Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG The things we do for our pets Dave Thompson Serviceman’s log, stardate 2023.12. We have encountered strange furry creatures that are demanding to be fed. It’s almost as if they believe they are our masters. I have beamed down to the planet to see if I can open a dialog with the three famished felines. The serviceman’s curse is indeed a curse! It is, of course, tied in with our peculiarly Antipodean DIY ethos and the ‘number 8 fencing wire’ lore of Australia and New Zealand. If we think we can do it ourselves, we will at least have a good go! Still, I like to think that most of us know our limits. I, for example, would not try my own dentistry or brain surgery. For one, I’d need someone to hold the mirror and/or torch, which precludes doing most common dental and medical procedures. I think that sort of thing is better left to the professionals. With just about everything else, though, I’m willing to give it a go. Whether it is building a guitar, installing an alarm system or replacing the main bearings on my car, I’ll give it a shot. I mean, what could go wrong? The guitar could be rubbish (I have made many, but only the first one was rubbish), or the engine overhaul doesn’t go as planned. They’re all pretty minor problems in the greater scheme of things, and there’s always the option to call in a professional, hopefully without them being annoyed that I have ‘had a go’ before bringing them in. 94 Silicon Chip I encountered this all the time in my line of work. Most people would only bring their computers in for repair after they’d had a go following some ‘tutorial’ on YouTube on how to fix it themselves. When I was trying to run it on the bench, they’d often chime in to say, “I tried that” or, “I’ve already done that”. I responded that I had my own troubleshooting processes, and I might very well replicate what they’ve already done. However, as they had brought it to me to fix, we could chat about what they’ve tried, or I could get on with my process for finding the cause of the problem. As we all know, there is so much misinformation on the web that it is almost impossible to find answers to even the simplest of questions without spending hours trawling through the clickbait, scams and people posting the same old erroneous rubbish just to scrape out some ad revenue. Hard drive on the rocks As an example, once upon a time, there was a data recovery strategy that required putting dead mechanical hard drives in the freezer to rejuvenate them just long enough to get the data off. In very specific and extremely limited circumstances, that method might work if the drive motor was seized. However, the way it was promoted on hundreds of sites was as if every failed drive could be recovered by doing this simple ‘hack’. As someone who has recovered data from thousands of drives over the years, I can say that it is not a reasonable strategy for recovering data. Yet, the number of people who brought drives in saying they’d tried that method was staggering. It is just one tiny example of how misinformation can spread and how it can also dramatically reduce the success of proper data recovery by messing with things people don’t understand. Many of those doing their own computer surgery are unaware that even tipping a drive over from standing on its edge to landing flat on the benchtop could cause platter damage. Handling drives and putting them in the freezer often precluded me from recovering any data from them because those people didn’t realise how fragile the drives were. Nothing in those tutorials mentioned static protection or physical vulnerability, so often, by the time I got them, they were already ruined. One guy drove hours to get to me with his hard drive floating around unprotected in the tray of his ute! Australia's electronics magazine siliconchip.com.au Items Covered This Month • • • What we do for our pets Repairing a Whirlpool washing machine The clock that was running fast 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 Alas, such is the DIY ethos we all have. It also encompasses devices that are often not really worth the time and effort to repair, and it is simple bloody-mindedness that keeps us trying to fix such things. I’ve spent hours repairing cat toys when it would be much more sensible to just go and buy another one. For me, it’s the principle of the thing; if I can fix it, I should. It’s my duty. And if I end up spending hours and hours on a job for someone else and can’t fix it, then I don’t charge. That’s the moral way of doing business, even if it is not a path to riches. High-tech moggie feeding solutions We have three cats, and they all eat different diets for various reasons. So we have different food bowls and feeding patterns, and this routine involves some highfalutin (read: expensive) microchip-based cat feeders for two of them. I have written about these units before. I’ve also broken one by trying to be clever; probably not the last time I’ll do something like that! These devices run on four C-sized cells, with no provision for external power. In this day and age, that’s an odd thing, or at least I thought so. I did what anyone else would do and tore it apart so I could power it with an external supply. I’ve done this with dozens of devices over the years, including guitar stomp-boxes. I have also converted primary-­cell-powered torches into rechargeable