly faster than conventional computers for certain problems and can provide real-time solutions that are difficult to achieve with conventional computers. Uses proposed for anabrid products are real-time flight wing adjustments, motion control in automation and energy-efficient supercomputers. Aspinity www.aspinity.com Aspinity produces ultra-low-power analog machine learning (analogML) chips for always-on edge AI applications. It claims to have produced the Australia's electronics magazine Fig.40: the Anabrid ABRGPX1 chip ‘floorplan’. Source: https://anabrid. dev/about/news/2024-09-19-tc01press-release world’s first fully analog machine learning chip, the AML100, from 2022. Its technology processes raw analog sensor data in the analog domain, detecting relevant events with nearly no power consumption, waking digital processors only when needed, thus achieving a 10-20× battery life extension. The chip’s current draw is only about 20-100µA. Applications include security monitoring of a parked vehicle using up to four sensors; monitoring of smart homes for events such as glass breaking, smoke, carbon monoxide, leaks, intruders, a baby crying etc; or waking IoT devices for voice/keyword, vibration, or other movements; all with low power consumption. Fig.44 compares conventional Fig.43: the redac analog supercomputer, which is used to solve differential equations. siliconchip.com.au Fig.41: the lucidac software editor. Source: https://anabrid.com/lucidac Fig.42: the lucidac analog/digital hybrid computer. Source: https:// anabrid.com/lucidac monitoring vs monitoring with the AML100 analog processor. analog AI inference is used to identify only relevant data to pass on to the digital processor. The avoidance of unnecessary analog-to-digital conversion saves a lot of power. Aspirare Semi www.aspirare.io This Canadian company has developed a range of hybrid neuromorphic AI accelerators using analog compute cores and digital components. Their Gen 1, Gen2 and Edge models and are commercially available. Blumind https://blumind.ai Another Canadian company that has developed all-analog neuromorphic processors such as the BM110 for low-power edge tasks like always-on voice (eg, keyword detection) and sensor processing. The BM110 is in volume production, while the BM210, intended for video image classification, is scheduled for volume production. BrainScaleS-2 siliconchip.au/link/aca8 BrainScaleS-2 from Heidelberg University is part of the EU Human Brain Project. It is a mixed-signal/ analog-emulated accelerated neuromorphic system using analog circuits to emulate neuron/synapse dynamics up to 10,000× faster than biology, with digital connectivity. It is ideal for large-scale spiking neural network simulations. DYNAP-SE2 siliconchip.au/link/aca9 DYNAP-SE2, developed by the Fig.44: a comparison of the AML100 senses (bottom) with conventional methods (top). With the AML100, unnecessary analog-to-digital conversion is avoided. Source: www.aspinity.com/aml100 siliconchip.com.au Australia's electronics magazine Institute of Neuroinformatics (INI) at the University of Zurich and ETH Zurich, is an experimental mixed-­ signal neuromorphic chip designed for real-time, low-power spiking neural network processing. It features 1024 analog neurons, each with up to 64 programmable synapses, combined with digital event routing for flexible connectivity. The chip is reconfigurable, supports adaptive and learning behaviours (eg, spike-timing-dependent plasticity) and operates at extremely low power, typically below 1mW for many workloads. SynSense markets and sells related development kits or boards featuring the DYNAP-SE2. EnCharge www.enchargeai.com EnCharge announced the EN100 commercial accelerator chip in 2025. It uses analog in-memory computing for high-performance AI applications. In the EN100, neural network weights are stored in digital SRAM, while computation is performed by charging capacitors to various levels and then connecting them together to redistribute the charge. This mechanism avoids the noise problems of other analog AI designs. The M.2 laptop chips can deliver >200 TOPS (trillions of operations per second) using just 8.25W of power. There is a PCIe card version of the processor for AI workstations that delivers >7.4 PetaOPS of capacity (1000 trillion operations per second). Compared to other forms of in-­ memory computing (see Fig.45 overleaf), EnCharge claims a superior signal-­ to-noise ratio, compatibility with standard CMOS 4nm technology June 2026  17 nodes, broad support for existing AI models and scalable technology. IBM siliconchip.au/link/acaa IBM is using phase-change memory (PCM) devices for analog in-­ memory computing (AIMC). These are nanoscale resistive elements that store neural network weights as varying resistance states and perform matrix-vector operations in-place. Prototypes include multi-core chips with tens of millions of PCM cells, achieving high accuracy on tasks like speech recognition and image classification while using far less power than digital equivalents. In PCM, an electrical pulse is applied to a material, which causes heating and changes its conductance by switching the material between its amorphous (glass-like) and crystalline phases. A small pulse results in there being more crystalline material and lower resistance. A large pulse results in The resistance value corresponds to a neural network weight. The PCMs are arranged in a crossbar configuration – see Fig.46. This enables analog matrix-vector multiplication in a single-­step as explained before. Fig.46: IBM’s PCM crossbar configuration that allows matrixvector multiplication in one step. Source: https://research.ibm.com/ blog/the-hardware-behind-analog-ai more amporphous material and more resistance. Between pure crystalline and pure amorphous phases, there is a mixture of both, representing a continuum of resistance values between 0 and 1. Imec www.imec-int.com/en They produced the experimental AnIA (Analog Inference Accelerator) chip in 2020, which uses analog in-memory compute (AiMC) architecture. The AnIA has reached a high efficiency of 2900 TOPS per watt for vector matrix multiplications. The technology is intended for pattern recognition with tiny sensors and other edge devices. Lightmatter https://lightmatter.co Lightmatter makes the Envise photonic analog AI accelerator chip, which is in a late prototype stage. It is a photonic chip like ACCEL, but while ACCEL relies on diffractive optics, Envise relies on ‘Mach-Zehnder inter- Fig.45: EnCharge’s comparison of traditional AI accelerators, other in-memory computing (IMC) and their own IMC model. NVM is non-volatile memory, MAC is memory and compute and SNR is signal-to-noise ratio. Source: www. enchargeai.com/technology 18 Silicon Chip Australia's electronics magazine siliconchip.com.au ferometers’ to perform light manipulation. Also, ACCEL is intended for research and is not a commercial product. Envise has a hybrid design, with photonic circuits handling the heavy analog computation and digital silicon parts managing control, memory (eg, SRAM) and digital tasks. Lightmatter has announced a multi-chip package with six dies, 50 billion transistors, and one million photonic components using 3D stacking. This chip achieves 65.5 trillion operations per second using just 78W electrical and 1.6W of optical power. Microsoft siliconchip.au/link/acab Microsoft produced an experimental Analog Optical Computer (AOC) in 2025. Like the products from ACCEL and Lightmatter, it is photonic, although it is not a chip but built with discrete components. It aims for a 100× improvement in energy efficiency for large language models (LLMs). The AOC combines 3D optics (lenses, micro-LED arrays) with analog electronics (eg, CMOS sensors from smartphone cameras) to perform computations directly with continuous light intensities, bypassing binary digital conversions and the von Neumann bottleneck. It excels at massively parallel vector-­ matrix multiplications (core to neural networks) and iterative operations, using physical properties of light for addition/multiplication via interference and detection. Nonlinearities are handled electronically, making it a hybrid analog-optical design that runs at room temperature with consumer-­ grade parts. Fig.47 shows a simplified view of the vector-matrix multiplication unit in the foreground. This consists of a linear array of micro-LEDs, a 2D modulator array (using display projectors), and a linear array of silicon sensors. Fig.48 shows the actual computer. For more information, see the video at https://youtu.be/cswAkdU_6yk Fig.47: a simplified diagram of Microsoft’s AOC setup. Source: www.microsoft. com/en-us/research/project/aoc Fig.48: Microsoft’s AOC computer. Source: https://news.microsoft.com/source/ features/innovation/microsoft-analog-optical-computer-cracks-two-practicalproblems-shows-ai-promise/ An analog computer kit from 1961 An article from Popular Electronics describes two simple analog computer kits, one from Edmund Scientific (Figs.49 & 50) and one from General Electric, both available in 1961. You can read it at siliconchip.au/link/acah The Edmund Scientific computer, based on a voltage divider circuit using three potentiometers, is described in more detail at siliconchip.au/link/acai That page describes how to build your own, with modern components; there is even a file to download for the front panel and potentiometer discs. Mythic https://mythic.ai Mythic produces the M1076 Analog Matrix Processor (Fig.51), a single-­ package analog AI accelerator. It is designed primarily for edge AI inference (running trained neural networks efficiently on devices like cameras, drones, robots or servers with low power consumption). Figs.49 & 50: an original Edmund Scientific analog computer kit and matching circuit from the 1960s. Source: www.servomagazine. com/magazine/article/ alternative-computingmodels-part-3-electronicanalog-computing siliconchip.com.au Australia's electronics magazine June 2026  19 Fig.51: the Mythic M1076 on an M.2 card, the same format used for many modern plug-in solid-state storage drives. Source: https://mythic.ai/products/ mm1076-m-2-m-key-card Fig.52: the operation of a Mythic AI processor. X and Y are row and column addresses. Original source: https://mythic.ai/technology/analog-computing It delivers up to 25 TOPS while typically consuming only 3-4W (a comparable GPU might use hundreds of watts). The chip can store about 80 million weight parameters onboard and performs computations without external memory. The M1076 contains 76 tiles (small chips inside the main package) each comprising one Analog Compute Engine (ACE). Each ACE has an analog flash memory array that stores neural network weights as varying resistances, and it performs matrix multiplications in-memory, using physical currents with low power consumption and at high speed. ADCs read out the results precisely. Each ACE has a small digital subsystem for support: a 32-bit RISC-V processor (for control/tasks), a SIMD (single-instruction, multiple-data) vector engine (for non-matrix ops), 64kiB of SRAM (local scratchpad) and a network-on-chip (NoC) router (to connect tiles efficiently). The flash memory cells are used as tuneable resistors to store the weights of a neural network. This can be achieved by controlling the charge stored in each cell. Input data is represented by voltages across the memory cells (in rows). These voltages across a known resistance produce a current determined by Ohm’s law, the product of the input voltage (the data) and the neural network weight. By summing all currents in a column, a vector-matrix multiplication can be performed. The result of the matrix multiplication is then read with the ADC (see Fig.52). Neurogrid siliconchip.au/link/acac Neurogrid from Stanford University is a mixed-signal experimental 20 Silicon Chip multichip neuromorphic system capable of simulating one million neurons and billions of synaptic connections in real time with very low power (about 3-5W). It is primarily used for large-scale brain simulations and employs analog computation to emulate synaptic connections and neuron dynamics. It is a landmark example of early mixed-signal neuromorphic hardware from around 2014. NeuRRAM siliconchip.au/link/acad NeuRRAM from UC San Diego/ MIT is an experimental neuromorphic mixed signal analog computein-­memory chip using resistive RAM (RRAM) for energy-efficient AI inference on edge devices. It runs a wide variety of AI tasks (eg, image classification, speech recognition and reconstruction) directly in memory with far less power than traditional methods. Okika https://okikadevices.com Okika acquired Anadigm and now produce field-programmable analog array (FPAA) chips (see Fig.53). These are not specifically designed for analog computing but can be used for such. FPAA chips contain configurable analog blocks like op amps, differential amps and programmable capacitor arrays that can be programmed as capacitors or resistors. Fig.53: an FPAA from Okika Devices mounted on a PCB. Source: https:// okikadevices.com Australia's electronics magazine A capacitor can emulate a resistor by switching it at high speed with two transistors or some other technique. This technique is called switched-­ capacitor resistor emulation. These chips can be used for sensor interfacing, audio processing, industrial control and low-power analog computing. They are the analog equivalent of a field programmable gate array (FPGA). There are a large number of analog components available for programming from the FlexAnalog FPAA Design Library – see Fig.54. Peking Uni siliconchip.au/link/acae Peking University announced a prototype analog AI chip using resistive random-access memory (RRAM). It achieved high accuracy with 24-bit precision (comparable to 32-bit floating-­ point digital systems) and claims 1000× faster speed and 100× less energy consumption than digital AI computing. Sagence AI www.sagence-ai.com This Silicon Valley company has developed analog in-memory computing chips, particularly for large generative AI models like Llama2-70B. The chips are currently in the customer evaluation phase. SynSense www.synsense.ai A leading manufacturer of ultralow-power mixed-signal neuromorphic chips and sensors, targeting edge applications such as audio processing and keyword detection, gesture and scene recognition, wearables, bio-signal analysis (eg, ECG, EMG, breath, and gait), behavioural monitoring, smart toys, home automation, security systems, industrial siliconchip.com.au Fig.54: the many different design elements that can be programmed into Okika FPAA, from the FlexAnalog FPAA Design Library. These include filters, op amps, differential amps, sample-and-hold circuits, differentiators, integrators and more. Source: https://okikadevices.com/collections/an231e04-reconfigurable-flexanalog-fpaa-chip-with-4-cabs testing, robotics, drones and more. AI research in Australia UWS Deep South (siliconchip.au/ link/acaf) is a neuromorphic computer to simulate the human brain, but is entirely digital and uses FPGAs. Conclusion We have traced the development of analog computing from its mechanical origins through the electronic era, siliconchip.com.au followed by a long period of dormancy as digital computing came to dominate. Today, we are witnessing a modern revival driven by the need to overcome digital computing’s limitations, particularly its high power consumption for AI workloads. The dramatically lower power requirements of analog and analog-­ inspired AI systems open the door to intelligent local ‘edge’ devices, such as smartphones and sensors. If these Australia's electronics magazine technologies fulfil their potential, they could usher in a new golden age of efficient, brain-like computing. For anyone interested in learning more about analog computers, we recommend checking out the two links below: • An analog computing book collection: siliconchip.au/link/acaj • Introduction to Analog Computer Programming by Dale I. Rummer: SC siliconchip.au/link/acak June 2026  21 FROM EVERYDAY PRINTS TO SHOW-STOPPERS. NEW TO JAYCAR Feed your printer what it deserves. Explore the extensive range of Filament THE RELIABLE PRINT CLASSIC 1995 . /RL ADD SPARKLE TO YOUR PRINTS TL6414 TL6440 TL6457 PLA+ ONLY $22.95/1kg Roll 16 Colours TL6457 - TL6472 TL6432 Rapid PLA+ ONLY $24.95/1kg Roll 11 Colours TL6435 - TL6445 HIDE THE LINES, KEEP THE STYLE Galaxy PLA ONLY $34.95/1kg Roll 3 Colours TL6431 - TL6433 DURABLE PRINTS, DONE FAST SILKY TEXTURE, GLOSSY FINISH EASY LIKE PLA, TOUGH LIKE ABS TL6485 TL6513 TL6418 Silk PLA ONLY $24.95/1kg Roll 13 Colours TL6418 - TL6430 MATTE PLA ONLY $24.95/1kg Roll 12 Colours TL6483 - TL6494 $ FAST PRINTS, STRONG RESULTS YOUR EVERYDAY PRINT HERO PLA ONLY $19.95/1kg Roll 18 Colours TL6400 - TL6417 FROM JUST TL6496 RAPID PETG ONLY $19.95/1kg Roll 13 Colours TL6509 - TL6521 PETG PRO ONLY $24.95/1kg Roll 13 Colours TL6495 - TL6507 TOUGH ENOUGH TO FLEX STRENGTH, PRECISION, & CLEAN FINISH PRINTS THAT BRAVE THE OUTDOORS TL6524 TPU ONLY $29.95/1kg Roll 7 Colours TL6524 - TL6530 TL6434 TL6522 ASA ONLY $24.95/1kg Roll Black TL6522 White TL6523 PLA-CF ONLY $34.95/1kg Roll Black TL6434 EXPLORE THE ELEGOO ® RANGE (AU) EXPLORE THE ELEGOO ® RANGE (NZ) Explore our great range of 3D Printing consumables, in stock on our website, or at over 140 stores or 130 resellers across Australia and New Zealand. jaycar.com.au 1800 022 888 | jaycar.co.nz 0800 452 922 Prices shown in $AUD, and correct at time of publication but are subject to change. Jaycar reserves the right to change prices if and when required. SHARP, SMOOTH, SPECTACULAR. Get the ELEGOO® Resin Range at Jaycar today. Smooth layers. Sharp details. Spectacular results. Saturn 4 TL4840 Mars 4 9k TL4824 Mars 5 TL4830 Saturn 4 Ultra 12K TL4842 Mars 5 Ultra TL4832 Saturn 4 Ultra 16K TL4844 Jupiter 2 16K TL4988* Build Volume (mm) Resolution Precision Layer Thickness Print Speed Special Features Cat Number Price Mars 4 153.36 x 77.76 x 175mm 9K 18µm 0.01–0.2mm 30–70mm/h • Laser-Carved Build Plate TL4824 $299 Mars 5 143.43 x 89.6 x 150mm 4K 35μm 0.01–0.2mm Max 70mm/h • One-click levelling TL4830 $349 Mars 5 Ultra 153.36 x 77.76 x 165mm 9K 18μm 0.01–0.2mm Max 150mm/h • One-click levelling • AI camera TL4832 $549 Saturn 4 218.88 x 122.88 x 220mm 12K 19×24μm 0.01–0.2mm Max 70mm/h • One-click levelling TL4840 $649 Saturn 4 Ultra 12K 218.88 x 122.88 x 220mm 12K 19×24μm 0.01–0.2mm Max 150mm/h • One-click levelling • AI camera TL4842 $849 TL4844 $999 TL4988* $1849 Saturn 4 Ultra 16K 211.68 x 118.37 x 220mm 16K 19×24μm 0.01–0.2mm Max 150mm/h • Auto-levelling • AI camera Jupiter 2* 302.4 x 161.98 x 300mm 16K 20 × 26µm 0.01 - 0.2mm Max 90mm/h Massive Print Volume Smart Auto-Levelling *Online only, available for order in store. GREAT RANGE OF ELEGOO® WASH & CURE MACHINES AT GREAT VALUE EXTENSIVE RANGE OF ELEGOO® RESIN AVAILABLE AT GREAT VALUE 2 MODELS AVAILABLE 13 TYPES AVAILABLE EXPLORE THE ELEGOO® RANGE (AU) Explore our great range of 3D Printing gear, in stock on our website, or at over 140 stores or 130 resellers across Australia & New Zealand. jaycar.com.au 1800 022 888 | jaycar.co.nz 0800 452 922 Prices shown in $AUD, and correct at time of publication but are subject to change. Jaycar reserves the right to change prices if and when required. EXPLORE THE ELEGOO® RANGE (NZ) FROM FIRST LAYER TO FINISHED MASTERPIECE. Explore the full range of ELEGOO® Filament Printers now at Jaycar. Neptune 4 TL4970 Neptune 4 Pro TL4972 EXPLORE THE ELEGOO ® RANGE (AU) EXPLORE THE ELEGOO ® RANGE (NZ) Neptune 4 Plus TL4974 Neptune 4 Max TL4976 Centauri Carbon TL4980 Centauri Carbon 2 Combo TL4986 OrangeStorm TL4982 Printer Type Build Volume (mm) Nozzle Size & Type Printing Nozzle Speed (max.) Temp (max.) Bed Temp (max.) Connectivity Cat Number Price Neptune 4 Cartesian Gantry Frame 225x225x265 0.4mm Standard 500mm/s 300°C 110°C USB, LAN, Wi-Fi^ TL4970 $399 Neptune 4 Pro Cartesian Gantry Frame 225x225x265 0.4mm Standard 500mm/s 300°C 110°C USB, LAN, Wi-Fi^ TL4972 $549 Neptune 4 Plus Cartesian Gantry Frame 320x320x385 0.4mm Standard 500mm/s 300°C 100°C USB, LAN, Wi-Fi^ TL4974 $649 Neptune 4 Max Cartesian Gantry Frame 420x420x480 0.4mm Standard 500mm/s 300°C 85°C USB, LAN, Wi-Fi^ TL4976 $799 OrangeStorm CoreXY Open Cube Frame 800x800x1000 0.4mm# Standard 300mm/s 300°C 100°C USB, LAN, Wi-Fi TL4982 $4,499 Centauri Carbon CoreXY Enclosed 256x256x256 0.4mm Hardened Steel 500mm/s 320°C 110°C USB, LAN, Wi-Fi TL4980 $699 Centauri Carbon 2 Combo CoreXY Enclosed 256x256x256 0.4mm Hardened Steel 500mm/s 350°C 110°C USB, LAN, Wi-Fi TL4986 $999 ^ Wi-Fi dongle sold separately. Also supports other nozzle sizes. See individual specs. # 12 FILAMENT OPTIONS: • PLA • GALAXY PLA • PLA+ • RAPID PLA+ • PLA MATTE • SILK PLA • RAPID PETG • PETG PRO • TPU • ASA • CARBON FIBRE SHOP ONLINE OR IN-STORE FOR OVER 115 ELEGOO® FILAMENT ROLLS ACROSS 12 MATERIAL TYPES AND A WIDE RANGE OF COLOURS. Explore our great range of 3D Printing gear, in stock on our website, or at over 140 stores or 130 resellers across Australia and New Zealand. jaycar.com.au 1800 022 888 | jaycar.co.nz 0800 452 922 Prices shown in $AUD, and correct at time of publication but are subject to change. Jaycar reserves the right to change prices if and when required. THINK GIGANTIC. PRINT GIGANTIC. Meet the OrangeStorm Giga by ELEGOO®. 800×800×1000mm of industrial-grade print capacity, now at Jaycar. ONLINE ONLY PORTABLE 7" HD TOUCHSCREEN 4499 $ Can be detached from the body for handheld operation TL4982 POWERFUL DOUBLE-SIDED COOLING FANS EXPANDABLE TO 4 NOZZLES * Can comprehensively enhance the model printing results Print four identical models at once, each in different colours. * Nozzle kit sold separately CATERPILLAR CABLE TRACKS Prevents the cables from being worn or broken and affecting your prints POWERFUL EXTRUSION SYSTEM High-quality nozzle capable of reaching 300°C. Dual-gear direct extruder. GIGANTIC IN SIZE UP TO 800 X 800 X 1000MM Print oversized models without splitting. Fast and efficient up to 300mm/s AUTO LEVELLING One-click automatic levelling and non-contact high precision sensor • EFFICIENT HEAT BED Four independent PEI magnetic high-temp platforms. Up to 100°C LEARN MORE: (AU) jaycar.com.au/p/TL4982 Explore our great range of 3D Printing gear, in stock on our website, or at over 140 stores or 130 resellers across Australia and New Zealand. jaycar.com.au 1800 022 888 | jaycar.co.nz 0800 452 922 LEARN MORE: (NZ) jaycar.co.nz/p/TL4982 SEE IN-STORE OR ONLINE FOR OUR EXTENSIVE RANGE OF FILAMENT AND ACCESSORIES AT GREAT VALUE Prices shown in $AUD, and correct at time of publication but are subject to change. Jaycar reserves the right to change prices if and when required. Part 1: introduction and overview T Machine Have you ever wanted to build your own pinball machine with a custom layout, artwork, sounds and lighting? It might seem like a daunting project, combining electronics, mechanics and software. In this series, we show how to design and build every part at home, from scratch! his major project, presented over the next few issues, will provide you with all the pieces to make your own pinball machine. You can pick and choose all aspects of it and ‘roll your own’ by 3D printing. Not only does that make it incredibly customisable, it avoids the expense of buying commercial pinball machine parts. They can add up fast! It seems like just a few months ago that I was talking with our most esteemed editor and asked what he thought a fun project would be. His answer was somewhat unexpected: a pinball machine. My immediate response was, “Sure, why not. I always wanted to build one of those as a kid.” I should have reflected on the reasons why I never went ahead with that before. In retrospect, that was a bit of a “... hold my beer” moment! The technology in a pinball machine is not advanced, but it is definitely in the complex electromechanical world. There are some very substantial forces involved in many of the mechanisms on a pinball deck, and controlling them reliably is not easy. There is a lot of electrical sensing and control, and therefore (in old-school machines, at least), a lot of wiring. Make no mistake, if you decide to follow my lead, you will be embarking on a major project, and there will be a lot of work. However, nothing about it is really difficult, especially since we have already done much of the hard work. We will be presenting triedand-tested electronic and mechanical designs. Critically, you will be able to download some files and start printing tested and verified 3D parts immediately. While the PCB and wiring are superficially impressive, the majority of the effort in this project has gone into realising a DIY-friendly approach to making the critical kickers, flippers, targets and bumpers. Of course, it is possible to buy parts to make a pinball machine; however, they are excruciatingly expensive. Also, many of the parts on the market are second-hand, which introduces reliability concerns. So what we did was draw on some experience and skills in 3D printing to bring more of a DIY approach to this project. That keeps costs down, as does using a Raspberry Pi Pico 2 microcontroller module as the ‘brains’ of the system. Australia's electronics magazine siliconchip.com.au Phil Prosser’s Phenomenal Pinball 26 Silicon Chip This also means that should anything break in the future, it will be easy to repair. Simply print a replacement part, swap it in, and away you go! We could even see Silicon Chip readers making and sharing their own elements to enhance this new ‘pinball ecosystem’. Don’t get the impression that following my instructions will completely remove the need for mechanical skills. Still, you will be making parts that we (and you) know will work when assembled and wired correctly. We spent a lot of time simplifying the wiring, with ribbon cables for most of the lights and sensors, and a control board that is ‘plug and play’. This was an attempt to keep the inevitable spaghetti wiring manageable! We will present all the component parts required for a decent pinball machine, including: Flippers Bumpers (that kick radially) Kickers (that kick outwards) Targets Rollovers (which detect the ball passing over) Loader and ball release A range of lighting parts A scoreboard A control system and a power supply You will be able to use some or all of the parts we used. You may decide to make more or fewer of some of them, or make your machine larger or smaller than ours. Before deciding if you’re going to build your own pinball machine, you will probably want to see what ours looks like and how it is to play. To that end, in addition to the photos in this and the following articles, we also have several gameplay videos you can view at the following link: siliconchip. com.au/Video/Pinball1 The controller includes more self-testing functions than we would normally include, allowing you to use a computer to monitor the state of the machine and test every component. So, we have some pretty solid built-in help for debugging. In a complex system like this, you want to be confident that you can track down any problems that arise. In addition to electronic and electromechanical parts, we will also provide several 3D-printable deck sections, especially around the reloader and ball release, which we think is pretty 📍 📍 📍 📍 📍 📍 📍 📍 📍 siliconchip.com.au cool. You don’t have to use these if you don’t want to; you may prefer to do some woodworking instead, or flex some of your other skills. The controller won’t care. So, you can make a deck that suits your needs; you don’t need to copy ours (but you certainly can if you want to!). We have included all the Fusion 360 CAD files in our download section, so if you want to modify these parts, you can do so. This software is free for non-commercial (hobbyist) use. The defining aspects of a pinball machine include a spring-launch mechanism and a playfield with obstacles and targets. Hitting the targets increases your score. The game ends when the ball goes into the ‘gutter’ a few times (usually three). Variants of the pinball game with these features have been around since the 1700s. In the 1900s, more complexity was added, including flippers and a coin operation mechanism, allowing pinball to become a commercial game. Post-WWII (the early 1950s) saw the widespread introduction of electrification and automated flippers, with the familiar placement of flippers near the bottom of the deck, allowing longer play and better control of the ball. Through to the mid1970s, all pinball machines remained electromechanical devices utilising relays for control and capturing scoring. In the mid-1970s, microprocessors were introduced, which enabled much more complex lights, sounds, scores and other functions. My rash promise to the editor was to bring the essence of this very long history of the game together into a project that you can make yourself. As implied above, you will be able to ‘mix and match’ and even customise our part designs into your own creation. As a final comment, this collection of the controller, electromechanical parts, example deck and software will do a lot of your heavy lifting. Even so, building a machine is a substantial undertaking. Keeping an oldschool pinball machine operational is a very technical job, and building one from scratch inevitably has some tricky parts. Photo 1: the general layout of the Pinball Machine. It’s highly configurable; you don’t need to do everything the same as us. You could have more or fewer bumpers, kickers, LEDs and so on. It’s designed to be flexible, to suit your idea of how a pinball game should work. June 2026  27 Fig.1: the system block diagram. Inputs are shown with red lines and outputs with blue. All inputs are active-low inputs, some from open-collector inductive sensors and some from pushbutton switches and microswitches. The outputs are direct drive for the 7-segment displays, opencollector outputs for the LEDs and open-drain Mosfet outputs for the high-power devices. For those of you with an artistic bent (or who know people with such skills), this will be an opportunity to let your creativity loose. We have made basic decorations for our machine, but we are sure that you, the reader, will come up with something even more unique! For those with a software background, we are also providing the full Visual Studio code, so you can hack into it and change the logic, tunes, sound effects, light sequences and whatever else you want. The world is your oyster. Let’s start with the system block diagram, shown in Fig.1. The central Control Board is a hefty 246.5 × 240.5mm, loaded full of through-hole parts. None of these are remotely fancy; the Raspberry Pi Pico 2 is the only high-tech part. This board has inputs for all buttons and sensors, LED drivers for the displays (a mix of 7-segment displays and individual LEDs), 12 Mosfet outputs that can drive very heavy loads, and interfaces for inductive proximity sensors. For the LEDs, the 7-­segment displays are driven directly by 74HC595 serial-to-parallel registers, while individual LEDs have bipolar transistor buffers to handle higher currents. 28 Silicon Chip All the inputs and LED connections are via 10-way ribbon cables, which run to small breakout boards local to where the LEDs and inputs are (see Photo 2). This simplifies the wiring considerably, as you can crimp a 10-way IDC ribbon connector in seconds using a vise or (ideally) a dedicated crimping tool. All power outputs are via pluggable headers using figure-8 cables. We used cheap speaker cable. This all operates at 24V DC and up to 6A, so we can’t use lightweight connectors. We recommend pluggable screw terminals, which again are quick and easy to assemble. If you are tempted to solder these wires to the board, have a bit of a think about servicing this machine later. Australia's electronics magazine We found the best arrangement was to have the power supply and controller behind the main display, with ribbon cables and high-current cables running from there to breakout boards at your sensors and effectors – see Photo 2. The Control Board is busy, as you can see from Photo b1 overleaf. Top deck We 3D printed the entirety of the top deck. We did this because we could dream up a rather complex layout, print it, try it, then tweak it so the game played better. The flippers can drive the ball in a range of angles, for example we wanted to have interesting things happen fairly often, so we moved the bridge entry to the left of siliconchip.com.au 3D-printed Electromechanical Parts Here’s an overview of the critical parts of the pinball machine that can be 3D-printed using files you can download from our website. In addition to the printed parts, some standard hardware is used (screws, nuts, washers etc) plus, depending on the part, pre-made switches, solenoids and other bits and pieces. 1. Flippers (Photo a1) – these are the way the user interfaces with the ball. They’re positioned near the bottom of the machine and pivot upwards when the user presses a button. If timed correctly, they will fling the ball up as it passes near them into the location desired by the user. If timed incorrectly, the ball may go somewhere you don’t want it – including in the gutter at the bottom, ending your turn. They use two solenoids to give enough speed and power to kick the ball properly. Photo a1: one flipper; a pair of these are what the player uses to move the balls around the deck and (hopefully) score points while avoiding the ball falling down the bottom, ending their turn. Note that this flipper is missing its rubber band. 2. Bumpers (Photo a2) – these are circular objects on the deck that detect when the ball hits them and push it away rapidly (driven by a solenoid). They make noise, flash lights and cause the ball to ping around rapidly, making the game much more exciting. 3. Kickers (Photo a3) – these are positioned touching a flexible band. If the ball hits the band, that is detected by a switch and the kicker then pushes the band out, causing the ball to fly away. They’re a bit like bumpers except they operate semi-linearly, rather than radially. 4. Targets (Photo a4) – these are basically labelled switches that detect when they are hit by the ball, usually increasing the player’s score. They’re typically placed against the side of the machine, where the player can fling the ball (if they have sufficient skill). siliconchip.com.au 5. Rollovers (Photo a5) – these are designs drawn on the deck with a corresponding inductive sensor underneath. The machine detects when the ball rolls over the image, increasing the score or having some other effect. Photo a2: bumpers like this are scattered (or clustered) around the deck, giving the ball a little boost while flashing lights and making noises (and possibly adding to the score). Photo a3: kicker(s) work somewhat similarly to bumpers but in a more linear fashion. They’re usually attached to taut bands that push the ball away when the kicker moves them. where it started and moved the targets up the deck. Making a ball game with a bridge is the one thing I wanted to do as a kid. You will probably form your own opinions on the layout; there is nothing stopping you from moving stuff around. Mind you, by the time we were satisfied with our layout, the sales staff in our local electronics shop were convinced we were 3D printing a battleship! While we used Fusion 360 to design all our parts, for those at the start of a 3D modelling journey, there are much simpler tools like Tinker-Cad that you could start with. Our experience is that you need to be willing to try different versions of the overall layout until you find one that plays well. Photo a5: inductive sensors mounted under the deck form ‘rollovers’, which can add to the score or trigger other actions. The zones are indicated on the deck with decorations like circles or starbursts. Photo a4: aim the ball perfectly at these targets for bonus points! They’re illuminated and trigger microswitches when hit. Photo 2: the machine’s wiring is greatly simplified by using 10-way ribbon cable for most runs. Breakout boards at the far end make it easy to connect to LEDs, inputs and so on. Australia's electronics magazine June 2026  29 Electronic Parts The Pinball Machine uses more than a dozen boards, most of them breaking out ribbon cable connections to simplify the wiring. 1. Control Board (Photo b1) – this hosts the Raspberry Pi Pico 2 and a lot of I/O. It senses all the switches and other inputs, then makes decisions to take action, triggering various outputs like LEDs and solenoids. It also keeps score, keeps track of the current player, and controls the flippers and ball launching mechanisms. 2. Power Supply (Photo b2) – the whole machine is run from a 24V DC 5A ‘brick’. This board derives +5.5V and +3.3V rails from the incoming 24V DC, then distributes all three rails to the Control Board. 3. Player & Score Displays (Photo b3) – these are mounted on the backboard and show the current player number and their score. They connect to the Control Board with one 10-way ribbon cable per digit and are driven from 74HC595 IC outputs with current-limiting series resistors. 4. Cascade & Bumper LED Boards (Photo b4) – these simplify the construction of groups of LEDs that shine through small windows on the deck. They connect back to the Control Board via one or two 10-way ribbon cables for each assembly. 5. Bumper & Kicker Interface Board (Photo b5) – this connects up to three bumpers and two kickers to the Control Board. Placing it near those devices simplifies the wiring. It connects the switches back to Control Board inputs via a 10-way ribbon cable, with additional figure-8 cables for each solenoid. 6. Power Distribution Board (Photo b6) – this makes it easy to connect multiple solenoids or other Photo b1: the Control Board. It’s a bit of a monster, but all the parts are cheap, commonly available and easy to solder. In a sense, it’s just a Raspberry Pi Pico 2 with a lot of I/O. Photo a1: one flipper; a pair of these are what the player uses to move the balls around the deck and (hopefully) score points while avoiding the ball falling down the bottom, ending their turn. They’re driven by two solenoids each to give good ball acceleration! Photo b2: the power supply simply derives 3.3V and 5V DC rails to power the Pico 2, LEDs and so on from the 24V supply. It also passes the 24V supply through to power the solenoids and audio amplifier. The final version moves the output connectors to align with those on the Control Board. Photo b3: the current player and score displays are simple 7-segment LED digits mounted on a PCB. The current-limiting LEDs are on the Control Board. If your woodworking skills are strong, you could start by making much of the deck out of timber, such as plywood. That may make it easier to experiment. You can either paint it to get a good finished product, or replace pieces with 3D-printed plastic parts once you have finalised the configuration. Bite-sized chunks There are quite a few parts to this project, which we will describe in sections. This article presents the overview and how it all comes together, the architecture and top-level software description. Over the next couple of months, we will describe the electronic modules and provide parts lists along 30 Silicon Chip with assembly and testing instructions. That will include the controller and I/O board, plus several other related parts If you’re planning to build the Pinball Machine, that will give you time to gather the components and start getting the electronics up and running, which will help with testing the electromechanical parts later. Once all the electronic parts have been described, we will provide the 3D printing files and describe how to put those parts together. That will include the electromechanical parts such as the kickers, flippers and suchlike. You can make these ‘standalone’, purchase commercial versions, or even design your own. When all the parts have been fully Australia's electronics magazine explained, we’ll give more details on our example layout for the machine we built. This allows us to present how to bring this lot together; we hope inspires you to develop your own deck layout, artwork and gameplay. That means the construction details for this new Pinball Machine will be spread out over the next few issues, finishing before the end of the year. Pico 2 software The processing is all done by the Raspberry Pi Pico 2. We gravitated to this as it is so easy to use, powerful enough for the job, and inexpensive too. The design does need to grapple with somewhat limited I/O, but by normal standards, the Raspberry Pi has plenty. siliconchip.com.au higher-power devices to a single output on the Control Board. 