Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Watch out – delicate repair in progress Dave Thompson It’s tempting for a serviceman to jump into anything that needs repairing, especially if we feel confident about ourselves. It’s one thing to repair the dishwasher or install a new cistern in the bathroom, but a different kettle of fish to rewire a switchboard or install a new gas line! I’ve previously mentioned that while sitting in my workshop a few years ago, I felt the ground shake with an accompanying “BOOM!” sound. Those of you who know where I live will realise that I’m used to the ground shaking – as of today, we’ve had around 24,000 quakes since September 2010, so we’re all pretty good at gauging how strong an earthquake might be from the sensation now. Generally, anything over magnitude five will cause mild concern, but anything under that is just annoying. Anything over six and I worry the house will fall down, but it has survived a 6.4 and a 7.1 so far. It’s only the ‘big one’ we’ve been expecting for the last 50 years that worries me. This shake, however, was different. It was very short and sharp, and the boom was unusual. Many quakes ‘roar’ but they don’t typically make a sound like this one. It turned out that a house about five kilometres away had literally exploded, which is what I’d heard and felt. There was a lot of speculation about what had happened at the time. It transpired that some maintenance had been done on the reticulated gas system (in one of the few suburbs in this city with built-in gas), and a leak had caused 90 Silicon Chip a gas buildup overnight. When someone lit a flame in the house, kaboom! It really made a mess, and of the neighbours’ houses too. Fortunately, nobody was seriously hurt. Another home that went boom! In a similar incident, a leak caused a gas explosion in my wife’s sister’s home in Croatia. They use bottled gas, and a line to the cooker had somehow worked loose. When she went to light the cooktop, it blew their doors off and the windows out. It left her hearing impaired but otherwise OK; the same couldn’t be said for their apartment. The place was rebuilt by the time we stayed there, and they now use an electric stove. Both these stories are cautionary tales about getting people who know what they are doing to carry out maintenance and repairs on systems within our homes. This principle can be applied to anything. The home mechanic working on the brakes on their car, the avid DIYer installing their own solar panels. Under normal circumstances, that is all fine because those things are relatively straightforward. The brakes will likely work correctly, the solar panels will soak the sun and all will be well. However, in some cases, such as those outlined above, it pays to get the professionals in to do the job. It’s an important skill to know when you’re in over your head and you need to call the experts! It’s better to have your car towed to a mechanic to fix your mess than to realise you forgot to reconnect the brake line when the pedal goes to the floor... Over the years, I have come to accept this. While it hasn’t always been an easy decision to make – in many cases, I tried to accomplish something before realising it was time to call in an expert – there are still some things I will try to do myself. In doing so, I hope I’m not making things tougher for the person who comes after me to pick up the pieces, but of course, that isn’t always the case. I’m all for suggesting people give things a go, but the problem is that doing so can hurt our chances of success on the other end. A classic example of this is data recovery. Many attempt to follow walkthroughs on the internet, only to make things worse. When they finally bring their computer to me to ‘fix’, they’ve damaged their data by their attempts. With that in mind, I’m very careful when doing anything a bit ‘out of my wheelhouse’ so as not to cause further problems. Australia's electronics magazine siliconchip.com.au Items Covered This Month • • The delicate act of repairing yourself Fixing the vacuum pump in an electron microscope • ATA automatic gate repair • Reviving an electric motor 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 The pitfalls of wearing a watch Recently, my watch started playing up. I’ve worn a watch all my life and have gone through a fair few over the years. I’m not particularly hard on them, but as an engineer, there are times when they take a beating. I also had many jobs where wearing a watch was not allowed. For example, when working in the battery section of the airline, it was strictly forbidden to have anything remotely metallic anywhere near the batteries. Back then, there were two different types of batteries: lead-acid and NiCad. They do not play nicely together, so there were two completely separate (but adjacent) rooms for maintaining them without cross-contamination. The NiCad batteries especially were quite dangerous because each 24V battery comprised 20 individual high-­ capacity cells. These cells are connected by heavy metal links in a set order, and once the battery cover is removed, this presents a very real danger should anything metal drop into them. The wall of shame in the battery shop boasted several blobbed shifters (“crescents” here) and half-screwdrivers that someone had let loose onto a battery. As these batteries are capable of delivering a huge amount of instant current, anything metallic going into them was spectacular! Getting a watch or band across any of the links could mean losing a hand, so jewellery