Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Charge of the light yardwork Dave Thompson Spring has sprung, the blossoms are out, days are getting longer and it isn’t as bone-crushingly cold as it was. It’s a time when the lawn starts growing like it’s on steroids and the garden is begging for a recharge of soil and nutrients. But the soil wasn’t the only thing that needed a recharge. Spring is also the time to break out the mechanised garden tools and lubricate chains, sharpen trimmers, recharge batteries, check oil levels and change plugs. There’s more than enough to do! Over the lockdown period, I built a garden shed to house all that stuff. I was sick of tripping over the mowers, chainsaws, spades and rakes stored in my main workshop, which I couldn’t use without removing all that stuff first. The shed was one of those build-it-yourself flat-pack kits you can buy at many of the big-box hardware stores. It came with about a million bolts and widgets, plus a booklet on how to assemble it, which is only helpful if you need to start a fire. None of the numbers on the bags of parts seemed to tally with the legends on the expanded diagrams of how to put each section together. It would say something like, “Take 15x A3 round-head bolts and 15x D5 square lug nuts, and using 15 x washer M6, assemble the door.” I should be so lucky! Nothing made sense at all. There are no prizes for guessing where this shed (and its user manual) was produced. I would usually have a very capable builder friend come and take a look at the whole thing and get him to suggest any improvements that could be applied during construction, but due to the lockdown, we couldn’t. Instead, I consulted with him using WhatsApp video calling. I ended up soldiering on with it myself, adding extra timber bracing, stronger door jambs and a few other ideas he pointed out. I also ditched the bolts and decided to use rivets instead. I had more than enough of all sizes and shapes that I had gathered over the years (I inherited about a gazillion from Dad’s estate). I also have a very handy pneumatic rivet gun, which saved the hand-crushing pain of doing it with a pop riveter. It also meant I could install the rivets from one side of the wall without needing to have someone holding a nut and spanner on the other side of the panel. That was going to make life much easier. The plan was also to use proper, heavy-duty Tek screws to hold the frames of the walls to the timber floor, which I’d already put in place before the lockdown. The kit came with inadequate (in our opinion) screws. I suppose the basic shed would be sound enough without adding the extras; after all, they sell them and people build them, but I felt better knowing it would remain standing in some of the gale-force winds we experience here at this time of year. I’ve already had one partial car-port-­under-construction wind up in the neighbour’s backyard. If it happens once, it could be considered an accident, but twice would make it seem deliberate! Note to prospective shed builders: having someone else to hold and steady the assembled walls whilst bolting and riveting them together is a real help. I managed it by balancing them on ladders and temporarily erecting scaffolds made from scrap timber. I also made sure to do it on days when the wind wasn’t blowing! So, now I have a shed to store all my tools, and they are handy to the garden as well. The weed wacker was knackered The other day, I went to get my weed whacker, a rather beefy, well-known brand yellow electric model and discovered that the 54V 9Ah battery was dead flat. I wasn’t too surprised; after all, it hadn’t been used for almost five months, although I had fully charged the battery before storing it. Pressing the button on the side, which usually shows the battery status via a three-stage LED display, resulted in three dark LEDs. Zero, zip, zilch, nada; nothing. So I took the battery to my workshop, where my array of chargers reside (there is no power in the shed) and plugged it into the matching yellow fast charger. The red charge light flashed briefly and went out. Usually, it would flash once a second, the internal cooling fan 82 Silicon Chip Australia's electronics magazine siliconchip.com.au Items Covered This Month • • • • • Not all instructions are created equal The malfunctioning security camera Solving TV program transmission problems Once upon a time in the Navy Repairing a weight scale 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 would fire up, and it would eventually turn solid red once the battery reached full charge. Hmm. I made sure the mains plug for the charger was seated correctly and that the four-way socket it was plugged into was switched on and had power to it, but I still got nothing. I removed the battery from the charger and replaced it again. No light at all. Uh-oh, this turned dark quickly! Anyone who knows anything about