devices by switching to NiCads and adding charging sockets. It’s usually all pretty straightforward stuff. This time, however... Getting these microchip feeders apart is easy; there are no stupid security fasteners or single-use break-away clips; just good, old-fashioned screws biting into solid plastic turrets and good-quality plastic mouldings that fit together seamlessly. A long-reach Phillips screwdriver is the only special tool required, as some of the holes are deep, and the screws are beyond the reach of a standard driver. Inside, they are surprisingly complex, but there was plenty of room for adding a socket in the rear corner, and the wires to the battery bays were easy to access. I dug through my boxes of power supplies, gleaned from years of collecting them and never throwing any away, and found a 6V DC supply that would do the job. This supply would determine what type of power socket I would install, because it is far easier to get a new socket (if needed) than to change the output lead of the supply (though I have done that many times in the past). I used to watch Dad changing a supply’s lead and/or polarity by opening up the sealed plastic case. I’ve tried to siliconchip.com.au replicate what he did but have never been successful. He would sit the supply on one corner on the workshop floor and, while holding it in position with his left hand, strike the uppermost/opposite corner of the case with a hammer. Every time I saw him do it, one quick rap with the hammer and the case would just pop apart down the seams like magic. Whenever I’ve tried it, I end up with a smashed case, typically parting everywhere but the seam and with crushed corners. It can usually be glued back together, but it is highly annoying that he made it look so easy, and I’ve never been able to do it, even with him teaching me! Since then, I haven’t even tried to do it, but I can do a plug swap or add an inline switch in the wire near the supply’s output. When done properly, it looks OK and functions perfectly well. Anyway, I found a suitable socket in one of my spares boxes and installed it into a space in the back of the feeder. I re-routed the battery leads to the socket, ensuring the polarity was right, and soldered it together. I used tape tabs to hold the wiring out of the way of the door mechanism and reassembled the whole thing. I plugged the supply in, connected it and tried the feeder using one of the RFID tags that came with it. Nothing; no lights, no response to button presses or programming functions. Bupkis. Well, that was disappointing. I checked the supply’s polarity and output, and though the measured 6.5V (unloaded) was a little high, I thought it was within limits. The polarity was correct, but it didn’t work. So, I took the whole thing apart again and rewired it back to battery power. This time, I got lights, but while the door tried to open, it wouldn’t go all the way, even though the motor tried to actuate it. Australia's electronics magazine February 2024  95 Bother! (or words to that effect) Try as I might, I could not get this thing to work correctly. It was the first device in decades that wouldn’t operate after a conversion. I was gutted, especially since this thing cost 300 Kiwibucks (about 280 Aussie dollarydoos) and it had lasted less than a week. It appears that some things are not meant to be mains-powered! I was 100% sure that I’d gotten polarities and wiring correct, and I’d double-checked it half a dozen times before applying power. I guess I was trying to be too clever by half. Lesson learned! That feeder still sits in my workshop today, dead and gathering dust. While covered by a factory warranty, I’d have to ship the unit to the USA for repair and pay all shipping costs myself. That would have cost almost the same as a new unit, and I think they’d ask questions about the hole I’d bored in the back for the power socket! So, an expensive lesson then. I still don’t know why it wouldn’t power up with mains power. It was an old transformer-style supply, not a switch-mode one, so I can see no reason that wouldn’t have worked. I wasn’t going to try it again, that’s for sure! I did what any self-respecting serviceman would do and went and bought another unit! That one is still going strong today, and the four C cells last almost a year, so it was a bit of a moot exercise anyway. The pitfalls of parts swapping However, the other feeder has now developed a problem. It sounds very rough when opening and closing and sometimes stalls partway through the door-open cycle. Either the motor was getting weak, or the gearbox driving the door assembly was wearing out. Not good either way. Now, I know what you’re thinking. Did he fire six shots or only five? Sorry, wrong movie. I meant to say I have a whole new spare, non-working device sitting in my workshop that I could burgle parts from. And you’d be right, except for one tiny prob. Between the time I got these and the new one, they’d changed the model slightly, including the folding door actuator mechanism, so they look completely different on the inside. Even the PCBs are different. You can rest assured there was some blue language flying about the workshop when I discovered that! The first thing I tried was changing the batteries. Though there is a low battery indicator on the feeder (a red flashing status LED once every minute or so), I swapped them out for some new, fresh