Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG The vacuum cleaner that didn’t suck Dave Thompson Anyone who works with particular tools or machines a lot gets to know how they sound and feel, and so can quickly tell when they are not working properly. A keen woodworker will know when his table-saw blade is getting dull simply by the noise it makes when cutting. An obvious sign of a dull blade is that it takes a lot more effort than usual to push the wood through the saw, or that the finish of the cut is not as good as it should be. But a perceptive craftsman will know well before that just because of the different sound the tool is making. As a musician, I also tend to notice melody and rhythm in almost everything. I know the regular rhythm of our dishwasher when it’s working correctly, and sometimes find myself humming along to the harmonic-rich tone our microwave oven emits when it’s cooking. If these aren’t operating correctly, I’ll know. While this might appear sad or a little weird, it means I can often pick up when something’s amiss just because it doesn’t sound normal. We rent our other house to shortterm tenants, and as I’m the one looking after it, I get to clean it from top to bottom on average every couple of days. It’s quite a time-consuming process. As a serviceman, I’m always looking for ways to improve how I do this job, and that includes improving the tools I use. If I can buy or make something to do things better, or quicker and easier, I will. pelled more dust back into the air than they vacuumed up! While modern designs and improved filter materials make new models more efficient, a lot of those older machines don’t make the cut. These days a good vacuum cleaner needs to be lightweight, efficient, clean, easy to manoeuvre and quick to empty. But models that tick all these boxes can be surprisingly expensive. Recently, I started noticing that my 10-year-old Bissell PowerForce Turbo bagless upright model sounded different and it took considerably more effort to push around the floor. I also found that I had to make more passes to pick up visible debris. When it was new, this cleaner glided over all surfaces and just about sucked the carpet up with the dirt. While it is tempting to just throw money at the problem by buying the most expensive cleaner on the market as a replacement, this wouldn’t necessarily result in the best tool for the job. We recently traipsed around the usual stores looking at the wide variety of new vacuum cleaners and weren’t that impressed by many of them, especially by some of the prices. The more we looked, the less inclined I was to buy new and the more convinced that I could refurbish our existing hoover and return it to rude health for far less than the cost of a new one. Fixing old faithful This machine has done us well; the only problem we had with it before Vacuum cleaners are my bag, baby Aside from the usual aids such as extendable dusters and good quality cleaning cloths and agents, one essential appliance is a vacuum cleaner. Not only does it keep the house free from dust and dirt, but it also makes it healthier. It’s best to use a vacuum with a HEPA-grade (High-Efficiency Particulate Air) filter or bags. Many vacuum cleaners I’ve used over the years ex68 Silicon Chip Australia’s electronics magazine siliconchip.com.au Items Covered This Month • • • The suckless vacuum cleaner An oven tripping the RCD breaker Fridge/freezer defrost repair *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz this recent loss of performance was a blown ‘headlight’ and a broken plastic height-adjustment assembly. The latter is mounted under the power foot. It alters the height of the brush and air intake above the cleaning surface using a simple mechanical adjuster knob mounted on top of the foot. The adjustment is meant to be used when moving from plain to carpeted floors. However, I’ve never changed this setting. We have a mix of short-pile carpets and vinyl floors, and the cleaner works just as well on both with the knob set half-way between the two extremes. This only adds to the irony that the adjuster is the only part that has broken. I don’t know how it broke; one day I noticed one of the two 25mm plastic roller wheels that form part of the adjuster had come away from its axle mount. The whole assembly is pretty flimsy, considering the strain it could potentially be under if one was to lift and drop the machine to the floor from more than a few centimetres. As is typical, the big-box store we bought the cleaner from doesn’t sell spare parts, fobbing us off instead to a vacuum cleaner speciality store. They didn’t carry parts for it either, even though this model was widely sold here. Nice one local stores, and you wonder why people increasingly buy online! I ended up sourcing and buying the part from Amazon; it only cost about thirty bucks delivered, so I was grateful I didn’t have to junk the vacuum for want of a cheap replacement part. The headlight hadn’t worked for about half the time we’ve owned the machine and was dim and next-touseless anyway, which is why I never bothered repairing it. It is one of those ‘features’ that