lithium-ion batteries will be aware of the ramifications of allowing them to fully discharge and fall below their minimum state of charge (SoC). That often results in one or more dead cells and, by extension, a dead battery pack. The minimum SoC is typically about 20% of a Li-ion battery’s capacity. If internal leakage drains it even further, such as when it’s being stored and not being charged, the cell will likely fail. The general rule is to only use the battery between 50% and 100% of its capacity, with cycling down to under 20% not recommended (as was the case with keeping NiCad batteries alive). For more information on this, see our series on batteries in the January-March 2022 issues (siliconchip.au/Series/375), especially p15 of February 2022 and p50 of March 2022. Still, let’s not get bogged down in the details. These 54V tools and batteries are not cheap, so just biffing it in the bin was out of the question, especially when I have the Serviceman’s Curse and ‘might’ be able to do something about it. And on closer inspection, the pack was only screwed together, unlike some which are glued, never to be opened (without some serious cutting, anyway). Bonus! Assaulting the battery So the obvious solution was to tear into this pack and see what we were dealing with. But I’ve been bitten before with battery packs. There could very well be a low cell (or more) inside, but I also knew there would be some electronics, like a BMS (battery management system) and perhaps also a thermistor, thermal cutout or a line fuse. The capacity of these batteries is approaching levels that shouldn’t be trifled with. There are now 60V and even 120V versions of this pack. They’re for a different family of tools, but are potentially dangerous to work with. My battery pack fits a wide range of 18V and 54V tools. It has a mechanical toggle switch that changes the output depending on what tool it is plugged into, and that switch is activated when you slide the pack into place. It’s a clever system. If I plugged this battery pack into an 18V drill, for siliconchip.com.au example, the cells are wired in series/parallel via that switching arrangement. If plugged into a 54V tool, they are all wired in series. It’s basic but clever. Of course, an 18V pack will do nothing in the 54V tools, but the 54V batteries are backwards-­compatible. So, it was time to open it up and see what was happening. In their wisdom, the company had screwed it shut with what looked like T9-sized Torx security screws. These are the ones with that annoying ‘nipple’ in the centre, so only a hollow-point tip would work. These screws really annoy me. Even many entry-level DIYers likely have a tip like this in their tool collection, so who is this security fastener folly meant to stop? Children playing with Torx drivers? I learned a little later via YouTube of a technique that involves using a pin punch or even a hardened concrete nail to ‘ping’ that little nipple out of the head of the screw (if you can clear gain access to the screw head, that is). Still, since I had the right tip, I decided just to use that. The heavy plastic side covers come off relatively easily once those screws are removed, revealing that the inside of the battery is built like a concrete bunker. The pack is quite heavy (1.4kg) and the construction is robust. There is virtually no free room inside, with the cells tightly packed. They all appear to be mounted into some kind of honeycomb-­style framework. The fifteen 18650 cells are connected in three groups of five, with the charging and output being controlled by a printed circuit board (PCB) that also varies the voltage output depending on the mechanical switch position. The cells are connected with spot-welded metal links, and I was very careful not to drop screws or bridge any of the links with a screwdriver. Significant flexible ribbon cables connect various points on the cell links on either side of the pack back up to the circuit board and charging socket. I could see I would have to do some research because it didn’t look easily repairable. If, for example, I had a dead cell (or several dead cells), I’d have to desolder the flexible connectors and break those splat-welded links. That’s likely easy enough to do, but putting it all back together again would be tricky given that I don’t have a spot welder that would handle links that size, and soldering directly to these batteries is usually fraught with problems. It might be possible to source new cells with solder tags already welded onto them, but I’m getting ahead of myself. A fatal flaw revealed I did what anyone would do and hit the interwebs. There’s a wealth of information on the ‘net, which presents a problem: sorting the gold from the dirt. Using very specific search words is the answer, and I found a lot of information on this particular battery pack, which proved incredibly helpful and saved me a lot of time and