7. Rollover Interface Board (Photo b7) – this simplifies the wiring to inductive rollover sensors, allowing up to eight to be connected to the Control Board via a single 10-way ribbon cable. 8. Input & LED Breakout Boards (Photo b8) – these simplify wiring to switches and LEDs, allowing up to eight of each to be wired back to the Control Board via a single 10-way ribbon cable. The three-pin headers are for inductive sensors, like with the rollovers. Photo b7 (above): this board makes it easy to connect up to eight inductive sensors to the Control Board via a single ribbon cable. Use more than one if you want more than eight rollovers! Photo b5 (lower left): this board provides a local connection point for the wiring of up to three bumpers and two kickers. All the inputs go to the Control Board via a single ribbon cable, while the solenoid drive comes from the Control Board via one figure-8 wire per bumper/kicker. Photo b6 (lower right): this allows you to connect several high-current devices to simplify the wiring. These are straightthrough connections with added flyback diodes. Photo b4 (above): these boards arrange 15 and 8 LEDs, respectively, and connect them back to 10-way header(s) for connection to the Control Board. Much easier than wiring them by hand! The software is essentially a state machine. There are two main modes: normal operation and Self-Test. If you hold the TEST button while it is booting, the machine goes into the Self-Test mode; otherwise, the system boots into normal operation. The software structure is shown in Fig.2. After normal initialisation, the system goes into the Idle state, monitoring the coin, player add and start inputs. The coin input can be a simple push button, but if you want to be a little bit fancy, you could create a coin slot. All it needs to do is pull the coin input low each time a coin is sensed. Once “Start” is pressed in the idle state, and assuming you have sufficient credit, the machine moves to the RunGame state. The system runs siliconchip.com.au Photo b8: these three input and LED output breakout boards simply break out the eight connections on the 10-wire ribbon cable to eight separate polarised headers for easy wiring and maintenance. Photos 4 & 5: the adjacent photo gives an idea of what the wiring is like on the underside of the Pinball Machine. The photo below and right is an expanded view of the Control Board wiring that is shown in Photo 2. Fig.2: the software has two main modes: test mode and normal gameplay. The code is written in Visual Studio C, and does nothing tricky, so most software-conversant people should be able to modify it. a loop that looks at all the inputs and, depending on changes to any inputs, triggers the required action: For the flipper, kicker, bumper and reload mechanism sensor inputs, this action is to trigger the associated solenoid for the required time, then to return to an idle state. Scores are incremented for some of these. The flippers are a little different; if the player holds these, the system continues to generate a solenoid output but with a lower duty cycle using pulse-width modulation (PWM) – there will be more on why that’s necessary later. For targets and rollovers, this action can be to light up LEDs and to increment scores. Sounds are triggered by these inputs or time passing. Because the majority of the pinball machine is fairly one-to-one causal input-to-output, you don’t need to use all the inputs or outputs. You have great flexibility in how you build your machine. We provide the source code as part of the download, so if this is your thing, you can modify the software to make it your own. Indeed, we encourage this. If you make something cool, please let us know and share it! Once the player loses and the ball 📍 📍 📍 32 Silicon Chip is sensed in the reloader mechanism, the system preloads the ball using a solenoid, and the player number increments. The system indicates the player whose turn it is on the main display and provides their score. Players have a total of three balls (programmable) and they take turns for each of their three balls in round-robin fashion. In the GameOver state, the system presents the score for each player in turn. Once this cycle completes, the system returns to the Idle state. We will describe the significant selftest capabilities of the software later, after we’ve built some of the hardware. During construction, this is probably the most important part of the code. Software development The Raspberry Pi Pico add-on for Visual Studio is a joy to use. Coming from a hardware engineer who turns the air blue every time software has to be written, this is high praise. While this author has generated many lines of code for this and other projects, every time I write code, it is an exercise in learning (and patience!). The Pi development add-on to Visual Studio with AI assistance is truly a generational step forward, and Australia's electronics magazine we encourage you to look inside the code and give it a crack. The best outcome is your own awesome machine, and the worst is you revert to the baseline code. Control Board The Control Board carries a lot of parts but is not that complex. Its heart is the Raspberry Pi Pico 2, mounted via a pair of headers. There is a small power interface and an audio section. The remainder is all buffered inputs and outputs (I/Os). This is required for the sensors and outputs such as LEDs and solenoids. The Control Board has six main functional areas, as per Fig.3. These sections are also delineated by silkscreen ink on the actual board; we will describe in general how the Control Board works below. Sensor inputs: the Controller monitors all inputs, including user controls like the flipper buttons, coin slot, start button and such. At the same time, it also watches gameplay inputs such as target sensors, rollover sensors and mechanism sensors on kickers and bumpers. State changes on these inputs trigger actions. This includes changes in the state of the game, the score siliconchip.com.au Pinball Machine Kits Note that we are supplying partial kits for this project, primarily for the Control Board and Power Supply - see the Online Shop on page 86 for details. and controlling outputs. The outputs affected might drive lights, bells, LEDs, or more active outputs like flippers and kickers. This would be really easy were it not for the sheer number of inputs and outputs. Low-power outputs: there are 40 of these, for driving the 7-segment LED displays that show the current player and score. They can source or sink about 8mA each. Medium-power outputs: there are 64 of these to drive LEDs for effects spread around the playing area. 16 of these sink around 30mA for driving white LEDs, while the other 48 sink around 20mA for red LEDs. You could change the current limit mix to suit your machine. High-power outputs: these are for driving solenoids and such, supplied from 24V, at up to an amp or more (although the 5A power supply limits the total current and thus the power of all loads). Control and audio section: this includes the Pico 2, a ‘heartbeat’ LED and the audio amplifier, which is driven from a PWM output on the Pico 2. Power supply interface: the separate Power Supply Board does all the heavy lifting here. The six-way siliconchip.com.au power connector on the Control Board matches that on the Power Supply Board. How will it all come together? The Control Board has a six-way pluggable connection to the power supply. There are seventeen 10-way IDC-style box headers for the inputs and LED outputs, and twelve high-­ current outputs to the solenoids and high-power devices via two-way pluggable terminals. The Control Board normally lives in the backboard cabinet on a pinball machine. The I/O wiring runs from this to distributed local boards, either for direct presentation of LEDs or local wiring to the sensors. By using lots of 10-way ribbon cables with IDC connectors at each end, we keep the wiring much simpler and neater than if single-wire spaghetti ran everywhere. That means your job will be much easier in building it. An initial prototype we built without using this system, instead using crimp connectors on individual wires, turned out to be quite annoying to build. Next month In the next issue, we will start by building the electronics, starting with the Control Board, then the Power Supply and the other electronics. After that, we will walk readers through the construction and testing of the bumpers, kickers, flippers and other 3D-printed parts. Then we’ll get to the fun bit – putting it all together into a cohesive game! Finally, overleaf you will find the part lists for all the electronics. If you’re keen to build the Pinball Machine, start gathering them and we’ll have the conSC struction details next month. Fig.3: the six main functional blocks of the Control Board, which you will find outlined on the board itself. Australia's electronics magazine June 2026  33 Parts List – Pinball Machine Electronics Modules Control Board (one required) 1 double-sided PCB coded 08107261, 246.5 × 240.5mm 17 2×5-pin boxed IDC headers (CON1-CON9, CON11-CON18) 4 2-pin polarised vertical headers, 2.54mm pitch (CON10, CON35-CON37) 1 4-pin polarised vertical header, 2.54mm pitch (CON20; optional, for future expansion) 12 2-way vertical pluggable terminal blocks, 5.08mm pitch (CON21-CON28, CON31-CON34) 1 6-way vertical pluggable terminal block, 5.08mm pitch (CON38-CON40) 2 20-pin header strips (for MOD1) 2 20-pin female headers (for MOD1) 1 SPST momentary PCB-mount tactile key switch (S1) 1 10kW logarithmic taper 16mm single-gang potentiometer (VR1) Semiconductors 1 Raspberry Pi Pico 2 microcontroller module (MOD1) 4 74HC165 8-bit parallel-to-serial shift registers, DIP-16 (IC1-IC4) 15 74HC595 8-bit serial-to-parallel shift registers, DIP-16 (IC5-IC9, IC11-IC18, IC22-IC23) 1 LM384N 5W power amplifier IC, DIP-14 (IC10) 12 IRLZ44NPBF N-channel Mosfets, TO-220 (Q21-Q28, Q31-Q34) 64 BC338 or similar 100-800mA NPN transistors, TO-92 ▲ (Q111-Q188 ■) 1 3mm or 5mm LED (LED1) 13 1N4004 400V 1A diodes (D1, D21-D28, D31-D34) 64 1N4148 or 1N914 75V 200mA signal diodes ▲ (D111-D428 ■) ■ not all numbers in the range are used ▲ it may be cheaper to buy 100 Capacitors 2 2200μF 35V electrolytic, 7.5mm pitch 4 470μF 25V low-ESR electrolytic, 5mm pitch 1 47μF 16V electrolytic, 2.5mm pitch 1 4.7μF 50V electrolytic, 2.5mm pitch 33 100nF 50V radial ceramic, 5mm pitch Resistors (all axial ¼W ±5% or better) 1 2.2kW 41 220W 3 100W 1 2.7W 145 1kW 48 150W 16 82W Power Supply (one required) 1 double-sided PCB coded 08107262, 106.5 × 83mm 1 miniature 5/5.08mm-pitch 2-way terminal block (CON41) 1 PCB-mount DC barrel socket (CON42) [Altronics P0620] 1 6-way vertical pluggable terminal block, 5.08mm pitch (CON43-CON45) 2 LM2576T-ADJ 3A adjustable buck regulators, TO-220-5 (REG1, REG2) [Altronics Z0589] 2 1N5822 30V 3A schottky diodes (D2, D3) 6 M205 fuse clips (F1-F3) 1 M205 5A fast-blow fuse (F1) 2 M205 2A fast-blow fuses (F2, F3) 2 100μH 3-5A toroidal inductors (L1, L2) [Altronics L6622] 3 2200μF 35V electrolytic capacitors, 7.5mm pitch 2 470μF 35V electrolytic capacitors, 5mm pitch 2 100μF 35V electrolytic capacitors, 5mm pitch 4 100nF 50V ceramic or film capacitors, 5mm pitch 1 3kW ±1% ¼W axial resistor 1 1.6kW ±1% ¼W axial resistor 2 1kW ±1% ¼W axial resistors Player LED display board (one required) 1 double-sided PCB coded 08107263, 37 × 52mm 1 red common-anode 12.7mm (0.5-inch) 7-segment single-digit LED display (DISP7) [Altronics Z0191, Mouser HDSP-511E] 1 2×5-pin boxed IDC header (CON107) Score LED display board (one required) 1 double-sided PCB coded 08107264, 142 × 51.5mm 6 red common-anode 12.7mm (0.5in) 7-segment single-digit LED displays (DISP1-DISP6) [Altronics Z0191, Mouser HDSP-511E] 4 2×5-pin boxed IDC headers (CON101-CON104) 2 BC338 or similar 100-800mA NPN transistors, TO-92 (Q1, Q2) 2 1kW ¼W axial resistors 12 220W ¼W axial resistors LED output board (one or more required) 1 double-sided PCB coded 08107265, 49 × 38.5mm 1 2×5-pin boxed IDC header (CON50) 8 2-pin polarised vertical headers, 2.54mm pitch (CON51-CON58) Bumper LED board (several required) 1 double-sided PCB coded 08107266, 84.5 × 84.5mm 1 2×5-pin boxed IDC header (CON48) 8 5mm ultra-bright clear-lens 20mA red LEDs (LED17-LED24) [Mouser 941-C503BRCNCW0Z0AA1] Cascade LED board (one required) 1 double-sided PCB coded 08107267, 89.5 × 99mm 2 2×5-pin boxed IDC headers (CON46, CON47) 15 5mm ultra-bright clear-lens 30mA white LEDs (LED1-LED15) [Mouser 941-C503DWANCCBEB151] Switch input board (several required) 1 double-sided PCB coded 08107268, 54.5 × 38.5mm 1 2×5-pin boxed IDC header (CON70) 2 3-pin polarised vertical headers, 2.54mm pitch (CON71, CON75) 6 2-pin polarised vertical headers, 2.54mm pitch (CON72-CON74, CON76-CON78) General Input board (one or more required) 1 double-sided PCB coded 08107269, 49 × 38.5mm 1 2×5-pin boxed IDC header (CON60) 8 2-pin polarised vertical headers, 2.54mm pitch (CON61-CON68) High-current interface board (one required) 1 double-sided PCB coded 08107260, 38.5 × 66.5mm 8 2-way vertical pluggable terminal blocks, 5.08mm pitch (CON91-CON98) 4 1N4004 400V 1A diodes (D4-D7) Rollover interface board (one or more required) 1 double-sided PCB coded 08117261, 65.5 × 38.5mm 1 2×5-pin boxed IDC header (CON80) 8 3-pin polarised vertical headers, 2.54mm pitch (CON81-CON88) Bumper driver board (one required) 1 double-sided PCB coded 08117262, 93.5 × 77.5mm 1 2×5-pin boxed IDC header (CON110) 11 2-pin polarised vertical headers, 2.54mm pitch (CON111-121) 10 2-way vertical pluggable terminal blocks, 5.08mm pitch (CON131-CON140) 5 1N4004 400V 1A diodes (D11-D15) 3 390W ±5% 1W resistors siliconchip.com.au By Steve Mansfield-Devine for PCBWay The benefits of quality Inspection Reports for PCBs However much you test your product designs, there will always be factors outside your control that affect your final product. One of the most important of these is the quality of the PCBs you receive from your chosen fabrication house. A ll PCB manufacturers make promises about quality control and standards. It’s crucial for you to know how well they live up to these assurances. That’s where quality inspection reports play a vital role – they are the proof that the PCB manufacturer is meeting its promises. Leading PCB fabs commission periodic, independent laboratory testing of their laminate materials and finished boards. These reports provide you with peace of mind and are also invaluable for ongoing process monitoring, quality assurance, regulatory compliance and certification. PCBWay has published 14 quality inspection reports from tests carried out by Centre Testing International (CTI), an accredited third-party testing laboratory (see the bottom of the page at www.pcbway.com/oem/ quality-control.html). Each examination is carried out with industry-standard methods so that customers can compare the results from fabricators worldwide. This also makes them directly applicable to compliance demonstrations for IPC acceptance standards. As a product designer or electronics engineer, you can use these reports to ensure that the physical hardware of the end product will conform to your 36 Silicon Chip stack-up and material set, delivering the necessary reliability, safety and longevity. Thermomechanical tests For example, it’s essential to know how boards will respond to heat, both during assembly and when in use. The glass transition temperature (Tg) is arguably the most important laminate specification for engineers designing boards that will be soldered with lead-free processes or that will operate at elevated temperatures. It’s the temperature at which the resin binder in the laminate stops being rigid, like glass, and changes to a viscoelastic state. This happens over a transition range, but the test results are usually reported as the midpoint. The higher the Tg temperature, the better, because exceeding it can mean the board loses dimensional stability and becomes susceptible to delamination and mechanical damage. The common Tg figure for standard FR4 boards industry-wide is 130140°C. However, PCBWay significantly exceeds this, with a figure of 169.61°C. Similarly, time to delamination (T260, T288 and T300) measures how long a laminate can withstand a specific temperature before the layers Australia's electronics magazine separate. It is a better predictor of assembly survival than Tg alone. The most commonly cited specification is T288, which tests the time to delamination at 288°C. The industry standard is typically 5-10 minutes for high-reliability boards, but PCBWay’s tests show times consistently above the top end of the scale. High temperatures also cause decomposition of the resin in the board, resulting in loss of mass and weakness. Most PCB fabricators aim for a decomposition temperature (Td) of 325°C, while PCBWay achieved just over 345°C in its most recent tests. Related to this is the coefficient of linear thermal expansion (CTE) – how much the board will expand or contract for each degree of temperature change, relative to Tg. As the laminate is reinforced with glass cloth, expansion in the in-plane axes (X and Y) are largely constrained. However, expansion in the Z-axis (the thickness of the board) can place significant stress on plated through-holes (PTH), vias and solder joints under reflow and thermal cycling, causing barrel cracking, pad lifting and intermittent open-circuits. For temperatures below Tg, the typical figure for CTE is 50-70ppm/°C. Again, PCBWay significantly improves on this, with a test result of 37.4ppm/°C. These better-than-usual results mean much greater safety margins in both manufacturing and end-use for PCBWay customers. Dimensional and structural integrity Heat is not the only factor. There are also mechanical and structural concerns, such as board flatness (bow and twist), the integrity of inner-layer copper and interconnects, which can be checked by microsection inspection, and solder mask adhesion, which is verified by lattice tests. These are physical aspects of the PCB that are often taken for granted. However, with bow and twist, for example, any board with SMDs that has distortion of more than 0.75% can experience problems such as tombstoning (components that should be flat on the board sticking up, attached only at one end) or the failure of solder joints on ball grid array (BGA) packages. This is something that needs to be kept well under control, and PCBWay’s siliconchip.com.au tests show twist results of only 0.16% and bow of 0.12%; well below the threshold where problems typically start. Dimensional tolerance is another factor that seems simple but can have subtle implications. You might expect that the board will be the exact size, with drill holes in the precise locations, as specified in your Gerber files. However, all manufacturing processes have certain tolerances, and you need to know that these are acceptable for your design. The size report covers finished board dimensions, hole locations, pattern shrinkage or expansion and, in some cases, layer-to-layer registration. This helps confirm that the PCB will fit the mechanical envelope, align with mounting hardware, suit its enclosure and mate correctly with connectors, card edges and components. In addition to reliability and manufacturing problems, poor sizing can also create signal integrity and soldering problems. There can be compliance and certification implications, too. Factors such as safety spacing – for example, between high-voltage and low-voltage sections of the board – need to exceed the minimum allowed values in the final product. Typical industry standards allow a ±0.1mm tolerance for the outline of the board, ±10% for the thickness, and ±50–75µm for drill position, finished hole size tolerances and layer registration tolerances. Electrical and chemical performance One key metric that may be on the minds of many engineers is how well their boards will stand up to high voltages. The voltage resistant test (also known as the dielectric withstanding voltage, or hi-pot test) verifies that the PCB laminate (or solder mask over conductor pairs) can sustain 1000V DC at a controlled ramp rate of 100V/s for one minute without breakdown, flashover, sparkover or excessive leakage current. It is a simple pass/fail test. This is an important test for the certification of mains-connected equipment. With medical devices, it’s essential to ensure compliance with Means of Patient Protection (MOPP) and Means of Operator Protection (MOOP) requirements. siliconchip.com.au Related to this is the breakdown voltage. An AC voltage is applied across the test sample, ramped at 500V/s, until there is a catastrophic increase in leakage current or visible arc discharge. It provides an upper limit to the laminate’s dielectric strength under AC excitation. For power electronics designs, such as motors, uninterruptible power supplies, inverters and other cases where high-voltage conductors must be isolated on the PCB surface, solder mask breakdown voltage is a critical parameter. A designer can use these figures to derive the maximum allowable electric field strength under the solder mask and determine the minimum safe conductor spacing. Stable layers A PCB is far from simple. It is a sandwich of laminate substrates, traces and copper layers. How well those layers remain together is obviously crucial. The bond strength is a measure of laminate-to-foil adhesion quality. Low adhesion is a risk that manifests as lifted pads during hand soldering or rework, delamination under thermal shock and trace peeling under mechanical vibration—all of which can result in product failures in the field. The test measures the force required to maintain the peeling of a copper foil from the laminate, the result being the average of multiple tests. For standard FR-4 (fibreglass) boards, the general standard is 100-120N, but PCBWay’s test samples averaged around 220N. The bond strength degrades at high temperatures, which is another reason that having a high Tg value is important. Moisture absorption also creates problems for manufacturing and longterm reliability. For instance, moisture trapped in the laminate can flash to steam during reflow, creating internal vapour pressure that causes popcorning (components flying off), delamination and blistering. In RF and microwave PCBs, moisture absorption can be a primary material-selection criterion because absorbed water changes the laminate’s dielectric properties. This can alter controlled impedances, detune RF structures such as filters and antennas, increase dielectric loss and affect phase stability, particularly in equipment exposed to humidity or temperature cycling in the field. When a PCB absorbs water, its mass increases. The IPC-4101 maximum for standard FR-4 is a 0.32% increase, although PCBWay’s figures are four times better, at 0.08%. Production quality There are also tests that relate directly to the production quality of the PCB. The first of these is porosity, an important factor for any board with a gold surface finish like ENIG (electroless nickel immersion gold), especially those that might be exposed to harsh environments. Microscopic pinholes in the gold surface can expose the underlying nickel or copper, opening up a pathway for corrosion, possibly leading to ‘black pad syndrome’, causing poor solder joints and BGA interconnect failures that may not be detectable until deployment. Sample boards are exposed to nitric acid vapour and then dipped in a An operator applies solder paste to a PCB panel using a stencil in preparation for assembly. Australia's electronics magazine June 2026  37 Above: a multi-head automated PCB drilling machine. Left: AOI (Automatic Optical Inspection) of assembled panelised PCBs. reagent solution. The latter reacts with any exposed copper or nickel to produce visible corrosion spots. These are counted and grouped into three diameter categories: ≤0.05mm, 0.050.51mm and ≥0.51mm. The lower the number, the better. PCBWay’s most recent report shows a perfect result of zero in all categories. Finally, the cleanliness test measures the total quantity of ionisable (ionic) contaminants on the PCB surface. These are the products of flux activators (organic acids, halide activators), electroplating chemicals, etching solutions and handling contamination. They can result in electrochemical migration (ECM), also known as conductive anodic filament (CAF) formation and dendritic growth. In the presence of moisture and a DC field, dissolved ions move under electromotive force and form conductive metallic filaments between adjacent conductors, causing leakage currents, intermittent short circuits and ultimately failure. The board is washed with a solvent mixture, after which its conductivity is measured, the result being normalised to the mass of salt per unit board area. The IPC requirement is a figure of less than 1.0µg/cm2. PCBWay’s report shows a significantly better value of 0.19µg/cm2. This is the result of a well-controlled cleaning process and minimal-residue flux due to high process control standards. How to use the reports These reports are not obscure technical documents but tools that exist to provide real-world data to back up assumptions engineers must make during the design process. Engineers can compare Tg, Td, CTE, thickness, 38 Silicon Chip dielectric data and weave structure against their simulation and reliability assumptions about power dissipation, impedance and via life, validating their stack-up assumptions. Size, bow/twist and micro cross-­ section dimensions confirm that the fabricator can meet the mechanical tolerances necessary for connectors, enclosures and heavy components. The reports also help engineers select suitable laminates. The size, micro cross-section, CTE/ Tg, bond strength and cleanliness reports provide essential data when seeking to qualify a new PCB or new supplier, especially in the automotive and aerospace sectors. The data can also support reliability assessments by helping engineers justify assumptions about PCB-related failure risks. Signal integrity (SI) engineers frequently make use of the water absorption and micro cross-section reports to be certain that the dielectric thickness doesn’t vary too much from the design specification. If it did, the trace impedance may shift, causing signal reflections and data errors in high-speed buses such as PCIe or DDR4. OEMs can share these reports between design, quality and procurement teams to maintain a documented quality history and to justify changes in materials or suppliers during regulatory audits. Evidence of such tests may be required as part of supplier qualification or periodic audits (for example, for ISO 9001 quality management documentation). Even when the final product is in production, the reports can aid in ongoing process monitoring. Periodic cross-sections, cleanliness and bow/twist measurements are used as process control metrics and logged in Australia's electronics magazine inspection reports to show statistical stability and trigger any necessary corrective actions. Regulation and certification Arguably the most significant benefit of these reports is in regulatory compliance and certification. For example, UL 796 (the Underwriters Laboratories standard for PCBs) requires fabricators to track Tg and bond strength. Automotive and aerospace projects need hard evidence for thermal cycling, vibration, humidity bias and high-temperature storage. PCB-level CTE, Tg, Td, bond strength, bow/twist, voltage and cleanliness measurements are key elements in qualification reports and control plans. Cleanliness reports are mandatory for Class II/III medical devices under ISO 13485. Ionic contamination can lead to leakage currents that interfere with sensitive bio-signals or, in extreme cases, affect patient safety. Similarly, safety standards such as IEC 62368-1 (IT/AV), IEC 60601-1 (medical) and their UL counterparts rely on factors such as voltage resistance/breakdown, dimensional accuracy and more to ensure that the PCB portion of the design is robust. The micro cross-section report is the primary evidence that a board meets Class 3 requirements (such as minimum copper wrap-around at the knee of a hole) for AS9100 certification in the aerospace and defence sectors. Conclusion For engineers, quality inspection reports offer confirmation of technical standards. They also provide confidence that the customer’s design will succeed, which is why PCBWay strives SC to exceed industry standards. siliconchip.com.au Better Builder Shop 24/7 <at> altronics.com.au Deals! Suitable for light to medium duty drilling and driving tasks. Removes screws with ease with its 4.5nm torque drive, adjustable clutch and variable speed trigger - it even has a battery readout on the back! Get building this winter with savings on tech, tools & more . T 2164A SAVE 20% 40 $ Pro 72 Piece Servicing Kit A premium quality driver set with a huge variety of driver 4x28mm driver bits, opening tools, spudger, curved tip tweezers and flexi extension. Includes bit types for latest phones & laptops. Easy USB recharging! 12V Lithium USB Cordless Driver/Drill T 2126 NEW! 69.95 $ 5W 80 Channel UHF Walkie Talkie X 0438 Superb build quality for a great price! Our new Powerhouse walkie talkie is great for recreational and commercial comms use with IP67 water and dust protection and charging dock SAVE 15% 75 $ Handheld LCD Magnifier Camera SAVE $40 Completely portable magnifier camera with 3” display with live view. It offers high levels of magnification with LED illumination and auto focus for a crisp, clear view. Record photos or video to internal SD card (sold separately). Recharge via USB C. 159 $ X 0664 s Work er w as a potoo! k ban D 0531 SAVE 25% X 0210B 30 $ NEW! 49.95 $ Ultra-Bright 1200 Lumen Torch Durable aluminium body with double sided adjustable beam and double sided lantern lights. USB C rechargeable. SAVE 15% 50 Iroda® Mini Jet Butane Blowtorch Produces a powerful jet like flame with temperature up to 1300°C! SAVE 20% T 2488A 17 $ $ T 1463 Ultimate Helping Hands This space efficient work station for your benchtop has a sturdy metal base, 4 flexible clamp arms, solder reel holder, dry tip cleaner, flux AND a magnifier lamp! SAVE 33% 40 $ D 2045 Great tunes on the go The speaker features dual 2” full range drivers and bottom mount bass driver for a full rich audio. Pair it with a second unit (TWS) function. Magnetic Wireless Power Bank Offers plenty of back up power for charging, 20W PD output, 15W wireless + Apple watch. 38Wh capacity. Your electronics supplier since 1976. Shop in-store at one of our 11 locations around Australia: WA » PERTH » JOONDALUP » CANNINGTON » MIDLAND » MYAREE » BALCATTA VIC » SPRINGVALE » AIRPORT WEST QLD » VIRGINIA NSW » AUBURN SA » PROSPECT Sale ends June 30th 2026. Build It Yourself Electronics Centre® Test & Measure Measures AC & DC Q 1089 59 This compact true RMS meter provides in depth functionality for troubleshooting circuits. Push button design simplifies operation and test jack indicators ensure you never plug a cable in wrong! $ 800A AC & DC Clamp Meter True RMS 19999 Count DMM Safe and easy measurement of AC & DC voltage/current. In-built non contact voltage detection indicates live AC wiring. Includes test probes, temp probe & carry case. SAVE $40 SAVE $40 159 $ Powerful Portable Testing. 50Mhz oscilloscope and digital multimeter. A powerful tool for field testing provides a 2 channel oscilloscope with 280MSa/s sampling and 50MHz bandwidth. Plus 25000 count 10A multimeter. Easy USB C recharging. ProsKit® Analogue Multimeter 70 $ Great for servicing, repair and design of electronics. Low noise switchmode design. Fine & coarse voltage and current controls. 3A or 5A max. Size: 85Wx160Hx205Dmm. M 8303 3A Extended resolution to 4 digits! Offers everything the serious enthusiast could need with auto ranging, min/max/rel modes, frequency, duty cycle and non contact voltage detection. $ SAVE $19 Low Current Lab Power Supplies M 8305 5A 250 70 $ 70 Autoranging True RMS DMM SAVE $65 SAVE $19 SAVE 22% $ Q 0102A Q 1135 Q 0965A SAVE $20 $ Value packed! T 2685 Includes multimeter probes USB Soldering Iron With Multimeter Ideal for observing constantly varying quantities such as cycling voltages or speaker resonance testing. Q 1026B Compact, metal-cased, heats up fast, runs on a rechargeable battery or power bank, and features a clear OLED display for easy use. Plus in-built multimeter for simple continuity, resistance and voltage testing. 139 Q 2100 SAVE $28 189 $ SAVE $29 SAVE $19 60 $ D 3011 80 $ D 3009A SAVE 20% 39 $ Peak® DCA55 Component Analyser This easy to use, component analyser is like having a library of electronic info at your fingertips! Saves hours of looking up specs. 2 year warranty. Made in the UK. Power Over Ethernet Integrity Checker Test the condition of 802.3af/at standard and nonstandard PoE cabling/ports on PoE equipment. Tests for power supply voltage, polarity and crossover mode, loopback & displays power consumption of devices. SCAN TO FOLLOW US! Stay up to date on latest releases, exclusive specials and news on our socials. PoE Network Tester Test live ethernet cabling for data and power. Works with any 802.3af/at ports/ cabling. Ideal for data/comms installers. Like our service? Review your store on Google. Every review helps us serve you better. Power Up! 240V power from a lithium battery! Portable power for any adventure. This portable power generator is fitted with 194Wh battery bank & 150W mains inverter. Allowing you cable free power for both AC and DC appliances anywhere! Plus 2.1mm DC power & USB charging for keeping everything charged on the go. 40W solar panel (N0040F) to suit $99.95. SAVE $50 M 8199B 249 $ SAVE $49 150 $ M 8194A Don’t get stuck with a dud battery! This jumpstarter provides up to 1000A peak output when cranking. Suits most 12V vehicles. It can also be used as a device power bank with 60W USB PD output. Great for camping! Now with 15W charging! Power all your devices from a single charger! SAVE $40 179 $ SAVE 17% SAVE $10 129 $ D 2322A* M 8871 Massive 188W USB Charging Hub This do-it-all GaN charging hub is great for use at home or while you travel. Keeps the entire family charged up thanks to two high output PD charging ports (up to 140W), plus low output PD port and dual USB A type QC ports. Build wireless charging into your desk Great for the workbench! An ultra-slim desk mount 15W wireless fast charger. Needs 60mmØ hole. Includes power adaptor & USB cable. Charge any device, anywhere. SAVE 34% Dual GPO Charging Station 39 $ P 8164* Great for the workbench! Two top mount GPO’s for appliances, USB charger (20W PD+2x USB A) & wireless charger. 46 $ This high capacity 12,000mAh battery bank has charging cables in-built for USB C, Micro USB and Apple Lightning SAVE 27% equipped devices. It even includes $ the recharging cable! D 0510A 36 T 5099 Size: 360W x 185H x 138Dmm. Connect-It-All DC Power Hub A complete central power hub for your 4WD, caravan or camping remote power set up. 6 contura style illuminated rocker switches, 3 car accessory sockets (12V), USB charger, volt meter, protected by a 6 way weatherproof fuse block. Switch & Charge With Ease! Switch all your gear on and off. Includes; 5 x Rocker switches, 1 x 12V cig. socket, 1 x 5V USB- A and 1 x 5V USB-C, voltage meter. Size: 150 x 10.5 x 60mm. SAVE $10 S 2753 Powerhouse® Watt Meter 130A Accurately displays real time input and power usage currents (and wattage) from solar panels or batteries. Peak current 200A. M 8636A See power usage at a glance NO STRESS 30 DAY RETURNS! GOT A QUESTION? Not satisfied or not suitable? No worries! Return it in original condition within 30 days and get a refund. Ask us! Email us any time at: customerservice<at>altronics.com.au Conditions apply - see website. SAVE 30% 35 $ 55 $ Audio Visual C 5162 SAVE $80 199 $ Watch on the go with a 12V telly. C 5161 Great picture quality! Bargain for your caravan or camper! SAVE $40 149 $ 23.6” 1080p screen and HD tuner. Designed for 12V vehicles, 4WDs, caravans and campers. Easily wall mountable with SAVE $100 VESA mounting points. USB PVR recording in-built . 299 $ S 8866 Why pay $300 or more? SAVE 20% 55 $ Boom Box & Wireless PA Systems Need instant sound for your next big get together? Pick up a new Bluetooth entertainer box - available in small or large systems. They not only sound great, but offer a wireless micro phone for PA use, plus TWS pairing to a second unit (of the same type) for added volume. Offers up to 3-8 hours use from a single charge (depending on volume). 