seems great in theory, but in practice, appears not very well thought out. siliconchip.com.au Actually, I think a headlight on a vacuum cleaner is a brilliant idea (LOL!). Not everywhere we want to hoover is well-lit, and illuminating that area can be very helpful. However, this light is mounted on the base of the moving handle assembly. So as soon as I stomp on the pedal to release the handle from its upright resting position, the whole thing tilts back and lights up the walls instead. This actually throws the area in front of the foot into a contrast shadow. It would have been better to mount the lamp on the foot, which always faces the same way and sits level on the floor. It makes me wonder whether Mr Bissell has ever used his own product; if he did, he would have seen in a flash (pun intended!) how useless this feature is. The latest repair This machine has done a lot of work over the years, but as it wasn’t some $99 special to begin with, I would only consider junking it if I had no other option. I noticed that if I removed the hose going from the cyclone to the power head, it had good suction. But for some reason, it just wasn’t ‘on song’ and picking up dirt the way it used to. I usually either wash its filters in the washing machine, as per the manufacturer’s recommendation, or simply blow them clean with my compressor and air gun. Overall performance is usually restored after filter cleaning, but lately, this hasn’t worked as well. Because replacement filters also aren’t available locally (noticing a theme?), I once again hit the Interwebs. I found and purchased a twin pack of replacement filters and a new drive belt. It seemed sensible to replace the belt as a precautionary measure, even though I hadn’t checked it for wear yet. These parts came to around $50 delivered, a relatively cheap fix if it got the performance back up to scratch. I’d much prefer to buy this stuff locally and support local stores, but if they don’t bother stocking parts, I can’t. Despite my servicing history, I have scant vacuum servicing experience. So I decided to take everything apart, inspect all the parts and replace or repair whatever seemed worn or broken. I would then reassemble the machine in the hope that what I had done would fix the problem. While this is a very Australia’s electronics magazine cowboy, shotgun approach, it does usually work. Dissecting the patient The PowerForce is an easy machine to work on; no lame anti-tamper or purposely-obfuscated screws, just oldfashioned, easy-to-access meat and three veg fasteners. I stripped this one down to spare parts in about 10 minutes with one medium-sized Philips screwdriver and a pair of pliers. There really isn’t much to it; the obvious things that could go wrong are the motor, the filters, the drive belt for the foot brush, and any bearings or bushes that could wear out. With the machine on the workbench, I took a good look at it. Most of the hard work is done by a highrevving, low torque motor mounted inside the body of the cleaner, near the bottom of the tilting handle assembly. This keeps the centre of gravity low. Many new models have the motor assembly at the top of the handle, which in my opinion makes swinging them around more difficult and harder on the arms. Above the motor is the ‘dirt cup’, the reservoir which collects the debris, then a clear plastic cyclone separator assembly sits on top of that. The dirtladen air is sucked from the bottom, through the ‘foot’ intake, and enters the cyclone at the top via a flexible hose running up the side of the handle. There it swirls around due to the cyclone design, and any dust and debris (hopefully) drops into the dirt cup beneath. The remaining air exits through various filters, which trap pollen and other pollutants, making the air (in theory at least) cleaner than when it went in. All components are easily removable; the dirt bowl to be emptied and the cyclone assembly to access three of the five filters. There is also a rotating brush in the power foot. This belt-driven cylindrical brush spans the whole front part of the foot, just before the air intake, and runs all the time. But it only makes contact with the floor when the machine’s handle is moved out of its upright resting position and the power foot sinks to the floor. It’s a basic but effective machine. Repairing the motor I learned as a boy playing in dad’s workshop that running vacuum cleaner March 2020  69 motors without load can lead to catastrophic failure, so I tested this one using my non-Variac branded variac. This allowed me to wind up the juice and check the operation of the motor assembly. Straight away, I could see a lot of arcing around the brushes, which also appeared quite worn down. On closer inspection, I could see the commutator had also been worn smooth. While the motor itself might be a common enough model, it appears to be built into an impeller and duct assembly specifically designed to fit this machine, so simply swapping one from another type of cleaner isn’t an easy option. This motor also has a longer shaft protruding from the rear for driving the foot brush, so swapping out an armature from another