Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG The accordion job Dave Thompson An unusual job turned up at the workshop the other day. Well, it didn’t just walk in; the owner brought it in after discovering it at an estate sale. The inheritors were going to throw it away, but my client saved it. It was a piano accordion, probably at least 50 years old, and this guy couldn’t bear to see it chucked into the bin. This client had played the instrument in various bands over the years and was always looking for a decent model to replace his existing ones because they eventually wear out with all that squeezing. Back in the ‘90s, when I was playing in a folk-rock band, the accordion player was always on the lookout for good working models, perusing second-­hand shops in towns we played because it was increasingly difficult to find a good working instrument. Life on the road is very hard on them. When we did find one, it was pressed into use, and as soon as the bellows blew out or the reeds went west, it would go in the skip because repairing or restoring them was just an exercise in frustration. There were no spare parts to be found, so it was just easier to get another one and put it into service. Now I know what you’re thinking: “did he fire six shots or only five?” Oops, sorry, wrong script. I meant to say: there’s nothing very electronic about a piano accordion. And usually you’d be correct, but this one had a unique feature. At some stage, someone had mounted a couple of microphones on the outside near the grille (where the treble sound comes out). These feed via some not-so-neat cables into a small Jiffy box, which I assume housed a preamp of some description, making it ready to be plugged in and amplified. Back when I played in the band, I was forever struggling to mic up the accordion properly. For one, the guy siliconchip.com.au who played it liked to move around a bit, and two, the microphones we were using (Shure SM57s) were very awkward to mount onto the instrument itself, so we inevitably ended up just gaffer-taping the mic in place. Not very elegant, but it worked reasonably well for what we liked to call “folk and roll”. One of the main issues is that the sound grille on an accordion is quite long, typically the entire length of the instrument and a single microphone is naturally going to pick up sound loudest from where it is placed on the grille. The other notes at the extreme ends of the scale will not be ‘heard’ as well by the mic. This created a headache for the sound guy because it would be very loud in the middle notes and buried in noise for the rest of the reeds placed farthest from the microphone. To work around this, we tried adding shrouds (usually made of folded and shaped stiff card) in an attempt to even out the audio, but with only partial success. Eventually, we settled on using two mics spaced out along the grille, and when mixed together, this provided the best solution. But it looked a right mess with the mics taped to the body and inconvenient cables dragging everywhere, making it a bit of a nightmare to play for the accordionist. Whoever modified this one had crafted two small ‘stands’ for the microphones, but they had ditched the bulky mic bodies and used only the dynamic capsule still mounted in its housing. It was a bit rough around the edges, but the mics were pretty sturdy and solidly mounted to the body. Australia's electronics magazine Items Covered This Month • • • • • The accordion job Brightening up a clock radio Unorthodox Porsche parts Mobility scooter repair The misattraction of a nuclear magnetic resonance machine Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz We apologise for the lack of cartoons in this issue. Our cartoonist, Brendan Akhurst, is currently trekking in the mountains of Nepal searching for evidence of past alien civilisations after their presence was revealed to him in a dream. Each capsule was permanently wired with shielded cables for the short run to the Jiffy box, which was taped onto one of the shoulder straps. There was an XLR connector mounted in the back end of the Jiffy box, and a single standard microphone cable would connect the whole shebang to the snake and off to the mixing desk. Apparently, this part of it was not working, nor were several of the bass buttons, which are mechanically operated by the player to open and close bass reeds on that side of the instrument. So there was a lot going on, and I decided to tackle the non-electronic part first. That was relatively easy; opening a hatch on the bottom of the accordion revealed all the mechanics of the bass buttons, a complicated system of springs, levers, actuators and pushrods. It was ‘literally’ choked with dust, grime, what looked like animal hairs and other detritus picked up over decades of being played in dingy lounges and smoky bars. A good going-over with a decent brush, a bit of low-pressure compressed air and a good lube job with February 2022  85 some light sewing machine oil soon had everything freely moving and ready to go. Now for the electronics The Jiffy box had simply been taped to the strap, and it had likely been there a long time. While the tape’s fabric came away easily enough, most of the adhesive stayed behind. Great, that was one more thing for me to take care of. The bottom of the box was held on by four screws that were easy enough to remove. Inside