349 $ A 2696A H 8182A SAVE $40 A 4860 299 $ Wall mount your telly and save space! A compact wall bracket for screens using 200x200 VESA mounting with a max weight payload of 25kg. Folds flat to the wall to just 52mm from the wall surface. SAVE $70 Internet radio, digital radio & audio streaming in one. 2x50W Stereo Bluetooth Amplifier Wi-Fi Internet Radio System with DAB+, FM & Bluetooth. A stylish, easy to use receiver with access to over 26,000 global internet stations, plus DAB+ digital radio, FM frequencies and bluetooth streaming from your devices. Digital S/PDIF and analogue RCA outputs. A compact bluetooth stereo amplifier for home and commercial sound systems. Features a high spec Class-D 2x50W RMS amplifier and optical input for connection to digital audio sources. Dual Mic Wireless Karaoke Mixer Offers UHF wireless mics, Bluetooth, HDMI/optical/3.5 mm connections, adjustable DSP (reverb, echo, bass, treble), and supports two wired mics — perfect for parties and home entertainment. SAVE $60 SAVE $30 89 199 $ $ D 0980 C 8825 A 1102B NEW! NEW! 19 29 A 1101A $ .95 Convert your car to Bluetooth. Designed to bring wireless audio to any vehicle with a 3.5mm input. Powered by USB $ .95 Mini Bluetooth Audio Adaptor Versatile design transmits or receives audio. Up to 10 hours portable use per charge. Bluetooth 5.4. Pro grade condenser mic. Clear, crisp sound Premium USB Desktop Mic A premium finish USB mic with all metal case & stand. Adds high clarity sound to your desktop for live streams & podcasts. Sale Ends June 30th 2025 Shop in-store at one of our 11 locations around Australia: WA » PERTH » JOONDALUP » CANNINGTON » MIDLAND » MYAREE » BALCATTA VIC » SPRINGVALE » AIRPORT WEST QLD » VIRGINIA NSW » AUBURN SA » PROSPECT Or find a local reseller at: altronics.com.au/storelocations/dealers/ Shop online 24/7 <at> altronics.com.au B 0006 © Altronics 2026. E&OE. Prices stated here in are only valid until date shown or until stocks run out. Prices include GST and exclude freight and insurance. See latest catalogue for freight rates. *Devices for illustration pursposes only. Image source: https://unsplash.com/photos/water-droplets-on-a-window-with-blurred-green-plants-Fr4as0DLD4E Human Comfort Indicator Whether it’s a couch or the environment, comfort is often subjective. That is true of temperature, too; you will feel a lot less comfortable on a warm day if the humidity is high. Thus, this Human Comfort Indicator is much more useful than a mere thermometer because it tells you how hot it feels. By Tim Blythman T his project comes about due to a request from a magazine contributor. He said it would be handy to have a device that shows whether an environment is comfortable or not, not just the raw ambient temperature. While comfort is subjective, this device deals with parameters that can be easily measured and quantified. It is well known that certain combinations of temperature and humidity can be uncomfortable to humans. Naturally, this can also apply to animals and plants. You would have no trouble identifying conditions that feel uncomfortable for yourself, but it’s handy to be able to put a number on it, so you can be alerted when others might not be comfortable. One suggestion we heard is that the Human Comfort Indicator would be well-suited to monitoring the conditions in a greenhouse. Comfort and dew point The parameter we are tracking with the Human Comfort Indicator or HCI is the ‘dew point’. This is the temperature to which air must be cooled for liquid water to start condensing from it. Being a temperature, it is measured in °C or °F, but it relates to both the raw temperature and the relative or absolute humidity. Let’s look at some theory to explain why dew point is important. If you are in a warm environment that makes you perspire, the dew point has come into play. As the perspiration (sweat) evaporates, it cools your skin, Features Displays temperature, humidity and dew point Historical displays for the last day, week and month Configurable units, display orientation and update frequency Battery-powered with USB charging for uninterrupted operation The ultra-low-power e-paper display is unobtrusive and easy to read at a distance Optional analog (voltage) dew point output Open-drain alarm output Specifications Displays temperature/dew point in °C or °F to the nearest degree Relative humidity shown to nearest % Based on the excellent BME280 sensor Average battery draw of around 300μA gives months of operation on a single charge Screen update interval: every five minutes siliconchip.com.au Australia's electronics magazine but it cannot cool any lower than the dew point, since that is the temperature at which the air is saturated with water. In this case, ‘saturated’ is used in the scientific sense; it means that the air is at 100% relative humidity. Intuitively, as the dew point approaches the ambient temperature (due to the dew point rising or the ambient temperature falling), the relative humidity rises. This can be an indicator of changing weather conditions such as rain. You might see some weather forecasts report a ‘feels like’ temperature. This takes into account the dew point, as well as factors like wind and sun. For indoor conditions, the sensation will be dominated by the dew point. At very low dew points, evaporation from the skin increases, which can cause problems like skin drying out and cracking. Important, this is not necessarily something you would notice, unlike high humidity. So it’s handy to have a device that can alert you to this condition, allowing you to do something like switch on a humidifier. Table 1 shows some typical ranges of interest for dew point. It can be quite subjective; if you live in a tropical area, you may be comfortable at ranges higher than those suggested. Measurements Dew point sensors are not common, but a figure can be easily derived if June 2026  43 the relative humidity and temperature are known. Thankfully, many modern sensor modules can read both. Historically, an arrangement known as a wet-bulb thermometer would be used. This is a glass thermometer that has its bulb surrounded by a piece of cloth soaked in water. The water evaporates, cooling the bulb and reducing the indicated temperature below ambient. It would often be used in conjunction with a drybulb thermometer to give the true air temperature. The ‘sling psychrometer’ is a device fitted with a wet-bulb and dry-bulb thermometer. It is spun around above one’s head for a minute or so, quickly bringing the wet-bulb thermometer to equilibrium. A chart was then used to determine the dew point from the two temperatures. Thankfully, you don’t need to swing the HCI around above your head! The equations for converting temperature and relative humidity to dew point are complex but well-­ established, so it is a simple case of performing readings from our sensors and then a few calculations to produce the desired figures. Design As you can see from the photos, the Human Comfort Indicator has a simple design that would suit being used around the home, similarly to a weather station. The case is 3D-printed, although the PCB is designed to be easily mounted inside any suitable enclosure with a few holes in it. Table 1: Dew point interpretation Dew point Subjective condition <5°C Very dry 5-10°C Dry 10-15°C Comfortable 15-20°C Mostly comfortable 20-25°C Muggy >25°C Uncomfortable Original source: https://media.bom.gov.au/social/ blog/1324/feeling-hot-and-bothered-its-not-thehumidity-its-the-dew-point/ We use an e-paper panel to display the readings. These draw close to zero power except when they are actually updating, so they are a good choice for a battery-powered device. They are also easy to read under a wide range of light conditions as they are similar to ink on paper. We have written a feature article with more details on e-paper in this issue. It includes some background on the technology and how we came to choose a specific panel. The person who suggested this project also asked for some extra outputs on the device. The first is an analog voltage that reflects the dew point temperature, which can be used as an input to another system, such as a data logger. The other is an open-drain output that can be triggered when certain conditions are met, such as the temperature or dew point falling outside preset ranges. It is controlled by a small Mosfet capable of sinking a few hundred milliamperes, so it can After the SMD parts have been fitted and the micro has been programmed, you can test the e-paper panel by supplying power via the USB socket. You will see this error message since the sensor has not been fitted. 44 Silicon Chip Australia's electronics magazine directly drive a buzzer or even a small relay if a larger load needs to be controlled. Circuit details Fig.1 shows the circuit diagram of the Human Comfort Indicator. The circuitry around CON4 and the MOD2 e-paper panel is virtually identical to that described in the feature article. It differs from the breakout boards we tested mainly in using larger components to simplify soldering. This part of the circuit generates the necessary voltages to drive the display panel. Mosfet Q1 is driven by the display controller on the e-paper panel to provide ±20V rails. BS1 is tied low to force the controller into 8-bit SPI mode. The circuit is driven by a 16-bit PIC24FJ256GA702 microcontroller (IC1) boasting 256kiB of flash memory and 16kiB of RAM. The large amount of flash allows us to store graphics, such as font data, while the RAM allows us to create a buffer large enough to store an entire screenful for display, something that would not be possible with most 8-bit microcontrollers. We’ve established that the PIC24­ FJ256GA702 is capable of low-power operation, having used it in the ESR Tweezers project from the June 2024 issue (siliconchip.au/Article/16289). It also has hardware multiply and divide functions, which will help performing the mathematical operations needed to process our readings. On top of all that, it’s relatively inexpensive. IC1 is supplied from the 3.3V rail and also has two 100nF supply rail bypass capacitors plus the necessary 10μF capacitor on its VCAP pin (pin 20). This bypasses an internal regulator used to power the processor core at between 1.2V and 1.8V. Pins 1, 4 and 5 are ICSP (in-circuit serial programming) pins connected to CON1, along with 3.3V and ground, allowing it to be reflashed after soldering. Pin 1 has a 10kW pullup resistor to allow normal operation unless a programmer is connected. IC1 controls the e-paper panel via six lines: three for the SPI interface and three more control signals. The PPS (peripheral pin select) feature of IC1 allows most peripherals (like the hardware SPI interface) to be directed to most pins, simplifying the PCB layout. One exception is the Vout signal, siliconchip.com.au Fig.1: the circuitry around Q1 and connecting to the e-paper display via CON4 is driven by the controller on the e-paper panel. These components generate the various supply rails needed to drive the display. Our feature article in this issue (see page 66) covers this in more detail. which is provided by the analog CVref peripheral, which is fixed at pin 25. The CVref output comes from a 5-bit DAC. When using the 3.3V supply rails as its inputs, we get 32 steps over 3.3V, or near enough to 0.1V resolution for the Vout signal. This is simply made available, along with a ground connection, at pin header CON5. The opendrain output is provided at CON6, is implemented using Mosfet Q2, driven from pin 11 of IC1. A bi-colour LED, LED2, is driven from another two I/O pins via a 1kW series resistor. The two pins allow it to be lit up red or green, or off entirely. Three tactile pushbutton switches are connected to three more I/O pins on IC1. These are configured as inputs with internal pull-up siliconchip.com.au currents, allowing the switch states to be detected. Each pin is pulled to ground by the associated switch when it’s closed, or held high by the pull-up current the rest of the time. The last component is sensor module MOD2, which is connected at CON3, a six-way header to suit its pinout. This module includes a Bosch Sensortec BME280 humidity, pressure and temperature sensor. We have chosen it because it contains only the bare minimum circuitry needed to operate the sensor chip. Specifically, it lacks a voltage regulator, so we don’t need to worry about a poorly designed module wasting power in an inefficient regulator. The sensor IC itself is designed for 3.3V operation, so it can run from the Australia's electronics magazine same rail as the microcontroller. It consumes just 0.1μA in sleep mode. The chip and module can work in either SPI or I2C mode; we are using I2C in this case. The module includes bypass capacitors and pullup resistors for the communication lines. It is an updated version of the GY-BM module described in our review of Pressure/Temperature Modules (December 2017; siliconchip. au/Article/10910). The similar circuit of that module is shown on p82 of that article. Power supply Like the e-paper controller, IC1 is a nominally 3.3V device (3.6V maximum), so the main logic and supply voltage is set at 3.3V. This comes June 2026  45 from REG1, an MCP1700-3302 LDO (low dropout) regulator. Helpfully, it also has a low quiescent current of around 2μA. This is in turn fed from one of two sources by a common-cathode dual schottky diode, D1. One anode connects to a Li-ion battery, while the other is a 5V supply from USB-C socket CON2. CON2’s CC1 and CC2 pins are connected to the requisite 5.1kW resistors to ground. These indicate that the device should be treated as a sink and be supplied with 5V when connected to a source. Two alternative locations are provided for CON2 so that the PCB can be constructed to suit one of two particular orientations. 5V from CON2 also supplies the charging circuitry based on IC2. The 10kW resistor sets the charging current to 100mA, the minimum permitted by this chip, while LED1 is another bi-colour LED that shows red while charging is occurring and the STAT pin is low. When charging completes, the STAT pin goes high and LED1’s green element lights. The arrangement of the two 1kW resistors allows this to work. While it might appear inefficient to have the resistors connected across the supply, they will only draw current while 5V is present at the USB socket and won’t drain the battery. The diode arrangement means that there is no current draw from the battery while charging is occurring, so the battery can charge fully. Software Software operation is focused on minimising power consumption where possible. This mostly consists of setting the external modules to low-power modes. The BME280 is checked every 5 minutes and the screen is updated at the same rate. A low-power RC oscillator (LPRC) in the microcontroller keeping time means that the processor can spend much of its time sleeping. The LPRC is configured to provide an interrupt at around 5Hz, quick enough that it can provide timing down to one second with reasonable consistency. Several counters are updated with this interrupt. Some counters keep rough track of hours, days and weeks, allowing daily, weekly and monthly averages to be accumulated. Another counter sets LED2 to light up for about one second every minute; it consumes a few milliamperes when on, so operating it with a low duty cycle reduces the average current consumption. The colour that LED2 shows when lit matches the alarm state; if red, then the alarm is active, Q2 is on and the ALARM output is pulled to GND. Otherwise, LED2 is green and Q2 is off. The sensor readings require a bit of processing. There are a total 18 Scope 1: the idle current of the Human Comfort indicator is around 100μA, with a peak below 5mA (during a full refresh, shown here). 46 Silicon Chip calibration parameters that are unique to each chip; they are used for compensating the raw 16-bit humidity readings and 20-bit temperature and pressure readings. The humidity and pressure readings are further compensated based on the measured temperature. Fortunately, Bosch Sensortec provides sample code to do this. The results are displayed according to user preferences, such as temperature display units. There is also a menu system that allows the preferences and settings to be changed. Since the e-paper display takes so long to refresh, we have kept the options and thus the menus fairly simple to prevent them from being unwieldy to use. We’ll explain more about the operation once construction is complete. Power Scopes 1 & 2 show typical current consumption. Scope 1 shows a full refresh occurring on the main page, with the low levels on each side of the peak representative of the normal idle state below 100μA. The refresh is no more than 5mA for no more than five seconds and occurs about once every five minutes, giving an average contribution of less than 83μA. Scope 2 shows a partial refresh occurring upon entry to the setting screen. Note that the current peak Scope 2: a partial screen refresh requires less current. The 1.5mA draw seen here is due to LED2 being lit while the settings screens are active. Australia's electronics magazine siliconchip.com.au Fig.2: we designed this 3D-printed case (above) to suit the Human Comfort Indicator. It comes in two variants, with this render showing the landscape version, with the USB socket coming out the side. Vents allow the sensor to sample room air. The portrait version is shown in the adjacent photos. is slightly lower due to the partial refresh. The higher level on the right is due to the LED2 lighting up and consuming 1.5mA. During normal operation, the LED is on for approximately one second per minute, contributing an average of 25μA. So we expect the long-term average consumption of the Human Comfort Indicator to be around 210μA. Given that the self-discharge of Li-ion cells is typically around 2% per month, a typical cell will lose 40mAh per month or around 53μA; a significant chunk of the usage! We have quoted 300μA to take into account some time spent on the settings or viewing different pages. With a nominal 2000mAh lithium cell, this equates to 6600 hours or around nine months of operation on battery power. Considerations Since the e-paper display panel includes look-up table (LUT) options for both full and fast refreshes, we have included an option to set how the refreshes occur. Broadly speaking, the full refresh will use more power, but will provide a clearer display. The fast refresh appears less distracting when it happens, but the resulting display has slightly poorer contrast. The above calculations assume that full refreshes occur at all times, so using partial refreshes should provide even longer operation than quoted. The outputs at CON5 and CON6 are nothing more than pin headers, siliconchip.com.au since we expect they may not be used in most cases. The analog output on CON5 is a voltage (in volts) that is one tenth of the dew point (in degrees). So a dew point of 16°C will result in an output of 1.6V; naturally, this is capped between 0V and around 3.1V, the upper limit of the CVref peripheral. It is not buffered and has an estimated output impedance of around 30kW. Unless you are connecting it to a high-impedance input, you may need to buffer it. We have chosen not to provide a buffer, since it would draw extra power that would be wasted if this output is unused. E-paper’s ability to show a display while using no power also has a subtle downside in that it can be hard to tell if the device is working or frozen. The main way we show the health of the Human Comfort Indicator is through the flashing of LED2. If you don’t see LED2 flash occasionally, the Human Comfort Indicator may have shut down due to a flat battery. The voltage of the regulated 3.3V line is also shown on the main page. If this gets near 3.0V, the battery may be overdischarged. There is no built-in overdischarge protection, so a protected cell must be used. That will prevent significant cell damage if you forget to recharge it. Any unit like the Human Comfort Indicator that measures ambient temperature is at risk of being affected by self-heating, where the heat dissipated by the unit’s own operation drives up the measured temperature. Australia's electronics magazine Thankfully, it has very low power consumption, and most of its power consumption occurs immediately after a reading has been taken (when the display is refreshed with the updated readings). So that should not be a probably, provided that the chosen enclosure does not trap heat. Enclosure We have designed a 3D-printed enclosure to suit the Human Comfort Indicator with a vent to allow air exchange. That should not only eliminate self-heating concerns, but it’s also necessary so that the humidity sensor (and to a lesser extent, temperature sensor) can respond to the ambient conditions properly. It will probably take a few hours to print, so we recommend that you start that while assembling the PCB. The two parts simply snap together, and they are designed to be printed without supports. There are two variants of the case, one to suit each position for the USB connector, CON2. One variant suits a portrait layout, with the hole for the USB socket on the long side. The other variant suits a landscape display and has the USB socket on the short side. Make sure that you choose two matching halves before printing them. The render in Fig.2 shows the two halves of the landscape version. We printed our prototypes on an Ender-3 V3-SE and, to ensure the best appearance, used a 0.08mm layer height and low speed, about 50% of June 2026  47 Fig.3: all the SMD parts are on one side of the PCB, with many of the through-hole parts mounted on the back. Try to keep the area under the e-paper panel clear on the back of the PCB. The entire Human Comfort Indicator is just under 9cm tall. It is fully self-contained and is powered from a rechargeable battery that should last close to a year between charges. normal speed. The job took about six hours in total: two hours for the front and four hours for the back half. The result should be usable with minimal post-processing. At most, you might need to remove small burrs or file down the parts if they have been heavily over-extruded. The bezels on the front half of the case (where the display mounts) are quite thin, so take care when removing them from your print bed. The vent holes are nominally 2mm in diameter if you wish to clean them up with a drill bit. The PCB also has mounting holes, allowing it to be fitted to just about any enclosure that is large enough and can have suitable holes made (see Fig.4). The PCB simply mounts to the front panel of such an enclosure using screws and spacers. A UB3 Jiffy box should be a good fit. The view shown in Fig.4 is from the outside of the enclosure. If you are using such an enclosure, we suggest using the bare PCB as a template for the mounting and switch holes. The USB socket will probably not align with the edge of the case, so you will likely have to cut a hole to allow a cable to pass through the case. Options The charging components are inside a white box printed on the PCB. Leaving these parts off will disable the charging feature. You could do this if you don’t want to use an internal 48 Silicon Chip battery; in this case, it will only operate when powered from CON2. We’ll describe the assembly including all components. Simply leave off the battery, battery holder and all parts in that box if you wish to pursue this option. You can also see this outline in the overlay diagram (Fig.3). We recommend using the battery, since this will allow the Human Comfort Indicator to operate during brief power interruptions. Importantly, it will be able to retain its long-term average readings in RAM. PCB assembly There are a few fine-pitch devices on the board, so you will probably need a magnifier and good lighting as well as the usual SMD gear such as flux paste, tweezers and solder-wicking braid. Start by fitting the SMD components, which are all on one side of the PCB. IC1 and CON4 have the finest lead pitches. Apply flux to the pads and rest these components in place, adding more flux on top of the leads. Carefully align the pins to their pads, then check they are correctly orientated before tacking one or two pins in place, making sure that they are flat against the PCB. We’ve made the pads quite long, so you can try applying the iron to the pad only; this should be sufficient to cause the solder to run onto the lead and form a proper joint. Check for any bridges and use the braid to remove Australia's electronics magazine excess solder, adding more flux as needed. When you are sure that CON4 is properly soldered, apply a solid fillet to each of the larger end pads for mechanical strength. Next, solder IC2. It has five pins, so it will only fit properly in one orientation. Follow with the three-lead SOT23 parts. Note that this includes a regulator, two Mosfets and a dual diode, which all look similar, so don’t get them mixed up. There are also three single diodes to be installed; solder them next, making sure to place the cathode stripes near the K symbols, as shown in Fig.3. Follow with the capacitors; they will not be individually marked. The 1μF parts are the most numerous, so we recommend starting with these. Ten are at the lower left near CON4, while two are near REG1. Most of the 1μF parts are in a single row that also includes a solitary 100nF capacitor, so watch out for the interloper. Two of the 10μF capacitors are near IC2, with the third near IC1. These will probably be thicker than the other capacitors, so you might be able to tell them apart. Follow with the single 4.7μF capacitor (or 10μF as supplied in the kit) and the two remaining 100nF capacitors, near IC1 and the interloper in the row of 1μF parts. The solitary inductor has large pads, but it doesn’t have a huge thermal mass, so soldering it should be straightforward; it is not polarised. siliconchip.com.au Follow with the 11 resistors, checking their markings (1003 or 104 = 100kW etc). To complete the SMD parts, fit CON2 to your preferred location. Be sure to add solid solder fillets to CON2 so that it is mechanically secure. That completes the SMD components, so clean up the residual flux with a suitable solvent and check for any bridges or dry solder joints. If you need to reflow any joints, repeat the cleaning process in that area. When finished, allow the board to dry. You can perform a brief check by applying USB power to CON2. You should see 3.2-3.4V on the second pad of CON1 relative to ground. Ground is the middle pad of CON2, or the marked pads on CON5 or CON6. If you do not see this voltage, disconnect power immediately and investigate before proceeding. Programming IC1 If you purchased the chip from the Silicon Chip Online Shop, it will be programmed and you can skip to the next step. Otherwise, fit a vertical pin header to CON1. The reverse of the PCB is where the e-paper module will sit, so it is best to minimise the amount by which components protrude into this area. Inside our 3D printed enclosure, the display panel sits 1.5mm from the PCB, so this is the absolute maximum by which items should extend behind the PCB in this area. One way to do this is to mount the header with its plastic block sitting off the PCB slightly. There should be enough clearance so that CON1 can be left in place afterwards. Connect your programmer to CON1; a Snap, PICkit 4, PICkit 5 or PICkit BASIC should all be suitable. You can use the CON2 USB socket to provide power. Open the Microchip MPLAB IPE program and use it to program the 2110526A.HEX file into IC1 and verify it. Checking the display Disconnect any power supply before proceeding. You can now check that the display panel is functional by plugging it into CON4. Pull the grey tabs outwards, parallel to the PCB, to release the catch. Slot the FFC (flexible flat cable) into CON4 with the gold contacts facing upwards. The gold part should not be visible when the FFC is fully inserted. Then carefully push the tabs back in. siliconchip.com.au Parts List – Human Comfort Indicator 1 double-sided 50 × 80mm PCB coded 21105261 1 3D-printed case (alternative parts listed at the bottom) [SC7453/SC7684] 1 single AA (14500 size) PCB-mount cell holder 1 Li-ion rechargeable 14500 (AA-sized) cell with built-in protection 1 5-pin header, 2.54mm pitch (CON1; optional, for ICSP) 1 SMD USB-C power-only socket (CON2) [eg, GCT USB-4135-GF-A] 1 6-way right-angle header, 2.54mm pitch (CON3) 1 24-way SMT FFC top-connect ZIF socket (CON4) [EastRising ER-CON24HT-1; www.buydisplay.com/24-pin-0-5mm-pitchtop-contact-zif-connector-fpc-connector] 2 2-way headers, 2.54mm pitch (CON5 & CON6; optional, for external signal connections) 1 47μH 6×6mm SMD inductor (L1) [LSXBD6060WHL470M from DigiKey] 1 3.3V 6-pin BME280 module (MOD1) [Silicon Chip SC5482] 1 EastRising ER-EPD029-2B 2.9in EPD module with 24-pin FFC connector (MOD2) [www.buydisplay.com/serial-2-9-inch-e-paper-screen-128x296for-electronic-shelf-label-lcd] 3 6×6×7mm through-hole tactile switches (~3mm actuators) (S1-S3) 1 piece of foam-backed double-sided tape or similar to secure e-paper panel to main PCB 4 small self-adhesive rubber feet (optional) Semiconductors 1 PIC24FJ256GA702-I/SS microcontroller programmed with 2110526A.HEX, SSOP-28 (IC1) 1 MCP73831T-2ACI/OT Li-ion charge controller IC, SOT-23-5 (IC2) 1 MCP1700-3302E/TT LDO 3.3V linear regulator, SOT-23 (REG1) 2 AO3400 30V 5.8A SMD N-Channel Mosfets, SOT-23 (Q1, Q2) 1 BAT54C 25V 200mA dual common-cathode schottky diode, SOT-23 (D1) 3 MBR0540 50V 0.5A schottky diodes, SOD-123 (D2-D4) 2 3mm bi-colour red/green 2-lead LEDs (LED1, LED2) Capacitors (all SMD M2012/0805 MLCCs) 3 10μF X5R 16V 1 4.7μF or 10μF X5R 16V 12 1μF X5R 50V 3 100nF X7R 50V Resistors (all SMD M2012/0805 ±1% ⅛W) 1 100kW 4 10kW 2 5.1kW 3 1kW 1 0.47W Alternative parts for non-3D-printed case 1 UB3 Jiffy box (see Fig.4 below) 2 M3 × 10mm panhead machine screws 2 M3 hex nuts 2 3mm-long, >3mm inner diameter untapped spacers CL Fig.4: if you don’t plan to use our 3D-printed case design, use this 34.5 34.5 diagram to cut and 22.5 drill a UB3 Jiffy 69 box lid instead. 32 DISPLAY Find the centre of WINDOW your panel by marking where 9.5 the two diagonals cross; this should 2 help to centre 8 12 7.5 5 12 12 Ø3 Ø4 Ø4 Ø4 Ø3 Ø3 the display and controls. Use the blank PCB to mark ALL DIMENSIONS IN MILLIMETRES the panel first. CL Ø3 10 SCALE: 100% SC7646 Kit ($60 + postage): includes everything except the case and battery Table 2: Settings summary Number Setting Options Notes 1 Dew point minimum Alarm on or Alarm off If this is off, the alarm is not triggered by a low dew point 2 Dew point minimum -10°C to 30°C or 14°F to 86°F Increments in steps of the currently selected units 3 Dew point maximum Alarm on or Alarm off If this is off, the alarm is not triggered by a high dew point 4 Dew point maximum -10°C to 30°C or 14°F to 86°F Increments in steps of the currently selected units 5 Sensor fail Alarm on or Alarm off If this is set off, then the alarm is not triggered by sensor failure 6 Units °C or °F All temperature settings and figures are shown in these units 7 Orientation Portrait, Adjust to suit the case or landscape, installation reverse portrait or reverse landscape 8 Text Black text or white text The background is the opposite of the text colour 9 Refresh Always, Hourly Daily This is how often a full refresh occurs; otherwise, a fast refresh happens 10 Flash options S3: Save to flash S2: Restore from backup The “Ready” message will change when a save or restore has completed This board has all the SMDs fitted but none of the through-hole parts (yet). The completed PCB is shown on the left before installation in the case. 50 Silicon Chip Australia's electronics magazine In use, the flexible cable bends 180° to put the display on the back of the main PCB, but for now, the whole assembly can lay flat on your workbench. These panels are quite fragile, so handle with care. Applying USB power at CON2 should cause the display to operate. If it flickers but turns solid black or remains white, it could be that one of the leads for CON4 or one of the components in that area of the PCB is soldered incorrectly. Disconnect the power and remove the display by pulling out the grey tabs, then investigate the fault and rectify it before proceeding. Last components Finish assembling the PCB by soldering the last few components. The three tactile switches and two LEDs should be soldered flat against the PCB. These are all on the opposite side to the previously fitted SMD components. The LEDs should be fitted such that the green cathode (shorter lead) is towards the nearest edge of the PCB. You can easily test this by applying power to the USB socket and carefully connect the LED between pins 1 and 3 of CON1, avoiding contact with pin 2. Whichever pin is connected to pin 3 when the LED lights up is the cathode – see Photo 2. The battery holder should also be on the same side as the SMD components. Trim its leads short so they don’t protrude too far into the area where the display panel will sit, and double-check the polarity. The AAA markings were intended to allow a pair of 1.5V AAA cells to power the Human Comfort Indicator, but we have not tested this configuration, and you should use the AA markings. The last component to solder will be the sensor module, connected via the CON3 header. Ideally, this should be spaced off the PCB as far as possible and near the vents in the rear of the 3D-printed case. Use the right-angle header to achieve this, being sure to maintain clearance from the battery and its holder. You can rest the PCB on the posts in the back half of the base to check the position and clearance. You can test these components similarly to before, by connecting the display panel and USB power. Check that the display updates to show our splash screen with an “OK” message below siliconchip.com.au it. If it says “Sensor error”, it has not been able to communicate with the sensor and you should check the circuitry around CON3. You should see LED2 flash red or green within a minute if all is well. LED1 might flash momentarily, but will probably not show a true indication until the battery is fitted. Completion Now is a good time to add the foambacked tape. We opted to attach some pieces to the PCB and leave the backing sheet on the side facing the display panel. This was still sufficient to hold the panel in place without having to worry about aligning the two permanently. Slot the display panel into the bezel window, with the FFC cable curving around at the notch near the USB socket. Carefully attach the display panel to the PCB and rest the PCB in place above it. The tape should apply just enough pressure to keep the display panel in place. Install the battery and check that the display operates as previously. Snap the back of the case on. You might like to add some small rubber feet to the lower four corners Screen 1: the default main screen shows information in portrait. This design was tested in late February and it is a bit sticky, as the dew point suggests! siliconchip.com.au If a bicolour LED, connected as shown here, lights up then the lead on the middle pad is the cathode (for the green element, in this case). of the case, since it is quite small and light. This should prevent it sliding around when connected to a USB cable. Using it For the most part, the Human Comfort Indicator should be working as intended from power-on. Screen 1 shows a typical display. As noted earlier, the Human Comfort Indicator can also show the daily, weekly, or monthly average statistics. This is done on the main page by pressing S3 or S2. The screen will refresh and show the respective averages if it has recorded enough valid values. Screen 2: one page of settings is for configuring display preferences. The last item allows all settings to be permanently saved to flash memory. If you see dashes displayed instead of numbers, there may not have been enough values recorded to make up the average. As expected, you will not see a monthly average until a month has expired. You might also see dashes if there is a problem with the sensor module. Changing settings Screens 2-5 show the settings and other options. All setting changes are effective immediately. They can also be permanently saved to flash memory, which will mean that those settings are restored after a power cycle. The main choices for the display are the orientation and colour scheme: black text on a white background or vice versa. The portrait and landscape settings should suit the two different 3D printed case variants. You can also choose a reversed (rotated 180°) option if you prefer. Another setting is a choice between °C and °F for temperature displays. Internally, all temperatures are stored as units of 0.01°C and converted where needed. You can also choose whether a full screen refresh occurs every five minutes (always), every hour, or every day (see Screen 2). Screen 3: the other settings page relates to the alarm outputs. Whether the alarm is triggered for a low or a high dew point can be set independently. Australia's electronics magazine Screen 4: the white text on a black background looks striking; it would be very impressive if used with a black 3D-printed case. June 2026  51 Screen 5: the landscape format lays out the main data screen in this fashion; the settings screens put the items into two columns. There are also some alarm settings: the minimum and maximum dew point and whether either threshold is enabled (or neither, or both). You can also set the alarm to trigger if a problem is detected with the sensor. Screen 3 shows these; they are all enabled by default. Because of the slow update rate of the e-paper panel, the settings menus operate slightly differently to those you might have seen in our other projects. Pressing S1 (closest to LED2) enters the settings menus and the alarms are shown first. LED2 is lit solidly while in the settings menus. Each button press is followed by a one-second delay before being acted upon. This allows multiple button presses to occur before the screen is refreshed, so it is easier to make bulk changes. LED2 switches off to indicate that a refresh is pending. Screens 2 and 3 show the cursor marker that indicates which item is being edited – S1 skips to the next option. For numeric values, S3 decrements them and S2 increments them. For binary options, either S2 or S3 will toggle the state. For example, to change the dew point maximum from 20°C to 15°C, you would press S1 once to enter the settings, then allow the screen to refresh. Press S1 three times quickly to jump to the DP maximum setting. Allow the refresh to happen to make sure you are changing the right item. Then press S3 five times to drop the setting by 5°C. Let the refresh occur and check that the value is correct. Then press S1 six times to jump forward to the flash memory options, then press S3 to save and check that you see the “Save done” message. Finally, press S1 to return to the main screen. Table 2 summarises the settings. If the settings appear to be corrupted, you can use Restore (S2) on the last settings item to reload the initial settings from flash memory. You should immediately use S3 to save these to flash memory to be sure that everything is as new. Since LED2 is on for the settings screens, it uses much more power than the main screen. So if 30 seconds elapses without a button being pressed, the Human Comfort Indicator returns to the main screen displaying temperature, humidity and dew point. Summary The Human Comfort Indicator will be a great addition to any household that needs to keep track of the local humidity and dew point. We’re sure readers will find interesting applications for the alarm and analog voltage outputs. The firmware and STL files can be found at www.siliconchip.com. SC au/Shop/6/3630 Back Issues The UK ’s Circui t Pr Electractical onics fT sTa r TE r premier electron ics so and Underst Surgery using anding and gyrato rs Make computin g ma ker ma Mrom Finishin h Mic E Th g gazine light con the PicoMite ite E ‘sP La troller T’ aT Audio softwar smart e Designi Out sw iTc h- on discreteng a practic ! al audio op am a Mi p GPS-Sy Analog nchronise ue Cloc d k it wit Ta 6- d Ec WIN M mTo ICRO A CH u Dev ch AR IP elo 10 adE pm 00 LLionPico r Es Kit ent PrEc MisE VaLU ite smis Ta n c Es Tolight E b ox finECont art YoUr TUnErolle dEsiG r ns WIN! Microch ip Inte grat Graphic ed sE s and Touch M CuriTE osity sT – Pa luation bUEva iLd Kit an rT 3 d Us siL ic E oU on ch r ULTiM Ec kE aT E r Jump start Mini LE Driver D Egg Tim er – eg to pe gcellen rfectio t brea n! kf ast, tim Compl PLUS! APRIL 13 Cover.in dd 1 ractica lelec practic ed etin g WideTechno range the Tal inter Cool Bea k – My tru face, Ohmmet th, you ne r truth techn Net Wo ns – Arduin er and AI o tal t work, cir o Boo rk – Rou k, pic ters, pow tcamp: new n’ mix cuit Surg www.e bo er sup Sep 202 ery, re lectron plies, TEM ards update 3 £5.9 adout, publish 9 ! U and ing.co 09 more m 9 7726 <at>p 32 5730 30 alelec APRIL 201 3 £4.40 tronics Practical Electronics is the UK’s premier electronics, computing & maker magazine. Each month has a wide variety of electronics projects suiting beginners and experts alike. It also includes many different features on topics like audio, radio, computers and more. 14/02/2 013 10:33:4 7 24 YEAR COLLECTION OF PRACTICAL ELECTRONICS Every issue of Practical Electronics published from January 2000 to December 2023. Covering 288 magazines & over 20,000 pages of content; see siliconchip.com.au/Shop/3 PDF Download SC7650 ▸ $165 The articles can be downloaded per month, year or the entire block. Some of the excellent columns of Practical Electronics include Audio Out, Circuit Surgery, Techno Talk and more. The download size is approximately 5.4GB. 52 Silicon Chip PDFs on USB SC7645 ▸ $180 plus postage cost Supplied on a 32GB Silicon Chip branded USB flash drive. Purchasing this also gives you access to the download version shown opposite. Australia's electronics magazine siliconchip.com.au Subscribe to MAY 2026 ISSN 1030-2662 05 The VERY BEST DIY Projects ! 9 771030 266001 $14 00* NZ $14 90 INC GST INC GST AMPLIFIER CL IPPING INDICATOR prot ect your loudspeakers from being overdriven Analog Computers how they differ from digital comp uters, and are they making a come back? microDCC Decoder our smallest decoder yet Australia’s top electronics magazine Despite its size, it has two 100mA Silicon Chip is one of the best DIY electronics magazines in the world. Each month is filled with a variety of projects that you can build yourself, along with features on a wide range of topics from in-depth electronics articles to general tech overviews. function outputs and sound output Rosehill Ga 3 – 4 June rdens Event Centr e 2026, se e page 40 Published in Silicon Chip If you have an active subscription you receive 10% OFF orders from our Online Shop (siliconchip.com.au/Shop/)* Rest of World New Zealand Australia * does not include the cost of postage Length Print Combined Online 6 months $77.50 $87.50 $55 1 year $145 $165 $105 2 years $270 $305 $200 6 months $95 $105 1 year $180 $200 2 years $335 $370 6 months $115 $125 1 year $220 $240 All prices are in Australian dollars (AUD). Combined subscriptions include both the printed magazine and online access. 