alkaline cells. There was no difference, as I suspected, so it was not caused by a lack of power. I pressed on regardless and took the covers off the grumbling feeder. It’s been opening and closing half a dozen times a day for years now, so I fully expected it to be worn out, especially given the current gear noise. The motor itself is the same in both the old and new versions, but the gearing and actuator assembly are slightly different, which is to be expected as the door opens and closes slightly differently in each version. This is a bit of a curse because swapping one assembly to the other unit would have been so easy. However, it’s rarely that easy! The only real option was to disassemble the grumbly gearbox and check it out internally for wear and tear. If anything inside the unit had totally failed and needed replacing, I’d be dead in the water, as the company does not provide spare parts. Oh, for a 3D printer! When I pulled it apart, I confirmed the motors looked the same. I guess that changing it between models would be pointless. The gearbox is a Nylon gear assembly that converts the rotary motion of the electric motor into a linear action to actuate the bi-fold door. If the door is closed, and the correct embedded microchip or RFID collar tag is detected, the motor runs in one direction to open the door. If the door is open, and the microchip or tag signal is no longer detected, after a switch-selectable preset time, the motor runs in reverse to close the door. The limits appear to be set electronically, similar to an electric window in a car, where the controller detects when the window reaches its maximum and minimum operating range by detecting the increased current drawn by the motor. This will also occur if someone gets their arm or fingers stuck in the window, or in this case, a cat gets a foot stuck in the closing feeder door. I tested it when we first got these units, and that safety feature works quite well. Fixing the feeble feline feeder In this feeder, sometimes the door won’t open more than a few centimetres. It always seems to close OK, though; it is just dodgy on opening, which likely takes more energy due to working against gravity. I guessed that something was fouling the gearing, or the grease had dried out, and the gears were binding up. In other words, I felt it was a mechanical fault rather than an electrical one. At least if the motor was failing, I have a spare one of those! I took the five screws out of the gearbox housing and carefully pulled the side clear. I’ve been caught with these sorts of things before, where some gears and shafts come out while others stay put, and half of it ends up falling onto the bench. I really didn’t want to have to work out what went where, having never seen it in place! I managed to get the case off and saw that while the gears had some noticeable wear, the light grease used was dry and crumbly and lying in the bottom of the gearbox. 96 Silicon Chip Australia's electronics magazine siliconchip.com.au I used a contact-cleaning spray to wash the box out and mixed bearing grease and oil (all I had) to make a light composite grease. I applied it liberally over everything I could see and added drops of oil on the shafts where they engaged with the housing. I reassembled everything and inserted the batteries. Without the bi-fold door fitted, I ran the actuator forward and reverse manually several times to coat everything well. It certainly sounded better! I then reassembled the whole thing, and it has been running quietly and smoothly for a few weeks with no stoppages. If it lasts another year, I’ll be happy – who knows, maybe I’ll have a 3D printer by then! Editor’s note: I’ve sometimes had luck spraying silicone lubricant into holes in gearboxes on small pieces of equipment with similar problems. While it’s rarely a permanent solution, it can save you from having to disassemble and reassemble the lot, and sometimes you can even get away with poking the straw through a hole in the case. Repairing a retro Whirlpool washing machine D. C., of Beachmere, Qld went on quite an adventure delving into the innards of his trusty old washing machine. It is a simple and robust design that has withstood the test of time and just needed a bit of TLC... We had an older model (non-electronic!) Whirlpool 6LSS5232 washing machine for well over 30 years. It is a two-speed model with suds return, which my wife loves. Washer technicians have told us that the machine is almost indestructible and to keep it going as long as possible. These machines are direct-drive via an ingenious gearbox attached to one end of the motor, with the pump attached permanently to the other end. This turns in one direction for washing, then reverses for spinning and pump-out. The design was so successful that many other manufacturers of the time either copied the gearbox design or used Whirlpool parts in their own machines. The drive motor is an open asynchronous AC motor with high/low speed windings, plus a winding for the capacitor start, which permits rotation in either direction. The usual centrifugal switch controls the start winding disconnection. However, the centrifugal switch does not operate reliably at low speeds, so the motor is always started with the high-speed winding. After that, changeover contacts on the centrifugal switch select the low-speed winding once it is spinning. Our machine was recently