was what appeared to be a preamp built onto a piece of veroboard. Several small trimmer-type pots were mounted on the board, along with the usual arrays of transistors, capacitors and resistors. I’ve made many preamps like this over the years, so I wasn’t too fazed by it; I’d simply reverse-engineer it to see what I was dealing with, and if I couldn’t get it working, I’d just make another one using one of my existing circuits. The interesting thing is that it had a 9V battery connector fitted, but no battery was present, so it might well be phantom powered. I’d know more once I had it out and under the light and magnifying glass. Once on the bench, I could see there were two channels involved – one for each mic presumably, and each one was identical, with the signals being mixed at the final stage. It was a relatively advanced preamp and appeared to be set up for phantom power, where 48V is sent along the XLR/microphone cable from the mixing desk to power the circuit. However, I thought I’d start things off by applying 9V from my bench power supply to the battery connector to see if there was any life in this thing at all. With power on, nothing happened. I used a signal generator at the mic input and listened to the output with my bench amp. Nothing. Zip. Nada. I drew up a circuit based on what I was seeing. The preamp used JFETs at the input stages, the classic MPF-102 types. With reasonably low noise figures and high input impedances, they were the go-to JFET for quite a few years. There was also a simple tone control circuit, which appeared to be of the Baxandall type, controlled by the trimpots. The output was buffered by a single transistor stage fed by both ‘halves’ of the preamp where the signal was mixed together; overall, it was a relatively straightforward preamp. Its gain and impedance could probably be changed by altering a few bias resistors here and there, but as it had obviously worked in the past, I thought I’d stick with the same values where possible. I used a similar design in a preamp I made many, many moons ago for my acoustic guitar. I’d modified the guitar for live use by including a so-called ‘thinline’ piezo pickup mounted under the bridge. Vibrations from the individual stings are detected by the pickup, and after piping it through to a preamp, the signal is fed to the outside world via a standard 6.3mm stereo output jack that doubles as both an on/off switch and the rear strap-fixing point. On my acoustic, the rear strap holder was on the centreline at the back of the main part of the body. Simply plugging a cable in switched on the electronics using one of the two contacts in the stereo socket, with the inserted plug shorting out the contacts like a switch. 86 Silicon Chip I mounted the preamp inside the guitar on the back side, near a handy timber strip brace, using stick-on Velcro, making it solid but easy enough to remove if I had to. I clipped a 9V battery into a holder using the same Velcro just under and inside the sound hole for easy access; while space was tight, I could change the battery without loosening any strings. The current draw was so low that a battery lasted me at least a year of regular live use. So I decided to use something similar here. All goes accordion to plan What I wouldn’t do is add the complexity of onboard tone controls. Not only is it pointless with them being inaccessible from the outside of the Jiffy box, but they are also redundant because the tone could be controlled by using the much more functional tone controls on the mixing desk itself. Someone can adjust these until the sound is pleasing and then leave them, or they can be adjusted in real-time if a sound engineer is present. I would also stick with the existing XLR output connecter, which would allow me to balance the output signal, with the downside being I couldn’t use the connector as a switch. As I mentioned, it appeared that the old preamp had been at least partly set up for using phantom power, which again complicates the circuit and requires extra components to step the supply voltage down from 48V to 9V. Since the phantom power function is controlled by a switch on the mixing desk, there would be no problem omitting it entirely and simply using a battery, which would last this client several years given the number of live gigs he plays. Then, it would merely be a matter of opening the Jiffy box and replacing the battery when required. The client was happy with all that, so I set about recreating the best parts of the original circuit. Finding components was not difficult, as I have plenty of new-oldstock transistors and FETs. I suppose I could have simply upgraded everything to modern parts, but this job was already eating into my time, and I didn’t want to have to research new values for different transistor types. The 2N3904 output transistor was modern enough, and I had dozens of MPF-102s that I’d likely not use in years, so I chose to use them. I assembled it on a piece of veroboard – designing and making a PCB for something this simple was beyond the scope of the job, but I gave the usual clearances for signal and power lines to minimise hum and RF pickup. Due to the size of the Jiffy box, I had plenty of room to play with. I could have used a new box with a battery compartment and all the usual conveniences, but that would mean lots of marking out and drilling holes and essentially