Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Salvaging a soggy ceiling circuit Dave Thompson Why is it that when a pipe leaks, it’s always in the most inaccessible location, and the water always ends up where you don’t want it? It must be one of the variations of Sod’s Law. A weird thing happened a while back on the way to work. All of a sudden, I was sitting in Airlie Beach, which, as all you boffins know, is in Queensland, in the stunning country of Australia. I’d had surgery on my legs a few weeks before that, and the recovery time they mentioned was only valid when wearing very rose-tinted glasses. So I was sitting a lot, but told to walk. We did a lot of sitting, watching thousands of tourists a day head out on huge boats to the Whitsundays and, of course, the south end of the Great Barrier Reef. However, I also hobbled around the town to sample the local fare (which is very good) and the local wines, which are also excellent. Then we took a trip to a pontoon hotel floating on the reef itself. So it was idyllic, and the first holiday we went on that didn’t involve pandering or catering to our families in Europe and Western Australia. Our time, as Agent Smith says in The Matrix. This was all very well until the neighbour back home, who had kindly looked after our place and our pets, messaged to say he had found water pooled on the floor of our downstairs bathroom. He’d found a bucket and put it down; as we were going to be back in a few days anyway, there was no panic. It was all in hand, or at least in bucket. I got home to find that the water was dripping, slowly but surely, through the bathroom light fitting. This is a combination heater, extractor and LED light (the motor runs both ways depending on the four-way switch on the wall). Now, given that this is electrical – a mains-­powered unit – and water was pooling in it before dripping through the grille into the container, I was quite worried about switching it on! Items Covered This Month • A soggy ceiling circuit • Calibrating a Silicon Chip Differential Probe • Repairing an LG air conditioner 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 86 Silicon Chip Fortunately, we have an upstairs bathroom as well, so we could use that. Unfortunately, the water was coming from somewhere up there. While it was clean water, which gave a clue as to which pipes to look at, it was still coming from somewhere that it shouldn’t. So, the only thing for it was to get my old gammy legs into the roof space and see if I could locate the source of the leak, all without electrocuting myself or falling through the ceiling (which might not take a lot of weight if it was soaked!). A job for the young and limber Luckily, the guys who added the top storey to this house in the 90s included several access doors. These are almost hobbit-sized small cupboard doors; so not overly easy for an old man to clamber through, but located in the main areas upstairs to access the crawl space that surrounds the upstairs rooms. They do come in very handy for running network cables and the like. Of course, one must be very careful not to step on anything but a roof joist, or one would find themselves sitting on the downstairs floor, confused and covered in Gyprock fragments! ‘Luckily’, the extractor fan was exactly halfway between one access cover and another, so I would have to put on my gloves, knee pads and headband LED lamp, along with a 3M particulate mask, to navigate there. The space up there is dusty, to say the least. It is also inhabited by spiders and the odd mouse. Come spring, birds find their way in and roost, as well as make baby birds, so it’s a menagerie up there. Australia's electronics magazine siliconchip.com.au As I said, the top level was added a while back, and they did a real bodge job moving the hot water cylinder up there. There was nonstandard plumbing and wiring all over the shop. I redid the wiring when I renovated this place before moving in, under the supervision of a sparky who had damaged his leg and couldn’t move very well (perhaps it was just so he could sit and drink tea on our dollar!) We also installed a new Earth rod as we changed all the underfloor, rusted steel pipes for modern butylene, which meant new Earth connections for everything that needed them, as this plastic doesn’t really conduct electricity that well. Fortunately, they had put plastic piping in the roof space, but much of it was non-standard. Since we didn’t have to touch much up there back then, we concentrated on renovating the ground floor. This meant that those old pipes, which criss-cross the space, and were routed wherever they could fit them for the upstairs bathroom, are 30+ years old. I made my way through the dust and the spiderwebs until I got to where the extractor unit sat. Sure enough, it was full of pooled water. The drips were coming from a small hole in the housing, likely an unused screw hole, and slowly making their way down to the floor downstairs. There were several pipes in the vicinity, but none quite near the fan, so I lifted some insulation (gloves essential) and tracked a damp timber beam back until I found a fitting almost inside the wall