effort. As I’ve learned so many times before, just rolling my sleeves up and piling into something can often lead to disaster, or at worst, failure, and a waste of time and money. I found evidence of a known problem with these battery packs: the middle cell in each group of five is likely to fail because of how the packs are wired and how the charger works and supplies charge to them. The result is that those cells are charged less consistently than the others in the chain. Australia's electronics magazine November 2023  83 That could have caused my problem. If it had, I thought that buying another battery pack would be the best option. However, given that mine was less than 18 months old, I thought it was a short time for a battery to fail from such a fault, as it would have likely seen only about 10 hours of actual use. While browsing the web, I also began seeing a lot of YouTube videos offering a ‘hack’ on how to resurrect these batteries, but only if the symptoms were the same as mine. That is, no LED indicators light on the battery when the test button is pushed, and when put in the charger, it won’t charge, with the red charging light flashing briefly before going dark. These videos are legion and of course of varying quality, both visually and with the information they communicate. The majority of people making those videos don’t know why this fault happens or why the ‘fix’ works. They only state that the hack ‘jump starts’ the battery into charging again. This is an old trick with NiCads and other types of batteries, and while it has varying degrees of success with those types, it is not a recommended practice for lithium-ion batteries. Typically, even if the Li-ion battery does get some ‘kick’ out of it, the capacity and charging capability are usually way down on what they should be. So, at best, if anything, this hack is a stop-gap measure that might or might not give me a little more time to use the battery in my tools before I’d be buying a new one anyway. Most of these demonstrations on the internet have a couple of flaws. The first is that many of these guys use an identical battery to ‘jump’ the dead one. That leads to the question: if you have another battery, why don’t you just use that one in your tools? If you don’t have another battery, you have to use either a car or a bike battery to do the ‘jump’. Still, not many of us have a spare one of these sitting around, and besides, I wouldn’t be too keen on wiring this pack up to my Suzuki Vitara battery without at least removing it from the car. What a faff all that would be. Fortunately, I have several good bench power supplies 84 Silicon Chip I’ve built from excellent designs in this magazine (and others), so I decided to use one of those to jump this pack. One of the bench supplies is a bit more ‘disposable’ than the others, and indeed, I have rebuilt it several times over the last 20 years! The ‘method’ used by most of these ‘job site’ type guys making the videos is to simply connect the positive and negative terminals of the donor battery to the same terminals on the dead one. Usually, there is a spark and a splat when connected, and they only leave it for a few seconds before pulling the wires free. They then place the dead battery into the charger, and voila! The charge light comes on, the angels sing, and they have resurrected the dead pack. What they don’t say or cover in any of these videos is how long the charge takes or how long it remains useful compared to a new battery. That’s what I’d be more interested in, which is why I kept looking past the cheap ‘hack’ for more information. I finally found some in a video put together by one of the more switched-on YouTubers, Matthias Wandel, who actually bothered to dig into the reasons behind the failure of this type of pack. He broke one pack down and explained how it worked, likely why the same three cells fail and much more good information. Might as well jump Regardless, I still had a dead pack, so I made up some leads with some heavy-gauge wire and cranked the bench supply flat out to 20V and as many amps as it could deliver (theoretically, five). I connected it, got the sparks and held it for five seconds. I removed all the leads and put the battery in the charger. Well, cue the angels because the charger kicked in. I left it to complete its cycle and, the next morning, plugged it into my chainsaw and weed whacker. It certainly powered them OK! As to how long it lasts, it would be hard for me to say as I have no control battery to try, but I’ll use this one until it dies, then buy another one. For now, it works, so ‘hack’ confirmed and now to the garden! Editor’s note: it is common for the BMS to disconnect the cells from the outside world if the battery voltage falls too low. This will often make the charger fail to sense the battery (and hence refuse to charge it). A proper BMS will still allow some current to flow