redesigning the wheel, so I left it all that as-was. What I did add was a low-profile toggle switch for turning the thing on and off. I mounted it next to the XLR socket, where it would be unlikely to be bumped but still handy to access. He’d just have to turn it on manually if he wanted to amplify the instrument through a PA system. I won’t bore you with the build, other than to say it is always the best part of the job for me, working out where stuff goes and what tracks to cut on the veroboard. Once it was done, I triple-checked it and powered it up on the bench using my power supply and fed in a signal. I was Australia's electronics magazine siliconchip.com.au greeted with a nice strong output signal in my ‘phones, so it was obviously working. The next step was to plug in the two mics and the output to my test amplifier and see what happened. I had a very clear output from the mics, with quite low noise, so I was pleased enough with that. The wires coming from the mic capsules were shielded but routed awkwardly over the accordion and simply held in place with strips of tape. As this wasn’t very elegant, I looked to see if I could improve on that somehow. As usual, getting the old gaffer tape adhesive off was a mission in itself, but some liberal use of isopropyl alcohol soon had it back to a natural finish. I wasn’t about to start drilling holes in the instrument’s body, and the only feasible way was along the edges of the moving parts and off up the strap to the Jiffy box. I’ve collected lots of those little square cable clips over the years – they used to come with some motherboards or computer cases, and I always ended up with way too many. They have a very low profile, with a small slot for a cable tie to pass through. I have both black and white versions, so I put each one on the bright red body to compare looks. I decided to go with the black ones since the cables were also black. I (literally) pressed them into service along the cable run, about every 60mm, using double-sided tape applied to the bottom of each holder. Once in place, it was a simple matter of running the smallest cable ties I could find in my drawer through the slot, around the cable and pinching them down snugly without the cut-off part of the tie being exposed. This can rip skin if that part sticks out and one rubs against it the wrong way. I also used longer Velcro straps to mount the Jiffy box to the accordion strap, in the position it was before, making it easier to remove to change the battery. I was pretty pleased with the result. It was not ideal, but a lot tidier than before and likely more robust as well. The only thing left to do was unclip the bellow straps and have a play through a proper amp. I’m no keyboard or piano player, so this test would just involve a lot of noise. Due to a few years of piano lessons, which ended about 50 years ago, I know a few scales, but that’s about it. And hefting accordions around, squeezing and pulling and hitting buttons and keys all at the same time is more complicated than drumming! While I couldn’t do it justice, it sounded pretty decent through the mic input on my guitar amp, and tone control was also broad and workable. I called the client, and he came around and put me to shame playing it but was very happy with the result. I hope he gets many good years of use out of it now. Brightening up a digital clock radio display B. P., of Dundathu, Qld is one of our most prolific contributors, and he hasn’t stopped yet. He doesn’t want a repairable device to be thrown away if he can help it... We have had this digital clock radio in our lounge room for longer than I can remember. I’m not even sure where we obtained it, but I think we bought it second-hand from one of the local op shops around the time we moved into our new home, in 1992. The clock has worked well over the years but lately, the siliconchip.com.au Australia's electronics magazine February 2022  87 time would start flashing even though it was still correct. I fixed this by incrementing the hours until I got it back to the right time. At first, I suspected it was caused by a power supply glitch, but it kept happening. After a while, the clock started going haywire and showing all sorts of random times. I ignored it for a few days, but then when I tried to reset the time, it was stuck flashing 12:00. I decided to replace the clock initially and have a look at it later. However, the replacement clock had a dull red display which was harder to see and is more suitable for a bedroom, whereas the original clock has a bright yellow display that was much better with the bright light in the lounge room. So it was time to have a look at the original clock to see what the problem was. I already had an idea what was causing the problem, as some years ago I’d encountered weird behaviour from a digital clock. I was unable to diagnose the problem until I built an ESR meter. I was then able to determine that the filter capacitor was faulty. Replacing it fixed that clock, and it’s still working well now. Suspecting that this clock had the same problem, I proceeded to dismantle it. This was quite tricky as, being a clock radio, it has the cable for the front radio display needle under the circuit board. That meant that I couldn’t take the circuit board out of the clock to work on it without making reassembly very difficult. After removing the