2 years $410 $445 Prices are valid for the month of issue. Try our Online Subscription – now with PDF downloads! Simple LC Meter; May 2026 Amplifier Clipping Indicator; May 2026 The History of Intel; Feb 2026 – Apr 2026 PicoSDR Shortwave Receiver; April 2026 An online issue is perfect for those who don’t want too much clutter around the house and is the same price worldwide. Issues can be viewed online, or downloaded as a PDF. To start your subscription go to siliconchip.com.au/Shop/Subscribe Installing a Whole-House Sound System Feature by Julian Edgar W hen my wife and I decided to build a new home, there were lots of decisions to be made – literally, hundreds of them. From décor to bathroom design, where to put power points and how high the ceilings should be. As part of the research to answer those questions, my wife and I went to a huge number of new display homes, looking at what features were available and how they were implemented. Unexpectedly, one display home really caught my eye – or more accurately, my ears. This home had ceiling-­ mounted speakers throughout the house, and additional speakers in an outdoor living area. They were playing quietly, and as we walked through the house, it started to change my mind about music in a home. Previously, in all five houses we’ve lived in, the main sound system has been in the lounge room – a traditional stereo hifi system. To get the best quality sound, you sat in a designated area and listened to the music. Except, we never actually did that! Instead, it was more likely that we’d crank up the system while entertaining or even when cleaning the house. We’d not be sitting in one place but instead moving around. It was bad for stereo imaging, and even for hearing tweeters, but it was how the sound system was actually used. So the idea grew in my mind: in the new house, let’s make the sound system a whole-of-house design. There would be inevitable trade-offs in sound quality – but the sound system would be much more practical and useful. Choosing the speakers Three amplifiers, one preamp, thirteen speakers, hundreds of metres of wiring – it all sounds a nightmare! So why do it? There are plenty of good reasons... Having decided that there would be speakers located throughout the house, the next decision was on the type of speakers to use. To minimise the use of floor space and to be aesthetically acceptable, the interior speakers needed to be mounted either in the ceiling or in the walls. These days, with home theatre systems often requiring numerous speakers, speakers designed to be mounted in walls and ceilings are widely available. Most have a glaring deficiency: they comprise bare drivers mounted on a faceplate, with that faceplate then fitted to the plasterboard. The resulting speaker ‘enclosure’ is just the random volume behind the faceplate. Australia's electronics magazine siliconchip.com.au Image source: https://unsplash.com/photos/a-room-filled-withlots-of-different-types-of-speakers-vAHw2myA0EM 54 Silicon Chip Photo 1: seven Bose DS 100F ceiling speakers are used throughout the house. They use a ported enclosure with a 133mm long-throw woofer and a 57mm mid-range/tweeter and suit both 8W and 70/100V systems. These are of much better quality than Bose consumer speakers. Photo 2: the wall in the lounge/kitchen backs onto the roof loft space. From left to right, you can see one Bose speaker, an air intake grille, the two vents for the subwoofers (with a ventilation intake grille in between), another (currently unused) air intake grille, and the second Bose speaker. Another five Bose speakers are distributed throughout the house. I would have chosen white speakers if I had bought them new, but I don’t mind the contrast. With wall-mounting, that volume might be quite small; just the volume between the noggins and studs. With ceiling mounting, that volume might be effectively infinite – the whole volume of the roof space! Thus, such speakers need to be designed with huge compromises, and they cannot use the most efficient common enclosure design: ported bass-reflex. Therefore, I resolved that any speakers would need to have their own acoustic enclosures, meaning that, because of the resulting greater depth, they would usually need to be mounted in the ceilings. I figured I needed up to ten high-­ quality ceiling speakers, and looking at new prices showed that this could very quickly become a huge cost. At this point, I started searching for what was available second-hand, and I found someone selling a complete ex-restaurant sound system. It comprised seven Bose DS 100F ceiling speakers (Photo 1), six Bose DS 16F ceiling speakers, a large Bose DXA 2120 amplifier, three smaller IZA 190-HZ Bose amplifiers, a line-level audio splitter box and various remote controls and cables. The price was very good – in fact, less than the new price of just the big amplifier! (As it later turned out, that was just as well.) To put it mildly, Bose has a variable reputation amongst audiophiles; many hate the brand due to its heavy reliance on processing and compensation to make small speakers sound good. However, in professional sound reinforcement, the story is different. I’ve had a lot of experience with different Bose gear, and I find two things. First, their professional gear is better-­ built than their consumer gear. And second, if you buy their upper-tier products, they can sound quite good. In this case, before committing the gear to the new house, I extensively tested the larger DS 100F speakers working with the DXA 2120 amplifier – and I was impressed. The speakers use a ported enclosure with a 133mm long-throw woofer and a 57mm midrange/tweeter. With a relatively small enclosure volume, and with those drivers, the speakers were never going to sound as good as full-size, dedicated hifi speakers – but then again, in this application they didn’t need to. Still, they had bass down to about 100Hz (more on bass in a moment) and the treble was adequate, although not wonderful. This was with the speakers tested not mounted in a ceiling: if mounted in a ceiling near to a wall, you’d expect the bass to be better due to the resulting acoustic loading. In addition to the ceiling speakers, we also wanted a pair of speakers mounted outside to service a large undercover deck. This area could also have used ceiling speakers, but the Bose speakers are not weatherproof, and while such speakers wouldn’t be subjected to direct rain, they would be subjected to fog and dew. siliconchip.com.au Australia's electronics magazine So I decided to keep an eye out for good quality exterior speakers. I eventually found a pair of second-hand Niles OS7.5 wall-mounted speakers for an excellent price. The amplifiers The Bose DXA 2120 amplifier was going to be used to drive the DS 100F speakers. But how many speakers could it power? These speakers have a rated impedance of 8W; the minimum impedance this amplifier is happy working with is 4W. Thus, each of the two stereo channels can drive two speakers in parallel, giving a total of four speakers. After a bit of thought, this worked out well. The rectangular-shaped house has two large rooms, one at each end of the house. These rooms have cathedral (raked) ceilings, with internal walls that back onto a roof loft space (Photo 2). If two of the speakers were placed through these walls, we’d have a stereo pair in each room. Within the previously discussed constraints, at least if sitting in these rooms, the stereo imaging should then be good. What about the other speakers? The two end rooms are joined by a wide, open corridor that passes down the middle of the house. Locating the remaining three DS100F speakers in this space would largely cover the rest of the house – there wouldn’t be speakers in every room, but you’d be able to hear sound everywhere. June 2026  55 But how do you drive three speakers from a stereo source? It didn’t make much sense trying to do so! So back to the remaining three Bose IZA 190-HZ amplifiers (Photo 3). These amplifiers all have 70/100V outputs. If you’re not familiar with this approach, each speaker to be driven from such an amplifier is equipped with a transformer with multiple taps, with each tap giving a different power rating (and so speaker loudness). The DS100F speakers were equipped with such transformers, and different power ratings could be selected by turning a knob. A conventional 8W speaker impedance was one setting, explaining how these speakers could also be used with the main amplifier. Multiple speakers can be driven from each output with such a 70/100V system; rather than worrying about impedance, you just need to ensure that the total selected speaker power ratings do not exceed the power output of the amplifier. The IZA 190-HZ amplifiers are also able to be configured with a mono output from a stereo source, making them suitable for driving the three speakers. Only one amplifier was needed to drive these speakers – which is just as well, as it turned out that two of the three second-hand IZA 190-HZ amplifiers were defective! A 70/100V system has a reduced frequency response, but in our case, where these speakers are used to ‘fill the hole’ between the speakers in the end rooms, their more limited frequency response is well masked by the 8W speakers, which are still audible from those locations. All about the bass Now, back to the topic of bass. For our previous house, I’d built two large subwoofers. Each used a JBL 15-inch (380mm) driver and a 200L ported enclosure (Photo 4). I’d mounted them beneath the floor of the lounge room, firing through floor grilles. They’d worked exceptionally well, so I resolved to move them to the new house. There, they could be mounted in the loft space, passing their sound through two wall-mounted openings. As they would also be used for the lounge home theatre sound system (not covered in this article), they would provide bass at only one end of the house – an acceptable trade-off as bass is largely non-directional. But what to power them with? In the previous house, I’d used a four-­ channel amplifier I had built. It had an output of 68W/channel and drove the subs and two full-range speakers. In that small lounge room, the subs’ output had been fine, despite the relatively low amplifier power. However, testing in the new house soon showed that the much bigger room seemed to suck all the bass away – the huge subs sounded quite anemic. So I then built a 200W per channel amplifier, and the subs came alive. Photo 3: a 90W Bose IZA 190-HZ amplifier drives three speakers using a 100V system. Photo 4: two subwoofers are used, each with a 200L custom-built enclosure and a 15-inch, 300W JBL driver. There is audible bass down to 25Hz and strong bass from about 35Hz. 56 Silicon Chip Australia's electronics magazine Photo 5: one of the two outdoor subwoofers for the deck; it is built inside a fibre cement stool. They provide bass support for two Niles wall-mounted speakers. siliconchip.com.au However, in all of this, I had forgotten the outside speakers – surely I didn’t need yet another amplifier to drive them, did I? Instead, I decided to use relay switching to select between driving the inside or outside speakers from the same amplifier. About this time, I developed the Outdoor Subwoofer published in the June 2025 issue (siliconchip.au/ Article/18313) and the improvement it gave to the outdoor sound system audio quality was major. So much so that I decided to build two of these subs (Photo 5). Now things were getting complicated – to switch from inside to outside sound, I needed to simultaneously switch the output of three different amplifiers! But again, Silicon Chip came to the rescue, and we developed the Remote Speaker Switch that was published in the January 2026 issue (siliconchip.au/Article/19561). It’s a very flexible system and does this job neatly using three interlinked switches (Photo 6). Photo 6: the Silicon Chip Remote Speaker Switch is used to simultaneously switch the output of three amplifiers between different speakers, with three modules required. The preamp The next step was to consider how the amps were to be fed signals, and where all these amplifiers were going to be mounted. Looking at the second point first, I decided to place the three amplifiers on a shelf in the roof space loft. The loft is accessed by a dropdown ladder, with the area used primarily for storage. If the amplifiers were remotely switched, and if volume control was also able to be achieved remotely, there was no need to have the amplifiers located in the living space. Making this decision easier was the fact that we would stream music from our phones via Bluetooth, with the volume also controlled by the phone. So, a series of plywood shelves were made, and the three amplifiers installed in the loft. How to feed signals to all these amps? The ex-restaurant sound system installer had faced the same dilemma and had used the powered line-level signal splitter to divide the source signal between the amplifiers. However, he or she had not had an extra subwoofer amplifier to contend with. Not only did the audio signal have to be sent to the different amplifiers, but a subwoofer crossover needed to be implemented as well. Here I took a step that is unusual in home sound siliconchip.com.au Photo 7: unusually in a home system, a Clarion EQS755 car sound preamp is used. This drives the three amplifiers from a single Bluetooth streaming input adaptor, provides the subwoofer crossover and allows frequency equalisation on seven bands. It runs from a 12V DC plugpack. systems and decided to use a preamplifier designed for a car: a Clarion EQS755 powered by a 12V plugpack (Photo 7). I’d used it in our previous house, and it had worked well. Now out of production, it has surprisingly good audio specifications, unbeatable at the price. It has a 7-band equaliser and a variable-level subwoofer output, with the sub crossover frequency able to be set at either 60Hz or 90Hz. In addition, it has a further two full-range outputs, for car use, dubbed ‘front’ and ‘rear’. However, in my application, I could use these outputs to feed the two Bose amplifiers. As a bonus, using the preamp would give me a seven-band EQ. Disaster! As an owner-builder, a lot of the house was being built by me, and I Australia's electronics magazine wanted to listen to music as I toiled. Therefore, the whole-of-house sound system was running quite early in the house build process – as soon as the house was made weatherproof. The ceiling speakers were mounted temporarily on plywood panels. It’s just as well the system was trialled before the house was completed, because I started experiencing major amplifier failures! First to go was the home-built subwoofer amplifier. It used two offthe-shelf modules, each comprising a switch-mode power supply and a Class-D audio amplifier. This amplifier lasted just weeks before it started blowing supply fuses at switch-on. I could have attempted to fix it, but truth be known, I was a bit suspicious of the quality and design of these modules, anyway. June 2026  57 Photo 8: two LD Systems amplifiers are used to drive the main and subwoofer speakers – an XS-400 (2 × 200W) for the main speakers and an XS-700 (2 × 350W) for the subwoofers. These fan-forced commercial amplifiers were purchased after multiple amplifier failures – the demands of providing a lot of power and working in a hot loft space proved to be extreme. When John Clarke saw them, he was scathing about their design layout! So I replaced this amplifier with another, more conventional design I’d built long ago, an amplifier that I was sure was of good quality. Next to go wrong was the big Bose DXA 2120 amplifier. This one developed a ‘splat! splat!’ in the audio output, then just died completely. What on earth was going on? My replacement amplifier for the subwoofers gave me the clue. Despite this amplifier having run in my previous home office for years, and despite my having used it to occasionally test loudspeakers at quite high power levels, the passively cooled heatsinks had never run warm. But here in the loft, driving the subwoofers, it was getting darn hot – over 55°C external heatsink temperatures... and still rising. Two aspects were at play. The first is that the ambient temperatures at which the amplifiers were working were high. In a normal domestic setting, ambient temperatures are seldom very high; after all, that’s why people have air conditioning! But in the loft, within the roof space, temperatures were up to 35°C; potentially even higher in some weather conditions. The second aspect was that the amplifiers were working really hard. Especially with the house unfinished, and so without plasterboard walls and ceilings in place, the house volume was very large. To get adequate sound levels in such a large volume meant using a lot of power. Put those two things together, and the amplifier working conditions were extreme. So I decided to buy two new amplifiers, both fan-cooled and both aimed at commercial (rather than domestic) use. After an extensive search, I found two LD Systems amplifiers – the XS-400 and XS-700. The XS-400 has an output of 2 × 200W into 4W, and the XS-700 develops 2 × 350W into 4W – see Photo 8. Both are Class-D amplifiers that have a maximum distortion of less than 0.1%. Not hifi, but good enough for a whole-of-house sound system. The amplifiers were bought second-hand, and while it was an additional cost I wish I didn’t have to outlay, they were much cheaper than buying similar quality amplifiers new. As you can imagine, I tested these amps very thoroughly, only to find that certain internal components were running very hot! However, a simple modification to the airflow path within the amplifiers, as described in my articles on amplifier cooling published in the August & September 2025 issues, improved this greatly (siliconchip.au/ Series/444). To give you an idea of the required power levels, I found that the 350W/ channel amplifier driving the two subwoofers was quite near clipping at times. The amp has a clipping indicator on its front panel, so it’s easy to see how hard it is working. The final system Fig.1: a Bluetooth adaptor feeds a preamp/subwoofer crossover that in turn feeds three amplifiers. These power internal and external speakers, both main and subwoofers, with inside/outside selection via three interlinked relay switches. A mixture of 8W and 100V systems is used for the main speakers. Fig.1 shows the layout of the final system. A Bluetooth adaptor feeds the Clarion EQS755 preamp. The preamp also drives the LD Systems XS-400, XS-700 and Bose IZA 190-HZ amplifiers. These power the four 8W speakers as two stereo pairs, and the three mono 70/100V speakers located down the middle of the house. The preamp’s subwoofer output also drives the LD Systems XS-700 subwoofer amplifier, which powers the two 15-inch subs. The speaker switch turns off all the interior speakers and swaps the XS-400 output to the outside wall speakers; simultaneously, the XS-700 amplifier’s output is switched to the two outside subwoofers. Power to the Bose IZA 190-HZ, LD Systems XS-400, preamp and the switching relays is controlled by a single wall-mounted conventional switch in the living area, with one LED indicator showing when they are active. That is, this part of the system is powered by a remotely switched power point. Australia's electronics magazine siliconchip.com.au 58 Silicon Chip Lessons learned Test at every step! Without testing, you cannot be sure what you’ll end up with. For example, until I had temporarily mounted the speakers in the half-built house, I wasn’t sure that seven interior speakers would be sufficient. Until I had tested over a period of many weeks in summer, I had no idea that amplifier cooling would prove to be so critical – and that so much power would be needed. Would the Bluetooth adaptor continue to work if I took my phone out onto the exterior deck? (It did.) Did all the gear purchased second-hand actually work? It took me many hours to discover that two of the Bose amplifiers were defective – I was chasing down blind alleys with their remote switching and volume control functions, sure that their lack of output was something I was doing wrong! (Their power LEDs came on and then slowly went off – the amps weren’t totally dead.) Finally, consider how you actually use a sound system in a house and develop a system to suit that application. Silicon Chip Binders REAL VALUE AT $21.50* PLUS P&P Are your copies of Silicon Chip getting damaged or dog-eared just lying around in a cupboard or on a shelf? Can you quickly find a particular issue that you need to refer to? Photo 9: this is a mock-up design shot of what I hope the finished house will look like, with the landscaping done. Another wall plate switch operates the LD Systems XS-700 subwoofer amplifier – it’s on a separate switch, so only this part of the system can be switched on for home theatre use. Finally, all the speaker wiring was made using 3.4mm2 cross-sectional area low-­voltage garden lighting cable – the cheapest cable I could find that had plenty of copper cross-section. It was much cheaper than dedicated speaker cabling of the same gauge. The results So, are we happy? Yes, we are. The interior system doesn’t have the subtle nuances of a good hifi system being listened to from the sweet spot – but then, with seven distributed speakers, it never could have. But it fills the house (or outside deck) with quality sound and has enough power to listen at ‘party’ levels. Conversely, it still sounds good at very low listening levels. siliconchip.com.au Despite the presence of the large subwoofers (and their 700W amplifier), the bass is set to give body and depth to the sound, rather than bump thump. In the two main rooms, the stereo imaging is clear, and when walking from these rooms to the hallway, the transition from stereo to mono is unnoticeable. The ability to balance the output levels of the separate amplifiers also means that the sound volume doesn’t change throughout the house. The outside speakers are ideal for listening at normal levels and, if required, can work at much higher levels (however, even in this semi-­rural valley, that can annoy neighbours). The best aspect of the system is that operation is seamless; you just flick the wall switches, connect your phone via Bluetooth, and play music. The system was complicated in development and implementation, but in use, SC it just works. Australia's electronics magazine Keep your copies safe, secure and always available with these handy binders These binders will protect your copies of S ilicon C hip . They feature heavy-board covers, hold 12 issues & will look great on your bookshelf. H 80mm internal width H Silicon Chip logo printed in goldcoloured lettering on spine & cover Silicon Chip Publications PO Box 194 Matraville NSW 2036 Order online from www. siliconchip.com.au/Shop/4 or call (02) 9939 3295 and quote your credit card number. *see website for delivery prices. June 2026  59 USB Power Monitor Kit (SC7683, $50): includes the PCB and all onboard parts U Simple S Power B Monitor By Richard Palmer This circuit uses just a handful of parts but it can measure USB voltage, current and power over a wide range (up to 36V, 3A & 108W). It displays the readings on an OLED screen. Displays USB bus voltage, current, power and energy delivered Simple and detailed display formats Supports USB 2.0 and 3.0 power delivery up to 3A/36V Display settings remembered between sessions ±0.3% voltage and ±1% current accuracy Display rotation at the touch of a button Resolution: 100μA, 10mV Serial data logging W hile USB-C is slowly taking over, there are still many devices using USB-A plugs and sockets. While our recent USB-C Power Monitor (August 2025; siliconchip.au/ Series/445) is also capable of measuring USB-A devices with appropriate adaptors, this project is substantially simpler – it is basically an update on the December 2012 USB Power Monitor (siliconchip.au/Article/460) with higher resolution, modern components and the ability to measure the wider range of voltages and currents when the device uses a USB power delivery (PD) mode. The monitor has two display modes. The default shows the USB bus voltage and current flow in large characters. A second mode, with smaller text, adds the power and the energy that has been delivered while the unit has been powered on. The display can also be flipped upside-down, if required for more convenient reading. The unit also provides a TTL-­ compatible serial logging output, shown in Screen 1, which avoids the need for copying down a long series of readings when monitoring over an extended time period. the many variations and how they evolved. In summary, a standard 4-pin USB 1.0 port can supply up to 500mA at 5V. USB 2.0 upgraded the 4-pin standard to 1A. USB 3.0 introduced power delivery and a 9-pin USB-A connector, which is backwards-compatible (ie, four of the nine pins are in the same locations and have the same functions). USB-C PD can deliver up to 5A at 20V after negotiation between the source and sink over the CC (Configuration Channel) pin using BMC signalling. When no negotiation occurs, Component selection USB connectors and power delivery There are a range of USB port types, each with its own power capabilities. The Wikipedia article on USB hardware (https://w.wiki/3oc8) details 60 Silicon Chip 5V is supplied. With USB 3.1, which is exclusively delivered via USB-C connectors, up to 5A at 48V can be delivered. In suspend mode, when the PC or laptop is asleep, the host controller stops sending keep-alive (Start of Frame) signals to devices. This tells the connected device to go into idle mode, with a reduced allowable current draw. A USB device may adopt very different power profiles, depending on which kind of power source and cable are used, and whether the host is active or asleep. Screen 1: example output of the serial logging feature in the Arduino Serial Monitor. Australia's electronics magazine The monitor uses just three ICs, an OLED display and a handful of passive components. The SSD1306 128×64-pixel 0.96inch (24.4mm) I2C OLED screen’s footprint determines the size of the PCB, with additional strips at the edges to accommodate the tactile switch (S1) and USB connectors. The USB connectors, CON1 and CON2, are soldered directly to the PCB. The 9-pin USB-A plug and socket are compatible with USB 1.0, 2.0 and 3.0 devices. SMD connectors were selected because very few 9-pin through-hole versions are available, making reliable component sourcing difficult. Voltage and current measurements are made by an INA237 power monitor siliconchip.com.au USB 3.1 PD Warning The unit may be damaged with USB 3.1 power delivery modes sourcing more than 36V. This can occur when connected to a USB-C power source designed to deliver more than 180W, which permits up to 48V to be supplied. chip, which has an internal 16-bit analog-­ to-digital converter (ADC), offering better than ±0.3% accuracy and providing up to 85V high-side voltage and current measurements. The higher-priced INA238 chip can be substituted, as it is pin- and code-equivalent. It has better (±0.1%) current accuracy, which will not increase the overall accuracy of the unit, as the shunt resistor’s 1% tolerance is substantially greater. Through-hole power resistors with resistances of less than 1W and tolerances better than ±5% are rare and expensive, so a 1W ±1% M6332 (2512 imperial) low-resistance metal alloy (LRMA) SMD resistor is used for the current shunt (R1). LRMA resistors use a cupro-nickel alloy and have a very low temperature coefficient, in the range of ±75ppm (±0.0075%) per °C, which translates to less than 0.4% variation across a working range of 10-40°C. As the stated temperature coefficient allows for either a rise or fall in resistance when the temperature rises, we have not included temperature compensation in the current calculations. The shunt resistor value is determined by the 3A maximum current requirement and the 163.84mV full scale of the INA237’s ADC. A value of 0.025W results in 75mV across the shunt and ¼W dissipation at 3A. While 0.05W could have been used, the lower value resistor means reduced power dissipation and a lower V+ voltage drop without significantly affecting the overall accuracy. To minimise the effect of any temperature rise on current readings, we have connected the resistor’s terminals to as substantial a PCB copper area as the USB signal traces allow. As USB 3.0 power delivery mode can negotiate anything from 5V to 20V on the V+ USB pin, a 3.3V regulator with a wide input range is required to reduce the voltage to drive the circuitry. While both the microcontroller siliconchip.com.au The INA237 DC power monitor IC The INA23x series of DC power monitors offer current, voltage, power and die temperature measurement (see the diagram below). It has a 16-bit ADC that is shared by voltage, current and temperature measurements. The chip can operate at 3.3V or 5V, is controlled over an I2C serial bus and consumes less than 1mA. Two address pins allow up to four chips to coexist on a single I2C bus. Other products in the INA2xx series come with higher-accuracy 20-bit ADCs or SPI control interfaces rather than I2C. The current shunt resistor can be located on the high or low side of the load, as the current input’s common voltage range is -0.3V to +85V. The chip has two ranges for measuring the shunt voltage, 42mV and 164mV full-scale, providing flexibility in choosing the shunt resistor’s value to best balance measurement accuracy and heat generation. As the chip has a very low input bias current, accurate current measurement from microamperes to kiloamperes is possible anywhere in the permitted input voltage range. A digital filter rolls off the ADC response at half the sampling frequency to avoid aliasing measurement errors. The sampling time is individually adjustable for the voltage, current and temperature measurements, ranging from 50μs to 4ms. The chip can average up to 1024 samples, further reducing noise. Any necessary calculations are undertaken in the background to minimise measurement lag. The INA237 also has an alert pin, which changes state when any desired combination of current, bus voltage, power or die temperature goes outside set limits. The INA237 chip includes a shared ADC that measures voltage, current and temperature. The embedded processor can average up to 1024 readings and calculate the power figure. and power monitor chip can operate at either 3.3V or 5V, 3.3V was selected to provide headroom for this regulator when operating from a 5V supply. The MIC5233-3.3 regulator was selected as it has a small SOT-23-5 footprint, can handle the required input voltage range and is readily available. It is connected to V+ on the upstream USB connector to avoid the unit’s current consumption being registered by the INA237. Like with the December 2012 design, no case is required. Instead, the unit is protected by enclosing it in a length of clear heat-shrink tubing. Circuit description The USB Power Monitor circuit is shown in Fig.1. Regulator REG1 reduces the input USB input voltage, Australia's electronics magazine V+ on CON2, to 3.3V for the ICs and display. Its heatsinking requirements are not substantial, as the circuit only draws a few milliamps. It can be disconnected at JP1 while the microcontroller is being programmed. Power and data signals travel between the two USB connectors, CON1 and CON2, with the V+ line interrupted by shunt resistor R1 (25mW). This resistor translates the current consumed by the device under test to a voltage, which is captured by the power monitor chip’s Vin+ and Vin− terminals. The USB V+ voltage is measured on the CON2 (output) side, so that the voltage available to the device under test is correctly displayed. An ATtiny85 microcontroller operating at 16MHz drives the power June 2026  61 A TTL serial-to-USB adaptor (with the black PCB) can be connected to the Power Monitor for logging power-on time, bus voltage, current and more. monitor chip and display. It has a Universal Serial Interface (USI) that can be configured as an I2C or SPI port. During device programming, the SPI mode is used, while under normal operation, the USI is in I2C mode to communicate with the OLED and INA237 ADC chip. The ATtiny85’s I2C interface has SCL on pin 7 and SDA on pin 5. Pull-up resistors for these pins are provided by the OLED display. The I2C bus runs at 400kHz, at which speed all the required traffic is completed well within the two-second display update cycle. The I2C pins are shared with the SPI interface used for in-circuit programming (via CON3), which is initiated when pin 1, RESET, is pulled low by the programmer. When the unit is booted normally, the code sets the pins to I2C mode. Once programming is complete, the MISO SPI signal (PB1 at ATtiny85 pin 6) is no longer needed in I2C mode. It is re-assigned as the unused RxD (receive) serial pin for serial log data, with the TxD (transmit) logging data coming from pin 2 (PB3). Software When power is applied, the microcontroller initialises the OLED display, which is expected at I2C address 0x3C. The power monitor chip, IC2, has an address of 0x40 with its A0 and A1 pins tied to ground. The resistance of the shunt and the maximum current to be measured are provided to the power monitor chip’s driver software, which calculates the calibration value for the chip’s SHUNT_CAL register using the value of the shunt resistor. Voltage calibration is inbuilt. The chip is set to sample the current and voltage every 280μs and average them over 1024 readings, which provides a final set of readings at approximately 300ms intervals. It provides measurements in signed 16-bit integer format, which are converted into floating-point measurements in software. The display updates once every two seconds. Fitting the display driver and font into the available memory proved challenging. The microcontroller only has 8kiB bytes of flash memory program space and 512 bytes of RAM. A full font would consume more than the total flash, but for this project, we don’t need the full ASCII character set. So large and small fonts were created that contain only the characters 0-9 plus the decimal point, space, “V”, “A”, “W” and “m” characters. The usual practice of creating a bitmap in memory and then copying it to the display would have taken 1024 bytes of RAM. Instead, each line of characters is converted to bitmap format and written directly, in rows of eight pixels at a time, to the display. These 8 × 128 pixel ‘pages’ create a restriction that a new line can only start at the top of a new page if the previously written data isn’t to be overwritten. In practice, this means eight rows of tiny characters, four rows of medium-sized characters, or two rows of large characters. Only the two- and four-row options are used. The tactile switch (S1) is sensed by the ATtiny85 on pin 3 with the GPIO port’s pull-up current enabled. It is checked once per display cycle. A press lasting one display cycle changes between the two display modes, while a two-cycle press rotates the display by 180°. Fig.1: the device uses three integrated circuits, one OLED display module, a shunt resistor and not much else. The circuit simplicity is mainly due to the features of the INA237/238 being a perfect fit for our needs in this application. 