Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Gaining a superpower, at least temporarily Dave Thompson I’ve always wanted to be able to see in the dark, but sadly, that is not among my superpowers (mainly, I’m just good at repairing stuff). But when the opportunity presented itself to try a ‘toy’ that could give me that power, if only briefly, I jumped at the chance. Every now and then, a job comes into the workshop that I find very interesting. Much of my work is boring computer stuff that any current 12-year-old can do, and is barely worth mentioning. But there is a wealth of projects out there built by keen hobbyists that sometimes don’t go to plan, and sometimes they need help getting them going. Any newly-built electronic device, powered up for the first time, might not work. At least not correctly. In the worst case, the magic smoke escapes in a catastrophic failure. I’ve had plenty of all of these scenarios in my time, but I have learned not to be so reckless when powering up newly built devices! I remember all too clearly that eagerness to solder everything in, wire it all up and just throw the switch (while throwing caution to the wind) without first checking thoroughly whether I have made mistakes. In that moment of excitement, the thought doesn’t even occur! Errors are not always lethal in hobby electronics, but caution is still more prudent than impatience. A blast from the past A while ago, a local guy brought in a device that I 92 Silicon Chip recognised immediately because I’d wanted to make one since I first saw plans and kits advertised in those small ads typical of late ‘70s to early ‘80s American electronics magazines. It was a ‘see in the dark’ “scope”, and back then, I thought it was merely a joke, like those X-Ray glasses you could buy for a buck. We all know they were a con – disappointingly, you couldn’t see the bones in your hand or see-through clothes like the ads promised. That was until, in the mid-1980s, I sent away for a book titled “Build Your Own Space-Age Projects” by a chap named Robert Ianini. Buying anything from overseas was a real mission in those days, before the internet existed or was widely accessible. I had to write to the company in America and enclose a money order, sourced from the post office, for an equivalent number of US dollars. Hopefully, after about six months, I’d receive the book. It did eventually arrive, and that ‘see in the dark’ project was one of the devices featured (along with such projects as anti-gravity machines and various home-built high-powered lasers and electron “ray” guns). It was then that I realised it was a legitimate electronics project that could be built by the home hobbyist. I’ve always been fascinated with night-vision stuff, and here was something I could potentially build myself. Although it was adequate for basic experimentation back in the ‘80s, it was nowhere near as good as commercial equivalents available at the time. Night-vision hardware – or, more correctly, the image intensifier tube inside the device – is typically classified in ‘generations’. This starts at Generation 0 and goes up to 3rd generation for modern starlight-amplified devices – at least for civilian use. These days, it is almost impossible to import any of today’s Gen3 night-vision devices from the USA or the UK without an export license from those countries, which of course isn’t easy (or cheap) to acquire. This DIY device would likely be somewhere between Generation zero and one, in that it requires an external illumination source to see anything at all. In other words, it doesn’t amplify available ambient light as the later generation devices do. The biggest problem for me back then, and what eventually stopped me from building one, was the requirement for a very specific type of vacuum tube called an “image tube” or “image converter tube”. This tube has a mirrored 28mm ‘lens’ at the front and a small, green, phosphor-coated Australia's electronics magazine siliconchip.com.au cathode-ray-tube style ‘viewer’ at the rear, about 15mm in diameter. These tubes were not available anywhere in New Zealand at the time (or even Australia as it turned out; I looked for one on a couple of my early visits there). So this project was dead in the water from the beginning. Even the supplier in the USA – the guy who wrote the book (still) runs a company there providing kits and plans – couldn’t supply the tube, so it had to be sourced separately. Oh well, just another dead idea among many others! And this is the way it remained until a few years ago. I was browsing an overseas auction site for valves for guitar amplifiers when, suggested to me at the bottom of the listings, was one of these image tubes, a Capehart Farnsworth 6302 Image Converter tube. I remembered, from all those years ago, that it was a direct substitute for the original IR16 type tube specified in the plans. This one was ‘new old stock’ (NOS), still in the box for only US$80, including shipping. As this triggered my memories, I thought I might just revisit this project after 30 years. So I snapped it up and hoped it didn’t get broken in transit. It arrived safely, and I put it with my other tubes in a drawer. There it sat, unused. I never did get around to building a ‘see in the dark device’ because, well, just because. to check it, but even with fresh batteries, the output (invisible to the naked eye) was pretty weak. I had an idea to fix Enter the customer this, but I would talk to the client about it later. Imagine my great surprise when a customer brought one I also put a fresh 9V battery into the handle of the ‘scope’ of these exact units in to see if I could fix it! He’d been and pressed the button, but there was no life from the tube given it by an uncle or some-such who had built it way at all. That didn’t bode well. The problem could be caused back when and he knew very little about it, except that it by the tube or any part of the power supply or oscillator used to work, but it didn’t anymore. boards. It looked almost identical to the project from all those The circuit is pretty straightforward; the power input years ago, so I was keen to get stuck in and see what was takes 6-12V DC (9V rechargeable battery preferred). A reawhat. sonably standard single-transistor, free-running LC oscilThe customer – as is typical – didn’t want to spend a lator drives the primary of a custom transformer. fortune on it, so I said I’d assess it and see what I could do. The secondary connects to a 12-stage full-wave voltage I started with the illuminator. This was a crudely-­ multiplier (in this case, a classic Cockroft-Walton arrangeconverted torch, with the reflector chopped up to accom- ment of diodes and capacitors) which supplies high-­voltage modate a small array of infrared LEDs. I used my camcorder DC ranging from 12-20kV (typically 15kV depending on the battery state) to drive the tube. There is also a tap from early on in the voltage multiplier that provides about 1/6th of the overall potential to connect to the tube’s focus ring, allowing builders to adjust focus within the tube. This is typically done once the rest of the circuit is operating. Various taps can be taken from different junctions on the multiplier and tried until the sharpest image displays on the viewing end. Items Covered This Month • • • • • Gaining a brief superpower Fixing a ducted gas heater Tektronix 556 oscilloscope repair The revolving door of PVR repairs Fixing an aircon with a faulty switch Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz Cartoonist – Louis Decrevel Website: loueee.com siliconchip.com.au Australia's electronics magazine April 2022  93 By itself, the tube image is inverted from one end to the other, so a sliding lens arrangement is used at the input end to correct this, allowing manual optical focus and also making things a bit larger. An eyepiece mounted at the back end of the PVC tube body is added mainly to protect the rear of the tube; it is basically a plain, unmagnified ‘lens’. Acquiring the right lens and eyepiece was another giant hurdle in the project back then, but obviously, the guy who made this one had purchased the short-form kit and sourced an image tube from somewhere else. Judging by the condition and colour of the PVC pipe used to build the device, it was likely quite old. Pulling it apart was easy enough – at least he hadn’t glued it all together. Pulling out the power supply from the handle and the tube-driver board from the main body of the viewer was also straightforward. The build quality was average, with some relatively sloppy perfboard point-to-point soldering employed. When boosting a voltage to this level, it is imperative that nice round solder joints are used on the multiplier, or at least normal joints insulated with enamel paint or corona dope, because arcs can form at the solder junctions if they are sharp and exposed. It all looked a bit rough and ready, but it obviously had worked at some point, so all I had to do was figure out why it wasn’t going now. Battery power was certainly getting to the board but stopped at the transistor, a classic TO-220 style MJE3055. From memory, this should have a heatsink, but it had none. I had several similar transistors in my parts bins, so I pulled this one and replaced it with a known-good one. This time, when I fired it up very carefully on the bench, I could hear the familiar faint HV crackle from the multiplier, which could indicate that something might be breaking down somewhere. The tube remained dark. It was good to know that at least the oscillator was working. The proprietary transformer used was also likely not open-circuit, but working around these Cockcroft-Walton circuits always makes me very nervous. I’ve experimented with them before many times, in the likes of air ionisers and various electrostatic experiments. It’s a case of once bitten, a hundred times shy! This one ‘only’ puts out in the region of 200µA at the nominal 15kV, but that’s enough to make someone jump and yell! flying leads, so a standard valve tester wouldn’t be of any help. I could find nothing about testing them online, so it was just a matter of swapping it out and hoping for the best. I temporarily put my tube alongside the unit. The IR16 has pre-connected wiring while the 6032 doesn’t, and that meant soldering directly to the metal body and rings around the tube itself. That also made me nervous, and the other thing was that the book plans (which I’d since dug out of storage) didn’t show which wires went where with this particular tube. There was nothing to do but try it, so I guessed where they were supposed to go by the physical layout of the tube itself. I mean, there are only three connections: one at the front, one as a ‘ground’ on the main metal body of the tube and one for the focus, which I assumed was the middle ring. What could possibly go wrong? As it turned out, nothing. I wired it how I thought, was über-careful soldering to the tube’s metal parts and used one of my trusty bench power supplies to power it all up. I started with severe current limiting, just in case, but gradually increased it until things started happening. It all looked OK, and the tube started glowing at the rear