Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Toys with a serious purpose Dave Thompson You might think I’m running out of things to repair because I’m working on toys again. But actually, while this device appears to be made of child’s toys, it has a serious medical use. It is used to check the hearing of young children who can’t yet talk (or don’t like to) but this particular example didn’t survive the tender mercies of one particular delivery service... It will come as no surprise to many that I sometimes get weird and wonderful devices through my humble workshop. Word somehow leaks out through the concrete-­jungle telegraph that I am willing to look at anything, always with a view to (hopefully) repair it. A few weeks ago, a long-time Silicon Chip reader from another part of the country contacted me about a device he had designed and built many moons ago that needed some electronics-based love and attention. Darryl was an audiologist in one chapter of his life, and as an electronics enthusiast, he had designed and built several VRA (visual reinforcement audiometry) devices to assist with testing very young children’s hearing. The usual traditional tone-testing we all know as adults is not so easily performed on toddlers. These so-called ‘puppet boxes’ are utilised as an audio-­ visual testing aid. In this case, the ‘box’ itself is in the form of a large, round, carpet-covered timber tube about 1.8m high and 50cm in diameter. The front ‘face’ of the tube is covered by a dark Perspex cover almost to the bottom, where there is a panel covering the electronics. 100 Silicon Chip There are three levels to this box, all isolated and separated from each other. In each of the levels is a different animatronic ‘toy’ which, when activated, is illuminated by an LED spotlight in the top-left corner of that section. A hand controller – a jiffy-type box with three press-tomake, release-to-break switches on a long lead activates each of the toys. The ant, the gorilla and the elephant In this box, starting in the top section, is a large ant in a forest-type setting. He speaks a phrase and his bug-eyes move and blink when the appropriate button on the hand controller is pressed. The animation and light activates for about 15 seconds before going dark again. The middle level contains a furry gorilla toy, set in a yachting scene; this toy dances and his mouth opens and closes along (roughly) to a popular 90s novelty song. The bottom section has a pink plush elephant, which animates with its legs moving in time to a typical child’s ‘crawling song’. In practice, I suppose it doesn’t really matter what toys are there and what they do; they are just something to grab the attention of the child. I was certainly entertained and admit to pressing those buttons a lot – once I’d fixed it. I now know the lyrics to a crawling song, so life is full of surprises! In use, a child being tested is ‘trained’ to respond to a test tone they hear through headphones by watching a visual stimulus every time they hear the tone. The audiologist plays a tone and activates the puppet box at the same time. The child looks at the animated toy and ‘learns’ that when they hear the tone, they’ll see the toy light up and move. As the audiologist changes the tone frequency and the volume level, they can build a picture of that child’s hearing and diagnose accordingly. It’s a time-honoured method of testing children’s hearing. The idea of the puppet box is relatively simple, and the implementation in this case very clever; I suppose any animated toy could do the job, and all we need is a suitable power supply, some lights and a timer board to control it all. The majority of us would likely never see such a thing, unless we had a very young child in need of having their hearing tested. I’ve certainly never seen one before, even though I’ve had many hearing tests over the years, so all this has been an education for me. Australia's electronics magazine siliconchip.com.au The VRA in my workshop is a classic example of someone with electronics knowledge and the vision to create something practical and useful using that knowledge. These things should have become an industry standard. They might very well be, for all I know – no doubt some company makes and sells them for exorbitant amounts of money. Still, for whatever reason, this box, and several others like it Darryl made, have stayed a relatively ‘local’ solution to this problem. Damaged in transit The problems with this unit started when the puppet box was shipped from ‘up north’ to ‘down south’. When it arrived, it no longer worked. According to the people involved, the packaging showed no signs of trauma, and the unit was intact, but there was obviously something quite wrong with it. The sections lit up with the button press, but the only toy moving was the elephant. While I got audio from the others, the ant and the gorilla were not animating at all, though their LED spot lights did activate. The recipient of the box got in touch with Darryl and then he got hold of me. Me being in the same town down here as the customer likely sweetened the deal. He asked me if I could take a look at it, rather than ship it all the way back ‘up north’. Of course, I said yes, and really, what else could I do? The chance to work on something new and unique is many a serviceman’s dream (well, it is mine anyway). The