of the upstairs bathroom. Reaching up the pipe, I found the fitting was wet, and my glove came away damp. It was dry above that, so this must be the source. The water was tracking down the pipe, dripping onto the joist and making its way to the lowest point, right into the extractor. The problem was that now I had to back out, go all the way downstairs and find the breaker for this unit, as it would likely have to come out to be dried and cleaned. That meant disabling the power, then getting up the ladder to take it out and clean it. Luckily, when I rewired this place, I made a map of all the breakers and circuits. Years of dust being drawn into it for the heating function and blown out of it in the extractor mode had left it looking almost flocked in a thick layer of dirt, which was now wet sludge in most places. The first thing I did was call a plumber, because plumbing work is above my pay grade and things could go seriously awry if I were let loose on it. He agreed to come that night, after his usual work, to have a look. Editor’s note: residents of New Zealand can legally do some of their own electrical work, including fixed mains wiring, with some provisos (eg, the work must be done to NZ standards). That includes disconnecting and reconnecting existing appliances. This is not permitted anywhere in Australia, where such jobs must be performed by a licensed electrician. And similar to NSW, you can only perform basic plumbing work by yourself in NZ. Anything more complex (eg, installing or replacing mains pipes) requires a licensed plumber. popped off with a knife blade, and I set them aside so I could refit them later, right at the end of the job. A few weak clips held the bottom ‘half’ of the unit on, and it came off easily. Half of that is a heater-style grille, and the other half is a large LED light; a flat, rectangular panel type. That had to be unplugged from the main body, but the plug is a standard barrel type, and it unplugged easily. I set that aside and looked inside the fan. I could see four ‘superscrew’ type fasteners had been used, two per side in each corner, to fix the main body of the extractor to the ceiling joists. The problem with this is that I was going to need three arms to get it down, and standing on a ladder makes it even trickier. I used a drill with a long-form Robinson bit in it to spin out the first three screws. I then held it up with my head and used the other hand to remove the last one. The assembly fell clear, dangling on the power cables. Before I went anywhere near them, I checked with my trusty mains detector pen to make sure I had switched off the right breaker. The light and fan no longer worked from the switch, but I don’t like to just assume there’s nothing still there that could kill me, especially as water poured everywhere when the unit dropped down. We all know that electricity and water are not good bedfellows! It all seemed good, so after taking a photo of it, I used my equally trusty electrician’s screwdriver to undo the terminal screws and free the wires. With the unit out, and now a large hole in the roof, I stuck a bath towel in the gap to stop any drafts and soak up any more water that might still be coming out. I set about trying to clean this thing up. The water had discoloured the plastic because it had run over treated timbers on its way down, leaving an almost tobacco-coloured brown stain on the plastic. I tried isopropyl alcohol and methylated spirits, but it looks pretty much permanent. It did clean some up, but not all. The unit’s chassis was mostly pressed or cast metal, so that was relatively easy to clean with damp rags and a bristle Meeting one of my fans Pulling the decorative bezel from the extractor unit body was relatively simple, after I had figured out that the screws were concealed under small plastic covers. The covers siliconchip.com.au Australia's electronics magazine September 2025  87 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. brush. Once as dry as I could get it, I left it – I’d blow out the fan with the air compressor tomorrow, and any other dust that might be left in the nooks and crannies. While I was there, I would also lubricate the motor bushes and check that everything else was in good shape. The fan motor does a lot of work, and sometimes it runs much of the day, in both directions. It’s actually quite a high-­quality unit; I was impressed with the build quality. Mind you, from memory, even eight years ago it wasn’t a cheap appliance, even though it came from one of the big hardware chain stores. It has been a very faithful unit, although we don’t use the built-in fan heater much, because that thing eats power like Homer Simpson at an all-youcan-eat seafood restaurant! A prompt tradie As promised, the plumber turned up on time (a miracle in itself!) and I pointed out the problem. While we could just see the pipe from the hole in the bathroom roof, there was no way to access it from there. I showed him where the access hatch was, and with the grace and flexibility mostly only young people have, he grabbed his headlight and ventured in. He made his way around behind the walls, and I went down and stood