three screws securing the board, I managed to lift one side of the board high enough to test the filter capacitor with my ESR meter, but I couldn’t get any reading from it. So the capacitor was basically open-circuit. I then managed to get my 25W soldering iron under the board and removed the capacitor. I re-tested the capacitor with the ESR meter while it was still warm from desoldering, and I got a reading of 88W. I tested it again later after it was cold and once again, I got no reading. This is one of the worst capacitors I have ever encountered that hadn’t blown its top; it looked like it was still good. This ESR meter has helped me greatly over the years to identify seemingly good capacitors as bad. It was marked as 470μF 16V, but a compact size. I hunted through my container of salvaged capacitors and I found a few around the same size. After testing them with my ESR meter, I selected the one with the lowest reading and installed it. This was quite tricky, trying to solder under the board with minimal room, but I managed to do it. Before reassembling the clock, I tested it to make sure that the repair had been successful. I set the clock up safely so that I was able to see the display and access the buttons on top of the top case. After plugging the clock in, it flashed 12:00, so I changed it to the correct time. This was successful, so I had obviously solved the problem. I unplugged the clock and then reassembled it carefully, ensuring that the power cable correctly looped around the post that acted as a cable restraint. I then returned it to its place in the entertainment unit, and it’s now working perfectly again with its usual nice bright display. This was another win for the environment and also my pocket. Classically unorthodox car parts D. T., of Sylvania, NSW ran into the bane of the classic car collector, non-standard parts that are hard to find (and often expensive). Thankfully, this one could be disassembled and fixed at a component level... This digital clock/radio had a few problem capacitors. 88 Silicon Chip Australia's electronics magazine During my spare time in COVID-19 lockdown, I’ve been restoring a 1982 Porsche 928. This is a nearly 40-yearold car, and parts are becoming scarce (read: expensive). I’ve been working my way through the car and came to the rear demister. Having resoldered the terminal to the back window (not as hard as it sounds), I connected the battery and switched it on, only to find no warmth at all. A quick check of the fuse box found the relay missing. The 928 is a complicated car by 1980 standards (not today’s, though!). The rear demister provides two power levels. A high heat ‘Boost’ mode operates for about 15 minutes when you push the (momentary) switch. A lower power ‘Maintenance’ mode runs continuously when the switch is on. Boost mode also activates the rear-view mirror heaters. The demister itself is the typical resistive type but is split into two halves – the halves run in series in Maintenance mode and parallel in Boost mode. When this car was designed, they didn’t have the integrated electronics systems that cars have now, so the timing and switching functionality was provided in a special double-width relay that plugs into the fuse panel. This relay also has start and ignition inputs to disable the demister during starting or when the engine isn’t running, and an output to drive the indicator light in the switch. I found a used relay online, and it wasn’t too expensive, so I bought it. It arrived a week later but, after plugging it in, I was disappointed to find Maintenance mode worked OK but Boost mode didn’t. I was about to contact the seller, but a check of the ad showed it was “for parts or not working” – I had missed that point. I decided to try to fix it myself. I thought it probably had a dried out electro. It wasn’t hard to open – I used a screwdriver to bend the aluminium case around the edge and removed the phenolic base. The base was part of an assembly that included the two relays and a phenolic PCB. The circuit consisted of two relays and three transistors plus quite a few resistors and diodes. It all looked pretty good – the tracks and soldering were OK with no apparent faults, nothing was scorched, and the electros hadn’t leaked or were siliconchip.com.au A redrawn circuit diagram of the demister from a Porsche 928, with the actual module shown in the photo below. bulging. I set it up on a bench supply and confirmed the Maintenance relay operated correctly but the Boost didn’t. I measured the relay coils and found the Maintenance coil to be about 60W but the Boost coil was way higher – in the kilohms range. I had a good look at the PCB – most of the soldering still looked OK, but the relay coil windings were very fine wire (0.1mm) and where they joined onto the PCB looked a bit sus, so I cleaned and resoldered them. It was tough to tell if the joint was OK because the wire was so fine, but now I measured something more reasonable for the Boost relay coil. Testing now showed it would latch for about three minutes, but nothing like the expected 15. To make matters worse, the time would get shorter each time I tried it, and after a couple of runs, it would only pull in while the Boost line was active (ie, while the button was being pressed). There were two electros – one of them was 470μF (clearly the main timing capacitor), so I measured voltage