box arrived at the workshop nicely bound up in a woollen rug, although that was not the original shipping material. After unwrapping, it seemed intact and undamaged when viewed from the outside. However, after plugging it in, connecting the hand controller and trying it, there was obviously something wrong. The ant and the gorilla toys looked to be sprawled on the floor of their respective compartments, and there was no movement at all visible when the compartment lights came on. The elephant at the bottom also looked a bit skewed, but it did at least have some music and motion when the relevant button was pushed. The first thing to do was remove the smoked black Perspex cover, exposing the toys. It was held on with standard super-screws, countersunk into the plastic. Some kind of sealant had been applied around the bevelled edges of the screen, perhaps to remove any sharp machining edges that might catch a small hand. With the screen off, I could see what was going on. The ant was meant to siliconchip.com.au Items Covered This Month • • • • • Toys with a serious purpose A Sony tuner with a side of gum Washing machine and mixer repair Repairing a Toyota wheelchair lift Restoring a Porsche 928 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 be standing upright and be glued to the back of the compartment, while his ‘feet’ were fixed to the timber ‘floor’ of the section to hold it into position. The glue holding it upright had torn away from the wrap-around cardboard picture background, so the toy was just folded up headfirst like a rag-doll on the floor. I unwound him and stood him up and pressed the button. This time, his eyes rolled and blinked and he said his lines quite happily. A simple fix for this one then; all I had to do was clean him up and glue his backpack back into place on the background to secure him in the right place. One sick gorilla The gorilla section was a bit different. He’d come well adrift and was crumpled onto the floor. He’d been tacked/ glued in three places to hold him upright – all these points had been torn away, as evidenced by the paper stuck to the glued areas. The box had evidently suffered quite an impact – at least enough to knock these guys free. The gorilla didn’t move at all, so something electrical had gone wrong. He did, however, emit the opening notes of that novelty song, but it sounded like a stuck record, playing the same sound-bite over and over. A check with Darryl confirmed that it should play the whole song for the allotted time, while the gorilla would usually dance and ‘mouth’ the words. Apparently, something else was amiss here. All the toys were originally battery-­ powered. Power for this box – and the toys and lights – was derived from a battery charger mounted into the cavity in the base of the unit. Australia's electronics magazine November 2022  101 Switched permanently to the 6V setting, the charger was working because all the compartment LED lights activated on the press of the button. Two of the toys moved, however weirdly, so the juice was making it through to all but the ape. The lights and animation were all controlled by a common timer, so they all operated for the same duration. It was mounted on top of the charger. The ammeter on the charger indicated when the toys were powered – again, except for the gorilla, which was not surprising as it seemed to be the most affected by the drop. The toy was easy enough to remove – one foot was still glued to the base, so I carefully got a box-cutter style knife blade in between the glue and card and gently separated the two. Once free of that, I then had to desolder the power leads, which went directly to the battery compartment and were soldered to the contacts there. With the toy removed, I pressed the button and my trusty multimeter showed roughly 6V getting to the now-bare wires, so whatever was wrong was inside the toy. Fortunately, the manufacturer had installed a zip in the bottom of the plush, and when opened, this revealed the internal mechanism. A large torso-shaped plastic clamshell box was held together with four screws, and with these removed, the two halves came away easily. Access was very tight – the plush would only pull back a limited amount, as a cable tie secured the neck area to the actuator case. I had to cut and remove that tie to allow deeper access, and as there were linkages to the arms, the head and the mouth, I had to be careful I didn’t damage those. More damage inside Inside the actuator was the sound and movement controller module. This was made from two small PCBs joined at right angles, screwed to the plastic case, with one containing a COB (chip-on-board) IC that likely had the song programmed into it. There were a few other surface-mounted components which I assumed controlled the animatronics. The bottom half of the case also contained a reasonably complex plastic gearbox and an electric motor. The motor is a typical small DC motor that powers countless toys. Its 102 Silicon Chip leads were adrift, explaining why there was no movement. I applied a bench power supply at 5V to that motor directly and sure enough, it fired into action, so that was the likely problem. A suppressor capacitor was also connected across the motor’s terminals – this had broken away on one side, so a quick hunt through my spares box had that replaced and the connections remade and solidly