into the battery so that you can recover it externally, although some possibly don’t. The recommended practice for a Li-ion/LiPo battery that has fallen to a low voltage is to charge it very slowly, over a few hours or a day at perhaps 100mA, until its voltage returns to something more normal. It should ideally be charged on a non-flammable surface like concrete. You can then attempt to charge it normally, but keep an eye on it and switch off charging if it starts to swell or get hot. That approach has a good chance of restoring most of the battery’s capacity, even if it fell to quite a low voltage, but it is certainly not guaranteed. A malfunctioning IP security camera G. C., of The Gap, Qld probably spent too long on fixing a malfunctioning security camera. Still, the perseverance paid off and the camera eventually returned to service... We have several security cameras at our home, enabling us to look at the images anywhere in the world if we have Australia's electronics magazine siliconchip.com.au an internet connection. A few years ago, one of the neighbour’s vehicles was stolen from his driveway in broad daylight. We were able to provide the police with useful information from the recorded video of the incident. In 2021, on an extended caravan trip to Far North Queensland, my wife monitored the cameras regularly to check on our house. Our newspaper was still being delivered even though it had been cancelled, and she was able to request the provider to take corrective action. An unexpected parcel was delivered to our front doorstep, so we asked a helpful neighbour to collect it. One night on this trip, my wife noticed that one of the cameras facing the driveway was showing a very dark image. Over the ensuing weeks, we noticed it occurred about two nights every three. So there was an intermittent fault with that camera. These Swann cameras have a ring of red LEDs around the lens to illuminate the surroundings at night. I suspected a fault with them. I presumed that the wavelengths of these LEDs extended into the infrared region and wondered what type they were. In my spare time, I emailed a couple of my learned friends to see if they had any security camera experience. One replied that the problem might be caused by a faulty infrared (IR) cut filter. I had not heard of these filters before, so I had to do some internet sleuthing. I discovered that camera sensors detect near-infrared light that is invisible to the human eye. In daylight, a security camera uses an infrared cut filter to filter out unwanted IR light to represent colours accurately. When the camera is operating in night mode, the IR-cut filter is switched out to allow the camera’s light sensitivity to reach very low lux levels. I thought these infrared filters would use some material that became opaque to IR with an electric field applied, but research revealed that these filters were much more primitive. They were usually moved in and out between the lens and the photosensor device mechanically, using a solenoid. When we finally arrived home, I established that the red LEDs on the camera in question illuminated at night, so the IR cut filter appeared faulty. Interestingly, the IR cut filter was always switched in correctly during daylight without exception. It only failed to switch out some nights. Dismantling the camera, I found two wires going to a layer between the lens and the photo sensor. After removing three tiny screws, I could remove the lens and reveal the IR cut filter. I could see that the solenoid was driven directly by an SMD. The IR cut filter consisted of a moving ferrite magnet with an attached arm that toggled the IR filter in front of or away from the sensor. The magnet moved one way or the other depending on the polarity of the pulse applied to the solenoid coil. I found many shards of ferrite material sticking to the magnet, and the cause soon became apparent. There was a fissure in the ferrite, causing it to shed material, I suspect due to a manufacturing defect. I gently removed the shards with a small stiff artist’s brush and a jeweller’s screwdriver. I wondered if the floating bits of ferrite were intermittently stopping the magnet from pivoting the full 45°. Still, I had a niggling doubt that there may be something else wrong, given that it always switched the filter in during daylight. siliconchip.com.au After reinstalling the camera, it was disappointing to find that the IR cut filter did not ever switch out at night now. I don’t know what I had done, but the camera did not like it! It was probably beyond repair; although various IR cut filters were available online, none were like the one used in this camera. A few days later, I had a thought. I removed a fully operational camera and dismantled it. As expected, the ferrite magnet was in pristine condition. With a magnetic compass, I carefully identified the north and south poles of the magnet. Then, with