motor isn’t viable either. Back in my aircraft engineering days, I refurbished a lot of motors. That typically meant replacing brushes and overhauling the commutator section of the armature as part of the process. Because many vacuum cleaner motors are similar, I hoped finding replacement brushes wouldn’t be too difficult. I removed the clamps holding the brush holders to the motor frame and eased the brushes out. I took them to a local appliance repair centre and asked the guys there (who all have incredible product knowledge) whether they had anything like them in stock. One guy came back with a few different types, which we compared on the counter, and I bought a set that was very close to the original’s dimensions. While the carbon composition might be different – ie, the brushes might be harder or softer – I didn’t have the luxury of choice so they’d have to do. Back at the workshop, the new brushes fit neatly into the brush holders, and while they were probably a little longer than the originals, there was plenty of room in the holders. 70 Silicon Chip The commutator was trickier. It was easy enough to get out, but I couldn’t find any information online about the depth of commutator undercut, or even if there should be any in these motors. However, I could see by the unworn part of the commutator, where the brushes hadn’t been contacting it, that it was originally slightly larger in diameter and that these segments were undercut. I’d have liked to have mounted the armature in a lathe and skimmed it flat, but there wasn’t a lot of meat left in the copper, so I made do with handrubbing it with 180-grit wet and dry sandpaper just to clean it up a little. I broke off a piece of a junior-hacksaw blade and wrapped lots of tape around one end to create a crude handle. I held the armature carefully in my bench vice and used this make-shift cutter to go around and ease out the areas between the commutator’s copper sections, being very careful not to slip and gouge any of the faces. Once I’d gone right around, I then cleaned up the copper again with sandpaper to knock off any sharp edges I’d created. Leaving them rough would chew the brushes out very quickly. I reassembled the motor, checking the sealed bearings at each end of the armature at the same time. They seemed OK, and the armature spun quietly and smoothly by hand. I powered it up and tested it; this time there was minimal sparking and it sounded great, so I considered that job done. While it was all apart, I looked at the headlight. The 12V, 11W bulb had blown. I rummaged around my bits boxes and found a 20W version; bigger is better, right? The shiny tape reflector behind the bulb had partially peeled back; a few dabs of superglue had that secured again. Australia’s electronics magazine The roller’s drive belt and bearings were next. The new belt was smaller and more pliable than the old one, so it turns out that it did need to be replaced. I spun the roller in my fingers, and the bearings ran smooth and quiet. I also took the opportunity to remove all the long hairs and threads that always seem to wrap themselves around these brushes. As I reassembled everything, I cleaned anything that looked dirty, removing years of built-up dust and trapped hairs. I found the centres of the main wheels had slogged out, so I cut some small strips of Teflon sheet and wrapped them around the axles before putting the wheels back on. I’ll eventually have to do something more permanent, but that’s a repair for another day. I installed the new filters, plugged the cleaner in and tried it out on the workshop floor. The difference was remarkable; not only does everything run much more smoothly, it is quieter, the suction more powerful and the motor sounds like it used to, all for a fraction of the cost of a new machine. The only thing that left me baffled was, despite having stripped down and disassembled the whole machine, even though this is the “Turbo” model, I never actually located its turbocharger. How strange! Oven tripping RCD J. L., of Toowoomba, Qld, had a very frustrating (and intermittent) problem with his oven tripping an RCD. False tripping of RCDs is unfortunately a common problem, but in this case, the cause turned out to be a bit unusual and unexpected. This is how he figured it out... A couple of years ago, our fairly new wall-mounted oven siliconchip.com.au started tripping the RCD on higher heat settings, typically around 220°C. This usually only occurred when the oven reached that temperature, so I surmised that it was happening when the element was switched off. The tripping was only occasional, ie, “nuisance tripping”, but it became more frequent over a year or two. I tried the usual approach: turn off every appliance on that circuit, then switch each one on in turn and see if the RCD would trip. I started with the usual suspects: fridges and dishwashers. The results were mixed. Every time I thought I’d identified the culprit, it would trip again without that appliance switched on, or even plugged in. This was becoming very frustrating, as each test required heating the oven to a high temperature. Eventually, it was decided