soldered in place. A press of the button proved the gorilla now jiggled and animated, but his mouth didn’t move (it should) and that song was still just a machine-gun-style delivery of one note. I didn’t think I’d be able to do anything about the lack of music, but I could look into that gearbox and see why the mouth didn’t articulate. After stripping the gearbox assembly from the toy (which meant pressing a few pin-hinges out), I could see an actuator shaft that should have operated the mouth had snapped at a join in the neck area. It was a clean break and there was plenty of meat there for a glue job, so I hunted through my adhesives collection for some plastic-welding cement. This material is standard PVC or similar, and a test drop of glue onto an open area confirmed that it would indeed weld those bits together. Once glued and held for a few minutes (nothing seems to go on as long as waiting for glue to harden), I carefully reassembled the gearbox, motor and linkages and left it overnight before powering it up. The next morning, I held the gorilla roughly in place and pressed the button. Success! He did a little dance, and his mouth snapped open and closed. The ‘music’ wasn’t happening properly, but Darryl mentioned that as long as there was some noise, it would suffice for the purpose. I suppose we could have sourced another toy, or even grabbed a sound module from a charity shop talking toy, but a quick look showed that animated toys are really expensive these days. I learned over the phone that most donated toys go pretty quickly from the Salvos and Vinnies stores. Shaving the gorilla (no, really!) After zipping him back up, I cleaned off as much of the old hot-melt glue originally used to hold the gorilla in from his fur coat as I could. I wouldn’t like to have to get this glue out of a real coat, because it makes a nasty mess! Shaving it partially was the only way to remove the bulk of it. I used the rest of the glue as a template to fit the limbs back into their original positions, which were pretty obvious due to the torn-away card. I also used the same glue to reattach the elephant toy to the bottom, which was a little askew on its mounts, but still working as intended. After a bit of tidying up of wire runs and the dog-eared and torn panoramic picture backgrounds, I was satisfied it was all working as well as it ever would. Due to the extra dollops of glue liberally applied to the toys, I’m reasonably sure they won’t be coming loose again. [I’m sure the post office would accept your challenge – Editor]. If they do, I think we’ll have bigger problems! The only thing left to do was to replace the front cover and give the box a good workout. As it has been in the workshop for a couple of weeks, all my visiting customers see it and, of course, ask what it is and how it works. I pass them the hand controller and everyone seems to have fun making the toys animate; it makes for a good conversation piece. Australia's electronics magazine siliconchip.com.au Keep your electronics clean, lubricated and protected. Service Aids & Essentials. GREAT RANGE. GREAT VALUE. In-stock at your conveniently located stores nationwide. 4 2 1 5 3 BUY IN BULK & SAVE!!! 1 Isopropyl Alcohol 99.8% 250ml Spray NA1066 BUY 1+ $7.95 EA. BUY 4+ $7.15 EA. BUY 10+ $6.35 EA. 99.8% 300g Aerosol NA1067 BUY 1+ $11.95 EA. BUY 4+ $10.45 EA. BUY 10+ $9.45 EA. 70% 1 Litre Bottle NA1071 BUY 1+ $19.95 EA. BUY 4+ $17.95 EA. BUY 10+ $15.95 EA. 2 Electronic Parts Cleaning Solution 1 Litre Bottle NA1070 BUY 1+ $15.95 EA. BUY 4+ $13.95 EA. BUY 10+ $12.45 EA. 3 Liquid Electrical Tape 28g Tubes, Red or Black NM2836-NM2838 BUY 1+ $19.95 EA. BUY 4+ $17.95 EA. BUY 10+ $15.95 EA. 4 175g Aerosols Contact Cleaner Lubricant NA1012 Electronic Cleaning Solvent NA1004 BUY 1+ $12.95 EA. BUY 4+ $11.45 EA. BOX OF 12 JUST $119.40 PTFE Dry Lubricant NA1013 BUY 1+ $15.95 EA. BUY 4+ $13.95 EA. BOX OF 12 JUST $149.40 5 J-B Weld Products J-B Weld Epoxy 28g NA1518 BUY 1+ $17.95 EA. BUY 4+ $15.95 EA. BUY 10+ $13.95 EA. SuperWeld Extreme 15g NA1539 BUY 1+ $14.95 EA. BUY 4+ $13.45 EA. WaterWeld 57g NA1532 BUY 1+ $23.95 EA. BUY 4+ $21.45 EA. BUY 10+ $18.95 EA. Shop at Jaycar for even more service aids & essentials: • Adhesives & Insulation Tapes • Solder & Soldering Aids • Wire & Heatshrink Tubing Explore our full range of service aids, in stock at over 110 stores, or 130 resellers or on our website. • Fasteners & Cable Ties • Ultrasonic Cleaners • Tools & Workbench Accessories jaycar.com.au/serviceaids 1800 022 888 It would have gone back by now, but the designer and the customer are talking about modding it for use with foot-­ operated switches as well. That would require finding some single-pole press-to-make, release-to-break stomp-box type switches (sometimes called momentary switches), some of which I think Darryl has already sourced. I initially thought readily-available musicians’ foot switches would be ideal, as they are designed to be trodden on. However, they are usually quite expensive and are typically multi-pole, press-to-make, press-to-break types used for switching effects in and out of the signal chain. Anyway, with the right switches in hand, it will be a relatively simple matter of taking off the power supply cover again and adding another paralleled multi-pin plug to the existing facia. As the audiologist may want to have the switches in a different room