end. I could vaguely see an image, but it was very faint in bright light. I’d need to mount it and adjust the lenses and the focus voltage to really test it properly, not to mention using it with a decent IR source. I called the client and told him what the costs would be, and as he was OK with it, I persevered with the rest of the job. First, I resoldered all the multiplier’s connections and any others that looked dodgy, then mounted a heatsink on the transistor. I didn’t have a lot of room, but a small Jaycar heatsink I had (HH8514) fitted in the case and should suffice. As I might need to have the tube in and out to set the focus voltage, I temporarily mounted it in the PVC body and held it in place with sponge wedges. With the workshop lights out, and the illuminator on, I could make out some outlines, but the focus was off. So out it all came, and I used another tap to test it. There is a provision in the plans for adding a resistive voltage divider network to further fine-tune it, but as it turned out, the image was pretty sharp with the next tap along, so I left it at that for now. The converted torch illuminator was very crudely made, Taking the tube I don’t have the gear to measure that kind of voltage, but it appeared to be working, so that left the tube itself. This one was the IR16 version of the tube, which was different in connections and size to my 6032 type, but if necessary, I could make it work – as long as my tube was functional. I wouldn’t know unless I tried it, so I set about removing the IR16 from the body tube. It was held in with three long screws, 120° apart, threaded into the PVC body. These pressed lightly on the tube and centred it. They had been coated in what looked like RTV or some other silicone sealant to stick it all together. Space was tight, but with patience, perseverance and a sharp hobby knife, I managed to get it all out without breaking anything. From the outside, there’s no way to tell if the tube is working or not. It has no pins like typical tubes, just three 94 Silicon Chip Australia's electronics magazine siliconchip.com.au and while it would work, I had a better solution. When dad was alive, he experimented a lot with then-quite-new LED torches, and he had several that were rechargeable and about the size of a three-cell Maglite torch. I inherited several working models and a few he had used for parts. The torches use an array of bright white LEDs mounted in a specially-moulded reflector and were very bright; all I’d have to do is swap the white LEDs for infrared versions, and I’d have a very powerful, self-­contained IR illuminator. I’d already factored in the cost of this to the client, and while I was happy to give him the torch, I did need to buy 25 IR LEDs for the job. I disassembled the torch and, using my trusty Goot desoldering pump(s) and lashings of solder wick, managed to extract the old LEDs without damaging the PCB they were all mounted on. It was then simply a matter of installing the IR LEDs and putting it all back together. Turning it on resulted in absolutely nothing because it is invisible. But my camcorder showed a powerful beam. That night, I fired up the whole thing and scanned our backyard. The output from the tube was patchy in darker areas, but everything was visible. I was pretty impressed and spent quite a while playing with it. Satisfied that it was operational, I used RTV to bog in the tube and buttoned it all up properly. So, after all these years, I finally got to play with one and didn’t mind losing my tube to a working model. Sometimes I love this job. Fixing a ducted gas heater which had a faulty ignition M. H., of Albury, NSW had a whole range of electronic appliances fail in a short time. Is he cursed? Probably not, considering that he managed to fix them all with just a few dollars’ worth of parts and some hard work... My pool chlorinator cell wore out. My attempts to repair it worked for a short time, but the plates were corroded away after seven years of hard work. A new one was the only option, and $650 later, it was back in service. At the same time, some small ants had entered the chlorinator supply box and destroyed the SMPS driver IC. After an eBay purchase and a few weeks delay, the supply was back in service. Then tree leaves got past the filters, entered the impeller and jammed the motor, and the pool started going green again. The motor is designed to be easily split to remove the obstruction, and the motor ran again without relying on the thermal cutout device to protect itself. If I had called the pool company to fix all these problems, I would have probably spent $1000 more than I did, given all the service call fees and the fact that they would likely replace all the parts rather than fix them. I realise that they have a lot of overheads, and quoting for a new part is the best option for them. In part, that’s because it moves the warranty for the repair restoration to the manufacture for 12 months (or more) and moves liability away from the serviceman. Next? Now my ducted gas heater would not start. With an ear pressed to the outer case of the in-ceiling heater, I could hear the combustion chamber fan start and run. After a short delay, the fan stopped and the unit smelled of gas with no ignition. The cycle repeated endlessly. siliconchip.com.au Australia's electronics magazine April 2022  95 I removed the power for five minutes and tried again. Success, the combustion chamber fan ran, the gas relay operated and ignition. But the ignition lacked the aggressive volley of sparks sound that it usually had. The Honeywell ignition box was