across it while I held the relay engaged. It discharged very slowly, as expected, but I didn’t know what the trip point was. I replaced it anyway, but it didn’t make any difference. I then started changing other parts – the other electro and the transistors siliconchip.com.au – all to no avail. I saw another solder joint that I didn’t like the look of, so I resoldered it, then decided to resolder them all. It still didn’t work. Next, I decided to trace out the circuit. This sounds easy, but the combination of non-standard part pin spacing, no overlay and some factory modifications meant it took a few hours before I had something that I thought was right. I’m quite amazed by electronic design engineers of these old eras – they did so much with minimal parts. Like old valve TVs – 10 or so valves to Australia's electronics magazine make a whole TV! These days you’d just pop in a microcontroller and be done with it, but that’d be a couple of hundred thousand transistors on its own. A 555 could do the timing, but that’s probably a hundred transistors, plus you’d need other logic. I tried monitoring voltage levels, but due to the very analog nature of the design and the pre-existing fault, I struggled to rationalise what was happening with what was on the schematic. In desperation, I measured the relay coil winding resistance again and found the Boost relay coil was February 2022  89 back where it was when I started, way too high. Thinking I still hadn’t made a decent connection, I fiddled around with it – sometimes it would measure OK and sometimes not. I couldn’t see anything wrong with the coil but nothing I was doing was working, so I decided to bodge in a temporary replacement. I grabbed a relay from an old motorised car antenna and wired it in place. Success! This worked for around 15 minutes every time. The next thing was to fix it properly. The antenna relay was too big, so I either needed a new, smaller version or had to fix the old coil. From the load resistance, I worked out it needed 20A contacts but I couldn’t find anything small enough, so I started looking inside old car relays. I found one with a coil similar in size and resistance to the faulty one, and with a bit of trimming, I got it to fit. My guess is the old relay coil has a break somewhere with the wire ends rubbing against each other to make a high-resistance joint. When I moved it around or some heat accumulated in it, the ‘joint’ would fail. Unfortunately, I’ll have to wait for a while before I actually use it as the car needs a lot more work. Mobility scooter repair B. G., of St Helens, Tas wasn’t content to simply swap a failed board. He decided to investigate and figure out why it failed. It turned out to be a simple but unexpected fault... My wife has a large second-hand four-wheel mobility scooter (she calls it her tractor). One morning when she went to power it up, it was dead; when switched on with the key, a small meter usually shows the relative battery condition and a power LED lights. I could see a bunch of cables running up the steering column, disappearing behind a cover. Removing that cover exposed a circuit board. This was easily removed by unplugging the cables. Close inspection showed a mixture of parts and no sign of heat or damage. We had the original operating manual with the agent’s number in Hobart. We rang him, and he very helpfully agreed to send several boards after paying a deposit. He suggested measuring the battery voltage and shorting the key switch, which I did to no avail. Starting with the easiest part to access, I decided to replace the control 90 Silicon Chip board on the steering column and was rewarded by the machine coming to life. I returned the remainder to the agent. He was surprised at the failure, saying they had never had a control board failure before. But the story doesn’t end there. When our family arrived for Christmas from the mainland some months later, lo and behold, the scooter failed again with the same symptoms. My son-in-law, a medical electrical engineer, decided to remove all covers and trace and check all the looms while I traced as much as possible on the new control board. There was no obvious damage on this board either, but the key switch track went through a plated-through via to a socket pin to the motor controller. The trouble was that there was no continuity from one side of the board to the other, so we used a small drill to open up the via and soldered a wire to the tracks on both sides. That fixed the continuity problem, and the scooter came back to life. For the other failed board, a simple wire link soldered between the socket contacts was an easier and quicker repair. So I now have a serviceable spare. I contacted the agent again. He seemed impressed, saying that they would not be able to fault-find to that extent, and they would email the manufacturer in Israel. Some weeks later, the agent rang again to say that they had agreed with our diagnosis and that they would modify all their boards with a wire link. I hope the brain keeps working; it’s satisfying when it does. Editor’s note: it seems that the via was too small and fused due to inrush current at switch-on. Larger vias or more vias in parallel would likely solve the problem, although a through-wire is a very robust solution. The misattraction of a nuclear magnetic resonance machine D. D., of Coogee, NSW recalls a