to the box (usually, they are in the same room), we’ll have to find a way to utilise jack panels on the walls. Still, that’s for another day – at least now, the box can be returned to normal service and delight kids of all ages. One Sony tuner with a side of gum B. H., of Little Mountain, Qld ran into one of those situations where, while diagnosing a fault, the measurements didn’t seem to make sense. Luckily, he managed to figure out the reason for those discrepancies... I inherited a Sony tuner that played for many years, but within the span of one week, its FM output diminished to zero. I thought the most likely culprit was the first RF transistor. This unit is beautifully made and is of the PLL type, which was very novel back in 1981. After googling for many days, I could not find a schematic, so I started from the beginning. This unit is not unusual; the RF input stage is followed by the mixer/oscillator, then a dual transistorised IF stage, followed by an IC that turned out to be the PLL FM IF decoder, which includes another six stages of IF amplification. Three ceramic filters are associated with the transistor amplifiers. However, I couldn’t find a bias voltage at the base or emitter of either transistor. I also checked the voltages on the connected resistors. They had used red gum to stop some flying leads from interfering with the station pointer dial mechanism. This blob covered a resistor and solder joint. I removed the red gum to access the resistor, which I believed to be the bias provider. The resistor had the correct value on the component side, but the solder joint on the copper side was peculiar. I could measure a positive 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. 104 Silicon Chip voltage on the resistor lead, but the surrounding solder tested 0V, suggesting a dry joint. Re-soldering it made no difference; the pad and resistor wire were still 14V apart. It all looked proper on closer inspection; the wire was surrounded by solder with the correct colour. But I could not find any continuity between the wire and the solder. I think the gum from the PCB component side must have corroded the tin coating of the resistor wire all the way down into the solder joint, isolating the resistor from the solder. Replacing the resistor with one soldered on the copper side of the PCB fixed it. Washing machine and mixer repair P. M., of Christchurch, New Zealand made three successful repairs on very different pieces of equipment lately... The first was my washing machine, which has been reliable for more than ten years. Recently, I switched it on and instead of the usual bright display of LEDs showing the wash program and time, the LEDs were dull and flashing slowly. Switching it off and on again didn’t help. A Google search first turned up some unhelpful suggestions, but then there was one that seemed to be on the right track. The post mentioned a faulty 10µF 450V electrolytic capacitor on the main board and that person had found a replacement at Jaycar. I was planning to go to Jaycar later that day, so I added the capacitor to my list. The post also mentioned cutting a hole through some plastic with a hole saw to get to the part, which sounded a bit extreme to me at first. The machine’s top must be removed to get to the main board. Once away from the wall, after removing a few screws and some plugs, out comes the main board. The board sits in a plastic tray filled with clear sealant covering most of the components. The capacitor was visible in the middle of the board and was obviously faulty as the top had bulged open. The problem was that the solder side of the board was face-down in the plastic tray, and I didn’t want to disturb the sealant holding it all there. The hole saw was starting to make sense now, but still seemed a little risky. Instead, I drilled a series of small holes to make a circle that I could nibble out to make a hole to access the solder joints of the capacitor. I was glad I did this as there were more components on the solder side of the board, and hitting any of them with the hole saw would not end well. I could now replace the capacitor and used silicone sealant to refit the plastic piece I had removed. As I was putting it all back together, I remembered the original poster had said to be careful not to pinch the rubber hose at the back of the machine when refitting the top cover. This was valuable advice, as when I looked at the hose, it was indeed pinched under the top. This hose is part of the water level sensor, and if I had not freed it, the machine would have overflowed the first time it was used. Unsurprisingly, when I fired it back up, it was working correctly again. Secondly, I repair a lot of analog audio mixers of various sizes; some of the most common problems are faulty faders. Those slots in the panel for the faders allow ingress of all manner of dirt and liquid spills. In most of these units, everything is mounted on one large circuit board, including Australia's electronics magazine siliconchip.com.au the faders, so access means removing all the knobs, nuts and screws to get the board out. I have found that in many cases, there is sufficient space at the front edge of the board to wiggle the fader out once it has been desoldered, and a new one can be refitted the same way. The mixer I worked on recently did not have this space, but I still managed to