that a ‘proper’ service agent had to be called. In due course, he arrived and tested the oven temperature, ramping it up slowly – which was precisely the condition under which the RCD would not trip. He determined that the oven should be on a separate RCD and charged us a $100 call out fee. Since this did not seem very helpful, I tried another approach. The RCD itself appeared to be the same model but older, than another in the same dwelling so I swapped them. But that didn’t help. So I tediously repeated the appliance removal tests. In the meantime, I built the SILICON CHIP Earth Leakage Tester (May 2015; siliconchip.com.au/Article/8553) to check the various appliances, and this worked as expected. Some, such as a dishwasher with ‘soft’ on, showed leakage of the order of 1-2mA when switched on at the wall but with the appliance off. This is not enough to trip a standard (30mA) RCD, even with several such appliances in-circuit. The next and most obvious culprit was the oven element. I thought this unlikely, but I know that elements can lose their insulation over time (especially if liquid is spilled on them), and can eventually trip the RCD. The only solution then is to replace the element. I removed the oven element, measured the resistance to work out its nominal power, and the dimensions of the attachment plate. After a week or two of unsuccessful searching, I decided that we could ignore the element for now, since I could not find siliconchip.com.au a replacement through the usual suppliers or even overseas, and that I had no proof that this was actually the problem. So then I thought back to what had changed since the oven used to work normally. Then I suddenly realised that the rangehood extractor fan over the stove top had never been turned off during my tests, since it has no separate switch (I wish it did). It has a pushbutton to turn on lights and a fan, but I had the impression that this was done via a controller board. Then I remembered that shortly af after the rangehood was installed, it was repaired under warranty. And some time after the repair (which I think involved replacing the main controller board), the halogen lamps exploded (thus ruining a meal). I had noticed that these two lamps ran very, very hot, and that did concern me. So the halogens were replaced with LEDs, which ran much cooler. But that led to another observation: a faint glow from the LEDs at night time. I found this strange, but did not get around to investigating it, and put it down to stray capacitance or inductive coupling. I removed the LEDs and replaced them with another brand of LED, and the faint glow also went. This initially appeared to solve the oven tripping problem, and it was put down to cheap LED drivers. But alas, the problem returned. I then replaced the LEDs with halogens, and the problem disappeared. But by now I was thoroughly perplexed. Removing the lights altogether also stopped the oven tripping. Testing the light fittings with a non-contact tester revealed a curious and worrying scenario; there appeared to be mains voltage at the socket with the light turned off, but no voltage when the lights were turned on. A few days later, suddenly it dawned on me that the mains wiring was almost certainly reversed; Neutral was being switched, rather than Active. That would explain the faint glow of LEDs in the night, with perhaps a few milliamps being inductively returned through Earth. When time permitted, and after suitable safety precautions, I removed the covers from the rangehood controller. The controller is a small box about 90 x 60mm, located in the exhaust path, visible when the grilles are removed. Australia’s electronics magazine MARCH 2020 71 Upon opening the cover, my theory was confirmed: Active (brown) was wired to N (Neutral) on the PCB via a screw connector, and Neutral (blue) was wired to L (for Line, which I think is the terminology used in the USA for Active). This is precisely the opposite of what it should be, and confirmed my suspicions. This should have been a quick fix, and the end of the matter. However, upon carefully unscrewing the PCBmounted connector, it seemed to be quite loose. This worried me, given that it was connected directly to the mains. Perhaps excessive force had been used in the previous repair to tighten the screw, or perhaps it was a cold solder joint, or both. So the unit had to come out, which entailed removing connections for the mains, the motor, lights, and an IDC connector for the switch panel. I managed to resolder the connector easily enough. However, I was not happy with the fuse arrangement. It was one of those barrel-type fuses, with a connector on the side and at the end. The soldering on these was, in my opinion, poor. But worse, the exposed metal meant that when the cover was replaced, the 72 Silicon Chip live parts on the fuseholder would be close to the main board and other terminals. That was fixed by resoldering the fuse connectors and applying heatshrink tubing to both exposed terminals. Upon reinstallation, I noticed that the push-on light connector had one wire connected by a copper strand. This was not visible before, as