62 Silicon Chip Australia's electronics magazine siliconchip.com.au Serial logging is accomplished by driving the serial port in software (bit-banging) as the hardware USI is occupied with I2C communication for the OLED screen. The signal is at 4800 baud with 3.3V TTL-compatible levels. A TTL serialto-USB adaptor and a suitable terminal program can be used to view and save the data. Only the RxD and GND pins on the serial adaptor need to be connected to the USB monitor. The data is comma-delimited, including the time since power-on, USB bus voltage and current written in the same format displayed on the screen. PCB design To ensure minimal voltage drop in the USB ground return path, the USB GND pins are connected to the PCB ground plane as well as having pointto-point copper traces. The ground plane has been removed under the USB signal traces to minimise parasitic capacitance, which can degrade high-speed signals. All the signal traces are of equal length to minimise relative phase shifts. Parts List – Simple USB Power Monitor 1 double-sided PCB coded 04104261, 44 × 29mm 1 128×64-pixel monochrome 0.96-inch I2C OLED module [SC6176/SC6936] 1 Würth 692112030100 9-pin USB 3.0 SMD plug (CON1) 1 Switchcraft RAHUA30E 9-pin USB 3.0 SMD socket (CON2) 1 2×3 pin header (CON3; optional, for ICSP) 1 3-pin header (CON4; optional, for serial logging) 1 4.5 × 4.5mm, 5mm tall SMD tactile pushbutton switch (S1) [Altronics S1112A] 1 6-pin vertical 2.54mm-pitch pin header (for mounting OLED module) 1 50mm length of 35-50mm wide (measured flat) clear heatshrink tubing 1 USBasp programmer with 6-pin adaptor and IDC cable (optional; for ICSP) [Jaycar XC4627 + XC4613] Semiconductors 1 INA237 or INA238 power measurement IC, VSSOP-10 (IC1) [Mouser 595-INA237AQDGSRQ1, DigiKey 296-INA237AQDGSRQ1CT-ND] 1 ATtiny85V-20PU microcontroller programmed with 0410426A.HEX, DIP-8 (IC2) [Altronics Z5105, Jaycar ZZ8721 (both supplied blank)] 1 MIC5233-3.3YM5 or MIC1793-330OT LDO 3.3V linear regulator, SOT-23-5 (REG1) [Mouser MIC5233-3.3YM5-TR] Capacitors/resistors 2 4.7μF 25V SMD M2012/0805 X7R multi-layer ceramic capacitors 1 100nF 50V SMD M2012/0805 X7R multi-layer ceramic capacitor 1 0.025W ±1% 1W+ M6332/2512 LRMA SMD current-sense resistor [Mouser LRMAP2512-R025FT4] Construction All components mount on the 44 × 29mm PCB, which is coded 04104261 – see Fig.2. The OLED screen, tactile switch and USB connectors fit on one side, with the remaining components on the other. The ATtiny85 chip comes in an 8-pin DIL package. While it could be socketed, that is not recommended as it will produce a bump in the heatshrink cover on the bottom of the unit. Leave mounting the OLED until last, as the USB connectors and in-circuit programming pins can’t be soldered in once it is in place. If you are programming your own ATtiny85, solder it in and follow the instructions below. Programming may be done at this point or after other components have been mounted. If the regulator is in place when programming, JP1 must be broken to prevent the regulator being reverse-powered and possibly damaged. Next, install the surface-mount parts. Begin with the three ICs and then follow with the passive components. Most of the SMD parts are big enough not to present too many difficulties. We’ve covered SMD soldering on many occasions in the past, so we won’t go into detail here. siliconchip.com.au Fig.2: like the circuit, the PCB is simple and assembly is straightforward. Make sure to mount the screen last, and carefully check all the SMD solder joints, especially on IC1, before moving on to the through-hole parts. The INA237/8 power monitor chip is in a VSSOP package with 0.5mm pin spacing. If you accidentally bridge any of the pins, simply use solder wick to clean it up. A dab of ‘no-clean’ flux paste applied to the bridge beforehand makes clean-up easier. Solder in USB connectors CON1 and CON2 next. I had to lever up the shield on CON2, as the pins were unreachable with a soldering iron. Don’t worry if the shield breaks off. Once testing is complete, it can be clipped back on and a couple of dabs of solder on the top edge will hold it firmly in place. At this stage, plugging the unit into a USB power source should produce 3.3V between JP1 and either of the grounded USB connector cases. Mount the OLED display using a pin header for the connections and two single pin pieces of header strip Australia's electronics magazine soldered through the mounting holes at the other end of the display to anchor it. Before soldering, make sure that the pins on CON3 don’t foul any components on the OLED module. Cut all the OLED pins off flush on both sides. If logging is required, trim off any protruding leads from pins 2 and 4 of the ATtiny85. Remove the middle pin from the three-pin header and mount it on the underside of the PCB, parallel to CON2. Extra solder pads have been provided to make the connections more robust. Programming the ATtiny85 If you haven’t purchased a pre-­ programmed ATtiny85, you will need programming hardware and software. While there are many options available, I have found the following to be straightforward and reliable on June 2026  63 Screens 2 & 3: the main Zadig screen with the USBasp device selected and libusbK as the target driver (shown at left). Device Manager showing that the USBasp driver has been successfully changed to libusbK (shown at right). Windows. For Mac and Linux users, there are several good online tutorials for ATtiny85 USBasp programming. First, purchase the USBasp programmer (see www.fischl.de/usbasp) from your favourite source. Make sure it has a 6-pin socket and IDC cable or includes a 10-pin to 6-pin adaptor. For Arduino users, complete code is also included in the download pack. Board and device settings are listed at the top of the main program. The programmer to select is the “USBasp (ATTiny Core)”. Otherwise, download and install AVRDUDESS (siliconchip.au/link/ acb4), which includes the AVRDUDE command-line programming software. Next, download and install Zadig (https://zadig.akeo.ie). Plug in the USBasp programmer and run Zadig (Screen 2). If USBasp doesn’t show in the device field, click on Options → List All Devices and select it from the list. Select libusbK from the dropdown list that the green arrow points to, and click on the Install (or Reinstall) Driver button. Wait for the process to complete. Now if you open Windows Device Manager, you should now see an entry for libusbK USB devices, similar to the one in Screen 3. Programming is undertaken with the ATtiny85 mounted on the PCB and the programmer connected to the ICSP header. No USB cables should be connected to the USB Power Monitor while programming. Everything is now ready to program the ATtiny85. If your USBasp programmer has a voltage selector jumper, choose 5V. Open link JP1 on the Power Monitor board to prevent reverse-­ powering REG1 during programming. Connect the USBasp and Monitor boards via the 6-pin connector. Pin 1 (red stripe on the cable) is marked with a white dot on the PCB. Plug the USBasp into a USB port on your computer and run AVRDUDESS (Screen 4). Select “usbasp-clone” from the list of programmers and select “ATtiny85” from the microcontrollers list. Locate the HEX file from the download package (siliconchip.au/ Shop/6/3621) using the “…” button to the right of the Flash field. Leave the EEPROM field blank but change the “Fuses & lock bits” settings to L = 0xF1, H = 0xD7, E = 0xFF, LB = 0xFF. The AVRDUDESS window should look similar to Screen 4. Click on the Write button next to the Fuses & lock bits settings, then click Program! The console panel should show progress, ending with a message indicating that the flash memory or fuse bytes have been verified. Disconnect the programmer and re-­solder JP1 to restore the power supply from REG1. The Power Monitor is ready for use. Testing The unit can now be fully tested. Plug it into a power source. The display should light up after a second or so, displaying close to 5V and 0.0A. Screen 4: this AVRDUDESS window shows the programmer type, target microcontroller and HEX file selected. Note the values in the “Fuse” and “lock bits” settings. These need to be written to set the clock speed correctly. The larger PCB is one version of the USBasp-based AVR programmer. Since this one has a 10-pin socket, you’ll need a 10-pin to 6-pin adaptor (shown adjacent), or to purchase one with a 6-pin socket. Australia's electronics magazine siliconchip.com.au 64 Silicon Chip Press and hold the tactile switch; the display should change format after a display cycle. Plug any USB device into the measurement port. The displayed voltage should drop marginally, and the current reading should show a non-zero value. Calibration is not required. If all is well, the heatshrink sleeve can be added and shrunk on the outside edges. If header pins were fitted to CON3, the pins should be trimmed down to the plastic retainer before fitting the heatshrink tubing. While large-bore clear heat shrink tubing isn’t readily available from the main Australian suppliers, eBay and AliExpress both have suitable products. Be careful not to shrink the tubing too tightly, as the screen rotation switch can become permanently depressed. If this happens, cut a circle in the heat-shrink tubing around the switch’s plunger. Operation Operation is straightforward. Simply connect the monitor to a USB power source and the device to test into the USB socket on the monitor. Initially, the unit will display the USB bus voltage and the load current in large characters. If the button is pressed for a display cycle (approximately two seconds), the display will change to also show the instantaneous power flow and the total energy that has been delivered since the unit was powered on. The display automatically switches between amps, watts and amp-hours and their milli- equivalents when the current/power/energy reading is low. If the text on the display is upside down, hold the switch down for two display cycles and it will flip. When logging, always connect the Power Monitor to its USB power source before connecting the TTL serial adaptor. Otherwise, the monitor may not operate correctly. The USBasp programmer connected to the underside of the Power Monitor (with a red colour PCB for the 6-pin adaptor this time. While an IC socket was used for IC2 here, we do not recommend using one, as it will produce a bump in the heatshrink. The Monitor attached to my vintage decade resistance box during testing. Conclusion The INA237 power monitor chip enabled this project to be developed with a high level of accuracy using only a few components. The ability to calculate the energy consumed and log readings on a computer extends its usefulness for devices such as power banks and battery chargers where the load varies SC over time. Here is the Power Monitor connected to my mobile phone from a power delivery capable charger. siliconchip.com.au Australia's electronics magazine June 2026  65 We have recently started using e-paper displays in our projects. We have looked at some modules incorporating them before, but this is the first time we’ve integrated a bare e-paper panel into a design. It was more involved than expected; this article explains what we did to make it work. By Tim Blythman F or our Human Comfort Indicator project, we decided to use an e-paper display to achieve low power consumption with a screen that can be read at any time. e-paper is readable in ambient light, so it does not require a backlight, which could otherwise have a substantial power requirement. These screens might also be called eInk or EPD, where EPD stands for electronic paper display or electrophoretic display. ‘Electrophoretic’ refers to the motion of charged or polarised particles in a liquid medium due to an electric field. The laboratory technique known as electrophoresis involves separating different molecules based on their size and electrical charge. The common factor is the presence of an electric field affecting charged or polarised particles. We reviewed a small e-paper module in the June 2019 issue (siliconchip. au/Article/11668) with a resolution of 200×200 pixels, measuring 1.54 inches (39mm) along its diagonal. It had an IL3820 controller and was configured to use an SPI (serial peripheral interface) bus. We created some demonstration code for the Arduino and Micromite. Display controller ICs for e-paper most operate similarly to LCD and OLED controllers. They have numerous internal registers to configure the device and some RAM (random access memory) that holds a representation of what needs to be displayed. The big difference with e-paper displays is that they are not constantly refreshed, like other display types. Instead, an explicit command is required to perform the refresh. It can take a second or longer, so it needs to be done under the control of the software. Since e-paper needs no power to maintain its display, the proportion of time it spends updating will dictate the total and average power consumption. The downside is that it is difficult to tell the difference between an e-paper device that is working and one that has not updated. Nearly all the e-paper modules we have seen use an SPI interface. Like many LCD and OLED controllers, many can also support I2C or parallel interfaces, although these may not be available due to the circuitry used on the module; they are usually designed for just one interface type. With low power consumption being an important aspect of the Human Comfort Indicator, we found that many modules included circuitry that made this target difficult to achieve, since they often contained unnecessary circuitry that would waste power. Thus we had to base our design on a bare display panel and provide the support Subcapsule addressing enables high-resolution capability Transparent Top Electrode Positively Charged White Pigment Negatively Charged Black Pigment Clear Fluid Bottom Electrode Fig.1: the operation of an e-paper display with particles having different colours, sizes and charges in a clear medium. Source: https://w.wiki/7qVD 66 Silicon Chip Australia's electronics magazine siliconchip.com.au Photo 1: the panel we selected includes a controller IC and a display panel integrated into a COG assembly. It connects via the FFC on the right. Source: www.buydisplay.com/serial-2-9-inch-e-paper-screen-128x296-for-electronicshelf-label-lcd circuitry needed to operate the controller. Fig.1 shows the operation of an e-paper display. With an electrode on each side of the panel, it’s easy to set a pixel by simply setting the electrode polarity. In practice, the electric field is pulsed and reversed several times to ensure that all the particles (which are smaller than the pixels) are not stuck. The type of display shown in Fig.1 uses particles with different colours, sizes and charges moving through a clear medium; some displays may use a coloured medium to further expand the range of colours that can be displayed. There are also variants that use multi-coloured, polarised particles. Since the panel uses electric fields to control the display, the support circuitry involves generating voltages above normal logic levels; the panel we are using spans +20V to -20V. Fortunately, the movement of the tiny particles in the display does not require much current. Controller operation Typically, once the circuit is powered up, you need to initialise the display controller with some configuration commands, then fill its RAM with data indicating what to display. Performing a refresh then sends the data to the panel. For low-power designs, the controller can be put to sleep between refreshes to ensure it uses the least amount of power. This requires it to be initialised again before the next refresh. The refresh process is controlled by a look-up table (LUT). This is effectively a simple program that describes how the panel electrodes are strobed, for how long, and with what voltages. siliconchip.com.au The values in the table depend on the characteristics of the panel. Some controllers can work with two-colour and three-colour panels; they will need different LUTs than monochrome panels. For the module that we reviewed in 2019, the LUT was stored by the microcontroller and sent to the controller over the SPI bus, as is fairly common. Some panels also have OTP (one-time programmable) memory that contain one or more LUT programs, which can be used directly by the controller or automatically loaded. Sometimes the controller is attached to the panel glass using a COG (chipon-glass) process. Connecting traces are made with transparent, conductive ITO (indium tin oxide) material. In this case, the OTP LUT parameters are typically programmed by the panel manufacturer, with the LUT optimised for the specific panel the controller is connected to. Many controllers that we investigated had multiple LUTs, each optimised for a specific temperature range. An onboard temperature sensor allows the controller to automatically pick the correct LUT. This is important because the temperature of the medium in the panel affects its viscosity and thus the rate at which the coloured elements move. Some (but not all) controllers also include LUTs for full and fast refresh. As the name suggests, a fast refresh is quicker than a full refresh, but it may show remnants of the previous display (a phenomenon known as ‘ghosting’). One trick that we saw in some epaper software libraries is to override the internal temperature sensor to make it use a faster LUT. In practice, we found this made a negligible difference to the speed we could Australia's electronics magazine achieve. Suffice to say that there are many factors involved in refreshing e-paper displays. The website www.buydisplay.com sells a wide range of bare display panels of various types, including LCD, OLED and e-paper, plus breakout modules. They also sell many of their items on eBay. Notably, they provide detailed data sheets and sample code for all their modules and panels, so we found it quite easy to get things working. We ordered several panels, breakout boards and modules from them to suit Arduino and Raspberry Pi boards. Many of the e-paper panels are fitted with a 24-pin flat flexible cable (FFC) connector, so they can be interchanged with other panels. With this option, we were able to try several panels easily to see what might work best in our application. Best-laid plans Several three-colour and four-colour e-paper displays are now available, and we thought they might be handy to show different states or conditions. However, these displays can take up to 15 seconds to perform a refresh, which we decided was too long. A multi-colour display consists of particles with different sizes or charges that move at different rates through the liquid layer. Thus, the controller must set each colour in turn and make sure that the other colours are not affected, significantly slowing the refresh process. A full refresh of a multi-colour panel typically involves the display flashing rapidly for a long period before settling on its final output, which would be quite distracting for a device that should sit unobtrusively in a home. Such a long refresh would also tend to use more power than a simpler panel. In the end, we decided to keep things simple and use a monochrome (black and white) display. The ER-EPD029-2B is a 2.9-inch (74mm) e-paper display using the SSD1680 controller IC. The panel has a resolution of 296 × 128 pixels, ample for the information we want to display. This controller has support for two sets of LUTs, allowing either a full or fast refresh without having to tweak any parameters. A 2.9in/74mm display is large enough to be clearly visible, with a 67 × 29mm display area (see Photo 1), and June 2026  67 the panel is quite well priced relative to its size. This controller can support multi-colour displays, but we are using the monochrome variant. There is also the ER-EPD029-2R version that supports a black, white and red display, which looks otherwise identical. e-paper circuitry The glass e-paper panel consists of the controller IC connected to the display matrix. There are 24 lines brought out via a 0.5mm pitch FFC. This is connected to a ZIF (zero insertion force) socket on both the commercial modules and our PCB. As well as providing the SPI interface for communication, there are some extra lines that need to be provided to the controller. Nine of these lines connect to 1μF capacitors and bypass various voltages for the controller. There are a few other components, too. The main reason they are connected this way is that it would not be easy to provide that amount of capacitance on the glass substrate. Fig.2 shows the circuit that is used on the breakout modules that we tested. This is about the minimum needed to support the display panel; the modules also include level-­ conversion circuitry that is not shown, so that the module can interface with Scope 1: the blue trace is the PREVGL line, the red trace is the PREVGH line, the green trace is the VGH line and the yellow trace is the VSL line (all shown in Fig.2) during a 1.5-second refresh period. microcontrollers running at different voltages. The circuit around the Mosfet and inductor is used to generate dual rails with voltages of up to ±20V. Scope 1 shows the measurements on some of these lines during a full refresh cycle on the panel. This circuitry is managed by the controller IC, including the drive to the Mosfet. The GDR line signal (not shown in Scope 1) starts out at around 300kHz with a duty cycle lower than Fig.2: the SSD1680 e-paper controller requires some support circuitry that can’t easily be mounted on the panel. That includes capacitors, a Mosfet and an inductor. 68 Silicon Chip Australia's electronics magazine 10%. Once the voltages have stabilised, the driver starts skipping pulses, presumably to limit and regulate the voltages. The PREVGH line is arranged in the standard boost configuration, with a 0.47W resistor allowing the controller to sense the rising current in the inductor to limit it safely. The PREVGL line is created by coupling the switching node to a point between the two diodes, acting as a charge pump, which brings the PREVGL line to -20V. Several internal regulators supply lines such as VSL (the yellow trace in Scope 1). You can see the various regulated voltages remaining steady despite PREVGL sagging under load. The data sheet for the SSD1680 controller indicates that the voltages are as expected and the internal regulator currents are less than 1mA. We noted that the total 3.3V supply current jumped up to about 5mA, which seems reasonable. With several internal regulators needing input and output bypassing, it’s easy to see why so many capacitors are needed. Fig.2 also shows the six control lines needed. As well as the three pins needed for a unidirectional SPI bus (SCK, MOSI and CS), there is a reset line and a data/command (D/C) selector. This is used to differentiate between commands and display data when addressing the controller. These are all common to other display controllers, such as those used for graphics LCDs. The sixth line is a BUSY signal that the controller drives when it is busy siliconchip.com.au Photo 2: nowadays there are even colour e-paper displays available. Some, like the Waveshare display shown here, only provide a select few colours, while others have access to the full spectrum. Source: www.waveshare.com/product/ raspberry-pi/displays/e-paper/3.97inch-e-paper-hat-plus-g. htm Photo 3: Amazon’s Kindle is one of the most well-known e-paper devices. e-paper devices like the Kindle have very low refresh rates, typically at 10Hz or less. Source: www.amazon. com.au/dp/B0CFPL6CFY refreshing the display and should not be interrupted. Once the refresh is completed, the controller can be shut down by taking its reset line (RES) low or sending a sleep command. While it might appear that the controller supports an I2C interface via the TSDA and TSCL lines, this is not used by a host to talk to the controller. Instead, these pins allow an external I2C temperature sensor to be read by the controller IC. The BS1 line can be used to select a 9-bit SPI mode that removes the need for a separate D/C pin, since the D/C bit is sent as the ninth bit. We have no shortage of pins, and the 8-bit SPI mode is much easier to implement, so that’s what we used. Summary Our experience with the Arduino display module gave us the knowledge and experience we needed to design the hardware and software to work directly with e-paper display panels. While it would have been nice to use a multi-colour display, we did not think they were suitable for our application; thus, we chose the monochrome panel seen in Photo 1. In another situation, such as a price tag in a shop, a slower/flickering refresh would not really be a problem as it would be updated so infrequently. A splash of colour would be nice there (eg, to separate the price from the product name), so colour e-paper displays clearly suit some applications. By designing our own hardware, we have been able to achieve our target of very low quiescent power consumption. You can see the result of this in the Human Comfort Indicator project, SC which starts on page 43. Dual-Channel Breadboard Power Supply Our Dual-Channel Breadboard PSU features two independent channels each delivering 0-14V <at> 0-1A. It runs from 7-15V DC or USB 5V DC, and plugs straight into the power rails of a breadboard, making it ideal for prototyping. Photo shows both the Breadboard PSU and optional Display Adaptor (with 20x4 LCD) assembled. Both articles in the December 2022 issue – siliconchip.au/Series/401 SC6571 ($40 + post): Breadboard PSU Complete Kit SC6572 ($50 + post): Breadboard PSU Display Adaptor Kit siliconchip.com.au Australia's electronics magazine June 2026  69 Background source: https://unsplash.com/photos/silhouette-of-trees-under-starry-night-O7FxxiZr-Hk By Andrew Woodfield, ZL2PD Micropower SSB Single Side BAND Transmitters This project is an attempt to see how small an HF SSB transmitter can be made with a minimal parts count. One of the three versions uses just three transistors while generating a surprisingly clean signal! D uring one of our local radio club meetings a little while ago, someone tossed a PTT microphone over to me. “Here, maybe you can use this!”, they joked. Lacking the usual curly cord and connector, it had clearly seen better days. Nevertheless, I took it home with me and overnight, an idea came to mind. The Pixie range of CW (continuous wave, ie, Morse Code) transceivers is very well known, as is Doug DeMaw’s March 1976 Tuna Tin Two CW 70 Silicon Chip transmitter and George Burt’s “OXO” three-transistor CW transmitter. Having more interest in QRP SSB transceivers (www.zl2pd.com), I wondered if an SSB transmitter could be shoehorned inside that microphone shell. Naturally, only a very low-power SSB transmitter would be possible, a design using an absolute minimum number of parts. As an extra challenge, I decided to avoid using SMD components or ICs. The result, for the cost of one more Australia's electronics magazine transistor than the famous OXO transmitter, is a small milliwatt-level LSB transmitter operating close to 3.7MHz on the 80m band. Named the “Mike-One”, it uses just four general-purpose NPN transistors and a set of low-cost crystals. In its present form, it will never achieve transoceanic communications. Instead, covering short distances across the shack or at the radio club, it’s intended to be a lighthearted example of minimalist analog design. It’s also quick and easy to build, so it can be used as a teaching aid to illustrate the generation of conventional SSB signals at a very low cost. Circuit description This transmitter is cut to the bone. As Fig.1 shows, it features a microphone amplifier (Q1), a carrier generator (Q2), a balanced modulator (diodes D1 and D2), a three-crystal ladder-type SSB filter, an unusual ‘autodyne’ oscillator-mixer (Q3) and a single radio frequency (RF) amplifier stage (Q4). All stages use the generic BC548 NPN small signal transistor or one of its equivalents. The relatively high level audio signal from the electret microphone allows a single transistor amplification stage (Q1) to generate sufficient audio from the microphone to directly drive the balanced modulator. The 18.432MHz carrier, set by a small 22pF series capacitor (Cx), is balanced out in the mixer using VR1, a 100W trimmer. This arrangement saved several bypass capacitors that are usually required in such stages. siliconchip.com.au The autodyne converter stage (Q3) allowed a further useful reduction in the parts count. This mixer is a 14.7456MHz Colpitts crystal oscillator, amplitude modulated by the 18.4320MHz SSB signal. The output includes the desired 3.7MHz lower sideband (LSB) output, the difference between these two frequencies. This type of mixer was very common in the first stage of cheap transistor AM broadcast receivers, and it was also briefly popular in a few early commercial and amateur radio VHF FM transceivers. While it saves a few parts, the output of this mixer demands a good bandpass filter (L1, L2 etc) to remove the other unwanted products, including the 14.7456MHz oscillator output. A single π filter at the transmitter output also contributes to the low spurious and harmonic products of the design. It serves a secondary purpose – the output load is unlikely to be a perfect 50W load. I’ve mostly demonstrated it with just a length of hookup wire, perhaps half a metre long. Of course, it will work perfectly into a good load, but the useful feature of a π filter is that it transforms the output impedance of awful loads, such as the very low impedance of my 50cm of hookup wire (less than 1W) or an off-resonant long wire (possibly a few thousand ohms) to an impedance of 25-200W at the collector of Q4. Q4 is most unlikely to suffer damage as a result. The bandpass filter (BPF) is designed to be as flexible as possible to allow for a variety of crystals, as the following Photo 1: MikeThree (40m band) is even less complex than the others, with just three transistors. sections will show. In the Mike-One, the BPF is arranged in a series-­parallel arrangement to reduce the loading on the autodyne mixer. The choice of carrier, filter and mixer crystals is dictated by the current selection of readily available crystals. If your parts bins are well-stocked, you may prefer to use other crystals. In the days of analog TV, 6.552MHz crystals were widely available, as were 10.245MHz crystals for converting 10.7MHz intermediate frequency (IF) signals to a second IF of 455kHz. This combination will also produce an 80m LSB signal close to 3.685MHz. In this case, the carrier frequency sits at the lower corner of the crystal filter passband. This means that capacitor Cx in Fig.1 is replaced by a 15μH RF choke (RFC1 in Fig.2). This lowers the carrier crystal frequency to 6.5500MHz. I’ve named this version the Mike-Two. The different array of outputs generated by the autodyne mixer requires a slightly different BPF, but the output LPF remains unchanged. The Mike-Three By this stage, I could see a way to further reduce the number of parts used in these first two versions. Mike-Three is an example of an ‘on-frequency’ SSB transmitter that avoids the need for the mixer stage. This time, it produces a 7.2MHz LSB signal on the 40m amateur band (Fig.3) with just three transistors. These crystals are also very widely available. Mike-Three uses the same PCB as the others, but with fewer components, as shown in Photo 1. Construction All three versions can be built on the same small single-sided PCB, which was designed to fit into the prototype push-to-talk (PTT) microphone case (see Photo 2). This style of PTT microphone has been made in very large numbers by many manufacturers over Photo 2: an otherwise useless PTT microphone lacking a cord was the inspiration for this tiny SSB transmitter. Fig.1: the Mike-One circuit features a basic three-crystal SSB filter, along with an unusual autodyne oscillator-mixer stage to minimise the parts count. siliconchip.com.au Australia's electronics magazine June 2026  71 the years. It’s fairly likely you can lay a hand on a suitable microphone without much difficulty. In case you can’t, I’ve created STL files so you can 3D-print one! To fit everything in the limited space, almost all resistors are mounted on-end. The PCB is coded 06103261 and measures 44.5 × 76.5mm. Refer to the overlay matching whichever version you are building – Fig.4 for MikeOne, Fig.5 for Mike-Two or Fig.6 for Mike-Three. Start by fitting all the resistors and capacitors, then proceed to fit the parts to complete each stage, one by one, testing each completed stage as you proceed. T1 is made by twisting three 120mm lengths of 0.25mm enamelled copper wire (ECW) together. Two or three twists per centimetre is all that is 72 Silicon Chip required. Wind eight turns of this ‘trifilar’ triple wire arrangement onto an FT37-43 toroidal core. The toroid may be replaced by a similar-sized toroid recycled from an old fluorescent lamp. I found a less expensive approach to winding T1: wind four trifilar turns of 0.25mm ECW on a low-cost ferrite bead. Avoid ferrite beads with a small 1.5mm hole. They can be used, but it’s quite difficult to get all that wire through the small centre hole. I had some ferrite beads with a 2mm hole, which allowed for the required turns to be achieved far more easily. The material used to make these ferrite beads can vary enormously, so this may not work for you with your parts and your carrier frequency. Using the FT37-43 toroidal core is the most reliable option. When the toroid or ferrite bead has Australia's electronics magazine been wound, identify the start and end of each winding. These are numbered on the circuit diagrams and PCB overlays to help with construction. Solder one set of three wires into the holes marked 1, 3 and 5 in any order. Now, using a continuity tester or ohmmeter with an audible continuity (‘buzzer’) function, identify each of the matching ends for each wire, one by one, and solder them into the correct matching holes, marked 2, 4 and 6. L1 and L2 are inexpensive 7×7mm unshielded variable inductors. These have 26 turns and a range of 3-6μH. The similar-looking inductors with only about 12 turns (0.6-1.7μH) cannot be used here. However, if you buy them by accident, just rewind them with the required number of turns. They will work just fine. The BPF has been designed to allow siliconchip.com.au Fig.2: Mike-Two uses a 6.552MHz carrier and filter crystals, along with a 10.245MHz mixer crystal to give SSB on 80m. Besides the crystal changes, some capacitor values have been altered, and the BPF has been reconfigured. Fig.3: the circuit of the ultra-simple 40m Mike-Three SSB transmitter. The crystals change again, plus some capacitor and inductor values. In addition, the autodyne mixer and its associated band-pass filter have been removed and bypassed. either a series-parallel BPF (Mike-One) or a coupled BPF (Mike-Two). The relevant PCB overlays show the location of the wire added to configure these correctly. Other arrangements are possible with this PCB layout for those wanting to experiment further. L3 is made by winding 27 turns (Mike-One or Mike-Three) or 15 turns (Mike-Two) of 0.375mm enameled copper wire onto a T37-6 core. However, a less expensive solution is to use a 2.2μH RF choke (Mike-One or Mike-Three) or an 820nH RF choke (Mike-Two). Both methods gave similar results for me. In the case of Mike-Three, no mixer or bandpass filter components are fitted or required. This time, a short jumper wire connects between two empty capacitor pads, as shown in Fig.6. However, the output pi filter siliconchip.com.au Figs.4-6: the PCB overlay diagram for ▶ the Mike-One (top), -Two (middle) and -Three (bottom) variants; use the component values and locations shown here to build each version. There are some small component differences between each diagram such as the crystals and lack of mixer circuitry. should still be fitted to ensure any spurious and harmonic products are minimised, and to deliver some useful impedance matching. The PCB mounts component-side down in the case, with the electret microphone (MIC1) soldered on the solder side of the board. I used several drops of hot-melt glue at the top edge and the PCB corners to hold it in place. 3D-printed microphone shell While the microphone shell shown is readily available, it’s likely some readers will still find it difficult to locate or expensive. For that reason, I’ve also designed a lowcost 3D-printed version, shown in Figs.7(a)-(d). In this version, the electret mic capsule mounts on the same side of the PCB as the other components. You can download the STL files from siliconchip.au/Shop/11/3582 The case is in four parts: lower, middle and upper sections, plus the ‘pressel’ lever. The lever’s hinge fits into the mating slot in the middle section. I used a 10mm-long scrap of copper wire to hold the lever in place on one version, and a cut-down 1.6mm panel pin (ie, a small nail) on another. This assembly is then placed on the back shell of the microphone. Three or four drops of hot glue will hold these together. Avoid getting any glue near the pressel. Australia's electronics magazine June 2026  73 The transmitter PCB can then be inserted into place – component-side up this time – and a drop of hot glue applied at the top edge to hold it in place. The battery, LED and related wiring can now be added, and the length of wire to be used for the antenna also connected to CON2 and fed through the antenna hole in the shell. The upper shell of the microphone may then be placed on top of the assembly. Three 20mm-long self-­ tapping screws hold the microphone case together. Alignment Depending on the crystals you use and the version of the transmitter you are building, you will need to mount either a small capacitor, Cx/Ca, or an RF choke, RFC1, in the top-right corner of the PCB. Nominal values for these parts have been shown in Figs.1-6. This allows the carrier crystal frequency to be at the upper or lower corner of the SSB crystal filter respectively. The values shown (22pF, 15μH or 18pF) were found to be best for the prototypes, and are likely to suit most applications, but your crystals may require slight changes. Values are likely to be in the range from 4.722μH for RFC1, and 10-33pF for Cx/ Ca. You can listen to your signals on an SSB receiver to confirm the audio quality is reasonable and the opposite sideband is nearly inaudible. L1 and L2 should be adjusted for maximum transmitter output. These have a reasonably broad tuning response. Of course, this step is not required for the Mike-Three. VR1 in the balanced modulator should be adjusted to give minimal carrier output in the absence of modulation. This setting is very sharp and will be close to the midpoint of the adjustment range of the trimmer. I built all the prototypes using a variety of crystals, which delivered about 0dBm into 50W with a carrier suppres- Photo 3: this version uses Cx to set the correct carrier frequency (PCB upper right), a ferrite bead for the balanced modulator (PCB centre), and an RF choke for the LPF (PCB lower edge). Photo 4: Mike-One uses more crystals than transistors! The narrow bandpass filter required for SSB demands at least three crystals, while the carrier oscillator and mixer add two more. sion of 40dB or better, and 30-40dB of unwanted sideband rejection. This latter value depends on the audio modulation frequency. All spurious and harmonics were attenuated by 50dB, and many by as much as 60dB. The second method I tested used a small A23 or A27 12V alkaline battery. An A27 battery has a diameter of just 8mm and a length of 28mm. The capacity of an alkaline A27 battery is about 30mAh. While modest, it proved ideal for the original microphone shell. I also tried fitting a slightly larger A23 battery (the holder is visible in Photo 3). The higher A23 or A27 battery voltage of 12V is perfectly OK for the transmitter. Usefully, it also allows a blue or red LED to be fitted in series with the supply wiring to CON1. The supply voltage at the transmitter is dropped by almost 3V by a blue LED and by about 2V by a red LED. The LED is lit during transmit mode and its brightness gives an approximate indication of the battery level. Power supply options If you are building the Mike-One in a small box rather than in a microphone, you can use a standard PP9 type 9V battery. The transmitter only draws 15mA, so the battery will last for quite a long time. Fitting a battery in the limited space available inside the microphone shell presented a challenge. One approach tested used a small recycled 3.7V 70mAh Li-Po cell and a tiny boost converter module. This gave a very reliable 9V supply. Figs.7(a)-(d): this 3D-printed microphone shell has been designed for those unable to locate a suitable microphone shell. 74 Silicon Chip Australia's electronics magazine siliconchip.com.au A23/A27 battery holders Battery holders for these tiny A23 and A27 batteries are not always readily available. Faced with this, I designed and printed a simple 3D-printed holder for each type, shown in Fig.8. The battery contacts were fabricated from a pair of M2.2 solder tags. I filled the hole normally used for a bolt with a film of solder. These were pressed through the 3D-printed battery holder from the inside and held in place by the battery and the slight tension of the holder. I placed some clear adhesive tape around the battery before inserting it to make it easier to replace. Parts List – Micropower SSB Transmitter Fig.8: the 3D-printed 12V battery holders provide a low cost solution to fitting a small battery inside the microphone shell. 