get the faulty fader out. Getting the replacement in would be a different story until I had an idea. I fed a piece of fine wire through the fader slot from the top of the panel and wiggled it around until I got it past the PCB. I then tied the wire to the fader’s shaft and used it to slowly guide the fader to its location, where I could then push it back into the holes in the board, resolder it and test it. That saved me a lot of work. Thirdly, a friend sent me his large stereo amplifier that had stopped working after a heavy session. After a bit of probing, I discovered the power transformer’s primary winding was open circuit. It is not uncommon for Japanese-­ manufactured equipment to have a thermal fuse inside the power transformer in series with the primary winding. These fuses are non-resetting, so the unit will no longer work after they trip. At first, I could not see the fuse; they are often buried inside the bobbin that holds the windings. After snipping away some small pieces of the bobbin, I could just see the fuse at the bottom of a slot roughly 5mm wide and 40mm deep. With a suitable probe, I could determine that this was where I needed to connect to the winding, but how could I connect to it when my soldering iron would not fit in the slot without melting everything around it? Then I remembered that my soldering gun uses a piece of looped copper wire as an element and it is not very wide. I dropped a small grain-of-wheat lamp into the slot and powered it from my bench power supply so I could see. I was then able to solder a wire to the fuse to complete the circuit. That got it back into business, but as I had defeated the thermal fuse, I checked to ensure a conventional suitably-­ rated fuse was in line with the transformer. The thermal fuse is an extra line of defence, but there are millions of power transformers out there that do not have one. I have even seen some transformers with an extra wire to bypass the thermal fuse should it fail. Repairing a wheelchair lift inside a car J. W., of Hillarys, WA is another regular contributor to this column. His latest entry involves repairing a wheelchair lift built into a Toyota Regius van... A friend rang recently and asked if I could have a look at his daughter’s wheelchair lift. He said that the lift was not going back to the home position in the back of the van. He had contacted several local auto electricians, but they were not interested in looking at it, as it was a Japanese import with no service information available. So I went around to have a look. The system was a genuine Toyota accessory and seemed well-built (much like their cars). If the carriage was taken too low, it would start to lift the van! The lift has large motor driving screws that lower and raise the platform, and that part was working as it should (see the photo at upper right). siliconchip.com.au The wheelchair lift attached to a Toyota Regius van. Once the platform was back at the level of the van floor, two smaller motors and a second set of screws was supposed to drive it into the home position. This was not happening. The system was all-electric and controlled by four relays & four microswitches. The main control assembly was easy to unbolt and inspect. It had three large onboard relays plus a smaller one. The large ones were not enclosed, so I could see them operating; they seemed to be working OK. I unplugged and tested the small relay, and it was also functional. Next, I decided to start checking the microswitches. Two were easy to access and tested OK. The other two, which switched power from the large motor to the smaller motors and stopped the platform in the home position, were challenging to access. After some deliberation and standing on my head in the back of the van, my friend figured out that we needed to take off the runner plate as that would let me access the microswitches and wiring. After that, I tested the final two switches, which were fine. I then lay on the van’s floor and used a torch to look at the wiring to the platform, which was hidden in part of the frame. It was encased in black plastic cable chain. Sticking out of the cable chain was a broken wire. I managed to effect a repair by joining the broken ends using a small section of new wire. The break had occurred because the plastic cable chain had broken at one end, so the wiring was not supported. I managed to cut some flexible conduit and fit it over the broken section, which did not actually need to be flexible. We then put the whole system back together, and it worked as it should. The whole process took many hours over several days. My friend and I were relieved finally to have the lift back in the van and working again. Restoration of a Porsche 928 (Central Warning System) D. T., of Sylvania Southgate, NSW previously wrote about repairing a Porsche 928 demister control relay (February 2022, pages 88-90). He’s still working on restoring that car, but this time, it’s the warning system that’s on the fritz... Australia's electronics magazine November 2022  105 During my restoration of a 1982 Porsche 928, I came to the Central Warning System (CWS). In addition to all the usual warning lights in the instrument cluster, the 928 CWS integrates all the warnings that the car generates into one central warning light, also in the instrument cluster. As well as illuminating this light if there is a major