showing three faults as well as quite a bit of cosmetic ageing and the usual rust from corrosive detergents, so it was time for an overhaul. The problems were timer unreliability, a leaking tub, and spinning a full tub of water while pumping out. The last one was worrying, as spinning a weight like that puts a lot of strain on the motor and clutch. To make a less rushed job of reconditioning the machine, I purchased a similar working machine, with a view to later use good parts from that machine as spares. While the spare machine was working, it had been sorely neglected and was seriously rusted, so it was of little ultimate use apart from spares. The timer is the typical rotary mechanical device of that era, with a “pull out to start” function and loads of cams and contacts to control the various functions, driven by a small timer motor. The fault turned out to be a broken siliconchip.com.au detent spring inside the timer which holds the knob in the start position. Of course, the spring is deep inside the timer mechanism, so I had to dismantle it completely. I was able to steal the spring out of the donor machine, but there were many hours of fumbling and cursing, getting all the bits back together (six hands would have been good) before I finally succeeded with repair phase one. I then started the real overhaul of the washer mechanism, learning as I went by making lots of mistakes! I managed to get it all apart and cleaned out the caked-on detergent deposits from 30 years of hard work. The tub seal is located deep inside the works and seals the agitator shaft as it enters the bottom of the tub. Amazingly, many parts for these machines are still available in the USA; some are even available in Australia, so there was no problem getting a new seal and fitting it. The spinning-while-pumping problem was a bit more involved. The gearbox allows agitation when the motor spins in one direction, and spinning and pumping out in the reverse rotation. To enable the reverse rotation to do two different jobs, spinning or pumping, the smart engineers at Westinghouse designed a device called a ‘neutral drain’. One of the ubiquitous YouTube videos helped me to understand how the neutral drain works. It is inside the gearbox and consists of a large metal wheel with several hard plastic cams and latches. The device counts up the wash agitator movements until a latch is set, and the latch prevents the drum drive dog from rotating when the motor is reversed, so the neutral drain only works immediately after a wash cycle. Thus, the water can be pumped out without the drum spinning. After a suitable time for pumping out, a short rest is included in the cycle, which allows the neutral drain latch to release, and regular spinning can resume. When the drum spins normally, the drum’s brakes are released by a cam, and the clutch allows slippage until full speed is reached. However, when the neutral drain is in, the drum is held firmly at rest by its brakes. Australia's electronics magazine February 2024  97 The Whirlpool washing machine shown with its AC motor and gearbox assembly on top. The hard plastic neutral drain components wear out with time, and a “neutral drain kit” is readily available, which completely solves the problem. I sent the required dollars, received and fitted the components, renewed the gearbox oil, then bench-tested the motor/gearbox assembly, which worked perfectly. To make bench testing easier, I built a small motor test unit that allows forward/off/reverse rotation and high/ low-speed selection. It was just a matter of mounting and wiring two switches and a start capacitor in a metal box and wiring it from the mains supply out to a motor plug. All that remained was to fit everything back into the newly de-rusted and spray-painted cabinet and do a test run with some real washing. All went well; my wife approved, and with any luck, we will get many more years out of the old machine. I learned a lot about the operation of our machine and now feel confident to solve any new problems. Finally, I dismantled the poor old donor machine, consigned useful spares to the workshop shelves and gave the remainder a decent burial at the local recycling centre. It started gaining time; after 20-odd years, you can expect some problems. Usually, when daylight saving changes, I don’t have to adjust the minutes, just the hours. However, last time, it had gained some time, then after setting it, I found it had gained again the next day. Rather than have the incorrect time showing, I switched it off, and when I got sick of walking into the bed in the dark, I decided to have a look at it. Clock radios of this era generally used the AC mains frequency as their time source, and this clock was no different. It uses an LM8560 IC as its clock driver. Plenty of data on that chip is available online, albeit as poor-quality scans. My Teac was very similar to the data sheet circuit as far as the mains cycle clock circuit goes, and the CRO showed signals that looked pretty much what you’d expect. The PSU main electro looked OK, so I rapidly came to the conclusion that the IC had probably developed a fault after 20+ years of service. I don’t blame it. I found a replacement on eBay from China for less than $2, including postage. I often wonder how they manage this – I would spend more than that on postage alone for the smallest domestic parcel, without even considering my