at the bottom bathroom gap as he explained that whoever had set this up originally had used a bad industry practice of putting two different plastic pipes together in an inline join. He said it was leaking from there. It was the clean water feed to the upstairs toilet and vanity. As is typical, the joint was buried behind some roof beams, and while he could see it from there, he couldn’t access it. So back he came, and while I went and turned off the water mains, and some outside taps on to drain as much as possible out of the pipes, he rummaged through the copious boxes of spare parts in his van and found a brass fitting specially designed for this type of job. 88 Silicon Chip Unfortunately, the toilet had to come out to get proper access. This meant undoing the screws and cutting the caulk away, but then it just slid out, and five minutes later, he redid the join (with a lot of water on the floor). Soon after, it was back in, and I would caulk it in a few days if no more water came out. So, it was a cheap and relatively easy fix, and he looked after us payment-wise (another miracle for a plumber!). Now, all I had to do was put the extractor unit back in, and once again, I’d need three hands to do it; two to hold the unit, and one to reconnect the wires to the terminal blocks. The obvious problem is that I’m not an octopus, so I had to just hold the unit in the gap with one hand, line up the mounting holes and use my driver to reinsert the superscrews in the same holes they came out of. Rewiring it all over again Of course, the wiring was just sitting up there disconnected. I would simply have to get all my gear on and venture back into the cave to re-terminate everything. First, I replaced the now-stained plastic bezel and used the four PK-style screws to fix it to the chassis. The screw covers popped back on without any hassles. It all looked pretty neat there, but now it was time to break out the hazmat gear and get back into the roof space. Fortunately, I had taken that photo before removing all the cables, as it was quite a complex setup, from the switch end to the unit itself. I took my phone up there with me and squeezed into the area. It wasn’t easy to get to, but possible, and with a bit of fettling, I managed to get the screwdriver in and attach the wiring to the terminals. After using some cable ties to anchor it all, I made my way back out and downstairs to reset the breaker. I went to the bathroom and tried the fan-only switch. The fan fired up and sounded good; the lubricant had done its job. I tried the light, and it lit up nicely. I noticed when I turned it off, though, that it seemed to linger and fade out over a second or two. Usually, it is instantly dark. So that was odd. I then tried the first heater setting, 1200W. I have to turn the extractor off first; otherwise, this switch reverses the fan motor, and just dumping it in reverse is no better for it than it would be for a car transmission. With the fan switch back on, sure enough, warm air blew into the bathroom. I switched on the second heater switch (2400W) and the breaker tripped instantly. That was new, too. I switched it all off, reset the breaker and tried again with the same result. Maybe I’d damaged something cleaning it, or the water had gotten into it somewhere. Either way, it didn’t bode well. I got all kitted up again and went back into the roof. I should install a walkway for all this traffic! I just wanted Australia's electronics magazine siliconchip.com.au to check I hadn’t been a dolt and wired something wrong – easy enough to do in cramped quarters with a mask and gloves on. It looked OK, but I redid it anyway. At least the pipe fitting was now dry, although it would take a few days for the damp up there to dry out. That was all I could do, so I once again made the knee-shattering journey back out. I really am getting too old for this, especially after my recent leg surgery. I went down and tried it again. Same result. So, during the coldest parts of winter, we could only use the low setting, by the looks of things. I was sure I’d rewired it the same as it had come out. Off came the bezel again, and everything looked OK there too. All I could do was consider buying another one. However, after a few days, I thought I’d try the second heat setting again, and this time, the breaker didn’t trip. Perhaps some water had encroached, and now it was dried out. I’ll take the win, but I’ll be keeping an eye on it! High-Bandwidth Differential Probe repair The High-Bandwidth Differential Probe published in the February 2025 issue (siliconchip.au/Article/17721) looked like a very useful addition to my toolkit, so I set about ordering the components, being very careful to procure the correct 0.1% tolerance parts. It made good sense to build three units to facilitate measurements on multi-phase applications. All up, I had to go to four different suppliers to get all the parts. After a few weeks, everything was in hand, and I proceeded with the construction as outlined in the article. A good friend was kind enough to solder REG5 (the SMPS chip) in his reflow oven, but it was still a challenging task. I used my hot air