1 single-sided PCB coded 06103261, 44.5 × 76.5mm 1 electret microphone (MIC1) [Altronics C0170, Jaycar AM4011] 1 PTT microphone shell, salvaged or 3D-printed 3 M3 × 20mm self-tapping screws (for 3D-printed case) 1 A23 or A27 12V battery 1 3D-printed battery holder 2 2.2mm solder lugs (for 3D-printed battery holder) 1 2-pin header (CON1; optional) 1 PCB-mounting right-angle tactile pushbutton, 6×6mm, 6mm-long actuator (S1) [Jaycar SP0607 or AliExpress 1005007559876628] 1 FT37-43 toroidal core (T1) [www.minikits.com.au/FT37-43, AliExpress 1005009245292057] OR 1 4mm OD, 2mm ID, 5mm-long ferrite bead (T1) [Altronics L5250A] 1 360mm length of 0.25mm diameter enamelled copper wire (T1) 1 100W top-adjust trimpot (VR1) [Altronics R2605] 3 BC548 30V 100mA 300MHz NPN transistors (Q1, Q2, Q4) 1 red or blue 3mm LED (LED3) 2 1N4148 75V 200mA signal diodes (D1, D2) various lengths of light/medium-duty hookup wire Capacitors (all 50V radial ceramic) 6 100nF 2 10nF 3 1nF 2 100pF 2 33pF Resistors (all ¼W axial ±5% or better) 1 1MW 1 2.2kW 1 220kW 1 1kW 3 10kW 1 470W 1 3.3kW 1 47W Extra parts for both Mike-One & Mike-Two 2 3-6μH 5-pin variable inductors on 7×7mm formers (L1, L2) [AliExpress 1005008114591102] 1 2.2μH axial RF choke (L3) [Jaycar LF1514, Altronics L7014] OR 1 T37-6 toroidal core (L3) [www.minikits.com.au/T37-6, AliExpress 1005005686909567] AND 1 400mm length of 0.375mm diameter enamelled copper wire (L3) 1 BC548 30V 100mA 300MHz NPN transistor (Q3) 1 100nF 50V radial ceramic capacitor 1 100pF 50V radial ceramic capacitor 2 56pF 50V radial ceramic capacitors 1 22kW ¼W axial resistor (±5% or better) 1 6.8kW ¼W axial resistor (±5% or better) 1 4.7kW ¼W axial resistor (±5% or better) Extra parts for Mike-One only 4 18.432MHz HC-49 crystals (X1-X4) 1 14.7456MHz HC-49 crystal (X5) 2 330pF 50V radial ceramic capacitors 2 47pF 50V radial ceramic capacitors 1 10-33pF 50V radial ceramic capacitor (Cx, nominally 22pF; see text) Extra parts for Mike-Two only 4 6.552MHz HC-49 crystals (X1-X4) 1 10.245MHz HC-49 crystal (X5) 1 4.7-22μH axial RF choke (RFC1, nominally 15μH; see text) 2 680pF 50V radial ceramic capacitors 1 330pF 50V radial ceramic capacitor 1 150pF 50V radial ceramic capacitor 1 22pF 50V radial ceramic capacitor Extra parts for Mike-Three only 4 7.2000MHz HC-49 crystals (X1-X4) 1 820nH axial RF choke (L3) 2 47pF 50V radial ceramic capacitors 1 10-33pF 50V radial ceramic capacitor (Ca, nominally 18pF; see text) siliconchip.com.au Australia's electronics magazine Wiring The wiring is straightforward; it’s shown clearly in all three circuit diagrams, Figs.1-3. In brief, run a red wire from the battery + to the LED anode (longer lead), an orange wire from the LED cathode (shorter) lead to the + terminal on CON1 and a black wire from the – terminal on CON1 to the battery – terminal. Also refer to Photo 3. Operation This is not a complicated transmitter to use! Just press the PTT button and talk. Your LSB signal will appear very close to 3.7MHz or 7.2MHz, depending on the version you’ve built. While the range is not massive when using a short length of hookup wire for the antenna, the signal is quite audible in nearby receivers. Usefully, the design is such that many popular data modes can also be tested with the transmitter, and further amplifier stages can be added if desired. In short, Mike-One (or Two or Three) will allow you to quickly, easily and inexpensively enjoy a short yet rewarding voyage on the QQRP ultralow-power HF seas. I hope you enjoy making and using one (or more!) of SC these little SSB transmitters. June 2026  75 SOnline ilicon Chip Shop Kits, parts and much more www.siliconchip.com.au/Shop/ Rotating Lights April 2025 USB-C Power Monitor August-September 2025 USB Power Adaptors May 2025 SMD LED Complete Kit SC7462: $20 TH LED Complete Kit SC7463: $20 Short-Form Kit SC7489: $60 siliconchip.au/Article/17930 siliconchip.au/Series/445 siliconchip.au/Article/18112 This kit includes everything needed to build the Rotating Light for Models, except for a power supply and wire. This kit includes all non-optional parts, except the case, lithium-ion cell and glue. It does include the FFC (flat flexible cable) PCB. You can choose from one of four USB sockets (USB-C power only, USB-C power+data, mini-B or micro-B). The kit includes all other parts. DCC Base Station Short-form Kit SC7539: $90 Complete Kit with choice of USB socket SC7433: $10 PICKit Basic Power Breakout Board September 2025 January 2026 siliconchip.au/Article/19558 This kit includes all non-optional components in the parts list (and the RJ45 socket, CON6). It does not include the case, DC power supply, glue, CON4 screw terminal and CON5 locking header. Mic the Mouse Complete Kit SC7508: $37.50 August 2025 siliconchip.au/Article/18637 It includes everything needed to build one Mic the Mouse, except for solder, glue and a CR2032 cell. Complete Kit SC7512: $20 siliconchip.au/Article/18850 Includes the PCB, all onboard parts and a length of clear heatshrink tubing. Jumper wire and glue is not supplied. → Subscribers receive a 10% discount on all purchases, except for subscriptions (postage is not discounted). → Prices listed do not include postage. Postage rates within Australia start at $12, rates are calculated at the checkout. CIRCUIT NOTEBOOK Interesting circuit ideas which we have checked but not built and tested. Contributions will be paid for at standard rates. All submissions should include full name, address & phone number. The iClap intelligent multi-clap switch I spent some time searching for a simple yet efficient clap switch circuit to build for some rooms in my house. I finally surrendered to the old proverb, “on n’est jamais mieux servi que par soi-même” (you are never better served than by yourself). It isn’t that I found no clap light circuits, but many had the same pitfall: false triggering from accidental sounds. Adding a microcontroller gives some intelligence to the whole circuit, as it provides more flexibility to adjust parameters without changing the hardware. Sensing more than one clap also reduces the probability of false triggering. One may expand this idea to use different numbers of claps to perform multiple functions. The circuit uses a small electret microphone, fed into a minimalist (yet very sensitive) discrete amplifier. The output of that amplifier drives the GP3 input of the PIC10F200, which has no analog-to-digital converter (ADC) or siliconchip.com.au comparator; it’s just a digital microcontroller. The PIC micro is normally in sleep mode and wakes up on a pin level change if the clap produces a pulse with an amplitude greater than 2.4V (as GP3 is a TTL-type input). Within the next two seconds, if another clap is detected, the light is switched on using relay RLY1. The same sequence is used to switch it off. Digital output GP2 drives the coil of the low-power 24V relay through transistor Q2, a general-purpose NPN bipolar transistor. The rest of the circuit is a classic transformerless capacitive power supply to produce the 24V and 5V rails. Capacitor C1 must be an X2 (or X1, Y2 or Y1) type for safety reasons. As it is somewhat bulky, we have chosen to use a 24V DC coil relay, rather than 12V or 5V, to reduce the coil current, hence reducing the current that C1 must supply. We may choose a relay with NO/ NC contacts to switch between a Australia's electronics magazine high-power bulb (active state) and a much lower power bulb (idle state) if we don’t like dark rooms. Another option is that, after activating the relay, an internal countdown starts, and after 45 minutes, it will automatically deactivate the relay (in case someone leaves the room and forgets to switch it off). The fully commented assembler source code takes less than 100 words. It is not optimised to make it easier to understand, in case it is worth being translated to another type of microcontroller. You can download the software from siliconchip.au/Shop/6/3567 Hichem Benabadji, Oran, Algeria. ($85) Note: this project operates at mains potential and must be fully isolated from contact. That includes the microphone that needs to be located well inside an enclosure and the LED shouldn’t protrude through the enclosure without a fully covering bezel. The enclosure must be Earthed if it is metal. For a plastic enclosure, no metal parts should be exposed on the outside of the enclosure. June 2026  77 Interruptible NiMH battery trickle charger with memory This circuit uses discrete parts to trickle charge a pair of AA NiMH cells from a USB source. One of its interesting features is that it remembers the charging time if you unplug it from the USB source but don’t remove the cells. If you plug it back in again, it resumes charging, meaning the total charge time will still be limited even if you charge the cells across several sessions. The timer is reset when you remove the cells and insert a different set, or re-insert the same ones later. It could be adapted to suit AAA cells by adding extra PCB mounting points for a 2 × AAA cell holder and changing the value of the 10W resistor that determines the charging current. It’s designed to be plugged into a USB socket for power (no data connection), and charges whenever that USB socket supplies power (as shown 78 Silicon Chip in the photo at upper right). After 27.5 hours of charging, it stops supplying current to the cells and flashes the LED. The timer has two chips that function as ‘memory’, remembering how long the cells have been charging for (IC3 & IC4). These two chips can be ‘battery backed’ by the cells being charged since they are 74HC logic chips that can operate with a supply as low as 2V. When plugged into USB, current is supplied to the cells via NPN transistor Q10. A 10W resistor in its collector circuit senses the charge current and the voltage across that is fed to the base of Q7, while similar transistor Q6 has its base set to the divided output of a 2.5V reference (REF1) that’s connected to the USB supply. The result is that the voltage across the 10W resistor is regulated to 1V. Australia's electronics magazine That means the charge current is limited to 100mA (1V ÷ 10W). Since the emitters of Q6 & Q7 are connected to ground through Mosfet Q9, charging only occurs when Q9 is switched on. That’s controlled by logic gate IC3b. Its output is high as long as either pin 5 or 6 of IC4a is low. IC4a is the end of a counter chain. When USB power is applied, IC1 is clocked by a 32,768Hz crystal oscillator. It divides that by 16,384 to produce a 2Hz signal at pin 3, which is used to flash the LED. However, the LED can’t flash during charging because it is forced on by Mosfet Q8, so it only flashes when the charge current has been switched off. Pin 1 of IC1 divides the crystal frequency by 4096 to produce an 8Hz signal that’s fed to the clock input of IC2. IC2 divides it by 4096 again, producing a signal at its pin 1 that has a cycle siliconchip.com.au Ideal Bridge Rectifiers Choose from six Ideal Diode Bridge Rectifier kits to build: siliconchip. com.au/Shop/?article=16043 28mm spade (SC6850, $30) The charger PCB uses a Type-B USB socket as they are robust. You can see the cell holder underneath the board. time of 512 seconds. This is gated by the cell charge voltage and then fed to IC3c and IC3d, into pin 13 of IC4b. This pin only toggles every 512 seconds if charging is occurring. IC4b divides this signal by 16 times to give a cycle time of 8192 seconds or about 2¼ hours. That’s fed into another counter, IC4a, and with the way IC3b is configured, the signal from its pin 6 to stop charging will go low after 27 hours and 18 minutes. Once charging stops, you have to power cycle the circuit by removing the cells and USB power before it’ll charge again. Russell Gurrin, Highgate Hill, Qld. ($100) siliconchip.com.au 21mm square pin (SC6851, $30) Compatible with PB1004 10A continuous (20A peak), 72V Connectors: solder pins on a 14mm grid (can be bent to a 13mm grid) IC1 package: MSOP-12 Mosfets: TK6R9P08QM,RQ 5mm pitch SIL (SC6852, $30) Compatible with KBL604 10A continuous (20A peak), 72V Connectors: solder pins at 5mm pitch IC1 package: MSOP-12 Mosfets: TK6R9P08QM,RQ IF signal injector To align certain radios (eg, the Sailor 66T navigation radio on p88), you need a way to connect an RF sweep generator to the first IF transformer without damping its primary coil. This injection needs to be done in a balanced/differential manner. The circuit shown here does just that, allowing you to sweep and check the IF bandpass frequency response. For many transistor radios, a sweep generator is not required because all the coils can be peaked on the centre IF frequency (455kHz in most transistor radios). The radio is often designed to have the correct overall bandpass response in that condition. However, this is not always the case for radios with double-tuned IF transformers, such as the Sailor 66T. The circuit for the adaptor I made uses a Jaycar LO1234 25mm toroidal ferrite core with two 12-turn windings made using enamelled copper wire approximately 0.5mm in diameter. Each winding has an inductance on the order of 200μH. In the unloaded condition (without the Compatible with KBPC3504 10A continuous (20A peak), 72V Connectors: 6.3mm spade lugs, 18mm tall IC1 package: MSOP-12 (SMD) Mosfets: TK6R9P08QM,RQ (DPAK) mini SOT-23 (SC6853, $25) 75W termination resistance on the secondary), the loading on the generator’s 75W output resistance at 470kHz is negligible. The 75W termination resistor on the secondary makes sure that there are no spurious resonances. The high-value series resistors ensure that the primary of the first IF transformer is not damped or affected by the connection of the adaptor. One thing that helps significantly with winding toroidal cores is to wrap Scotch 27 cloth (fibreglass) tape around the core. It stabilises the position of the wires and makes for a much better result. Dr Hugo Holden, Buddina, Qld. ($80) Width of W02/W04 2A continuous, 40V Connectors: solder pins 5mm apart at either end IC1 package: MSOP-12 Mosfets: SI2318DS-GE3 (SOT-23) D2PAK standalone (SC6854, $35) 20A continuous, 72V Connectors: 5mm screw terminals at each end IC1 package: MSOP-12 Mosfets: IPB057N06NATMA1 (D2PAK) TO-220 standalone (SC6855, $45) 40A continuous, 72V Connectors: 6.3mm spade lugs, 18mm tall IC1 package: DIP-8 Mosfets: TK5R3E08QM,S1X (TO-220) See our article in the December 2023 issue for more details: siliconchip.au/Article/16043 Australia's electronics magazine June 2026  79 SERVICEMAN’S LOG Music to my ears Dave Thompson It may seem obvious to many regular readers that I am ‘into’ music. Or perhaps it isn’t that obvious, but rest assured, I am into it! So, not surprisingly, a lot of the gear I’ve fixed over the years has been related to musical instruments and hifi equipment. I both listen to and play music, and I have done so since I was eight years old, when I was taught piano by an elderly Hungarian Jewish man who lived just around the corner. He was a very good player and a very good teacher. He had some stories to tell about the war years – which were likely heavily sanitised for my tender young ears! Sadly, he passed away a few years after I started, so I changed teachers. That killed my piano-playing career. What it did teach me is that having the right teacher is crucial. While I still play instruments today, I don’t play any of them anywhere near as much as I used to, or should. As for listening to music, I have my favourites, of course, but am not averse to listening to anything new coming along. I know what I like (usually within seconds of starting), so these days I stick with what I know, or I check out recommendations from others I know who have similar tastes (that includes some of the folks producing this magazine). All that said, I am definitely not one of those guys with a library-sized collection of vinyl, playing them on an audiophile-­ level home audio system. You know 80 Silicon Chip the ones, with those thousand-dollar-per-metre speaker cables made from unobtanium and Mars dust. My earliest dabbling in electronics – with a lot of help and encouragement from my dad – was with simple oscillators, crystal radio sets and very basic amplifiers. It fast became a keen interest, and the more I got into it, the more I saw music-related potential. I don’t think I was alone with these interests either, if the number of music-related projects in the magazines of the day was anything to go by. I took it as a sure sign that others were as keen as I was, and I was proven right. There was even a relatively short-lived magazine out of England that featured just that: electronics, current music technology and really cool projects for musicians. I have every copy stored away somewhere, but there weren’t that many with that first-generation theme before it morphed into something that wasn’t of much interest to me. Publishing is tough in any field, time period, or country! It seemed every month, one magazine or another would have an interesting music project. I must have built many dozens of them (mostly on prototype boards); some useful, some just gimmicks. Still, it was all educational and much grist for my mill back then. Making projects kept me busy while I should have been practising, but I had a lot to learn, and so I was very eager to do both as much as I could. I saved all my newspaper delivery money (then my parttime after-school job earnings) to buy these magazines and then the relevant components for projects. I still have large stacks of the mags in storage (and components!). Sadly, the magazines will likely end up in the recycling bin if I can’t give them away. I’ve seen many people trying to pass on complete collections of magazines and libraries worth of books over the years, but the modern age, with much of it digitised retroactively anyway, what’s the point in keeping vast collections, other than just that – collecting them? It seems such a waste, but it’s just reality. Years ago, as a wee tacker, I made everything from tone boosters, fuzz boxes, wah pedals and metronomes until, after a while, I was making increasingly complex projects such as my own stereo amps, preamps, full-blown PA amps, speaker cabinets, mixing consoles, guitar/bass/ keyboard amps, microphones, wireless systems and guitars themselves. I wanted to be able to record music at home, on a computer, and that’s where this music rabbit-hole led me. Australia's electronics magazine siliconchip.com.au Back in those days, the go-to home recording device, if you could afford one, was something like a Tascam or a Fostex four-track, which recorded onto standard cassette tapes. They were pretty good for their day, but as we know, cassette tapes aren’t the best for audio quality, especially if they were used to ‘dub’ or ‘bounce’ many tracks down into one track a few times. The step up from there was into a reel-to-reel four-track or, if you owned a corporation, an eight-track reel-to-reel that used half-inch tape. They had much better audio quality. Fun fact: The Beatles’ album Sgt. Pepper’s Lonely Hearts Club Band was recorded on just two Studer J37 four-track recorders (using one-inch tape), synched together electronically by audio engineers, at what later became Abbey Road Studios. This shows what could be accomplished in 1966 with some very good gear and much lateral creative thinking, producing and mixing. Those cassette-based four-track recording machines were very much a 1980s child and, as they were relatively affordable, they launched the imaginations of many and the possibility of a viable home recording studio. All was well then, but recording at home is not just a matter of grabbing some digital audio workstation (DAW) software from the interwebs, plugging in your microphone, keyboard or guitar to your computer and laying down a new number-one hit single. While you’d think that the computer already has analog inputs to allow instrument/microphone/line connections, the CPU has to work hard to crunch that signal into a digital form so the computer can process, store and manipulate it. You might be surprised to know that most built-in ‘sound cards’ - while most are built onto the motherboard now, people still call them sound cards – are pretty much useless for doing any studio-level sound processing. Yes, they can record speech and basic mic audio, but there are inherent problems. For example, if you want to monitor recorded audio through headphones while multitracking, there is a delay between input and output signals as the computer struggles to keep up. This is called latency and makes onboard sound unsuitable for multitrack recording, where there must be minimal latency between in and output signals. So, an external sound interface is usually the way to go. These typically connect via USB or Thunderbolt, operating similarly to an accelerated graphics card in that the interface handles all the audio heavy-lifting for the computer. To do this, a special driver is needed, called an ASIO driver (Audio Stream Input/Output). This was developed by the Steinberg company in the late 1990s specifically to Items Covered This Month • Musical instruments and USB ASIO card repair • Repairing a Simpson Contessa washing machine • A powerless Li-ion charger • Small replacements for a Dyson fan • Fixing a Samsung TV Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz Cartoonist – Louis Decrevel Website: loueee.com siliconchip.com.au reduce latency and allow the processing of multiple concurrent audio streams. Usually, these drivers are supplied with the interface, or available as a download or, at a push, a generic freeware ASIO driver. Interfaces cost anywhere from under a hundred bucks to several thousand, depending on what you want. My Line 6 Toneport UX2 is almost vintage now, yet still cost me 500 bucks second-hand in 2003. It still works perfectly for my needs – or it did until recently. It is connected permanently to my PC and acts as a normal sound card when just listening to music or watching YouTube, but it’s when all the audio inputs are utilised with a DAW that we really see and hear the difference. The problem now is its age. All the pots are scratchy, some of the sockets are intermittent, and it is just not right. So I decided that it was finally time to do something about it. The beauty of this era of music gear is that it really is ‘analog’ and it was designed with repair in mind. Typically, clamshell plastic or metal cases were screwed together with normal screws, meaning easy disassembly. All the sockets utilised in most pedals and music gear are standard components, most of which are still available. However, this unit is interesting in a few different ways. Disassembling it was odd from the start; after turning it over multiple times (no doubt with a puzzled look on my face), I simply couldn’t see how it was held together. There are no obvious screws underneath. I even popped one of the four glued-on rubber feet off to check for hidden screws (as is typical of many laptops) but I found none. I thought maybe that the four rotary controls (two independent mic input gains, headphones volume and output level) located on the top half of the case might actually be holding it all together. However, when I popped the knobs off, I saw that Line6 had simply extended the pot shafts up through the top of the case, so they actually didn’t hold down anything at all. Interesting... It turns out there are no screws and no clips; just a friction fit that holds the top half of the case to the bottom, although the front and back panels also play a vital part. Australia's electronics magazine June 2026  81 It’s always a little scary when applying pressure to something that might be clipped or screwed (oh for a small X-ray machine!) but it eventually let go and eased apart with gently increasing pressure. This left the top of the PCB exposed, with the front and back panels still engaged in a slot that runs around the edges of both halves of the case. It’s a purposely tight fit and, when the case is assembled, it all holds itself together. Given that this is designed to sit on a desk and not go on the road, this is an elegant (and likely cheaper) way of holding it all together. Both the front and back panels are an odd but very cleverly implemented two-layer lamination. These consist of the screen-printed, outward-facing coloured fascia, and behind it, a very thin metal laminate, which is no doubt added for shielding and grounding. Between them, they fill the slot around the case edge perfectly, with a pleasing, ever-so-snug interference fit. It has obviously been carefully thought out by the designers and well-implemented by the manufacturers (what we pay the bigger dollars for, I suppose). The back panel is held onto the PCB by a raft of 6.35mm 82 Silicon Chip (¼-inch) jack sockets and a sole RCA socket. These are the line in, footswitch, S/PDIF and analog output sockets. I never use those anyway, except for the USB connector (which utilises the older, but still widely used USB-B type connector), so I left that panel alone; there was no point in taking it off. The multi-layer PCB itself is packed with assorted sockets, pots, a couple of backlit, olde-worlde analog VU meters and what looked to be about a million SMD components distributed in between the bigger parts. I decided to remove the front panel, mainly for easier access to some of the sockets and switches that live right behind it. The PCB was held to the panel by the barrel nuts and black plastic washers of three 6.35mm jack sockets and four (surprisingly long) machine screws holding on to two Canon XLR sockets. Both socket types are the universal standard for plugging instruments and microphones into audio gear. There are two guitar inputs (one normal and one ‘padded’ for hotter inputs) and a headphone jack output. Removing the barrel nuts on the jacks exposes some interesting, thick, multi-pinned plastic locating washers that hold the jack sockets fast into the front panel without any risk of Australia's electronics magazine siliconchip.com.au the socket turning (as often happens with chassis mounted versions of these same sockets), as shown in the photo. More importantly, this locking system prevents the transference of any cable pressure or tension directly onto the fragile PCB. One common mistake I’ve made myself many times before is to forget I’m plugged into something and walk away, blissfully unaware that when the length of the cable runs out, some huge stresses are going to be inflicted on those poor sockets at either end. I’ve done this with power tools, laptops, mobile phones, headphones and guitars. We’ve likely all done it! The two XLR connectors are solely for microphone inputs and have two screws per socket. The only other control on the front panel is a push-on/push-off phantom power switch for sending 48V DC to the microphones (if needed). They must have a decent DC/DC converter in there to take 5V from USB and be able to send 48V (admittedly only up to a maximum of 10mA of current) to power up to two connected active microphones. Once it was all open, I sat it on a plastic box, plugged the USB cable into my workshop computer, and my headphones into the output to monitor what was happening. Then I worked on the potentiometers. All four were scratchy and felt gritty, so I gave them a good squirt of contact cleaner while rotating them from end to end, making sure the entire track got a decent wet wipe. They quietened down a lot, and as there is no obvious dust inside the interface, it was likely worn pot track debris. I gave them another few blasts once this lot had dried. I moved on to the jack sockets. These are mostly enclosed by hard plastic mouldings, so getting access to the contacts inside was a problem. The barrel nuts on the outside of the socket are easy enough to tickle up, as they need to be taken off anyway, but I replaced them before cleaning. If these ferrules aren’t electrically clean, there’s not a lot of point doing the rest. What I needed was a means of cleaning the contacts inside, at least where jack plugs would touch. To this end, I took a timber dowel and glued very fine (1200 grit) wet and dry sandpaper around it, making it a snug fit inside the socket. After spraying the sandpaper with contact cleaner, a few gentle ins-and-outs and a twist or two cleaned the relevant surfaces, at least as much as I wanted to use abrasives on them. Then I blew out any debris after the cleaning rod work and sprayed with more contact cleaner to flush things out. Left: a Line 6 Toneport UX2 audio interface. Removing the barrel nuts on the headphone jacks reveals multi-pinned plastic antirotation washers. siliconchip.com.au Australia's electronics magazine June 2026  83 I repeated this on all the sockets and manually polished anything I could see and access with my diamond contact file. I cleaned the XLR connectors (not that I use them) with a bamboo skewer soaked in contact cleaner. It was the perfect size to fit in the holes in the socket. The probe came out clean anyway, so I considered that done. That was about as far as I could go with it. Other than the pots and sockets, there really are ‘no user-serviceable parts inside’. The audio tests after that were a lot better, but not perfect as far as the pots went. However, I wasn’t too keen on the prospect of desoldering anything from this circuit board, especially a through-the-board pot. For the amount of use this device gets now, I can live with it. At least now I know that if the pots continue to give me trouble, I can get the device apart easily enough and replace them. Fingers crossed it doesn’t come to that. The sockets all seem to be working well now, so I reassembled everything in reverse order for a final test. All was good, which was music to my ears! Simpson Contessa 425 washing machine repair I refurbished a Simpson Contessa 425 washing machine about four years ago. Since then, it has worked well. I had to repair the spin solenoid around two years ago (described in Serviceman’s Log, October 2022). Lately, it has been leaking water. This was not a major drama, as the machine is on the back verandah and the small amount of water just ran off the concrete onto the grass. It was more of an annoyance than anything. I had been meaning to check where the water was leaking, but every time I noticed it leaking, the machine was in use with water in it. I finally co-ordinated troubleshooting with my wife. I needed to pull the machine out and remove the back panel so that I could see exactly where the leak was. With the machine pulled out and the back panel removed, we added some water, and I could see that it was the water pump that was leaking. Water pumps are one component that I do have spares of, so I was confident I could find a replacement. I recalled that we had previously used a Simpson 7.5kg machine that was computer controlled. We used it for several years until the computer failed, so that was not economically repairable and it went to the boneyard. I located the machine and removed the pump. I then removed the pump from the 425 and compared them; they were completely different. The original pump was much longer and had the inlet on the top, whereas the possible replacement pump was much shorter and it had the inlet on the end. Therefore, the original drain hose was too short. But maybe I could use the hose from the 7.5kg machine. I went back to the boneyard and retrieved the inlet hose. Then it was time to see if the later-model pump and hose would fit in the older machine. I first checked to see if the bolt holes in the pump would line up with the holes in the back of the machine where the pump mounts. On the 425, there were four holes for mounting the pump; two of them lined up with the bolt holes in the replacement pump, so that was OK. After bolting the pump into the machine, I fitted the inlet hose to the pump and then to the bowl. It fitted correctly, with plenty of clearance to the pulley on the bottom of the bowl, so the pump transplant was a success. I decided to change the outlet hose at the same time, as the old hose had previously had some leaks in it, which had been repaired with duct tape. While this had been successful, I thought that the newer hose looked better, so I would use it. I connected the replacement hose to the pump and noticed that the new hose was a slightly smaller diameter, so I had to change the fitting where the outlet end of the hose is connected to the grey water disposal pipe, but that was easy and I then replaced the back panel. With the machine back in use, there were no more water leaks. The reason these old pumps leak is that the seal where the motor shaft enters the pump wears out. This is a special type of seal that is not available as a spare part. I have previously repaired these old pumps by swapping parts between them, like swapping a good motor onto a good pump, but there’s only so much I can do without being able to obtain spare parts. Bruce Pierson, Dundathu, Qld. Ozito QL09009A 5-cell lithium-ion battery charger The original (longer) Simpson Contessa 425 washing machine pump is shown above. The replacement pump is below, with the inlet located at the end. 84 Silicon Chip This Ozito battery charger had no output. It is a conventional flyback design using an AP8263 IC. The secondary side has an LM358 dual op amp that monitors the charging current and alters the drive to the feedback optocoupler to vary the output voltage of the charger. Australia's electronics magazine siliconchip.com.au From left-to-right: the AM05 & AM07 motor control boards; a brushless DC hair dryer motor; the drive signal from this hair dryer motor. There is a P-channel Mosfet in series with the output, which was not being switched on, hence the lack of output. There is also an unmarked 8-pin IC that seems to monitor the battery charging and drives a three-pin LED, giving a red or green light. Another output of this IC drives an NPN transistor that should switch on the Mosfet. I decided that the mystery IC was faulty, so I forced the Mosfet on by shorting out the NPN transistor. That got the charger working again. I assume the Mosfet was supposed to be turned off if the charging current got too high or charging took too long. It would help if the IC could be identified. Roger Sanderson, Fig Tree Pocket, Qld. Dyson bladeless fan faults (AM05 & AM07) The brushless DC motor circuit drive boards in these Dyson fans use a Power Integrations LNK304D buck regulator to produce 15V or 5V for the control ICs from the 350V DC rectified mains. Unfortunately, this part fails, feeding a destructive voltage through to the low voltage parts, particularly the microprocessor. Power Integrations now has an LNK3204D, which hopefully is more reliable. Replacement boards for these fans cost around $200, so it is questionable whether it is worth repairing them. Of course, the microprocessor that holds the software is not available as a spare part. I am looking at using an alternative brushless DC motor control board to power the fans. There are low-cost controller boards available, but they are designed to drive low-voltage motors. They have a built-in switch-mode converter to provide a low voltage for the motor and drive Servicing Stories Wanted Do you have any good servicing stories that you would like to share in The Serviceman column in SILICON CHIP? If so, why not send those stories in to us? It doesn’t matter what the story is about as long as it’s in some way related to the electronics or electrical industries, to computers or even to cars and similar. We pay for all contributions published but please note that your material must be original. Send your contribution by email to: editor<at>siliconchip.com.au Please be sure to include your full name and address details. siliconchip.com.au circuit from the AC mains. To use them, the motor would also need replacement. Attached are photos of the AM05 and AM07 control boards. The AM05 uses an SMA6823 to drive the motor, while the AM07 uses individual Mosfets. I am also investigating using a brushless DC hair dryer motor from AliExpress to replace the motor and driver in a Dyson bladeless fan. Note that the drive signal on one of the three phases is not continuous, but is in bursts. The driver board provides three levels of airflow. As the airflow is increased, the driver pulse width increases, but the pulse rate stays the same. Roger Sanderson, Fig Tree Pocket, Qld. Samsung TV fixed the IT way We have a Samsung TV (UA55CU8000) and sound bar (HW-Q600C). One evening, as my wife settled down to watch one of her programs while I was occupied in another room, she called out that the sound was “not working properly”. When I came to check, I could see that using the remote would bring up the numerical data on-screen for the volume setting but it was obvious that the volume was not changing. As a quick fix, I jumped into the TV menu and enabled the TV speakers while disabling the sound bar. This worked, seemingly indicating a problem with the sound bar. The next day, I switched on the TV and re-enabled the sound bar. The remote still had no effect, but now the volume was changing, slowly upwards, without me doing anything! As part of my fault-finding process, I decided to transfer the sound bar to the small TV in the kitchen and see what happened. Well, of course it worked properly, didn’t it! I put the sound bar back while glaring sideways at the now-suspect TV. While I was doing this, the old computer adage of “Have you tried turning it off and on again?” came to mind. Since no modern piece of electronics is ever turned completely ‘off’, I pulled the power cord out of the back of the TV, waited a while and plugged it in again. When I turned it on again, everything worked correctly and has continued to work since. When we thought about it, earlier that evening we had a momentary mains power glitch, probably a little less than half a second. This appears to have addled the TV set’s brain, and a full power reset got it straightened out again. SC Ian Malcolm, Scoresby, Vic. Australia's electronics magazine June 2026  85 SILICON CHIP .com.au/shop ONLINESHOP HOW TO ORDER INTERNET (24/7) PAYPAL (24/7) eMAIL (24/7) MAIL (24/7) PHONE – (9-5:00 AET, Mon-Fri) siliconchip.com.au/Shop silicon<at>siliconchip.com.au silicon<at>siliconchip.com.au PO Box 194, MATRAVILLE, NSW 2036 (02) 9939 3295, +612 for international You can also pay by cheque/money order (Orders by mail only) or bank transfer. Make cheques payable to Silicon Chip. 06/26 YES! You can also order or renew your Silicon Chip subscription via any of these methods as well! The best benefit, apart from the magazine? Subscribers get a 10% discount on all orders for parts. PRE-PROGRAMMED MICROS For a complete list, go to siliconchip.com.au/Shop/9 $10 MICROS $15 MICROS ATmega328P ATtiny45-20PU ATtiny85-20PU 110dB RF Attenuator (Jul22), Basic RF Signal Generator (Jun23) 2m VHF CW/FM Test Generator (Oct23) Graphing Thermometer (Mar26), Simple LC Meter (May26) Simple USB Power Monitor (Jun26) PIC12F617-I/P Active Mains Soft Starter (Feb23), Model Railway Uncoupler (Jul23) Battery-Powered Model Railway Transmitter (Jan25) PIC16F1455-I/P Battery-Powered Model Railway TH Receiver (Jan25) Dual Train Controller (Transmitter / TH Receiver, Oct25) PIC16F1455-I/SL Battery-Powered Model Railway SMD Receiver (Jan25) USB Programmable Frequency Divider (Feb25) Dual Train Controller (SMD Receiver, Oct25) PIC16LF1455-I/P New GPS-Synchronised Analog Clock (Sep22) PIC16F1459-I/P K-Type Thermostat (Nov23), Secure Remote Switch (RX, Dec23) Mains Power-Up Sequencer (Feb24 | repurposed firmware Jul24) 8CH Learning IR Remote (Oct24), Heat Transfer Controller (Aug25) Vacuum Controller (Oct25) PIC16F15214-I/SN Silicon Chirp Cricket (Apr23), Mic The Mouse (Aug25) PIC16F15214-I/P Filament Dryer (Oct24), Tool Safety Timer (May25) PIC16F15224-I/SL Multi-Channel Volume Control (OLED Module; Dec23) NFC IR Keyfob Transmitter (Feb25), Rotating Light (Apr25) PIC16F18126-I/SL RGB LED Star (Dec25), DCC/DC Stepper Motor Driver (Apr26) μDCC Decoder (May26; bell [G] or whistle [W]) PIC16F18146-I/SO Versatile Battery Checker (May25), RGB LED ‘Analog’ Clock (May25) USB-C Power Monitor (Aug25), DCC Remote Controller (Feb26) DCC Booster & Reverse Loop Controller (Mar26) PIC16LF15323-I/SL Remote Mains Switch (TX, Jul22), Secure Remote Switch (TX, Dec23) STM32G030K6T6 Variable Speed Drive Mk2 (Nov24) PIC16F1847-I/P PIC16F18877-I/PT Digital Capacitance Meter (Jan25) Dual-Channel Breadboard PSU Display Adaptor (Dec22) Wideband Fuel Mixture Display (WFMD; Apr23) PIC16F88-I/P Battery Charge Controller (Jun22), Railway Semaphore (Apr22) PIC24FJ256GA702-I/SS Ohmmeter (Aug22), Advanced SMD Test Tweezers (Feb23) ESR Test Tweezers (Jun24), Human Comfort Indicator (Jun26) PIC32MX170F256D-501P/T 44-pin Micromite Mk2 (Aug14), 4DoF Simulation Seat (Sep19) PIC32MX170F256B-50I/SP Micromite LCD BackPack V1-V3 (Feb16 / May17 / Aug19) Advanced GPS Computer (Jun21), Touchscreen Digital Preamp (Sep21) PIC32MX170F256B-I/SO Battery Multi Logger (Feb21), Battery Manager BackPack (Aug21) PIC32MX270F256B-50I/SP ASCII Video Terminal (Jul14), USB M&K Adaptor (Feb19) STM32L031F6P6 SmartProbe (Jul25) $20 MICROS ATmega32U4 ATmega644PA-AU PIC32MK0128MCA048 PIC32MX270F256D-50I/PT Wii Nunchuk RGB Light Driver (Mar24) AM-FM DDS Signal Generator (May22) Power LCR Meter (Mar25) Digital Preamplifier (Oct25) $25 MICROS PIC32MX170F256B-50I/SO + PIC16F1455-I/SL Micromite Explore-40 (SC5157, Oct24) PIC32MX470F512H-120/PT Micromite Explore 64 (Aug 16), Micromite Plus (Nov16) PIC32MX470F512L-120/PT Micromite Explore 100 (Sep16) $30 MICROS PIC32MX695F512H-80I/PT Touchscreen Audio Recorder (Jun14) PIC32MZ2048EFH064-I/PT DSP Crossover/Equaliser (May19), Low-Distortion DDS (Feb20) DIY Reflow Oven Controller (Apr20), Dual Hybrid Supply (Feb22) KITS, SPECIALISED COMPONENTS ETC HUMAN COMFORT INDICATOR (SC7646) (JUN 26) Kit: includes all parts, except the case and battery (see p49, Jun26) - white 3D-printed case: portrait (SC7453) or landscape (SC7684) version - 3.3V GY-BME280 module (SC5482) PINBALL MACHINE KITS (JUN 26) $60.00 $12.50 $10.00 (JUN 26) Kit: includes the PCB and all onboard parts (see p63, Jun26) - 0.96in 128x64 cyan OLED screen (USB Power Monitor, Jun26; SC6176) - 0.96in 128x64 white OLED module (USB Power Monitor, Jun26; SC6936) μDCC DECODER KIT (SC7617) SIMPLE LC METER COMPLETE KIT (SC7657) (MAY 26) POWER AMPLIFIER CLIPPING INDICATOR (SC7649) (MAY 26) Includes all the parts and the 3D-printed enclosure (see p67, May26) $50.00 $10.00 $10.00 (MAY 26) Includes all the parts and the optional piezo (wire not included). Specify if you want a bell or whistle sound for the microcontroller (see p88, May26) $25.00 $45.00 Short-form kit: includes the PCB and all onboard parts, the case and power supply are not included (see p35, May26) $95.00 - pair of red & white PCB-mounting RCA sockets (SC2615) $4.00 STEPPER MOTOR DRIVER KIT (SC7601) (APR 26) CALLIOPE AMPLIFIER PARTS (SC6021) (APR 26) DCC BOOSTER / REVERSE LOOP CONTROLLER KIT (SC7579) (MAR 26) Includes all required parts for DCC or DC mode (see p55, Apr26) Includes some of the harder-to-get transistors, resistors and a capacitor Includes all required parts, except for the Jiffy box, OLED screen (see below), power supply and front panel (see p58, Mar26) - 0.91-inch OLED screen (SC7484) (FEB 26) MAINS HUM NOTCH FILTER (SC7598) (FEB 26) Includes all required parts, except for the case and wire/cable (see p63, Feb26) $35.00 Includes everything except for the case and power supply (see p53, Feb26) $150.00 DCC BASE STATION KIT (SC7539) Control Board (SC7659): includes the PCB and all non-optional onboard parts Power Supply (SC7680): includes the PCB and all onboard parts $50.00 Cable & Connector Set (SC7681): includes 17 10-pin box headers, 34 10-pin IDC connectors, 10m of 10-way ribbon cable, 30 2-way pluggable terminal blocks and 20 2-way polarised headers $65.00 SIMPLE USB POWER MONITOR (SC7683) siliconchip.com.au/Shop/ DCC REMOTE CONTROLLER KIT (SC7552) $35.00 $15.00 Includes everything but the plastic case, power supply and some optional parts. The Pico 2 is supplied but not programmed (see p39, Jan26) $90.00 RGB LED STAR KIT (SC7535) (DEC 25) DCC DECODER KIT (SC7524) (DEC 25) EARTH RADIO KIT (SC7582) (DEC 25) RP2350B COMPUTER (NOV 25) Includes the mostly-assembled board and all non-optional components except the power supply (see p43, Dec25) Includes everything in the parts list (see p73, Dec25) Includes everything to build the radio itself except the case and battery, plus the plug for the antenna (see p65, Dec25) $80.00 PICKIT BASIC POWER BREAKOUT KIT (SC7512) (SEP 25) RP2350B DEVELOPMENT BOARD (AUG 25) Includes all parts except the jumper wire and glue (see p39, Sep25) Assembled Board: a pre-assembled PCB with all mandatory parts fitted, optional components are sold separately below (SC7514; see p49, Aug25) - 40-pin header (two are required, SC3189) MIC THE MOUSE KIT (SC7508) (AUG 25) USB-C POWER MONITOR KIT (SC7489) (AUG 25) Includes all non-optional parts except the case, cell & glue (see p39, Aug25) *Prices valid for month of magazine issue only. All prices in Australian dollars and include GST where applicable. # Overseas? Place an order on our website for a quote. $25.00 $55.00 Assembled Board: a fully-assembled PCB with all non-optional components, front and rear panels are sold separately below (SC7531; see p28, Nov25) - front & rear panels (SC7532) - 8MiB APS6404L-3SQR-SN PSRAM SOIC-8 IC (SC7530) Includes all parts except a CR2032 cell (see p64, Aug25) $45.00 $7.50 $50.00 (JAN 26) $90.00 $7.50 $5.00 $20.00 $30.00 $1.00ea $37.50 $60.00 PRINTED CIRCUIT BOARDS & CASE PIECES PRINTED CIRCUIT BOARD TO SUIT PROJECT SECURE REMOTE SWITCH RECEIVER ↳ TRANSMITTER (MODULE VERSION) ↳ TRANSMITTER (DISCRETE VERSION IDEAL BRIDGE RECTIFIER, 28mm SQUARE SPADE ↳ 21mm SQUARE PIN ↳ 5mm PITCH SIL ↳ MINI SOT-23 ↳ STANDALONE D2PAK SMD ↳ STANDALONE TO-220 (70μm COPPER) RASPBERRY PI CLOCK RADIO MAIN PCB ↳ DISPLAY PCB KEYBOARD ADAPTOR (VGA PICOMITE) ↳ PS2X2PICO VERSION MICROPHONE PREAMPLIFIER ↳ EMBEDDED VERSION RAILWAY POINTS CONTROLLER TRANSMITTER ↳ RECEIVER LASER COMMUNICATOR TRANSMITTER ↳ RECEIVER PICO DIGITAL VIDEO TERMINAL ↳ FRONT PANEL FOR ALTRONICS H0190 (BLACK) ↳ FRONT PANEL FOR ALTRONICS H0191 (BLACK) ARDUINO FOR ARDUINIANS (PACK OF SIX PCBS) ↳ PROJECT 27 PCB WII NUNCHUK RGB LIGHT DRIVER (BLACK) SKILL TESTER 9000 PICO GAMER ESP32-CAM BACKPACK WIFI DDS FUNCTION GENERATOR 10MHz to 1MHz / 1Hz FREQUENCY DIVIDER (BLUE) FAN SPEED CONTROLLER MK2 ESR TEST TWEEZERS (SET OF FOUR, WHITE) DC SUPPLY PROTECTOR (ADJUSTABLE SMD) ↳ ADJUSTABLE THROUGH-HOLE ↳ FIXED THROUGH-HOLE USB-C SERIAL ADAPTOR (BLACK) AUTOMATIC LQ METER MAIN AUTOMATIC LQ METER FRONT PANEL (BLACK) 180-230V DC MOTOR SPEED CONTROLLER STYLOCLONE (CASE VERSION) ↳ STANDALONE VERSION DUAL MINI LED DICE (THROUGH-HOLE LEDs) ↳ SMD LEDs GUITAR PICKGUARD (FENDER JAZZ BASS) ↳ J&D T-STYLE BASS ↳ MUSIC MAN STINGRAY BASS ↳ FENDER TELECASTER COMPACT OLED CLOCK & TIMER USB MIXED-SIGNAL LOGIC ANALYSER (PicoMSA) DISCRETE IDEAL BRIDGE RECTIFIER (TH) ↳ SMD VERSION MICROMITE EXPLORE-40 (BLUE) PICO BACKPACK AUDIO BREAKOUT (with conns.) 