time, packaging materials and international freight. Perhaps the Chinese Government subsidises it, or maybe its one of the reasons Australia Post is losing money. Maybe both. A month or so later, it arrived, and after walking into the bed again one night, I was inspired to fit it. It didn’t take too long, but I was both unsurprised and disappointed that the display showed gibberish with the new part. It wouldn’t respond to the time set buttons either. I checked the orientation and soldering and neither looked like the source of any problems. After muttering some profanities and vowing never to buy cheap electronic bits on eBay again, I picked it all up and unceremoniously shoved it onto the e-waste pile. A few weeks later, I saw that my son had left his Panasonic clock radio in his room when he moved out, and later that evening, it was in place next to the bed with the correct time. It wasn’t until the next day that I noticed the display was dim – so dim that I first thought it was turned off and had to shade it with my hand to read it. Ahead of its time D. T., of Sylvania Southgate, NSW has experienced that strange feeling when you manage to fix a faulty piece of electronic equipment, only to be baffled how it ever worked in the first place... I bought my Teac clock radio over 20 years ago. At the time, I wanted a combined phone-plus-clock-radio to minimise space used on the bedside table, and this Teac was the only one around. While the display was a bit bright at night when it was new, it wasn’t long before the brightness ‘wore off’, and it has been a good product overall. At night, the LED clock is easy to read; much easier than finding my smartphone and figuring out where the on button is. It is also good to navigate by – late at night, I can walk toward the display in the dark to find my side of the bed without disturbing my wife. 98 Silicon Chip This early-2000s era clock radio uses a couple of singlesided PCBs and mostly standard through-hole parts. Australia's electronics magazine siliconchip.com.au I tried it like that for a while, thinking I didn’t need to see it during the day, but I found it annoying looking over and not being able to read it. Obviously, I look at it more during the day than I thought. At least it was keeping good time. I thought this might be a case of bad electros causing the high-voltage feed to a vacuum fluorescent display to fail, but I quickly ruled that out when I opened it up and found the display used LEDs. Interestingly, I noticed it used the same LM8560 IC, with the same package and everything. I rescued the Teac from where I left it, then desoldered both ICs, and soldered the one from the Panasonic into the Teac. I gingerly switched and was greeted with a nice, clean, non-gibberish “12:00” flashing in bright green digits. I set the time and let it run. It would be great if that were the end of the story; however, I’m slightly embarrassed to say that my diagnosis of a faulty LM8560 was clearly incorrect because, by the next morning, the time was out again. So, back to the drawing board. I dug out the data sheet again. For an LM8560 to gain time, you’d think it would need to receive extra pulses, causing extra counts. Given the rate of time gain, I calculated it was receiving probably one or two additional pulses per second. I measured the main supply electro again – it was marked 470μF but measured around 350μF. It looked OK and the rail looked OK on the CRO too, but I replaced it anyway. I also checked the clock line. It was a half-wave rectified line from the transformer via a 100kW resistor and looked good. However, I noticed that while the data sheet showed a 1nF cap across the input pins of the IC, the Teac didn’t have one, nor did it have pads for one. This cap and the 100kW source impedance would have formed a simple lowpass filter with a time constant of 100μs. I thought about adding that capacitor. I don’t like second-­ guessing design engineers; I worked as one for many years, and usually, the design you choose is what’s ‘needed’. Clearly, this design hadn’t needed that capacitor for 20 years. But I was getting tired of this repair dragging on for so many months, so I soldered in a poly I had handy. It was then a matter of putting it all back together and trying it out. The initial results were promising, and I was relieved to find it still showing good time the next day, and it continues to do so. So what was it? I’d guess the main supply electro. It was definitely low in value, and perhaps something was intermittent about it, or its ESR had crept up over time. It reinforces my distrust of old electros. There could also be a latent fault somewhere else, like a shorted turn in the transformer reducing the clock voltage margin. I’m not sure. Interestingly, I note the Panasonic, like the Teac, had no capacitor on the clock input pin, so it’s probably optional. I don’t think having the poly cap there will cause any harm. I’m just glad I can find my way to bed now without injuring myself! Editor’s note: increased ESR in the power supply filter capacitor might not be obvious (unless you use an ESR meter) and could lead to reduced noise margin on the timing input. The manufacturer no doubt figured they could get away without the extra filter capacitor, and they were right, but sacrificed some reliability in saving that cent. SC siliconchip.com.au Australia's electronics magazine February 2024  99