workstation for the remaining components. As per the instructions, I successfully tested the power siliconchip.com.au supply before proceeding to install the remaining SMD components. I arrived at the point where the instructions say to apply the conformal coating. While I was questioning if I should really coat the components without first testing the circuit, I decided to go with the instructions and apply the coating. Before doing so, I used a masking paste to cover the remaining through-hole parts so that they didn’t get lacquered. Many years ago, I purchased a can of Circuit Board Lacquer NA-1002, which had a sufficient quantity left in it. I applied three coats of this lacquer and let it set for 24 hours before installing the remaining components. Now for the fun part – calibration. Step 1 was easy; adjust VR2 to read <10mV. Step 2 (the CMRR adjustment) was asking to adjust for a reading under ±20μV. Using a bench meter with the required resolution, I was very surprised to see the voltage measurement fluctuating in the mV range, some magnitudes above my target. The power supply I was using to feed 64V DC into the measuring inputs was a lab-grade Rigol unit with excellent specs, and it checked out OK. My oscilloscope showed this ‘noise’ to be at 50Hz, and it was equally present on all three PCBs. To allow me to progress to the next step, I decided to use the graph display of the bench meter and to get the fluctuating voltage around the 0V line; the screenshot below shows the changing measurement as I adjusted VR1. I knew I had to come back and find the problem, but at least I could move onto the final calibration of offset voltage trim and frequency compensation, both of which worked as advertised. Now onto testing all three units in full operation. Using a variac connected to an isolation transformer, I was able to adjust the voltage being fed into the probes. That’s where I discovered the three probes provided different readings, and none of them were correct when compared to my R&S ScopeRider measurement! The difference between the probes was over 20V looking at a peak-to-peak measurement with an oscilloscope. I went through my components purchase details to double check that I had ordered 0.1% resistors and I couldn’t find any obvious errors. The difference I was seeing was just not possible with those components. I had to put this project aside for a few days, as it was driving me crazy. Nothing seemed to make sense. After a few days of thinking about this, I decided to tackle this Adjusting VR1 let me change the DC offset but a significant AC noise voltage was superimposed on the output signal. Australia's electronics magazine September 2025  89 project again and to take some more measurements to discover the source of the difference between the three probe boards. The challenge was the circuit board lacquer all over the critical resistors. I shouldn’t have lacquered the board until after circuit testing, and with the variac in place, I could keep the voltage low enough so the lack of coating wouldn’t cause any issues. Hindsight is a great thing! So I scratched off the coating around all components of the input circuit voltage divider. I did this on all three boards – it really tested my patience! However, the result of the hard work was well worth the effort. After recalibration, I compared the measurements with all three boards, and they were practically the same. So it appears that removing the coating fixed the problem. The only conclusion I can draw is that the conformal coating provided a high-resistance path in parallel to the 1MW resistors and therefore affected the voltage divider ratios. Due to the variation in thickness and uneven application, the resistance provided by the coating was different between the boards, hence producing different readings on each board. I have now ordered a different conformal coating spray, one where the data sheet states “Surface Insulation Resistance 1 × 1015W”, I couldn’t find any “Surface Insulation Resistance” rating for the NA-1002 spray (now consigned to the recycling bin). The CMRR calibration problem remained unresolved, but without the conformal coating, it became easier to take measurements. The measurements still indicated a fluctuating voltage of ±2mV where I need to adjust to within ±20μV. I disconnected the 64V DC supply from the measuring inputs, but the fluctuation remained. Even after I switched off the probe using S1, the fluctuation was there. This was a clear indication that I was measuring an external interference signal which was being induced in the components on board. Since the measuring point for CMRR calibration is across the ~20kW part of the divider, I took a 22kW resistor and connected it to the input of the bench meter. You guessed it, I still saw ±2mV, obviously from the workbench environment where I was performing my calibration! To convince myself that my assumption was correct, I removed power from everything except the bench meter and the fluctuation was reduced to <2μV with the 22kW resistor. Upon restarting everything, the measurement returned to ±2mV. I am an Amateur Radio operator and my workbench is my radio shack, so