was forbidden. The old salt who ran the place would slyly ask for the time, and if I’d forgotten to take my watch off, I’d be dressed down a peg or two! The problem with taking a watch on and off all the time is that it wears everything out. The pins, the clasp and the strap all fail eventually from wear and tear. I went through many watches for this reason, and probably also because I was banging them against airframes and workbenches. I eventually bit the bullet and decided to buy a proper watch, a Tag Heuer Professional. It wasn’t cheap, but it was rugged, water resistant to way deeper than I’d ever swim. It also had a 1mm-thick sapphire crystal on it, which means it should be impervious to scratches and abrasion, something all the cheaper watches had succumbed to as I scraped and smacked them during my career. Long story short, I still wear this watch today after 30+ years. It looks as good as the day I bought it and has used a total of six batteries. I had it serviced every time the battery has been replaced. Usually, I’d take it to the place I purchased it from – which has since had to move location siliconchip.com.au because their original store was trashed in the quakes – and the same guy would look after it, as he has done for the last 25 years. This time, when I went into their new store, I learned that my guy had retired, and his son had taken over the business. I was assured everything would be the same: the same fine service, the same warranty and the same level of craftsmanship, yadda yadda... I was quoted a price for the service that was in line with what I’d paid over the years, allowing for the usual price increases. However, when I went to pick it up, the cost had ballooned. When I queried this, I was told that the bezel spring had worn out and needed replacing (this is a ‘dive’ type watch with a ratcheting bezel holding the crystal on). This cost an extra 80 bucks, and if memory served, had also been done before on a previous service. Fair enough; I trusted them to do what was best. They also replaced all the seals and O-rings, and pressure tested the watch (how? I don’t know) to ensure it really was sealed. This was important because if I go surfing or swimming, I don’t take my watch off, and I like to know it isn’t going to fill up with water. No time for my watch to die I got it home, and two weeks later, I woke up to it showing the incorrect time. I usually set it to an atomic clock app I have, and it is always within a second or two after three months, so I know it is an accurate timepiece. That morning, it was reading some two-and-a-half hours slow. This was the first time in 30 years that the watch had been wrong. I also noticed that the third hand, ticking away the seconds, no longer lined up with the markings on the watch face. When I first got this watch, I marvelled at how amazingly precisely the hand hit each second marker perfectly. I concluded that the people who’d serviced it, and who’d had the crystal off, had altered something, by accident or otherwise, and now I was seeing the results. Australia's electronics magazine October 2023  91 they were and what a useless klutz I must be – the usual factory-­floor hazing. The other guys there smirked knowingly because they’d been through it, too. My next exercise was to use a microscope and tweezers to re-bend this coil spring into a usable shape, or the airline would go broke because of my ineptitude! I spent the next hour sweating and getting the spring back into a proper shape, which is evidently impossible for anybody with brains. To my credit, I almost got there, and earned the foreman’s grudging respect. Later, he told me that most apprentices gave up after 10 minutes, but I’m stubborn like that! With that in mind, I had no doubt I could have this watch whipped into shape ‘tout suite’. Watch this... I reset the time and resolved to keep an eye on it over the next few weeks to see what would happen. The time didn’t change again, and it seemed accurate, but the third hand not hitting the marks really bothered me, so I did what anyone else would do and went back to the service agent. They looked at it, hummed and hahhed about it, and grudgingly agreed to check it out. I left it with them for another few days, after which they called to say it was ready. When I picked it up, they said they’d found nothing wrong with it and that the third hand issue was likely ‘wear and tear’ on the watch, as it was getting on a bit. I commented that it had been fine when I first took it there, but now it wasn’t. Again, the ‘old watch’ excuse was trotted out. I doubted they had done anything or even had it apart. I took it anyway and went on my merry way. A few months after all this, I was getting more annoyed with the hand not lining up. I don’t think I suffer from OCD, yet this was really bothering me. Timekeeping seemed fine, but I thought, how hard can it be to open this up and have a look? (Famous last words...) Fortunately, when I was going through my ‘buy everything I see from AliExpress’ phase, I bought one of those small watch vices and a kit of various watchmaker’s tools. No, I don’t know why either, other than to have them. So, I broke them out, blew the dust off them and set about seeing what I could do with this watch. Back in my apprenticeship days, I spent six months in the instruments workshop at the airline. This was in the days before avionics cockpit panels were ‘glass’, so plenty of analog instruments needed repairing, maintenance or calibration. As a rite of passage, on my first day