8-CHANNEL LEARNING IR REMOTE (BLUE) 3D PRINTER FILAMENT DRYER DUAL-RAIL LOAD PROTECTOR VARIABLE SPEED DRIVE Mk2 (BLACK) FLEXIDICE (RED, PAIR OF PCBs) SURF SOUND SIMULATOR (BLUE) COMPACT HIFI HEADPHONE AMP (BLUE) CAPACITOR DISCHARGER PICO COMPUTER ↳ FRONT PANEL (BLACK) ↳ PWM AUDIO MODULE DIGITAL CAPACITANCE METER 5MHZ 40A CURRENT PROBE (BLACK) BATTERY MODEL RAILWAY TRANSMITTER ↳ THROUGH-HOLE (TH) RECEIVER ↳ SMD RECEIVER ↳ CHARGER USB PROGRAMMABLE FREQUENCY DIVIDER HIGH-BANDWIDTH DIFFERENTIAL PROBE NFC IR KEYFOB TRANSMITTER DATE DEC23 DEC23 DEC23 DEC23 DEC23 DEC23 DEC23 DEC23 DEC23 JAN24 JAN24 JAN24 JAN24 FEB24 FEB24 FEB24 FEB24 MAR24 MAR24 MAR24 MAR24 MAR24 MAR24 MAR24 MAR24 APR24 APR24 APR24 MAY24 MAY24 MAY24 JUN24 JUN24 JUN24 JUN24 JUN24 JUL24 JUL24 JUL24 AUG24 AUG24 AUG24 AUG24 SEP24 SEP24 SEP24 SEP24 SEP24 SEP24 SEP24 SEP24 OCT24 OCT24 OCT24 OCT24 OCT24 NOV24 NOV24 NOV24 DEC24 DEC24 DEC24 DEC24 DEC24 JAN25 JAN25 JAN25 JAN25 JAN25 JAN25 FEB25 FEB25 FEB25 PCB CODE 10109231 10109232 10109233 18101241 18101242 18101243 18101244 18101245 18101246 19101241 19101242 07111231 07111232 01110231 01110232 09101241 09101242 16102241 16102242 07112231 07112232 07112233 SC6903 SC6904 16103241 08101241 08104241 07102241 04104241 04112231 10104241 SC6963 08106241 08106242 08106243 24106241 CSE240203A CSE240204A 11104241 23106241 23106242 08103241 08103242 23109241 23109242 23109243 23109244 19101231 04109241 18108241 18108242 07106241 07101222 15108241 28110241 18109241 11111241 08107241/2 01111241 01103241 9047-01 07112234 07112235 07112238 04111241 9049-01 09110241 09110242 09110243 09110244 04108241 9015-D 15109231 Price $5.00 $2.50 $2.50 $2.00 $2.00 $2.00 $1.00 $3.00 $5.00 $12.50 $7.50 $2.50 $2.50 $7.50 $7.50 $5.00 $2.50 $5.00 $2.50 $5.00 $2.50 $2.50 $20.00 $7.50 $20.00 $15.00 $10.00 $5.00 $10.00 $2.50 $5.00 $10.00 $2.50 $2.50 $2.50 $2.50 $5.00 $5.00 $15.00 $10.00 $12.50 $2.50 $2.50 $10.00 $10.00 $10.00 $5.00 $5.00 $7.50 $5.00 $2.50 $2.50 $2.50 $7.50 $7.50 $5.00 $15.00 $5.00 $10.00 $7.50 $5.00 $5.00 $2.50 $2.50 $5.00 $5.00 $2.50 $2.50 $2.50 $2.50 $5.00 $5.00 $2.50 For a complete list, go to siliconchip.com.au/Shop/8 PRINTED CIRCUIT BOARD TO SUIT PROJECT POWER LCR METER WAVEFORM GENERATOR PICO 2 AUDIO ANALYSER (BLACK) PICO/2/COMPUTER ↳ FRONT & REAR PANELS (BLACK) ROTATING LIGHT (BLACK) 433MHZ TRANSMITTER VERSATILE BATTERY CHECKER ↳ FRONT PANEL (BLACK, 0.8mm) TOOL SAFETY TIMER RGB LED ANALOG CLOCK (BLACK) USB POWER ADAPTOR (BLACK, 1mm) HWS SOLAR DIVERTER PCB & INSULATING PANELS SSB SHORTWAVE RECEIVER PCB SET ↳ FRONT PANEL (BLACK) 433MHz RECEIVER SMARTPROBE ↳ SWD PROGRAMMING ADAPTOR DUCTED HEAT TRANSFER CONTROLLER ↳ TEMPERATURE SENSOR ADAPTOR ↳ CONTROL PANEL MIC THE MOUSE (PCB SET, WHITE) USB-C POWER MONITOR (PCB SET, INCLUDES FFC) HOME AUTOMATION SATELLITE PICKIT BASIC POWER BREAKOUT DUAL TRAIN CONTROLLER TRANSMITTER DIGITAL PREAMPLIFIER MAIN PCB (4 LAYERS) ↳ FRONT PANEL CONTROL ↳ POWER SUPPLY VACUUM CONTROLLER MAIN PCB ↳ BLAST GATE ADAPTOR POWER RAIL PROBE RGB LED STAR EARTH RADIO DCC DECODER DCC BASE STATION MAIN PCB ↳ FRONT PANEL REMOTE SPEAKER SWITCH ↳ CONTROL PANEL DCC REMOTE CONTROLLER MAINS HUM NOTCH FILTER MAINS LED INDICATOR DCC BOOSTER / REVERSE LOOP CONTROLLER ↳ FRONT PANEL SOLAR PANEL PROTECTOR (WHITE) GRAPHING THERMOMETER PICOSDR CONTROL PCB ↳ RF PCB ↳ FRONT PANEL (BLACK) DCC/DC STEPPER MOTOR DRIVER CALLIOPE AMPLIFIER MICROMITE AUDIO PLAYER ADD-ON ↳ ALL-IN-ONE μDCC DECODER SIMPLE LC METER WIFI ALARM MONITOR POWER AMPLIFIER CLIPPING INDICATOR DATE MAR25 MAR25 MAR25 APR25 APR25 APR25 APR25 MAY25 MAY25 MAY25 MAY25 MAY25 JUN25 JUN25 JUN25 JUN25 JUL25 JUL25 AUG25 AUG25 AUG25 AUG25 AUG25 SEP25 SEP25 OCT25 OCT25 OCT25 OCT25 OCT25 OCT25 NOV25 DEC25 DEC25 DEC25 JAN26 JAN26 JAN26 JAN26 FEB26 FEB26 FEB26 MAR26 MAR26 MAR26 MAR26 APR26 APR26 APR26 APR26 APR26 APR26 APR26 MAY26 MAY26 MAY26 MAY26 PCB CODE Price 04103251 $10.00 04104251 $5.00 04107231 $5.00 07104251 $5.00 07104252/3 $10.00 09101251 $2.50 15103251 $2.50 11104251 $5.00 11104252 $7.50 10104251 $5.00 19101251 $15.00 18101251 $2.50 18110241 $20.00 CSE250202-3 $15.00 CSE250204 $7.50 15103252 $2.50 P9054-04 $5.00 P9045-A $2.50 17101251 $10.00 17101252 $2.50 17101253 $2.50 SC7528 $7.50 SC7527 $7.50 15104251 $3.50 18106251 $2.00 09110245 $3.00 01107251 $30.00 01107252 $2.50 01107253 $7.50 10109251 $10.00 10109252 $2.50 P9058-1-C $5.00 16112251 $12.50 06110251 $5.00 09111241 $2.50 09111243 $5.00 09111244 $5.00 01106251 $5.00 01106252 $2.50 09111245 $5.00 01003261 $7.50 10111251 $2.50 09111248 $5.00 09111249 $5.00 17112251 $7.50 04102261 $3.00 CSE251101 $5.00 CSE251102 $5.00 CSE251103 $7.50 09111242 $2.00 01111212 $5.00 01110251 $2.50 01110252 $5.00 09111247 $1.50 04103261 $2.50 01304261 $2.50 01104261 $15.00 PINBALL MACHINE CONTROL BOARD ↳ POWER SUPPLY ↳ PLAYER LED BOARD ↳ SCORE LED BOARD ↳ LED OUTPUT BOARD ↳ BUMPER LED BOARD ↳ CASCADE LED BOARD ↳ SWITCH INPUT BOARD ↳ GENERAL INPUT BOARD ↳ HIGH-CURRENT INTERFACE ↳ ROLLOVER INTERFACE ↳ BUMPER DRIVER SSB TRANSMITTER (MikeOne/Two/Three) SIMPLE USB POWER MONITOR HUMAN COMFORT INDICATOR JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 JUN26 08107261 08107262 08107263 08107264 08107265 08107266 08107267 08107268 08107269 08107260 08117261 08117262 06103261 04104261 21105261 NEW PCBs $25.00 $7.50 $2.50 $5.00 $2.50 $5.00 $5.00 $2.50 $2.50 $2.50 $2.50 $5.00 $2.50 $5.00 $5.00 We also sell the Silicon Chip PDFs on USB, RTV&H USB, Vintage Radio USB and more at siliconchip.com.au/Shop/3 Vintage Radio The Sailor 66T Navigation Radio This radio was very popular in the 1960s, 1970s and 1980s. Primarily, it was deployed for radio navigation in the North Sea between Norway and Scotland, as well as being used in the North Atlantic prior to modern GPS navigation systems. It was made by Danish company SP Radio Aalborg. By Dr Hugo Holden T his article is more about the radio itself than its radio direction finding (RDF) applications. However, numerous radio navigation and radio direction finding systems existed in the past that no longer do because satellite navigation (eg, GPS) has taken over. The RDF system briefly discussed here was called Consol. The navigator takes a direction reading by rotating their radio’s direction finding (DF) antenna to receive and null signals emanating from a specific fixed radio beacon on land. By taking bearings from only two known radio beacons, then plotting those on a chart, the navigator can determine the vessel’s position. The DF antenna typically consisted of a ferrite rod and tuned coil that could be rotated manually. Loops can 88 Silicon Chip also be used. When the long axis of the ferrite rod is aligned with the axis of the beam from the beacon, there is a signal null. This is called a ‘relative bearing’. However, the rod could be pointed either towards or away from the beacon due to its figure-8 sensitivity pattern. To get around this 180° ambiguity and make a ‘relative determination’, there is a “Sense Switch” on the radio’s front panel. When deployed, the switch combines some of the received signal from the main antenna, which is omni-directional. This creates a cardioid sensitivity pattern. Therefore, when the ferrite rod is rotated 90° to a signal maximum, the combined result is more sensitive in one direction than the other. The navigator can combine the information with compass readings too. Australia's electronics magazine For North Sea navigation, one radio navigation beacon was on 266kHz in Bushmills on the north coast of Northern Ireland, and another on 319kHz at Stavanger, on the coast of Norway – see Fig.1. They had a typical transmission range in the order of 1000 nautical miles (about 1850km). The beacons transmitted their carrier waves as dots in one sector (sector A) and dashes in another (sector B) during the direction-finding transmission period. The transmission period for direction finding is 60s with a one-second pause, then the station call sign is transmitted for six seconds. Most of the remainder is a long dash (heard as a long tone due to the radio’s BFO) for 50 seconds, followed by a three-­ second pause. The DF information repeats again, siliconchip.com.au Photo 1: the connection panel for the radio is utilitarian, but it provides everything you need. Photo 2: when the optional speaker box is mounted, the connection panel is inside it. Photo 3: the loudspeaker is a quality unit. so the entire cycle takes 120 seconds. The speed of the rotation is one sector width per 120 seconds. For example, if you were in the position marked × in Fig.1, you would hear 48 dots of the remaining A sector and 12 dashes from the B sector from the Stavanger Beacon in Norway as the beam passes by your location. On the other hand, you would hear 28 dashes and then 32 dots when the beam from the Bushmills beacon passes. Aside from the internal battery pack, three more DC power options can be selected using an internal rotary switch for 12V, 24V or 32V operation – see Photo 4 overleaf. The radio sports a nice audio amplifier with two good-sized Philips transformers, each with a 10 × 10mm cross-sectional core area, visible below the switch. Just beside the upper (output) transformer on each side are the two AC128 output transistors mounted to heat-conducting fins. The audio output is rated at 1.8W. Also in Photo 4, at the bottom, is a large power stud-type 9.1V zener diode (BZZ19) mounted on a 4mm-thick, 6 × 8cm black-painted aluminium heatsink. This is because the voltage regulator design for the external power Features of the 66T receiver Apart from its RDF capabilities, it is a highly sensitive and capable superheterodyne radio receiver. It can be powered from an internal battery pack of six D cells in a battery box on the right side of the radio housing. When the attached speaker box option is not used, the radio’s power input panel, shown in Photo 1, is simply screwed onto the left-hand side of the radio. However, when the accessory speaker box is used, this panel sits on standoffs that attach the speaker housing to the radio’s housing and is attached with thumb nuts, as shown in Photo 2. To remove these, the speaker and the front panel retaining it must be removed. The speaker box has a hole in its rear to allow the main antenna connection to pass through. The speaker is a high-quality four-inch (102mm) unit – see Photo 3. Interestingly, it is mounted on a timber baffle, which likely improves the damping in the cabinet a little. siliconchip.com.au × Fig.1: the locations of the two main radio beacons for the North Sea area. The radial sectors provided a way to determine the ship’s location based on the signals received from both stations. Australia's electronics magazine June 2026  89 Photo 4: from top to bottom, you can see the power selector switches, output transformer flanked by the output transistors, phase splitter transformer and power regulator zener diode. Photo 5: the dial is large and clear. Note the index mark and calibration marks. option for this radio is a shunt regulator design. The excess input voltage is dropped across a substantial ceramic wire-wound power resistor. While that might seem inefficient to some, the beauty of it is that it makes the power supply and radio highly resistant to electrical abuse such as high voltage transients on the DC supply, because the zener diode snubs them off. It also prevents accidental reverse polarity accidents because it conducts in the forward direction in that case. More complex series pass voltage regulator circuits are more easily damaged, often requiring TVS protection devices or other protective parts. The radio can operate on four bands: • Long-wave (LW): 150-285kHz • Navigation band (NW): 255425kHz using the DF antenna input • Medium-wave (MW): 5251600kHz • Short-wave (SW): 1.6-4.2MHz The 66T has a very attractive glass dial with a well-calibrated scale for each band (see Photo 5). The dial contains some additional markings that are very helpful in performing an alignment (calibration) of the radio. While the calibration frequencies were mentioned in the manual, I could not find any mention of the critical index mark in the text. Having said that, I did not have a fully translated manual. The index mark controls the mechanical relationship of the dial pointer to the three-gang variable capacitor. That relationship in my radio was badly off, making calibration 90 Silicon Chip and tracking impossible until it was corrected. Photo 6 shows the general architecture of the radio. The rear section of the 3-gang variable capacitor is the one that tunes the set’s local oscillator and its associated inductances for each band. The middle section tunes the inductances associated with the RF stage, and the front section tunes the inductances associated with the antenna circuit. Notice the bends in the outer adjustment wings of the rear (oscillator) section of the variable capacitor; these are discussed later. Another notable feature of this radio is the 470kHz intermediate frequency (IF) amplifier board. This uses double-­ tuned IF transformers. On some versions of this IF board, the first IF transformer had an additional small coil added to its primary. It was a signal injection point labelled H. This was so that a low output resistance sweep generator could easily be connected without damping the tuning on the first IF coil. However, in later versions, such as my radio, that coil was dispensed with, and instead, two test points, corresponding to “Test point H”, were provided across the first IF transformer primary. This is a very high-­ impedance zone. I had to make a special adaptor to drive it, as will be discussed briefly in the alignment section. There was a note in the manual: “Never touch the intermediate frequency alignment unless proper Australia's electronics magazine measuring equipment is available” (by this they mean a sweep generator and scope). In this radio, if the IF tuning slugs are simply peaked at 470kHz, the overall bandwidth is far too low and the recovered audio modulation is therefore very muffled and lacking in high-frequency components. General specifications The radio weighs in at 8kg. The sensitivity of this radio is very good on the SW band, giving 50mW output for only 3μV RF input, specified with 30% audio modulation. The IF bandwidth is specified as 6.5kHz. This can only come about with correct tuning of the double-tuned IF transformers, as will be outlined in the alignment section. The image suppression was specified as an excellent 50dB or better at 2.2MHz. The audio frequency response is stated as 100-3000Hz (without the filter switched in). I found this a little restrictive. I reduced the value of a filter capacitor to widen the frequency response in the audio section, which made music listening better. In keeping with many transistor radios like this, the current consumption is amazingly low at around 40-150mA depending on the volume setting. The six D cells in the battery carrier have a very long life. On external power, due to the nature of the shunt zener voltage regulator, the current consumption is 400mA. The radio’s signal-to-noise ratio was specified in the manual. To measure siliconchip.com.au this requires an output signal from a low-impedance source (25W) and the use of a dummy antenna. This will be described in a later section. The radio also contains five quartz crystals for fixed-frequency reception. In my radio, the crystals fitted were for 2182kHz, 1792kHz, 1834kHz, 1841kHz and 1848kHz. The crystals are housed in a row on the lower rear chassis (see Photo 7). Circuit details and factory modifications There were a number of revisions of the circuit by the manufacturer. The first 66T set was series A, then going all the way to series K with small changes. After series A, most of the schematics are very similar except that after series A, an extra switch gang was added to switch the two local oscillator signals (crystal versus the four tuned bands) as separate signals into the mixer circuit. After series B, the RF amplifier was modified. It turned out that either electrostatic discharges (lightning) or RF output power from the ship’s transmitter via the antenna could fry the BF115 RF amplifier transistor. The general approach to this sort of problem is to use diodes to protect the transistor. Interestingly, in some of their earlier versions, they had a diode in series with the base of the RF amplifier transistor, likely to augment the AGC rather than to protect the transistor. Although a large positive signal impulse would tend to reverse-bias the diode, the diode’s reverse breakdown voltage is not enough to protect the transistor from very high voltages. No such diode was used in series B. Then they added a diode across the base-emitter junction of the BF115, as explained in the manual (translated from the original Norwegian): “Since we have received complaints that the RF transistor in some receivers burns out due to static electricity on the antenna or RF voltage from the transmitter, we have introduced protection for the transistor in future production runs. The protection consists of a silicon diode mounted across the base-emitter junction of the RF transistor.” The RF sections of the series C circuit are shown in Fig.2. This one corresponds to my radio. I have highlighted important sections with boxes. siliconchip.com.au Photo 6: the chassis layout is neat. You can clearly see the three-gang variable capacitor on the left; I bent some of the plates on the lower gang to improve the tracking. Photo 7: the five plug-in crystals for fixed-frequency tuning. They have matching coils and capacitors. Photo 8: the capacitors that match the crystals shown in Photo 7. Potentiometer R2 is the “Sense– Balance” preset, which is accessible through a hole in the front panel, just beside the Sense Switch. In my radio, a green LED power light had been placed in that hole. Its appearance and that of the wiring to it suggest it was done by the manufacturer, but it was only operational on external power. Australia's electronics magazine June 2026  91 RF amplifier Diode mixer RF ▲ Sense preset potentiometer (R2) ▲ OSC Channel selector switch* 92 Silicon Chip Crystal oscillator Oscillator for SW, MW, NW & LW bands Band selector switch* I switched the LED over so that it runs whenever the radio is powered from any source. It is good to have, because when the radio is powered by batteries, it is all too easy to leave it accidentally switched on. As noted previously, the Sense system creates a mix between the signal received by the DF antenna and the main antenna to create an asymmetry in the reception sensitivity so it can be used for unambiguous direction-­ finding. The Test H inputs are used to couple in the sweep signal for aligning the IF amplifier. This arrangement is not as ideal as the earlier version with the small coupling coil. Fig.2 shows the whole circuit for the 66T. The left-hand of the circuit shows the coil sets and the top end tuning capacitor for the SW- band. The other coil sets and tuning capacitors for the other bands (three for each band: the antenna coil, RF coil and oscillator coil) are connected to the empty positions on the rotary switches. For the fixed crystal reception frequencies, there are five crystals with an antenna coil and RF coil associated with each crystal channel. In this case, fixed polystyrene tuning capacitors are used for the antenna and RF 470kHz double-tuned IF amplifier and AM detector stages. The arrangement here accounts for 10 coils in total and 10 fixed tuning capacitors associated with them. These capacitors, two for each of the crystal channels, are mounted vertically on the side of the chassis, as shown in Photo 8. BF115 silicon transistors are used in both of the oscillators and the RF amplifier. AF127 germanium transistors are used for the IF amplifier. The specifications of both parts are excellent. The BF115 is a spectacularly good silicon planar epitaxial transistor. Its transition frequency is in the order of 230MHz, and it has a low noise figure of 1.2dB at 1MHz. The AF127 belongs to a family of parts that replaced the AF11x series of transistors, which are now prone to failure from tin whiskers. Fortunately, the AF12x series of parts does not suffer from these problems. The AF127 is a diffused-alloy transistor, with a transition frequency of 75MHz. It is a very capable part for RF and radio work, with a low noise figure of 1.5dB. One of its very useful features in IF amplifier applications is that it has a very low feedback capacitance, only in the order of 1.5pF. This means it can work as a stable IF amplifier Australia's electronics magazine without requiring neutralisation. On the other hand, older-generation germanium RF parts, such as the OC45, had feedback capacitances in the order of 10pF and always required neutralisation feedback components added to be stable in an IF amplifier application. One would therefore expect the performance (especially noise and sensitivity figures) of a radio such as the 66T to be very good on account of the very capable RF, IF and oscillator transistors. Certainly, the sensitivity figure specified for the SW band being less than 3μV input for 50mW output is very good. The signal-to-noise (S/N) ratio is specified at 10dB below 1MHz with a 10μV signal and a dummy antenna. The AGC system is shown on the right of Fig.2. It feeds a separate line to the RF amplifier and the first IF amplifier. A separate preset is used to adjust the AGC for the RF amplifier. The front-panel RF gain control affects both the AGC to the IF and the RF amplifier. The specified performance is that an increase in RF input voltage from 31μV to 100mV will increase the output by less than 10dB. The radio’s metal chassis is independent of the actual positive and negative siliconchip.com.au AGC amplifier and signal meter driver Audio preamp and audio output see text ★ ★ Front panel RF PCB preset gain (R46) (R49) ◀ ★ ◀ nfb AGC RF Signal meter BFO Shunt zener regulator (D7) Dial lamp * Crystal array, select switches and 10 associated coils and five crystal – two coils (L15 & L16) and one crystal E shown * three-gang 500pF variable capacitor, nine coils and nine top-end tuning capacitors: coils L1-L3 and capacitors C1-C3 shown – SW band supply power system, only bypassed to those with capacitances. So the radio could be mounted in a vessel that had either a positive or negative ground power supply system. In my radio, I made some modifications to three capacitor values in the audio system, shown with purple stars in Fig.2. One problem I encountered was a noisy volume control, which persisted even after substituting in a new control and renewing the coupling capacitors. I changed both the capacitors around the volume control to low-­ leakage 1μF axial tantalum capacitors. I could have used film capacitors, but I had no axial types of that value that fit the PCB well. Reducing the 10μF capacitor, leading away from the control, to 1μF substantially reduced low-frequency noise with control rotation. It did not degrade the audio low-frequency response. In this circuit, the resistances are such that the frequency response, even with a 1μF capacitor (rather than the 10μF value), does not reach -3dB until it is below about 20Hz. The 22nF capacitor in the base circuit of T9 resulted in fairly heavy audio high-frequency roll-off, muffling siliconchip.com.au Input power conditioning Fig.2: the radio’s circuit. It uses a 470kHz IF and has five sets of crystals/coils/capacitors for quick fixed-frequency tuning. The audio stages, AGC system, BFO and power supply regulator sections of the 66T radio are shown on the right-hand side of the circuit. the sound somewhat. That may well have been OK for voices but not so much for music. For a better tone balance, I reduced that value to 1.5nF. I could not find anything else that required changing. In terms of faulty parts, the only capacitor that I had to replace was C100, a 400μF electrolytic that I replaced with two parallel axial 220μF parts. The original had gone high in ESR, resulting in a motorboating effect with low-frequency oscillations in the audio. The audio amplifier in the radio is quite capable and gives a good sound with the speaker in the speaker box. The amplifier has negative feedback to reduce distortion, and audio transformers with good-sized iron cores. Conveniently, SP Radio provided the results of injecting signal voltages at different test points in the radio, under the condition that 50mW is being produced with a 30% modulation at 400Hz. This is helpful in verifying that the radio is working to specifications. Dial miscalibration When I checked the radio initially, I found that none of the received station frequencies were close to their Australia's electronics magazine Photos 9 & 10: after resleeving the knobs, they are close to the original but not identical. I made the sleeves by boring out a plastic donor knob. June 2026  93 Scope 1: if all the IF adjustments are tuned for 470kHz, the result is a response that’s too narrow and peaked. Scope 2: how the IF response should look like with correct alignment, with each element tuned to a slightly different centre frequency. Mechanical repairs One curio is that this model of radio has a peculiar failure rate of the black phenolic sleeves that were placed over the chromed knobs. They have a habit of splitting, falling off and getting lost. They were all missing on this radio, except one. This required some donor phenolic knobs. I machined them out with an internal taper to create a sleeve or shell so they would slip over the original chrome metal knobs to act as a reasonable replacement. As can be seen in Photos 9 & 10, these have finer finger-grip grooves than the original sleeve, but they were as close as I could find. Another mechanical problem that cropped up related to the signal meter; it was sticking. Investigation revealed that some rust crystals had projected out of the side wall of the laminated iron pole pieces and were catching on the meter’s coil form. I removed these by slipping in strips of sticky tape to extract them. Some people have attempted to blow debris out of meter movements with compressed air. This is better avoided, as it normally destroys the hairsprings and movement. Photo 11: I use this RF signal generator and this frequency counter for aligning radios. 94 Silicon Chip dial markings. Sailor went to the trouble of making a very precise-looking dial, suggesting the unit should have good calibration. This is unlike some domestic radios with poor dial markings, without graduations between them, and somewhat loosely spaced dial legends. On the SW band, it was not possible to receive the frequencies at all above about 3.6MHz. The radio was significantly out of alignment. One of the true arts of radio restoration is in the radio’s alignment. It might not be so important in some radios, such as pocket transistor radios, with single tuned IF transformers and limited dial markings. Still, in commercial types such as communications receivers, calibration is very important. It is often clear from inspecting a radio’s tuning dial whether the manufacturer thought of it as more of a domestic product, or more of a scientific instrument, where the dial information was expected to be reasonably accurate and meaningful. Australia's electronics magazine Alignment tools I have several useful tools to help siliconchip.com.au with radio alignment. One is the Philips PM5326 RF Generator, which includes an accurate frequency counter. It puts out exactly 50mV RMS into a 75W load on 0dB attenuation. It has an excellent shielded RF attenuator that goes beyond -80dB. A -80dB output corresponds to 5μV RMS into a 75W load. In most radios, the local oscillator (LO) runs at the intermediate frequency above the received station frequency. To examine the LO, I have a frequency counter with a programmable offset value, in this case set to subtract 470kHz, the set’s IF. As Photo 11 shows, with no signal input applied to the counter, its display reads 99.5300MHz. It has an input sensitivity in the range of 10mV to 40mV and its maximum counting frequency is about 48MHz. The input impedance is too low (in common with many counters) and its input capacitance is too high to directly connect to a radio’s oscillator circuitry without loading it and causing a large frequency shift. To solve that, I designed a buffer circuit that is described in Circuit Notebook on page 79. IF alignment It is the gain and bandwidth of the IF stages in a superhet radio that confer much of the radio’s selectivity and sensitivity. There is less selectivity in the RF and antenna circuits, as these need wide enough bandwidth to accommodate tracking errors. One basic principle of superhet radio alignment is to make sure the IF stages are correctly set up with the correct centre frequency and bandwidth (if the latter is adjustable). With the IF amplifier in the 66T, if all the IF slugs are peaked to the same frequency, the bandpass response is far too narrow. The result is a muddy sound with a loss of treble. The sweep result shown in Scope 1 occurs when all the IF transformers are peaked at 470kHz. The resulting narrow bandwidth response has an asymmetrical skirt. The manufacturer specified a bandwidth of 6.5kHz, meaning that the response should be 3dB down at ±3.25kHz around the 470kHz centre frequency. This is easily achieved by adjusting the IF slugs while using a sweep generator, with the result shown in Scope 2. siliconchip.com.au Indexing Prior to any other alignment processes, as well as the IF being correctly adjusted, it is important that when the dial pointer is pointing to a specific legend at the low-frequency end of the dial, the variable capacitor is in the correct position. This is so that over the tuning range, capacitance varies over the correct range to suit the coil set. The question is, where is that position? It may not be explicitly stated, or for that matter even present on some dials. It was once a custom for a manufacturer to put an index mark on the dial. In most cases, the pointer should be aligned with this mark with the variable capacitor fully meshed, or close to that. I found no mention of this mark in the Sailor 66T manual, although one was evident on the dial. This setting for my radio was so far off that the variable capacitor had completely unmeshed by about 3.6MHz on the shortwave band! It was therefore impossible to tune in to any frequencies above that. The dummy antenna The Sailor 66T manual says to use an IEC dummy antenna to interface the RF generator with the radio. It says to use it with a generator with a 25W output resistance (eg, a 50W output with a 50W terminator applied to bring the output impedance down to 25W). This is to be used for the LW, NW and MW bands but not the SW band, where they suggested using the 25W signal source directly. A typical dummy antenna is meant to be driven by a low source resistance of 25W or less, but there is little practical difference in using a 37.5W source, ie, a terminated 75W source. Fig.3 shows the American IRE dummy antenna circuit. Its performance is shown graphically in Fig.4. Unfortunately, the IRE version of the dummy antenna does not suit the Sailor radio, especially on the LW band, Fig.3: a standard IRE dummy antenna circuit. It will work for the MW & SW bands but is no good on the LW range. 150-280kHz. The IRE circuit mainly suits radios with MW and SW bands. Unfortunately, the circuit for the recommended IEC (not IRE) dummy antenna is not readily available. If attempting to align this radio with the IRE dummy antenna, the LW antenna coil would not come into range on its tuning slug because, at 17kHz, the IRE dummy antenna does not apply enough load to the primary circuit of the antenna coil. Its load is in the order of nearly 5kW capacitive reactance at 170kHz, on account of the 220pF capacitor. As a result of this, and the mutual coupling, the tuned secondary resonant frequency of the LW antenna coil was too low. Even with the ferrite slug removed from the former, it still only came up to a maximum of around 160kHz. As a solution, I made an adaptor to emulate the capacitance of the realistic antenna system shown in Fig.5(a). This capacitance is present regardless of what band on the radio is being used, so it is suitable to adapt the generator to the radio for alignment purposes on all bands. Since the length of the wire antenna is relatively short compared to the Fig.4: an impedance chart for the IRE dummy antenna shown in Fig.3. Australia's electronics magazine June 2026  95 Obtaining The Best Possible Dial Calibration Textbook alignment Once the IF and dial alignments are correct, the oscillator coil’s tuning slug (or its adjustable padder capacitor, if there is one) is set to receive the tone modulated test frequency for maximum signal out of the IF’s detector or the audio amplifier stages. The dial pointer is then moved to an instructed position near the high end of the band, and the generator set to that frequency. The oscillator’s trimmer capacitor, which is in parallel with the oscillator’s variable capacitor gang, is moved to tune that in for a peak signal. These two steps are then repeated a few times because they interact. After that, the antenna and RF coil slugs can be peaked at the low end of the band at the same dial locations, and the trimmer capacitors associated with those coils are peaked at the recommended high-end frequency. With this common alignment method, the tracking of the oscillator’s frequency is exactly correct at the upper and lower points and at some intermediate point. These three frequencies are called ‘crossover frequencies’. Photos 12 & 13: the oscillator tuning gang in the radio as I received it (top). The oscillator tuning gang in the radio after I minimised the tracking errors (bottom). 