servicing problem he encountered many years ago. At first, it seemed that something was wrong with the electronics, but the fault was traced to another nearby source... Two articles in the August 2021 issue prompted me to write to you: Advanced Medical & Biometric Imaging (siliconchip.com.au/Series/369) and the History of Op Amps article (siliconchip.com.au/Article/14987). Both brought back fond memories of my long-lost youth and reminded me of a story that might amuse your readers. The top and underside of the control board of a mobility scooter. A simple wire link as shown on the underside fixed the continuity problem that was found. Australia's electronics magazine siliconchip.com.au In the mid-1960s, I worked at a university chemistry department in the UK, looking after electronic equipment. The story involves NMR (nuclear magnetic resonance) machines and valve-based op amps. NMR machines were highly prized (and very expensive) in those days, and the chemists loved them because they could get a beautiful paper chart output showing the exact chemical composition of a sample. Not long after I started, we got an NMR machine. It was installed during a holiday period when the university was very quiet, in a small room on the lower ground floor of the building. One of the lab technicians, Archie, was ‘promoted’ to work as the machine operator and given the necessary training to use it. All went well for a few weeks; academics and researchers brought samples down to be analysed, and Archie duly provided the relevant chart outputs. However, it was not long before things started to go awry. One day, I got a call from a harassed Archie asking if I could go and see what was going wrong with his machine. He showed me charts where the trace had started normally and then suddenly disappeared. “It happens at random,” he said, “and usually when I am just doing something very critical, it is driving me mad. Do you think you can fix it?” I was a bit dubious as it was a very complex machine, and I only had the vaguest idea how it worked, but I took the manuals back to my workshop to study and promised to come back the next day. I could see that it had a huge magnet, and the manual made it clear that the stability of this magnet was of paramount importance, within a few parts per million. I also saw that the output peaks could be integrated to indicate the quantity of each element in the sample. This was done using a valve-based op amp integrator. My first thoughts were that either the magnet or the integrator were drifting randomly. I wasn’t game to go anywhere near the magnet as the manual had lots of dire warnings, but I thought I could have a look at the integrator. This was a plug-in module; I pulled it out and saw it had a row of valves and an impressive looking feedback capacitor, among other components. I could see no obvious signs of a fault. Ordinarily, I would have suspected the feedback capacitor and replaced it, but I could not find a suitable part, and I was reluctant to ‘hack into’ this new and expensive machine. So I admitted defeat and said I would call the company. Soon, the rep turned up and of course, Murphy being alive and well, the machine behaved perfectly. He said that the problem was probably caused by large metallic objects moving in the magnet’s fringe field. Maybe it was cars passing by in the car park, right outside the wall, or the lift next door. He said the magnet fringe field could extend several metres, and the solution was to install steel sheets in the walls of the room to screen the magnet. The estimated cost was thousands of pounds. At this point, the Professor was called, and a discussion ensued as to what to do. As a true academic, he decided that an experiment must be conducted to find the actual cause of the problem. One of the junior lab techs was summoned and asked to drive his car past the NMR room, jump in the lift, go up to the top floor, then come back down. Archie started a scan, and we all waited to see the results. Sure enough, both things caused the machine to go haywire. The Professor was very annoyed and puzzled, and demanded to know why this had not been observed when the machine was first installed. Of course, it was now term time, and hordes of students were around, going up and down in the lift and driving in and out of the car park. The Professor said he was not going to pay thousands for screening the room. His solution was to paint an exclusion zone outside on the car park tarmac and instigate times when the lift could not be used. Poor old Archie then had to put ‘out of service’ signs on the lift whenever he quickly did a batch of scans. The situation still exists with modern NMR and MRI machines, but proper installation planning involving medical physicists can eliminate the problems (see www.aapm.org/ SC pubs/reports/RPT_20.pdf). U Cable Tester S B Test just about any USB cable! USB-A (2.0/3.2) USB-B (2.0/3.2) USB-C Mini-B Micro-B (2.0/3.2) Reports faults with individual cable ends, short circuits, open circuits, voltage drops and cable resistance etc November & December 2021 issue siliconchip.com.au/Series/374 DIY kit for $110 SC5966 – siliconchip.com.au/Shop/20/5966 Everything included except the case and batteries. Postage is $10 within Australia, see our website for overseas & express post rates siliconchip.com.au Australia's electronics magazine February 2022  91