there, I was given a gauge to ‘repair’ that I had to remove the bezel from. It was almost impossible to remove without distorting a coil spring sitting right behind it. This is, of course, a consumable part and must be replaced anyway. Still, as a n00b, I had no idea. And when I bent it, the foreman made a song and dance about how expensive 92 Silicon Chip Still, I had to be careful! Having skills 40 years ago doesn’t necessarily mean I have skills now. I used the tools I had to remove the back, then searched the web for how to remove the bezel, which required a bit of salt and pepper to pop the spring and detent. I’m always wary of just ripping into things like this, but that’s what it took in this case. As it turned out, I didn’t need to take the back off, but it did give me a chance to work with the watchmaker’s tools I’d bought, and they worked fine. With the bezel and crystal off, I could now gain access to the watch hands. The main hands were obviously OK because they worked, but that third hand still irked me. I asked myself: why would they take that off, anyway? Did they knock or bump it by accident? Perhaps it really was just worn out, as they had claimed. Still, I had a tiny hands puller (which is like a bearing puller, only much smaller), so I stopped the watch first by pulling the adjuster knob two clicks out, then took the third hand off, noting where it was sitting and being extremely careful not to touch the others. The hand itself is so tiny and thin that I was worried about wrecking it – it certainly wouldn’t take much to do that. Fortunately, my hands were still capable of some finesse, and I did all this while using my headset magnifier and a decent LED bench light; without those tools, I wouldn’t be able to see a thing! With the third hand now off, I could see it was a simple interference (friction) fit onto the shaft. There were no splines or flat sections for locating it, so it seemed a simple task to line it up properly and press it back into place, which is precisely what I did. When I’d stopped the watch, it was almost to the 18-­minute mark on the face, so I lined it up exactly with that, pushed the hand carefully home and restarted the watch. This time, the hand aligned perfectly, and I watched it go around a couple of times and saw the other hands responding at the correct times, so it must’ve been in the right place. I replaced the bezel and spring and ensured the crystal was clean before putting it back on. I didn’t want to be taking this section apart again. I also ensured the battery was installed correctly and seated – I didn’t want it losing time. I put the seals back into position and reinstalled the backplate. Having the right tools certainly makes this task much easier than trying to use a pair of pliers to grab hold of the indented areas on the back of the watch, all while not being able to hold it all steady. Australia's electronics magazine siliconchip.com.au It has been fine for months now, so hopefully, that’s the last time I’ll have to take it apart! Electron microscope vacuum pump repair M. C., of Leonards Hill, Vic runs a repair business that specialises in keeping unsupported and otherwise obsolete high-value equipment up and running (website at: www. technicalmayhem.com.au). Clients so far have mainly been universities, but he is hoping to expand into other fields. Here is the story of one repair undertaken... One Tuesday morning, I received a call from a major Melbourne university. One of their 1990s-era JEOL electron microscopes had developed a startup error after it had been left switched off over a long weekend. It was complaining that one of the vacuum pumps wasn’t starting. These microscopes are complicated beasts that take up a small room and require chilled water, several bottled gases and an extremely low vacuum inside the main unit. The vacuum system in this particular unit comprises five different pumps to achieve a high vacuum to avoid contamination of the sample or electron gun. In this case, the fault was reported to be in the second pump, a turbomolecular pump that looks similar to a truck turbocharger. Once I arrived, I confirmed that was the problem – the fault light on the rackmount pump controller was glowing red, and the user interface listed the fault in the startup sequence. Seiko Seiki in Japan manufactured the pump in question. Makers of complicated equipment like electron microscopes often use equipment from other manufacturers to avoid the huge expense of designing it themselves. These pumps run on a magnetic levitation bearing to achieve the super-low friction required to spin at up to 90,000RPM, undoubtedly a significant design challenge. The pump controller manual revealed that the fault light could be triggered by three different faults to do with the pump itself and one in the controller, unfortunately omitting any detail about how to narrow it down. The pump faults were the usual overspeed, underspeed, overload etc; the controller fault was a flat backup battery. This seemed easy – it must be a flat backup battery! However, the battery had been replaced recently and tested 100%. The battery is required in case the controller loses power without being shut down nicely, allowing the pump to spend a leisurely 15 minutes spinning to a stop on its frictionless magnetic bearings. I explained to the client the difficulty of troubleshooting the pump and controller unit without a schematic diagram or service manual. Still, such an investigation was probably the only reasonable course; a replacement was simply