problem like oil pressure, high temperature or low fuel, it will also light up with less common things like a low washer bottle level or worn brake pads. This warning light repeatedly flashes a specific number of times, with the number indicating what the problem is. Low-importance warnings can be cleared by pressing a button, but high-importance warnings can’t. The central warning light is also duplicated in the centre console – the two lights are connected in parallel. I guess that way, if the driver doesn’t notice it, the passenger might. Finally, the CWS also turns all the instrument cluster warning lights on each time you start the car, so you can verify they’re working correctly. Electrically, the CWS sits between the sensors and the dashboard warning lights. The less-expensive model of the car had a simple adaptor fitted in place of the CWS that connected the sensors directly to the dash warning lights in the traditional way. With it plugged in, there was no response from any of the instrument cluster warning lights or the light in the centre console. I found that the light in the centre console was shorted out with a piece of wire deliberately soldered across it – I guess at some stage in the past, the CWS has failed, and this was the easiest way to eliminate the warnings! The indicator panel cleaned up with the short removed, but I wasn’t surprised when it didn’t change anything – with a short across the lamp, the driver was probably toast. The next step was to open the CWS unit. I’ve found the Porsche 928 modules to be quite serviceable in the past, and this one was no exception. After I bent a few tabs, the internals slid out of the aluminium housing to reveal a sandwiched pair of PCBs with plenty of parts – all through-hole, and all discretes except for one 14-pin DIP IC. The IC was an SN8400 with a TI logo, but I couldn’t find any data on it. I decided it’s nothing like an embedded microprocessor – this car was designed in the mid-1970s, and they didn’t exist then. There were plenty of transistors onboard, of seven or eight different types, with all but two being in TO-92 plastic packages. Some I could find data on, but most not. When I tipped the housing over, a blackened 1W resistor fell out. It only took a minute to figure out where on the board the resistor came from, and tracing showed it drove the two central warning lights – no doubt it overheated and melted the solder because of the external short. The driver transistor connected to it also had a crack in it. I didn’t recognise the number on the transistor and couldn’t find any data on it, but I measured a few others with the same marking and decided a BC547 would probably do. Having fixed those two items, I now had the central warning lights stuck on, but no instrument cluster warning lights. Next, I looked at the electros. There were some standard-­ looking aluminium electros which all looked OK, but there were also some plastic-cased 22μF units that had brown residue around the top. Brown residue is never good on electros. Funnily enough, it wasn’t anywhere near any 106 Silicon Chip seam in the case, but all the plastic-cased electros had it to some degree. I tried measuring them in-circuit with my Fluke, but I haven’t really used the capacitance measurement facility on the meter before and was dubious about it, especially since I was measuring in-circuit. I wasn’t surprised when I measured something other than the marked value. I decided to remove one and measure it out of circuit and it measured low – about 8μF. Not really being sure how they were used, I thought I might as well replace them all. I now had a flashing central warning light, but the flashing was inconsistent, bordering on gibberish. I thought it might be because multiple valid warnings were present (for example, the fuel tank and washer fluid were both empty). However, posts on a Porsche forum said that it managed this OK when it was working correctly. The next step was to check all the transistor junctions using the Fluke’s diode tester. I felt more confident about this than I did with the capacitors – the voltage across a PN junction should never exceed about 0.7V at low currents. Plus, there were enough of each type of transistor on the boards, so I could guess what was normal and what wasn’t. Some types gave high base-emitter readings of about 1.3V. I decided these were Darlingtons with two diode drops between the base and emitter. I found two transistors I didn’t like. One I replaced with a BC547, but the other was one of these Darlingtons. I spent a bit of time looking around to try to find something to replace it with, but TO-92 Darlingtons aren’t that common even when there isn’t a silicon shortage, so in the end, I replaced that device with a pair of BC547s connected as a Darlington. After that, I had a consistent flash indicating I had no petrol, and the instrument cluster lights started working. Afterwards, I measured the aluminium electros I’d replaced on the bench and saw a range of values, some very low, under 1μF. Interestingly, when I measured the new ones in-circuit, they all measured about 22μF – chalk another one up for the Fluke. Editor’s note: the SN8400 is likely a quad NAND gate like the SN7400 but with a different (possibly wider) operating SC temperature range to better suit vehicular use. Australia's electronics magazine siliconchip.com.au