there are lots of mains cables and many devices located in very close proximity to my workbench top. In fact, when the measuring leads with the resistor accidentally dropped onto the floor, the ±2mV fluctuation disappeared, so it’s just the top of my workbench, which is exposed to a 50Hz field. It looks like the solution is to perform the CMRR calibration under the workbench instead of on top of it! Erwin B., Wodonga, Vic. LG aircon repair My wife informed me that the kitchen air conditioner would not work. This is a small window-mounting unit I installed around 2000; it has never been touched except for cleaning the air inlet screen. It’s the type that has a very simple control system with no electronics; a ‘vintage aircon’, if you like! It has a temperature knob and a power switch with a couple of modes of operation, fan or cool. With those, you get two speeds of running and that’s it, simple. No remote control or fancy computer stuff. I switched it on to test it. It made a humming noise only; neither the fan nor the compressor operated. Should I bin it and buy a new one at about $600, or try to work out what’s wrong and fix it? That’s a no-brainer! I always try to fix things even if they’re past their ‘use-by date’. I found a postage-stamp sized circuit diagram on the frame under the front plastic cover. That confirmed there are not many electrical parts to go wrong! I then pulled the main cover off the unit, which took some time because a lot of the screws were rusted solid; some had to be cut out. Once inside, I traced out the wiring by colours and took The 1.5μF + 25μF 400V AC dual capacitor unit shown at left was open-circuit; the LG aircon is shown on the right. 90 Silicon Chip Australia's electronics magazine siliconchip.com.au photos to be sure. The wiring looked the same as the circuit, and since both motors would not run, this indicated maybe a common component had failed. I measured the resistance of the motor windings and insulation-tested them, just to make sure nothing was shorted or grounded. The compressor motor (COMP) had about 6W in each winding and the fan motor (MOTOR) measured hundreds of ohms. Both motor windings were infinity to chassis. The capacitor (C) was a dual 1.5μF + 25μF 400V AC unit. The capacitors inside the can were simply open-­circuit; I couldn’t get a capacitance reading anywhere. Most likely, the internal fuse link in the common leg had failed. I wanted to run the unit to test the motors before buying a new capacitor. In my junk box, I found an oil-filled 2μF capacitor and a similar 30μF capacitor, both with 250V AC ratings. I patched them temporarily into the circuit in place of the original dual unit and, when powered up, the fan and compressor promptly started! After a few minutes of pulling 2.3A (the nameplate rating is 2.8A), the back air got hot and the front air cool. That looked promising, so I left it for half an hour. Nothing blew up or overheated and the temperature differential between the evaporator and the condenser (front and back) was about 23°C. The open-circuit capacitor sections are in series with the startup winding of each motor. Without any power to the phase-shift windings, neither motor would spin and they just sit stalled and humming. I checked eBay for 1.5μF + 25μF motor start capacitors and there were plenty available. I ordered one from a local air conditioner supplier at $30, including postage. The new capacitor arrived in a couple of days, so I fitted it and repeated the test run. The operation was normal. This unit is probably past its service life, as the base frame is rotting away with pinholes, but the gas circuit is still charged and electrically it’s OK. Fixing it was the easy part. The hard part was cleaning out the rust from the bottom water tray and evicting the spiders and webs! I just cleaned it up as best as possible. I am very careful about not moving any of the gas pipe work. On an old unit like this, the pipes will be brittle and have no ‘give’ in them; moving them, you risk cracking a pipe or joint. If that happens, the unit loses its gas and is scrap. It would have been nice to drop the bottom base tray off, de-rust it and seal up any leaks. Still, that would mean moving pipes, so I just cleaned the loose stuff out and flooded the area with rust killer/undercoat to slow the rot down. I took a fair bit of time cleaning the fins on the heat exchangers with mould detergent and removing any build-up on the air duct surfaces and fan blades. All that was left was to refit the cover and strong-arm the unit back into the frame. As we all know, there is a phenomenon where any cabinet unit grows slightly in all outside dimensions after servicing on a bench! That caused required a bit of wriggling, and the cabinet needed a few healthy thumps to slide it back into the frame. A quick run showed normal performance, so the repair was a success. This fix illustrates that it is always worthwhile looking at a faulty appliance for an obvious failed component that can be easily replaced and the whole unit made serviceable again. SC Fred Lever, Toongabbie, NSW. siliconchip.com.au Australia's electronics magazine September 2025  91