unobtainable. The pump and controller are matched to each other, and if they couldn’t be repaired, the alternative was a newer model pump and some work designing an interface to the microscope. We decided that the most reasonable course of action was to spend some time trying to diagnose the existing problem further. Opening up the controller case revealed many modules and PCBs squeezed neatly into the case. The front cover of the unit folded down to reveal a card cage with many PCBs that could be unplugged. The original service techs would have had a kit with an extender card for measuring test points and making adjustments, but I would have to improvise. siliconchip.com.au Australia's electronics magazine October 2023  93 I traced the fault LED wiring back to the card cage interconnecting backplane and onto a logic board with many 4000 series CMOS chips; a 4-input NOR gate drove the LED. This made sense; each input would indicate one fault. I couldn’t get the DMM probe into the unit with the PCB in the cage, so I soldered four numbered wires onto the gate inputs and re-inserted the card. With the unit powered up again, I checked each wire until I found the one sitting at +5V, narrowing the fault further. The input with the fault travelled off the board, back into the backplane and onto another PCB with a lot of analog circuitry onboard. The fault signal traced back to a comparator that measures the input of a voltage divider. Once again, I used the trick of soldering three numbered wires onto the top of the voltage divider and the two comparator inputs. The divider input measured 13.3V, with the comparator inputs measuring 0V and 2.47V. 13.3V seemed suspiciously like a fully charged 12V battery voltage – this was the backup battery voltage monitor circuit! What was going on? I removed the PCB again and measured the divider resistors. The lower measured 10.7kW while the upper, marked 47kW, measured open-circuit. This was the problem; a humble 0.25W resistor that looked perfect! All this was very strange, but I didn’t stop to think about it for too long; I quickly fitted a replacement. That fixed the controller; the system got through its startup sequence and the pump started. As the system crept towards its operating vacuum, I did a quick calculation. The result showed that the resistor should dissipate about 2.5mW in this application, roughly 1% of its rated maximum value. Don’t ask me to explain why it failed! Regardless, the client was very happy and there have been no more faults for several months. ATA automatic gate opener repair G. C., of The Gap, Qld went through quite a few trials rejuvenating a failed swinging gate controller. His story demonstrates how helpful it would be to have circuit diagrams of your equipment to help with repairs... 94 Silicon Chip Almost four years ago, I installed an ATA swing gate opener that used a 24V DC motor linked to a gearbox which transferred power to an articulated drive arm attached to the lower edge of the gate. The DCB-05 controller was mounted in its own plastic housing together with a solar controller board. A 30W solar panel was provided to charge the 24V 12Ah battery in a separate box. There were two failures in the first few months. One was caused by the normally-closed contacts of the limiting microswitches not making, and the other by an enormous ant infestation in the battery box, resulting in significant corrosion of the terminals and connectors. The gate was left open during a long renovation, and a large bush progressively enveloped the solar panel. When the renovations were finished, the gate opener was not working. However, the battery voltage measured 22V, which surprised me. The message on the controller’s LCD indicated that the limits needed to be set. I suspected that meant the battery voltage had dropped so low that the system required re-initialisation. I thought there was Buckley’s chance of the four-year-old lead-acid battery being salvageable, so I purchased a new pair of 12V 38Ah batteries. After installing the batteries, I was gobsmacked that when I went to re-initialise the controller, the display was showing gibberish with a continually changing pattern. Only a week before, it was perfect. What had happened? Was the microprocessor sending the display rubbish? If it was, there was no way I could fix it. I powered the control board directly from a 24V transformer (it accepted 24V AC or DC). When I looked closely at the display, I saw that the pattern was scrolling from left to right. When either the NEXT or PREVIOUS buttons were pressed, the unintelligible pattern remained stationary, and there was a confirmation beep. So, it was a fair bet that the LCD was faulty. Searching online, I found that Jaycar sold a display that was a close match and appeared to have the same pin-out. It was a discontinued line, selling at only $9. Within a day or so, I had purchased one, installed it, and it worked perfectly. Originally, double-sided tape had been used to adhere the display to its driver board. Prising them apart, I found a small area about 3mm in diameter of corrosion on the circuit board. The tape was so firmly stuck to the board that it was hard to see how any water/condensation, let alone an insect, could have gotten in there. There was no sign of corrosion anywhere else. After reinstalling the controller, when I tried to set up the limit switches, I got a “Limit Switch Not Activated” message every time. I found that the CLOSE microswitch was faulty. Fortunately, I had a spare with me, but substituting