sad; most likely, it uses capacitor discharge via an SCR. The unit was easy to remove, easy to open, but impossible to repair. The manufacturer had covered the EHT section with epoxy resin. A lot of heat, wiggling and cutting eventually got the single-sided PCB out of the case to reveal a dry joint on the discharge capacitor (1μF 250V polyester). The remainder of the circuit design looked (to me) straightforward and expected, with a thin, cheap singlesided PCB manufactured by solder reflow. The capacitor measured close to 1μF, but I replaced it anyway with a 1μF 2kV polyester capacitor pinched out of a plasma TV. The unit produced the clearly audible volley of aggressive sparks while the flame established itself. Success! A professional serviceman (in my opinion) would not be inclined to diagnose the fault. After a quick assessment of the age of the unit, they would give the expected “I will get a quote for a new one” and “that model is not made anymore”. The replacement ignition system would come in about $500 plus the hourly service rate and call out fee. Again, I was greeted with a small pop and two blown 30W resistors. This had me quite confused, as all components for that rail tested good, and no other faults were obvious. These 30W resistors are no slouches, being 5W wirewound types, making me think there must be a catastrophic short somewhere. I then realised that testing the transistors out-of-circuit was a mistake; when installed back onto the heatsink, there was a short from collector to ground. On one of the little boots that insulate the transistor screw from the heatsink and allow a connection to the collector, there was a burned carbonised track from a previous arc. This was impossible to see as the boot is black. All that separated the 225V rail and ground was 1mm of burned plastic. Upon replacing this and the two 30W resistors again, the scope powered up as it should. I left it for half an hour before trimming the -150V, 100V and 225V rails. I’ll need to recalibrate the timebase as I fiddled with the voltage rails, but that can wait for another day. I can only assume that a build-up of dust and condensation from recent cold days caused the insulator to arc over. Evidence of greasy, dusty grime was present. So before reassembling the scope, I gave it a thorough cleaning throughout in the hope that this never happens again. Fixing a fried Tektronix 556 oscilloscope The revolving door of PVR repairs D. V., of Hervey Bay, Qld got a shock when one of his prized possessions had a minor explosion when he powered it on. The cause appears to be age-related, but perhaps not in the way you might think... I have a collection of old Tektronix oscilloscopes; the latest acquisition was a mint-condition 556. Even though I had switched it on several times before, on this occasion, I was greeted with a loud bang followed by what could only be described as the sound an egg makes when frying on the barbie. Reaching to switch it off felt like an eternity, but in reality, only a few seconds passed. However, the damage had been done. On inspecting the underside, I found two 30W resistors had burned out. These are part of the +225V circuit, and the fact they were damaged at all surprised me, as the 225V rail is individually fused and the fuse was intact. This led me to believe the fault was within the power supply unit (PSU). Scopes like my 556, while discontinued mid-1970s, are marvels of engineering. It is a true dual-beam scope and boasts tunnel diode triggering, dual plugins, individual timebases and 50MHz bandwidth. It is a monster weighing 40kg, with 34 valves and sinks 840W when in operation. No wonder it was the last of the 500 series scopes to be made! The regulator circuits in the 556 are semiconductor-based whereas the previous 500-series scopes used valves. Transistors T03 and T02 in the PSU are mounted on a heatsink directly behind the fan assembly because Tektronix had difficulty keeping these components cool. T03, the main pass transistor for the 225V rail, had gone short-circuit. It was a 2N4348, so I substituted a 2N5672 from the junk box. I removed all the other transistors from the defective rail and they tested OK. I replaced the two 30W resistors and wondered if this will be the magic bullet, but I had doubts. So with the 225V rail fuse removed for posterity, I proceeded to switch the unit on for just a second to see what would happen. 96 Silicon Chip B. P., of Dundathu, Qld has had to fix the same devices multiple times due to similar faults. It seems that they were made with poor quality components... We use two Beyonwiz DP-P2 Personal Video Recorders (PVRs) to record and play back TV programs. I originally bought both of these units on eBay as “not working, for parts”, both with an ERROR 0000 fault. In both cases, the cause of the faults were bad electrolytic capacitors in the power supply. I fixed both these units when I got them a few years ago by replacing the bad capacitors, and both worked well for some time, although I had the same fault return in one unit when another capacitor failed. Recently, my son told me that the PVR in the camper Australia's electronics magazine siliconchip.com.au was playing up. Sometimes it would work correctly; other times, it would show the ERROR 0000 and yet other times, it would be on when it should be off. Usually when it’s off, it shows the time on the front panel, and everything else is on standby. But sometimes, it would be off with the hard drive still running. Removing the lid, I could see a bad 3300μF 10V capacitor. I looked through my salvaged capacitors, found a suitable