an unmagnetised small screwdriver, I got some idea of the magnet’s strength – very subjective, but better than nothing. I again removed the IR cut filter from the faulty camera. This time, I endeavoured to clean out any debris between the magnet and the coil former with part of a razor blade. With a strong ferrite magnet recovered from a loudspeaker, I attempted to strengthen the magnetisation of the filter’s magnet, taking care not to reverse its polarisation. I suspected that the loss of magnetic material had reduced its field strength, resulting in marginal performance. After reassembling and re-mounting both cameras, it was pleasing to see both cameras working properly. Unfortunately, after six weeks, the camera failed again. This time, after much internet searching, I found a Chinese supplier of IR cut filters of similar dimensions. The new IR cut filters were not physically identical to the original, but after shaving off bits of plastic, I made one fit. The new filter was thicker than the original, so it was necessary to re-focus the lens. The camera manufacturer had been over-zealous with the glue used to stick the lens to its mount. I had to carefully scrape it off before the lens would budge. When the camera was reinstalled, I found, much to my chagrin, that the IR cut filter did not work. On one of the seller’s descriptions, I found that the filter should operate at voltages from 3.5V to 5.0V, and I verified that it worked after Australia's electronics magazine November 2023  85 completely disassembling the filter. However, it refused to operate when I reinstalled the cover over the mechanism and the lens mount. Pulling it apart again, I found that a plastic ridge around the optical opening was impinging on the operation of the moving arm. I gingerly pared away some material with a small wood carver’s tool and, when I reassembled the filter, it appeared to work satisfactorily on my workbench. However, when reinstalled, objects on the right side of the image were tinged in pink with the lights on, while at night, the image was black. I purchased an RJ45 coupler so that I could connect a long Cat5 cable to the cable going back to the security cameras’ recorder, which was hidden in an inaccessible place. With the camera plugged into the extension cable, I could conveniently work on it at a table. I removed the outer shroud, the lens mount and the plastic cover over the filter’s mechanism. By simulating day and night-time conditions, I could directly watch the operation of the filter. If only I had done this earlier, I would have saved myself much angst! Immediately, the problem was obvious. The filter plane was toggling perfectly, but it was entirely out of phase, ie, it was switching the filter in when it was dark and vice versa. It was a simple matter of cutting the wires to the filter and transposing them. Finally, after such a protracted period, the camera was working perfectly. Solving problems in TV program transmission G. G., of Macleod, Vic relates a servicing story from nearly 50 years ago. He likely remembers it because the cause was so unusual... From 1963, television stations in Melbourne and Sydney often shared program material via the three pairs of ‘tubes’ in the interstate coaxial cable. They would book time on the limited resource and, when it was their turn, the coax was connected with patch leads onto the ‘tail’ from the city out to their studios. The analog baseband signal in the tail required repeaters every few miles; in Melbourne, the longest tail was to ATV0 at Nunawading. When a program came from elsewhere, the receiving 86 Silicon Chip studio would synchronise with the source studio via the incoming program feed. This incoming signal became the master for the whole studio and its activities. In the early 1970s, ATV0 was having problems with their synchronisation late in the afternoon. They were taking children’s programming from Sydney and (apart from inserting local ads) were passing it straight to their Mount Dandenong transmitter (via a private microwave radio link). The synchronisation ‘hits’ were causing interference to the viewers’ pictures. That didn’t matter too much for the junior viewing audience. Still, other programs were being made in the Nunawading studios at the same time, and these synchronisation ‘hits’ were upsetting the studio recorders and ruining those recordings. The Sydney to Melbourne coax was checked out and found to be OK. The city to Nunawading link was found to be introducing spikes into the program material, but only between 5:30pm and 6:00pm. There were several repeaters on the tail, and the program was monitored on these sections progressively out from the city and was clean until the section into Toorak. Fortunately, there was a spare coax tube in that link, which we could monitor during the troublesome period. With no equipment connected except for our monitoring storage