it made no difference. While the gate was closing, I could operate the microswitch manually, but the gate kept moving. I checked the wiring continuity from the microswitches back to the main board but found no problems. It looked like another fault in the control board! I also noticed that the battery voltage was dropping slightly, and when I measured the current from the solar controller to the battery, it was zero – not even a microamp. Clearly, this board was also faulty. Australia's electronics magazine siliconchip.com.au A close-up photo of the solar controller section of the gate controller. I emailed the manufacturer’s technical support guru. He responded quickly and said to ring him the next time I was on-site and he would lead me through setting up the limit switches. I followed his steps, conveying voltage readings to him. Unsurprisingly, he confirmed that both boards were faulty. He thought they might have been hit by lightning, but I could not see any evidence of that. He pointed out that it was possible to dispense with the microswitches by setting the controller to switch off the motor when the current started ramping up when an obstacle was encountered. I found that the limits for the controller could be easily set up by using large potted plants to constrain the gate’s travel. Returning to the main controller board, tracing the tracks from the terminal block, I found that the microswitch signals went through a resistor network and then a surface-mount IC. I could not find any data sheet, but I assumed it was a buffer. When I simulated the operation of the CLOSE and OPEN microswitches, I could see the output of this IC responding accordingly. Its outputs were connected to the inputs of the microprocessor by short tracks and there was no sign of corrosion. I did not try to look for these signals at the microprocessor as it was too risky; the multimeter probe was bound to short pins with my clumsy fingers. It was frustrating that I could not find any fault with the microswitches, the wiring or the main board. I noticed that the display was sometimes warning that “Service is due”. Being a born optimist, I set the service counter to 60,000 operations before this message would re-appear. It is doubtful that this service required warning would interfere with the operation of the controller, but I was not sure. Regardless, I was at the end of the road with this controller. Now the solar controller... I tried to “recondition” the recovered 24V battery using a smart charger, but it was too far gone. At least it charged to 24.4V, sufficient for testing the solar regulator. On the small solar controller board was an LM2588 adjustable flyback regulator that was delivering 27.3V after the output filter. The guru told me this module was designed to charge the battery at 27.5V, so that was close to the expected value. When I measured from the board ground (same as the solar panel negative) to the battery’s positive terminal, the reading was 27.3V. However, when I measured across the battery terminals on the board, the reading was 24.4V. Where had 2.9V disappeared? I found that a TO-220 package Mosfet (IRLZ44N) was between the board ground and the battery negative terminal, which was connected directly to the drain pin that siliconchip.com.au measured 2.9V. Its source pin was connected to ground and the gate to the battery’s positive terminal via a 1MW resistor. I therefore expected to measure 27.3V at the gate, but it was 0V; no wonder it was not conducting. I wasn’t sure of the purpose of this Mosfet; my friend said it was to protect the board in case the battery polarity was reversed. I note that no such precaution was taken with the solar panel. After removing the Mosfet, it appeared to test satisfactorily. However, it seemed to be a very strange circuit as the specifications of the Mosfet give a maximum allowable gate voltage of ±16V but, in this circuit, it appeared to be hit with 27.3V. I wondered if there was a breakdown between the gate and source terminals. Replacing the Mosfet with one with slightly better specifications, I found that the voltage measurements stayed the same. Removing the device from the board. I noticed a thin track from the gate terminal that disappeared under the edge of a large surface-mounted diode. I soon discovered that this track came out under the diode and led to a 1MW resistor in parallel with a capacitor to ground. At last, it all made sense. When I got a measurement of 1MW across the resistor, I was measuring through the switching regulator to ground with one probe and through the 1MW resistor, which was connected to ground with the other probe. Clearly, the 1MW resistor to battery positive was open-circuit. Unfortunately, this tiny resistor had doomed the battery. With the original Mosfet reinstalled and the open-circuit resistor replaced, the solar regulator was back in action. After reinstalling both boards, I attempted to set the system up using the microswitches to determine travel limits. The gate CLOSE limit was set immediately, but then the gate refused to open – no drive whatsoever. Now the normally-closed microswitch contacts for the gate OPEN limit were open-circuit. This second set of microswitches had also failed prematurely when the manufacturer’s specification was for an expected life of 200,000 operations. The metal enclosure for the motor/gearbox and microswitches did not have any sealing gaskets and, when I initially opened it, I was staggered to find a fair amount of sand and dirt inside. Perhaps the unlocked cover had not been put back properly, and sand, cement dust, sawdust etc had found their way into the box during the house construction. I think these contaminants must have compromised the microswitches. I didn’t waste any more time and set the limits using current sensing. The gate opener finally worked as it should. 