replacement and fitted it. While looking over the circuit board, I spotted a small capacitor that looked suspicious. It was a 330μF 25V capacitor, so I removed it. Then I noticed another one of these capacitors that looked suspicious, and this kept happening until, in the end, I had removed at least six of the same value electrolytic capacitors. I later tested these with my ESR meter, and all read well above what they should have. I found replacement capacitors in my salvaged capacitors collection, installed them, and then put the power supply board back in the PVR. After buttoning it up again, it was working well. Not even a week later, I turned on the other PVR in the lounge room, and it showed ERROR 0000. This PVR had been working well since its original repair, apart from Channel 7 being corrupted during the day, although it was usually mostly OK at night. There was also occasional corruption on SBS. None of the other channels had this problem. After removing the lid, I could see a really badly bulged 3300μF 10V electrolytic capacitor. Not only had the top bulged, but the seal on the bottom had been pushed out, and the capacitor was sitting at a significant angle. This was obviously the cause of the ERROR 0000 fault. I removed the defective capacitor and I found a replacement Nichicon capacitor in my salvaged capacitors. I scanned the PCB, but I could not see any other problems. All the rest of the capacitors, including the small ones I’d replaced in the other PVR, were fine. A quick test again showed the unit to be working. Since the repair, the corruption on Channel 7 seems to have disappeared. It’s really handy being able to make these repairs; otherwise, taking the unit(s) to get repaired could easily run into hundreds of dollars, and purchasing replacements would be similarly expensive. Aircon repair reveals a faulty switch R. W., of Hadspen, Tas offered to fix his friend’s air conditioner (which was said to be unrepairable) and traced the fault came back to poor installation practices... Many years ago, I was asked by a friend whether I knew anything about air conditioners. He had 5kW and 2.4kW units from a reputable manufacturer installed in an innercity apartment in Brisbane, and the 2.4kW unit stopped working after a year or so. An air-conditioning tech looked at the unit and told him it needed replacement, as the boards and refrigerant were no longer available. A quick check on the internet revealed that was not the case; while there would be benefits in replacing the unit with an inverter system, the cost seemed unwarranted on such a relatively new unit. I thought it was worth a look. During the installation, the electrician had routed the single-­circuit power cable through the downstairs ceiling and had cut several holes in the plasterboard, which were siliconchip.com.au 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. later covered by four blank switch-plates and a snap-in ventilator. The smaller unit had a square section of adhesive conduit emanating from a bedroom power point, then passing through the external wall to the outside isolator switch. My friend was not happy with the blank plates, and had a plasterer make good the ceiling. He put up with the conduit as it was largely hidden by a bedside table. This suboptimal installation should have given me a clue. The 2.4kW unit responded to the remote commands, and when the internal fan started the louvres were able to be adjusted. About three minutes after powering it on, I could hear a relay click, but no cold nor warm air emanated. The problem seemed to be in the outdoor unit. This air conditioner had all of the control electronics on a board in the indoor unit with two switched Active wires going to the outdoor unit, one for the compressor and outdoor fan and one for the reverse-cycle solenoid – all fairly simple. I monitored the outdoor unit, and after the compressor timer had run, the outdoor fan ‘kicked’ but that was all. It was time to look at the circuit board. After removing some connectors and prising some clips off, the board was easy to remove. I looked for the usual suspects like dry joints, bulging capacitors and burnt components but found none. I decided to connect the infrared sensor and bench test it with mains applied, taking the usual safety precautions. The board behaved faultlessly. The relay clicked in, and power was available to the outdoor unit terminals. I connected a fan heater to these terminals in case the relay contacts had failed, but it sustained a 10A load. It had to be something in the outdoor unit. I was thinking possibly a failed compressor or motor run capacitor, but this did not explain why the outdoor fan would not run. I reinstalled the board and put it through its paces again while up on the ladder. This time, I noticed something that I should have realised earlier. The relay clicked in then dropped out, and the indoor fan lost speed when it clicked in. I connected a voltmeter to the input mains and noticed it drop to less than 100V AC when the relay energised. This drop was not apparent at the power point. I flipped the power circuit breaker and went to recheck the outdoor unit. It was then that I noticed a rust stain down the wall behind the isolating switch. I took the switch cover off, and rusty water poured out. The switch was not sealed against the wall; water had entered, rusted the mounting screws and caused a high impedance path within the switch. My friend engaged a better electrician to replace the switch, and the unit is still running some ten years later. SC Australia's electronics magazine April 2022  97