CRO, this raw tube suddenly showed significant but random spikes. The cable followed a tram line, which was a suspect as the return path for the 600V DC traction currents was their rail. If there were any broken rails or joints, it was typical for the telephone cables’ heavy lead sheath to become a convenient Earth return path for that fault current. But why only between 5:30pm and 6:00pm? An investigation began into what else was carried in the cable that could be a source of the interference. A bunch of coaxial tubes always had spaces between them, and those spaces were filled with copper wires called interstitial pairs. These pairs were considered premium as they were larger (less lossy) than regular pairs and shielded by the heavy lead sheath. They were usually first assigned as audio program lines for TV and radio stations connecting studios and transmitters or outside broadcasting locations. However, somehow many of these pairs in a section of the cable had been assigned to the basic alarm circuits for the shops in the Toorak shopping strip. These simply provided a DC loop back to the alarm company. Any break in the current would be treated as an alarm. When the shopkeepers shut up for the day soon after the standard 5:30pm closing time, they activated their alarms, which turned on their monitoring current. This instant step in current induced a voltage spike into the adjacent coaxial tubes. These alarm circuits were quickly transferred out of the coax, and the problem disappeared. Seven is greater than five G. D., of Glen Iris, Vic was a Navy repairman for many years. This incident must have stuck in his memory for him to remember it so clearly decades later... A young sailor with HMAS Torrens emblazoned around his cap walked into the Radio Workshop at Williamstown Naval Dockyard with a box under his arm. He plonked it down on the workbench and said it was a... (I have forgotten Australia's electronics magazine siliconchip.com.au what it was, but I couldn’t tell you even if I remembered. It came from the crypto room, so it was very hush-hush). He proceeded to tell us that the box was US – not American, unserviceable – which was the official Royal Australian Navy’s term for a bit of kit that doesn’t work. The sailor said his captain was desperate for us to fix the box because they were going to sea in a few days. As he disappeared down the stairs, he yelled that he would be back tomorrow morning to pick up the repaired box. It appeared that the sailor didn’t understand that asking Willie Dockies to fix something in less than 24 hours was wishful thinking. But my boss Bruce took the request seriously and handed the box over to his top technician – a man with many years of experience. The technician unscrewed the top cover and couldn’t believe what it was. Intrigued, we wandered over to see what he was moaning about. He pointed at dozens of black plastic rectangles with little legs that looked like caterpillars, which were soldered onto a green board. Having just finished a stint at RMIT (Royal Melbourne Institute of Technology), I recognised the caterpillars as integrated circuits. Old Max, a radio tradesman with years of sea trial experience, was outraged that our latest naval equipment didn’t have glowing glass bottles with 300V running to the anodes. Bruce sighed. Then he looked at me and reminded all that I had made a radio control encoder with integrated circuits, so I was nominated to have a crack at fixing the secret box. Not having a circuit diagram, I noted that the integrated circuit numbers started with the prefix 54. There were 5400s, 5404s and 5408s. It dawned on me that the 54 prefix was the military specification for TTL or ‘transistor-­ transistor logic’ digital integrated circuits. 500 I pulled out my blue National Semiconductor TTL Data Book, and there it was – a 7400 quad NAND gate was functionally equivalent to a 5400 chip. Armed with our Tektronix 465 storage oscilloscope, I powered on the box and started looking at the 0V/5V signals going in and out of the gates. Then it happened – a 5404 hex inverter was changing state on the inverter input but not the corresponding output. It had to be blown. Off I went to the store to see Old Jock. It reminded me that when I worked in the mines in the Pilbara, our German foreman Klaus told us that the best storemen were from Scotland because they treated the store’s contents as their private property and didn’t give supplies out without a fight. Our storemen didn’t have to fight me because, when I asked for a 5404 integrated circuit, he looked at me as if I had predicted that Collingwood would win the premiership! He consulted his DSN books and pronounced there was no such thing as a 5404 (there were no computerised stores systems back in the day). I politely asked if he could order two 5404 chips and when they might arrive. That really made him laugh – he mumbled that he would