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. Australia's electronics magazine October 2023  95 The downside of this method was that the gate closed and opened more slowly. Still, the troublesome microswitches were not needed, and the current to the motor was throttled back before the end stop, so the gate glided into the stop position without any clunk. Reviving an electric motor B. P., of Dundathu, Qld is a prolific repairer. This time he’s tackling an electric motor that he got for a song. It was in bad shape but just needed a bit of care before it was functional again... I was setting up a piece of equipment that used to be powered by a three-phase electric motor. I don’t have a three-phase supply here, so I decided to replace it with a single-phase ¾ horsepower (~550W) electric motor that I had picked up at one of the local tip shops. When I checked it, I found it was seized, so I dismantled it. It was difficult to get apart, but I eventually succeeded. It was obvious that the motor had been flooded at some stage because the rotor had a thick coat of rust, and the stator laminations weren’t much better. One bearing was utterly seized, and the other was not turning freely either. I started by removing the rust from the rotor and stator laminations with a rotary wire brush on my electric drill. I then tried to make the bearings usable so that I could test the motor before investing in new bearings. Both bearings were double-sealed, so I prised the seal off one side of each bearing. I sprayed them with lubricating spray and eventually got both running freely, so I oiled them. The bearings were not in a good enough condition to be reused but were good enough for testing, so I reassembled the motor. The good old electric motor shown in full along with the troublesome centrifugal switch contacts. 96 Silicon Chip I plugged the motor in and it tripped the safety switch after a quick flick of the shaft. I was not entirely surprised, as I’d previously worked on an irrigation pump with a leaking seal, which had caused the winding insulation in the stator to deteriorate and cause an Earth fault. I was about to scrap the motor when I decided to test the windings. There were four wires connected to a terminal block, so I removed them all, then got out my multimeter and turned it to the 20MW range. While this was not a Megger, it would at least give me an idea of where there might be an Earth fault. I tested each wire in turn and got no reading on any, so I turned my attention to the terminal block. I tested each of the four terminals, and one showed conductance to Earth. I thought that was strange, so I dismantled the motor to have a closer look at the terminal block. Behind the terminal block is a contact operated by the centrifugal switch, which switches in the capacitor to start the motor and then switches it out once the motor speed is high enough. I tested the terminal block, and I could find no fault with it, but I determined that the arms that ride on the centrifugal switch were slightly bent, which was causing them to contact the metal part of the rotor. I straightened the arms, reassembled the motor, and retested it for Earth faults. This time there was no fault, so I plugged the motor in again, and it sprang to life, but with a horrible bearing noise, which was no surprise. However, it was turning in an anticlockwise direction, whereas I needed it to turn in a clockwise direction. After unplugging it, I swapped over the two wires for the start winding and tried again. Now that the motor turned clockwise, it was time to see if I could fit it to the equipment. I removed the mounting bracket from the old motor, fitted it to the new one, and tried it on the equipment. The pulley did not align with the pulley on the equipment. I changed the bracket to the last two holes on the motor and the alignment was close enough that I would be able to adjust the position of the pulley on the shaft. But now, the bracket was only held on with two bolts. I dismantled the motor again to see if I could drill into the case to fit another two bolts. Luckily, there was enough clearance between the inside of the case and the windings on the stator to do that. I placed some timber between the windings and the case to avoid drilling into the windings. With new holes drilled and bolts fitted, the motor was ready to use after it got new bearings. I suspected I would have problems getting the bearings because when I measured them with my vernier caliper, they were both imperial sizes. Imperial bearings are now less common than metric. While shopping, my wife took them to the local Bearing Service in town but only returned with one new bearing. The shop got the other bearing for me in about a week, and after collecting it, I reassembled the motor. After a full service of the equipment and a few minor things replaced, I could use it again. The equipment now ran smoothly under heavy load with no indication of stalling. The replacement motor only cost me a few dollars, with $31 spent on new bearings, for a total of under $40. I was very happy with the outcome; an otherwise piece of useless scrap metal now had a new purpose in life. SC Australia's electronics magazine siliconchip.com.au