96 Silicon Chip In the tracking zones around the crossover frequencies, the local oscillator runs a little faster or slower than ideal. These errors, shown in Fig.a, are called ‘tracking errors’. They are intrinsic to a superhet radio where the variable capacitor’s gangs have the same capacity and the oscillator gang has a required padder capacitor in series to reduce its overall capacitance. When the padder capacitor is the correct value, the magnitudes of the + and – tracking errors, at their worst, are about equal. If significant enough, they can result in a reduction in the sensitivity of the radio and/or a reduction in the image rejection in those zones. Tracking alignment There is an alternative method to adjust a radio to ensure that the dial markings match the received frequencies as closely as possible and that the tracking errors are minimised. No modulation is required on the carrier from the RF generator in this process. This method does not rely on the IF amplifier, and it can also be used to measure the magnitude of the tracking errors. However, the IF amplifier must be properly set up for a final result. Once the IF has been set up correctly with the sweep generator and marker generator, and the mechanical relationship of the variable capacitor angle and dial pointer are set, the oscillator transistor stage is disabled. I do this by shorting the base to the emitter with a 100W resistor in the case of a separate oscillator transistor. For designs with mixer/oscillator stages, shorting out the oscillator’s resonant coil also works. Then, for each band, the cores in the RF and antenna coils are peaked on the manufacturer’s recommended low frequency, and the capacitor trimmers at the upper band end in the usual way, but in this case by monitoring the output of the antenna/RF tuned circuit (if present) where it feeds into the mixer, or on the RF amplifier’s variable capacitor gang. You need a low-capacitance (<1pF) probe to monitor the RF output; I have Australia's electronics magazine designed one that is presented in Circuit Notebook this month, on page 79. This allows the tuned carrier to be seen on a scope with negligible detuning effects on the RF stage’s resonant circuit. In essence, this part of the radio is being treated as a TRF circuit. Once the upper and lower frequency points are set for the antenna & RF stages, the positions of all other dial markings with respect to the dial pointer (representing the variable capacitor’s angle) can be checked to see how closely the dial markings and pointer match the applied carrier frequencies. This is why an RF generator with a built-in frequency counter is very helpful. In the Sailor radio, it turned out that the antenna and tuned RF stages were closely correlated with the variable capacitor’s pointer and the dial markings. In this case, there is no requirement to adjust the wings on the variable capacitor gangs associated with those two radio frequency stages. The glass dial’s markings had clearly been created from a law defining the tuned frequencies when a straight-line wavelength (SLW) variable capacitor was used. Rather than calibrating the dial in wavelength, which would have more evenly spread the values, it was calibrated in frequency. This is very convenient because any adjustments to the oscillator’s fine tuning can be targeted to match the dial markings as closely as possible too. This way, both tracking errors and dial marking/pointer errors for tuned stations can be simultaneously minimised. In the case that the dial markings closely follow the oscillator’s tuning, the wings of the oscillator’s variable capacitor should never be altered from standard. Only the wings of the RF and antenna sections should be adjusted (with the oscillator disabled) if required, to better match the dial/ pointer relationship. Setting up the oscillator After re-enabling the oscillator, adjust it at the low and high recommended frequency points on the dial siliconchip.com.au with the tuning slug and trimmer capacitor, respectively. It pays to do it a few times because they interact. Then the points on the dial in the tracking error zones can be checked by disabling the oscillator and tuning the antenna and RF system for a peak, then re-enabling the oscillator. If required, the adjustment vanes on the oscillator gang of the variable capacitor can be altered to improve the tracking. Due to the fact that the capacitance of the variable capacitor gang can only be reduced by bending the wing outwards, to gain full control, all the vanes will have to be bent outward initially. After that, you can bend one in to increase the capacitance or bend it out to decrease it further (see Photos 12 & 13). But this does not always need to be done. Every time the wings are adjusted, both the oscillator’s tuning slug and the trimmer capacitor have to be reset at the upper and lower frequency points to correct the upper and lower set frequencies. Ideally, before starting, you create a tracking error map. This is done by disabling the oscillator, as noted above, tuning the RF signal for a peak at each major dial frequency step, then re-starting the oscillator and inspecting its deviation. This is much easier if you have a counter that subtracts the intermediate frequency for you. As one might expect, the ideal adjustment of the wings would be an S curve, reminiscent of the tracking error curve itself. However, it is compressed due to the nature of the SLW variable capacitor’s vane profile. To get an actual curve, the vanes would require twisting as well as bending outward. Another approach, as shown in Fig.b, is to bend them directly outwards. There are not enough adjustment wings to acquire a super accurate result; 10 wings would be better. With the wings untouched and matching those of the antenna and RF stage gangs, in the zone between 1.8MHz and 2.7MHz, the oscillator is running a little more slowly than ideal and requires a little less capacitance; this is why the wings in that zone are better bent outwards. In the siliconchip.com.au Fig.a: typical tracking errors across the dial of a correctly adjusted superhet. Fig.b: by correctly bending the ‘wings’ on the variable capacitor, it is possible to minimise the tracking errors. Fig.c: the configuration of the tuning capacitor wings in this set. zone between 2.7MHz and 36MHz, the oscillator is running a little faster and requires more capacitance. With this radio, after adjusting the wings on the SW band, tracking was as ideal as possible on the other three bands. The SW band was the most convenient one to use for the tracking adjustment because the outer dial scale and pointer have a larger range of relative motion for a small change in angle of the variable capacitor, and the dial marking details are very helpful. The maximum mechanical error in the position of the pointer with respect to a dial marking in the tracking zones in SW mode, when all is set in proper alignment, is in the order Australia's electronics magazine of about 1-1.5 times the width of the pointer. Fortunately, in the scheme of things, the effects of tracking errors in single-conversion superhet radios are generally small. This is because of the relatively wide bandwidth of the tuned RF and antenna stages, being significantly broader than the oscillator’s tracking errors, so there is no significant loss of sensitivity or image rejection. It is the IF stage in the radio that confers the selectivity to the receiver as a whole. Still, it is good to have the radio in good alignment, as well as the dial pointer giving a good representation of the received station’s frequency. June 2026  97 Silicon Chip PDFs on USB with its own case The USB also comes ¯ A treasure trove of Silicon Chip magazines on a 32GB custom-made USB. ¯ Each USB is filled with a set of issues as PDFs – fully searchable and with a separate index – you just need a PDF viewer. ¯ Ordering the USB also provides you with download access for the relevant PDFs, once your order has been processed ¯ 10% off your order (not including postage cost) if you are currently subscribed to the magazine. wavelengths involved, there is no requirement to model the antenna’s inductance or the transmission line properties of the coax. The antenna system is essentially a capacitive load. The relatively low load of 37.5W (the terminated generator), placed in series with the load capacitance, has negligible effects on the total load, but it allows signal injection in series with the 680pF load capacitance. With this arrangement, shown in Fig.5(b), the antenna coil’s tuning slug positions closely matched the manufacturer’s slug positions (locked with red paint) on all bands. At 170kHz, the reactance of the 680pF capacitor is in the order of 1.4kW. At 4MHz on SW, it is quite low, around 58W. This low load is as recommended on the SW band by Sailor. Summary The Sailor 66T is a remarkably wellmade radio. It lives up to its sensitivity specifications on testing and sports a very attractive, well-calibrated dial. Its direction-finding capabilities are quite remarkable. In the days prior to satellite navigation, it probably saved a number of sailors’ livelihoods, and lives too, in the treacherous North Sea. It has a very satisfactory speaker and audio system, and with only a very small change to a capacitor value, makes a very pleasant sounding radio to listen to music stations on the MW SC band. Receive an extra discount If you already own digital copies of the magazine (in the block you are ordering). THE FIRST SIX BLOCKS COST $100 OR PAY $650 FOR ALL SEVEN (+ POST) NOVEMBER 1987 – DECEMBER 1994 JANUARY 2005 – DECEMBER 2009 JANUARY 1995 – DECEMBER 1999 JANUARY 2010 – DECEMBER 2014 JANUARY 2000 – DECEMBER 2004 JANUARY 2015 – DECEMBER 2019 OUR NEWEST BLOCK COSTS $150 → JANUARY 2020 – DECEMBER 2024 WWW.SILICONCHIP.COM.AU/SHOP/DIGITAL_PDFS 98 Silicon Chip Australia's electronics magazine Fig.5: (a) the recommended antenna for using the 66T radio on a ship; (b) a dummy antenna circuit that provides similar operating conditions for aligning the set. siliconchip.com.au ASK SILICON CHIP Got a technical problem? Can’t understand a piece of jargon or some technical principle? Drop us a line and we’ll answer your question. Send your email to silicon<at>siliconchip.com.au DCC Base Station likely has a short circuit I am building the DCC Base Station from a kit (January 2026; siliconchip. au/Article/19558). When testing the REG1 circuitry using a 9V battery, as per the printed instructions, I am only seeing 4.00V DC at the anode of D1. The instructions say to expect 5.2-5.4V at this point. I have checked every solder point for dry joints, bridges etc. Any advice on troubleshooting further? (J. L., Mudgeeraba, Qld) ● There isn’t anything that immediately comes to mind that would cause it to be near 4V, as long as the 9V battery is, say, 8V or higher. So it seems that there may be a short circuit on the board, causing the 9V battery output to sag. That can be easily verified with a multimeter. You say you’ve already looked for bridges. We’d also check if any components have been reversed, although we can’t think of any specific reason why that would result in 4V at the output. If you have already soldered the Pico, you could try removing D1 to isolate the Pico and see if that narrows down which part of the circuit is causing problems. Building model trains from scratch to start or get an idea of the full process. I really appreciate any reply. (N. G., Qld, via email) ● Model railways are a very broad topic and people take an interest for different reasons. Where to start will depend on what you want to achieve and what you enjoy doing. Some prefer working with electronics, scenery, painting, construction or operation (and probably other different things, too). Being an electronics magazine, the most relevant articles that we have are related to electronics projects that can be used with model railways. The DCC (Digital Command Control) series is a good example of this. You can view free previews of the articles in this series at siliconchip.au/ Series/455 – that might give you an idea of what we have written about. As far as building trains and tracks from scratch, that isn’t something we’ve heard of many people doing, although it must be possible. We have built (in HO and N scale) our own vehicles and some track items, but these have all used commercially available parts like motorised chassis and track profiles, combined with 3D-printed parts and then painted. While researching this answer, we found a few open-source projects that are focused on making functional 3D-printed model railways. We suggest making contact with a local model railway club (they often have their own layouts) or visiting a model railway exhibition. The clubs and exhibitions will have people with different interests. If you have a specific question about how to use electronics, we might be better equipped to help you with that. Analog Devices placing restrictions on parts I have finally gotten around to gathering the parts to build the Versatile Waveform Generator (March 2025; siliconchip.au/Article/17792) but I have run into a major problem! I ordered the Analog Devices op amps that are central to the project from element14 and got a message that they won’t send them to New Zealand. So I put together a big order from Mouser, including all those devices and switches etc. Back came a message that the manufacturer will not allow shipping to New Zealand. What?! How did Silicon Chip obtain the chips to build the project and how can I obtain them? If I purchase them Finding an LCD screen for the Graphing Thermometer I would like to build my own model railway from scratch. While I was able to find a 1999 magazine issue that appeared to have some relevant information, I was having trouble finding where to start or if there is an existing kit. I’m generally a ‘techy’ person, having studied game development and computer science, but electronics will be new to me. I intend to build the trains and tracks as well (not using off-the-shelf products). I do miniature building and painting, so I’m excited to bring that together with some of my other interests. There’s a lot of information out there on the internet, but even with AI assistance, it’s difficult to know where I want to build the Graphing Thermometer from the March 2026 issue (siliconchip. au/Article/19833) and have bought the PCB and programmed ATtiny85 from your web store. Now I am collecting the other parts and I am having problems searching AliExpress for the LCD screen, boost regulator and the 2-way barrier screw block connector. Any assistance would be appreciated. I have been a long-time subscriber and enjoy every issue. (D. M., Geraldton, WA) ● Maintaining current links for parts from these sources continues to pose challenges despite checks shortly before publication, as AliExpress items often change links or are discontinued by sellers. The LCD can be found by searching for “GLX12864” on supplier websites. One current item number is AliExpress 1005006267035677 (try putting that number in the search box). There are many equivalents for the boost converter module used in the design. Search for “0.9 - 5V in / 5V out boost regulator”. One current item on AliExpress is 1005006809303748. Other supply options exist, as noted in the article. The 2-way barrier strip used was a standard 7.62mm/0.3-inch pin-to-pin part for which there are many equivalents. Other metric types can be readily adapted to fit the PCB. One of many sources (for now) is AliExpress 1005003429295570. siliconchip.com.au Australia's electronics magazine June 2026  99 directly from Analog Devices, it will be an additional $52 in postage, making this an expensive exercise. (G. D., Burleigh, NZ) ● We have never had trouble getting Analog Devices parts in Australia. We suspect that Analog Devices has an exclusive deal with some NZ distributor, so other vendors are restricted from selling you those parts. However, we don’t know who that is. You might have to contact Analog Devices to try to find out. It will probably be easier to simply select similar parts from another manufacturer. We haven’t tested these, but we suspect they will work as they have very similar specifications: • Analog Devices AD8065ART = Texas Instruments OPA810IDBVR • Analog Devices AD8091ART = Texas Instruments (National Semiconductor) LMH6642MF/NOPB or LMH6642MFX/NOPB As for Analog Devices, the phrase “shut up and take my money” comes to mind. What do they have against New Zealanders buying op amps? with an extra row or two of width, but have not been able to locate one. Can you point me to a supplier, or am I mistaken? (K. W., Newport, Vic) ● We think the trick is to connect two of those breadboards side-by-side (some are designed to be expanded that way). With the power rails down the middle, the sets of five holes should be further apart, so you will be left with more rows to plug into. Some larger breadboards come with the power rail down the middle (or can be reconfigured that way), for example, the one from Altronics shown in the photo below. Questions about the Calliope Amplifier I have some thoughts on the Calliope Amplifier (April 2026; siliconchip.au/ Article/20084) as I’m looking at building some to use at a higher power. I recognise that they’re already very low distortion, but I can’t help myself. Can I suggest that you look at the KSA992FBTA (PNP) and KSA1845 (NPN) for the low-power stages? These are going obsolete, but are still availUsing RP2350B Dev able from the usual suspects. They Board on a breadboard will allow the amplifier output to go I have a question about the pre-­ up to possibly 150W into 8W, maybe assembled RP2350B Development more if parallel output transistors Boards (August 2025; siliconchip.au/ are used. Article/18635). The magazine article They are low-noise types, so will states: “...designed to suit solderless have low Rbb figures. This means that breadboards with two rows of 32 pins”. the amplifier should be fed from a low If I try to plug it into my bog-­ source impedance, which can be prostandard breadboard, it is too wide to vided by an op amp used as a unity-­ leave any empty pin holes along one gain buffer in the amplifier, or by an edge. I have tried to find a breadboard active crossover. 100 Silicon Chip Australia's electronics magazine Your use of the BF722 has solved a problem that I’ve had for a while, of finding a good VAS transistor. Thank you for that. I’ve tried finding KSC3503 units, but most I find are on eBay. I have some doubts about them. In the original Hummingbird, you used bench power supplies for distortion testing. What are the figures when standard power supplies are used? I’m looking at running them in Class-B only as I can measure the distortion and optimise the bias to minimise that distortion to meet the Oliver criteria (about 26mV across re + Re where re is the transistor’s internal emitter resistance and Re is the emitter resistor). Douglas Self recommends the MJL3281 and MJL1302 output transistors as they are ring emitter units and thus suffer less beta droop than other transistors. Your thoughts? If they’re run in Class-AB, the amplifier will spend a lot of time switching between Class-A and Class-B. Is this a concern? The value for Cdom seems to me to be high and could have a significant effect on slew rate, particularly as the LTP current isn’t very high. Thank you most kindly for the work you’ve put into both the amplifiers and the digital preamplifier. (K. J., Cleveland, Qld) ● The Calliope designer, Phil Prosser, responds: Regarding the KSA992/A1845 transistors, I have come across them before. When I design projects, I go out of my way to use parts that are commonly available. I recall these being lownoise types; I had not looked at their voltage ratings and did not consider them over the ‘decent’ devices that can be bought from the likes of Altronics and Jaycar. If we specified a hard-to-find part, there would be a lot of people asking how to get them and what else they could use. We note that Mouser stocks the KSA992, though. There is nothing stopping you from using these; just note that the pinout is different. The voltage rating would allow higher voltage operation, but more on that later... The lack of reasonable VAS devices on the market is what drove me to go back to the design. It is really frustrating that so few through-hole parts are even half-decent. I am not fussed by SMDs, but a lot of people are. I bought a stack of various VAS devices from eBay years ago and tested a bunch of siliconchip.com.au them to convince myself they were OK. To be honest, the BF469/470 remain a favourite, but these are totally unavailable through normal channels, and if I saw them online, I would not buy them. The tests I have performed on fully built amplifiers with a conventional power supply were totally consistent with the measurements running from the bench supply. The way a typical audio amplifier works means it really doesn’t care that much about the ripple on the power supply. The caveat is that you get the grounding right. That star ground is critical; even minor errors on that will noticeably increase the distortion levels. As the article noted, even running the output wire over the input stages causes havoc with the distortion numbers. There are similar problems if using cruddy connectors. The truth is that you would still need test equipment (not your ear) to detect the problem, but that is kind of what it is about, isn’t it? Regarding Class-AB vs Class-B, I played with the bias level a lot and made many measurements. If the bias is grossly maladjusted, you will get substantially elevated distortion (from memory about 10dB worse). Anything in the ballpark of the recommended bias point, which isn’t that different to what you are shooting for, gives the measured results. I built about a dozen of these amplifiers in various configurations as part of testing. In one set of tests, I couldn’t work out why one amplifier module was giving more distortion than I expected. I was checking parts and wiring until I realised that the bias was turned right down. It is surprising how forgiving the ‘blameless’ architecture is. On the output devices, the “Frankenamp” I mentioned in the article used an MJL21193 on one side and an MJL3281 on the other, making it about as mismatched as you can get. The input was a BC54x and a BC33x. While the DC output offset was borderline dangerous, the distortion was still very low indeed. I was frankly surprised at this, but the point is that there are many things that are ‘technically better’, such as matched output devices, but make a surprisingly small difference to the performance. Don’t get me wrong, I match input devices, but it is less important than many other aspects of the design. I would choose the MJL21193/4 for the outputs if I were going to give the amplifier a hiding. Those devices are incredibly hard to kill. I have tried and failed several times. I have a stack of MJL3281/1302 devices, which I also use. If I were pushing my luck and wanted to wring every watt out of the board at elevated voltages, I would go with the MJL21193/4 as their SOA is better. If I were worried about getting the distortion down, noting that the measurements didn’t show a glaring difference, I would choose the MJL3281/1302. A less obvious thing is that the MJL3281/1302 have an ft of 30MHz. This is not a problem, but it is almost 10 times that of the MJL21193/4. Some people see this difference and jump to the conclusion that a higher ft is better, but it is a double-edged sword. The increased ft does mean that the high-frequency gain is elevated, and this can lead to oscillation on recovery from clipping, unlike the ‘slower’ TEST MANY COMPONENTS ESR TEST T Measures ESR/resistance from 0.01Ω to 1kΩ Measures capacitance from 100nF to 50μF Can perform in-circuit testing as long as capacitors are discharged Compact Tweezers format makes probing parts easy ITH OUR EEZERS Runs from a single 3V lithium coin cell Will operate down to a cell voltage of 2.4V Displays results on a clearly visible OLED screen Typical accuracy better than 10% Adjustable sleep timeout and brightness Display can be rotated to suit left- and right-handed use Simple calibration of most parameters The standby cell life is close to the cell shelf life Complete kit for $50 (SC6952; siliconchip.com.au/Shop/20/6952) The kit includes everything pictured, except the lithium coin cell and optional programming header (CON1). The three resistors and single capacitor needed for calibration are included. See the article in the June 2024 issue for more details (siliconchip.au/Article/16289). For testing other components like capacitors and diodes, check out our Advanced SMD Test Tweezers from the February & March 2023 issues (siliconchip.com.au/Series/396). We sell a kit for those Tweezers for $45 (SC6631). siliconchip.com.au Australia's electronics magazine June 2026  101 Getting tachometer working with Ignition System I recently built your Programmable Ignition System (March-May 2007; siliconchip. au/Series/56). It works great except I can’t get my rev counter to work. The vehicle is a 1987 Ford Courier ute with a points distributor. I have connected the programmable ignition tacho output but the rev counter doesn’t work. If I connect an external aftermarket rev counter to the tacho output, it works, so I know I’m getting a signal. I have also tested the tacho in the ute, and it’s working fine. I tried making an amplifier circuit that I found online, but that doesn’t work either. Any help or advice would be appreciated. (G. N., Lincoln, NZ) ● Presumably, the original tachometer is an impulse type that relies on the high primary coil voltage to operate rather than a low-voltage signal such as the tacho output from the programmable ignition. Check the source of the original tachometer signal. If it was at the points, then the tacho may work when connected to the transistor coil driver that connects to the coil. 21193/4. The oscillations are usually in the MHz region and not reflective of the devices’ behaviour at audio frequencies. When you roll back the frequency to audio performance in an amplifier, with that Miller capacitor providing the dominant pole, the ft difference is not significant. So I agree with the preference on the basis of distortion, but would advise you to be careful with safe operating areas if you want to crank the voltage. You really ought to use extra output devices and resistors, but then you are basically building something like the SC200 amplifier. Keep that protection circuit in there, as if your speaker dips in impedance, as many do, you may find your nominal 8W load presents a problem at some frequencies. In that case, the load-line protection circuit would possibly save you repair to the amplifier and/or your speaker. Unless you have a compelling reason that demands such a tiny amplifier board, if you want more than 100W then go to one of the Ultra-LD amplifiers, as it saves you rat’s-nesting the extra devices in, and also gives you more separation of the input stages from the high currents in the output stage, which has distortion benefits. I would prefer to run the amplifier in Class-AB, as the notion of ‘being exactly on the edge of Class-B’ sounds both optimistic and not likely to be true over temperature variation, voltage rail variation and the life of the amplifier. The expense of a bit more dissipation is fine by me, and the only argument I would easily accept is that running a BJT amplifier in Class-A is ‘obviously’ going to result in lower crossover distortion. I have tried this, and I kind of liked the massively hot heatsink, but was unable to actually hear any difference except for the very slight buzz of the power supply cranking out hundreds of watts into a big heatsink 24/7. At night this annoyed me, so I turned that bias down again. Cdom is set higher than absolutely necessary in the Calliope to allow a ridiculous range of parts to be used without the amplifier giving stability problems. The article makes brief mention of me adding a two-pole Miller arrangement. This reliably and substantially reduced high-frequency distortion, but was really hard to squeeze in, and when I started throwing random parts in for output devices etc, I decided that 220pF was a safer choice for your average constructor. There is nothing stopping you from building and tuning your own Cdom or even adding that second pole; it isn’t hard to fit on the back of the board. Just remember that you need to test for stability not only at low power levels; you really need to look at stability running into a low-impedance load near clipping. Watch the waveform as it goes into and comes out of clipping near the negative rail. Bursts of high-frequency oscillation at this point are a characteristic of a marginally stable amplifier using this topology. It is seen in pretty well every conventional BJTbased amplifier when you start pushing the slew rate. So in the end, I went for a value of Cdom that was reliable, safe and gave good results. Reliability in my mind trumps slightly better performance when it comes to the range of readers and constructors the magazine has. One great aspect of this project is that people can do what they want. If you want to use the KSA parts, go for it. I am confident they will work well. The MJL1302 transistors are known good too. Feel free to adjust Cdom if you have the means to verify the amplifier is still stable, and if you want to try a two-pole version, that definitely has benefits. A couple of extra components on the back of the board will not hurt. Just be careful turning the voltage WARNING! Silicon Chip magazine regularly describes projects which employ a mains power supply or produce high voltage. All such projects should be considered dangerous or even lethal if not used safely. Readers are warned that high voltage wiring should be carried out according to the instructions in the articles. When working on these projects use extreme care to ensure that you do not accidentally come into contact with mains AC voltages or high voltage DC. If you are not confident about working with projects employing mains voltages or other high voltages, you are advised not to attempt work on them. Silicon Chip Publications Pty Ltd disclaims any liability for damages should anyone be killed or injured while working on a project or circuit described in any issue of Silicon Chip magazine. Devices or circuits described in Silicon Chip may be covered by patents. Silicon Chip disclaims any liability for the infringement of such patents by the manufacturing or selling of any such equipment. Silicon Chip also disclaims any liability for projects which are used in such a way as to infringe relevant government regulations and by-laws. Advertisers are warned that they are responsible for the content of all advertisements and that they must conform to the Competition & Consumer Act 2010 or as subsequently amended and to any governmental regulations which are applicable. 102 Silicon Chip Australia's electronics magazine siliconchip.com.au MARKET CENTRE Advertise your product or services here in Silicon Chip KIT ASSEMBLY & REPAIR FOR SALE PCB PRODUCTION DAVE THOMPSON (the Serviceman from Silicon Chip) is available to help you with kit assembly, project troubleshooting, general electronics and custom design work. No job too small. Based in Christchurch, New Zealand, but service available Australia/NZ wide. Email dave<at>davethompson.co.nz LEDsales PCB MANUFACTURE: single to multilayer. Bare board tested. One-offs to any quantity. 48 hour service. Artwork design. Excellent prices. Check out our specials: www.ldelectronics.com.au KEITH RIPPON KIT ASSEMBLY & REPAIR: * Australia & New Zealand; * Small production runs. Phone Keith: 0409 662 794 keith.rippon<at>gmail.com LEDS, BRAND NAME AND GENERIC LEDs, filament LEDs, LED drivers, heatsinks, power supplies, kits and modules, components, breadboards, hardware, magnets. Please visit www. ledsales.com.au Silicon Chip Binders REAL VALUE AT $21.50 PLU S P&P PMD WAY offers (almost) everything for the electronics enthusiast – with full warranty, technical support and free delivery worldwide. Visit pmdway.com to get started. Order online from www.siliconchip.com.au/ Shop/4 or call (02) 9939 3295. ADVERTISING IN MARKET CENTRE Classified Ad Rates: $32.00 for up to 20 words (punctuation not charged) plus $1.20 for each additional word. Display ads in Market Centre (minimum 2cm deep, maximum 10cm deep): $82.50 per column centimetre per insertion. All prices include GST. Closing date: 5 weeks prior to month of sale. To book, email the text to silicon<at>siliconchip.com.au and include your name, address & credit card details, or phone (02) 9939 3295. up, and if you do, then watch the SOA. There is not a lot of ‘room’ there for speakers that do surprising things, so if you can’t resist the temptation, then add an extra set of output devices. That is a little untidy, but it is only six more wires per amplifier. That buys you a lot of SOA headroom. ● We published a white noise generator in September 2018 (siliconchip. au/Article/11225). You could feed its output into an amplifier and then to speaker(s). You would probably want very directional speakers, perhaps using a parabolic reflector, to confound the camera without making a racket. White noise generator wanted Using a differential ADC for Reflow Oven My daughter unfortunately lives next door to the area snoop and gossip. She recently discovered that the neighbour’s security camera, which is allegedly aimed at the footpath, actually points over their deck and picks up what is said on the deck. It is probably illegal, but hard to prove, so I thought a white noise generator will probably solve the problem legally. Have you produced a suitable circuit? (J. A., Healesville, Vic) I am putting together all the parts to build the DIY Reflow Oven Controller project (April & May 2020 issues; siliconchip.au/Series/343). I have a question regarding the thermocouple amplifier section. I purchased the same module you used from eBay. It looks like it is a copy of an Adafruit design. Why does CON10 not indicate an input for the reference voltage? I would have thought it better to use a differential ADC measurement (direct or two siliconchip.com.au Australia's electronics magazine individual channel readings) so that the reference/offset voltage did not necessarily have to be known, as long as it provided a large enough positive temperature span. As you mentioned, 2.5V would be too large for a 3.3V system. What CON10 pin would you use/suggest? I will update the software to handle this. I am going to make my own special board to take the CON10 ribbon cable direct and have a separate twopin connector for the SSR connections. (M. V., Taree, NSW) ● The designer, Phil Prosser, responds: if I understand your intent, you are asking about using a second ADC channel to make the sensor temperature measurement independent of the temperature sensor’s DC offset. If your sensor output voltage range does not saturate the ADC across the full measurement range, you could certainly do that. You will need access to June 2026  103 Advertising Index Altronics.................................39-42 Blackmagic Design....................... 7 Dave Thompson........................ 103 DigiKey Electronics..................OBC Electronex..................................... 5 Emona Instruments.................. IBC Hare & Forbes........................10-11 Jaycar............................. IFC, 22-25 Keith Rippon Kit Assembly....... 103 LD Electronics........................... 103 LEDsales................................... 103 Microchip Technology.................. 9 Mouser Electronics....................... 3 PE Back Issues........................... 52 PCBWay....................................... 35 PMD Way................................... 103 SC Bridge Rectifiers.................... 79 SC ESR Test Tweezers............. 101 Silicon Chip Binders.......... 59, 103 Silicon Chip Kits........................ 76 Silicon Chip PDFs on USB......... 98 Silicon Chip Subscriptions........ 53 Silicon Chip Shop.................86-87 The Loudspeaker Kit.com............ 8 Wagner Electronics..................... 83 Errata and on-sale date Airzone 6552A, Vintage Radio, May 2026: in the circuit diagram, C4 is pointed to the wrong capacitor, it should point to the capacitor directly below C3. Some components also had incorrect values, the correct values are – R10 is 250kΩ; R12 is 500kΩ; R14 is 250kΩ; and capacitor C1 is 250nF. Digital Vehicle Compass, Circuit Notebook, April 2026: on p17, the SDA and SCL connections going down vertically from the Arduino’s SDA and SCL pins are swapped. All the SDA pins should be joined together, and similar for the SCL pins. PicoSDR, April 2026: in Fig.3 on p39, pin 39 of the Pico (VSYS) should be connected to +5V. Next Issue: the July 2026 issue is due on sale in newsagents by Monday, June 29th. Expect postal delivery of subscription copies in Australia between June 26th and July 13th. 104 Silicon Chip the reference voltage on the thermocouple interface, and would need to pay attention to limit cases such as low-temperature operation. The current calibration process zeroes out the absolute offset. The actual precision of the thermocouple really limits what you should be expecting to achieve; you probably want to balance the effort you put in with the benefit you would get. The PIC microcontroller is hugely programmable in its I/O capability. Look into the data sheet for your options, which you need to configure with the other inputs and outputs. Depending on how you implement this, you will need to check the temperature scaling, but I suspect will be OK with the calibration. To be honest, I wonder if the rework of the code will deliver a great benefit. Knowing how agricultural my coding style is, you may end up tearing a few handfuls of hair out, though. Converting Nano Pong to HDMI output I built the Nano TV Pong kit a while ago (August 2021 issue; siliconchip. au/Article/14988) and it works well plugged into a TV with component inputs. But I want it to output HDMI. I’ve tried to use a component-­toHDMI converter (Jaycar Cat AC1722) and the screen comes up, but it goes away after a few seconds with “No Signal” being displayed for a few seconds, then back to the game display. I’ve tried different TVs, and the No Signal is the same display on each of them, so I assume it’s coming from the converter box. The sound works fine throughout. Any ideas? I am a long-term reader back to Radio, TV & Hobbies. (M. H., Parkinson, Qld) ● The signal from the Nano Pong is composite video. Since the AC1722 Converter states that it supports composite video, it should work. It seems like the converter thinks that the signal is going away and coming back for some reason. Have you tried changing the resolution setting on the Converter? We are not sure that will help, but it would be worth trying. We found some forum posts that suggested the 720p setting would work better (despite 1080p sounding like it should be ideal). The Converter may be a bit fussy about the Australia's electronics magazine signal it expects; we didn’t have any trouble with any of the TVs or capture cards that we tried. The Converter mentions features such as “black/white level expansion, color transition improvement, dynamic range expansion”, so we wonder if it is not handling the twolevel monochrome image well. It may be better to produce a native HDMI signal rather than try to convert it. That could be done using one of our small computers that have an HDMI output, such as the RP2350B computer project from November 2025. It would just need software to play Pong. We found a BASIC version that seems promising, but we have not tested it: https://github.com/jmdeejay/ mmbasic-picomite/blob/main/games/ pong.bas Where to obtain VOC sensors? I have a query regarding the volatile organic chemical (VOC) sensor mentioned on page 45 of the February 2020 issue for the Indoor Air Quality Monitor (siliconchip.au/Article/12337). This MOX sensor worked very well. In the University lab where I worked previously, it notified our staff that we had a methanol leak (odourless but harmful), and also that a fume cupboard was malfunctioning (it used a three-phase motor with two of its poles inadvertently reversed, blowing down instead of up). In addition to these accomplishments, I used the ‘nose’ to estimate the range of VOCs in pharmaceutical products, which we tested on a gas chromatograph. There is a window in which to measure such a signal, thus the appropriate dilution was employed, and it was spot-on every time. However, both the CCS811 with onboard HDC1080 I have no longer give readings on both Micromite BackPacks. My guess is that the heaters have failed. Core Electronics has discontinued this line. Do you know of any other more reliable VOC sensors that are as good as that one? (G. A. D., PhD, Biochem). ● The sensor module is still available from sellers on AliExpress, see: www.aliexpress.com/w/wholesaleccs811-hdc1080.html w w w. a l i e x p r e s s . c o m / i t e m / SC 1005006603898777.html siliconchip.com.au “Rigol Offer Australia’s Best Value Test Instruments” New 2026 Products Oscilloscopes New 12Bit Scopes New Up To 13GHz RIGOL DS-1000Z/E - FREE OPTIONS RIGOL DHO/MHO Series RIGOL DSO-8000/A Series 450MHz to 200MHz, 2/4 Ch 41GS/s Real Time Sampling 424Mpts Standard Memory Depth 470MHz to 800MHz, 2/4 Ch 412Bit Vertical Resolution 4Ultra Low Noise Floor 4600MHz to 13GHz, 4Ch 410GS/s to 40GS/s Real Time Sampling 4Up to 4Gpts Memory Depth FROM $ 649 FROM $ ex GST 684 FROM $ ex GST 13,191 Multimeters Function/Arbitrary Function Generators RIGOL DG-800/900 Pro Series RIGOL DG-1000Z Series RIGOL DM-858/E 425MHz to 200MHz, 1/2 Ch 416Bit, Up to 1.25GS/s 47” Colour Touch Screen 425MHz, 30MHz & 60MHz 42 Output Channels 4160 In-Built Waveforms 45 1/2 Digits 47” Colour Touch Screen 4USB & LAN FROM $ 713 FROM $ ex GST Power Supplies ex GST 604 FROM $ ex GST Spectrum Analysers 632 ex GST Real-Time Analysers New Up To 26.5GHz RIGOL DP-932E RIGOL DSA Series RIGOL RSA Series 4Triple Output 2 x 32V/3A & 6V/3A 43 Electrically Isolated Channels 4Internal Series/Parallel Operation 4500MHz to 7.5GHz 4RBW settable down to 10 Hz 4Optional Tracking Generator 41.5GHz to 26.5Hz 4Modes: Real Time, Swept, VSA, EMI, VNA 4Optional Tracking Generator ONLY $ 799 FROM $ ex GST 1,321 FROM $ ex GST 3,210 ex GST Buy on-line at www.emona.com.au/rigol Sydney Tel 02 9519 3933 Fax 02 9550 1378 Melbourne Tel 03 9889 0427 Fax 03 9889 0715 email testinst<at>emona.com.au Brisbane Tel 07 3392 7170 Fax 07 3848 9046 Adelaide Tel 08 8363 5733 Fax 08 83635799 Perth Tel 08 9361 4200 Fax 08 9361 4300 web www.emona.com.au EMONA