have to send a signal to naval headquarters in Canberra, and it would take at least six months! I wasn’t happy with that news. Then an evil thought crossed my mind. On the way home after work, I made a detour to the new Dick Smith store in Melbourne. I parked on the footpath directly in front of the store (which is legal in the great state of Victoria, of course), walked in and purchased two 7404 chips. The next morning, I gingerly soldered a 7404 chip in place of the dud 5404 and powered the secret box up. It POWER WATTS AMPLIFIER Produce big, clear sound with low noise and distortion with our massive 500W Amplifier. It's robust, includes load line protection and if you use two of them together, you can deliver 1000W into a single 8Ω loudspeaker! PARTS FOR BUILDING: 500W Amplifier PCB Set of hard-to-get parts SC6367 SC6019 $25 + postage $180 + postage SC6019 is a set of the critical parts needed to build one 500W Amplifier module (PCB sold separately; SC6367); see the parts list on the website for what’s included. Most other parts can be purchased from Jaycar or Altronics. Read the articles in the April – May 2022 issues of Silicon Chip: siliconchip.com.au/Series/380 siliconchip.com.au Australia's electronics magazine November 2023  87 he was stunned by the Captain’s knowledge of digital ICs. “You went to RMIT, didn’t you, Gerard?” He had me – I confessed that I told the sailor about the study. Mr Caton shook his head and smiled, and we all went back to work. Fixing a clever weight scale circuit came up, blinking lights – the lot! But since a non-­militarygrade component was used to repair the box, we had to wait until a suitable replacement came before it could officially be used. Shortly after, I had returned with a naval officer in tow, wearing a white uniform with a lot of gold on his cap – he was the commander of HMAS Torrens and asked to speak to Mr Caton, the Radio Workshop Manager. Eventually, the Captain and Mr Caton walked out of the office and came into the workshop. Mr Caton announced that the captain was sending a naval signal to the Radio Workshop approving a temporary repair to their box using non-military-specification components. Mr Caton was impressed with the knowledge of digital integrated circuits the captain possessed. The captain said that he understood that studies at RMIT had revealed that commercial-grade 74-series digital integrated circuits were more reliable than their 54 military specification series equivalents. Evidently, the extra stress testing of the 54 series during manufacture can cause premature failure. So we agreed to the temporary fix using the 7404 chip. Afterwards, Mr Caton approached me and told me that Like many, M. H., of Albury, NSW prefers to fix faulty appliances rather than discard them. It’s worth giving it a try when the fault appears to be a simple one... Like many weight scales, you start ours by lightly stomping on it and waiting for it to complete an automatic zero calibration. You can then stand on it to get the bad news for the day. Over time, that stomping action increased to a jumping action and then to a lift and drop action. The wife had enough of this noisy forceful operation and bought a replacement set of scales. The faulty scales were then forced to the second bathroom to collect hair and baby powder, doomed to be discarded. My thinking moved to how its internal processor would be started with a stomping action. Maybe the sudden change in the load cell output produced the reset action, and a sad capacitor was to blame. However, that would consume the battery when the device was idle. So I had no choice but to open it up and see what was going on inside. Four small screws exposed the LCD screen and a small PCB. The four load cell wires went to the microcontroller and, as expected, it was a blob of black epoxy on the PCB. I was about to declare it beyond repair when I noticed a disc piezo element wired to the PCB and questioned why a set of scales would have a buzzer. Then it dawned on me – it wasn’t a buzzer! The buzzer disc operates as a microphone to ‘hear’ the stomping and wake the microcontroller up. Also, the solder joint that held the disc’s outer edge to the PCB had broken away. That explained the final lift-and-drop requirement. Using a piezo buzzer disc as a microphone is a wonderful idea. It is suspended by the outer edge into free space to amplify the stomping action, generating the impulse voltage required to start the processor. The disc generates its voltage from the kinetic thump and does not impose any battery drain when idle. I added a dob of solder to suspend the disc off the side of the PCB, and it was back in action. I will remember this idea to wake microcontroller projects where a tap to the SC side of the box lights up the screen. Photos of the piezo disc connected to the weight scale PCB. 88 Silicon Chip Australia's electronics magazine siliconchip.com.au