it was shrouded in plastic. I replaced this with a PCB-mounting screw-type connector, as I thought it would be much more robust. The motor connector terminal also had a dry solder joint, like the mains connector. In all, I had to take the PCB out three times to fit new sockets and resolder dry joints. This was all complicated by one “recycled” connector which would not accept a wire. This sounds trivial, but remember that reinstalling the controller requires one to have their head in the rangehood itself, bent over a stovetop, and twisted around. The non-compliant connector was high up and difficult to access. Also, some of the mains wires had been tinned before being inserted into the screw terminals. I don’t think this is good practice, as it necessitates a tighter turning of the screws to obtain a good mechanical contact, and that could be why the connector came adrift. Furthermore, temperature and time can make the solder flow, leading to a loose connection. So I cut off the ends and re-stripped those wires. Eventually, all was reassembled and checked, and everything worked perfectly, even with LED lights. Most importantly, the oven no longer tripped the RCD. But I was unhappy to have to do all this work when the board supplied with the rangehood should have been built to a better standard in the first place. It came in what is supposedly an up-market kitchen fitting. So I surmised that a small leakage current was always flowing due to the reversed polarity of the controlAustralia’s electronics magazine ler. This reversal did not affect the operation of the unit as such, but it did mean that the external lamps were always at mains potential. This in itself was not enough to trip the RCD, but apparently, it was when combined with the high current pulse at oven switch-off. It puzzled me why the unit was incorrectly wired. Perhaps the installer did not know the difference between L and N. Perhaps the installer did know, but didn’t care because it worked anyhow. So beware of incorrectly wired appliances and sockets – I have subsequently heard of people switching off appliances, but not unplugging them, and receiving a shock due to similar wiring problems. Fridge/freezer defrost repair T. M. retired from the refrigeration industry a few years ago and after 40 years working with commercial refrigeration equipment – he knows a thing or two about fridges. He never liked working on domestic units but faced with warm beer, he had no choice but to delve into such a repair. It was an interesting experience, as he narrates... I noticed the temperature in the fridge side of our 10-year-old Whirlpool two-door fridge/freezer wasn’t quite what it used to be. I had set the temperature on the front keypad to 4°C but a measurement indicated that it was actually 9°C! I used a datalogger to check the temperature over the next couple of days and it was gradually getting higher. But the freezer seemed to be working fine, maintaining around -16°C. My first thought was that the motorised damper that allowed the cold air from the freezer into the fridge was not opening. This damper opens/closes to allow cold air from the freezer section into the fridge, maintaining the desired temperature. But on inspection, the damper was fully open but there was minimal airflow into the fridge. Air is channelled from the freezer by a fan located above the finned evaporator, which also circulates air in the freezer. I could hear the fan running so this was not the culprit; it was evident that the evaporator was iced up and after removing several panels, that proved to be the case. But this is a frost-free unit so that should not happen. A frost-free system works by automatically defrosting several times a day, thereby preventing siliconchip.com.au ice buildup. The resulting meltwater is channelled away into a tray in the bottom of the fridge to be evaporated; usually, the hot discharge piping from the compressor passes through the tray to accomplish this. My next thought was that the defrost element was faulty but a resistance check gave a satisfactory reading for a 750W element. The next possibility was the defrost thermostat (or “Klixon”) attached to the evaporator. This has normallyopen contacts when warm, closing when cold. It’s a mechanical safety device that terminates defrosting, should the controlling defrost timer device fail, thus preventing a mini Chernobyl! This also tested good, however, I was somewhat mystified by this encapsulated device as it had six wires coming out of it; usually, only two are required to perform the safety function. I identified the two wires that were open when warm. So what were the other four wires for? These remaining wires were attached to the fan and a multi-pin plug that exited from the freezer compartment and went down to the controller PCB. I thought there might be a problem with this board, as I expect the compressor and evaporator fan to switch off for several minutes each time the unit performs a defrost cycle and I was not observing this. Nor could I measure any voltage across the defrost element. A fault with this control board would be bad since it doesn’t lend it- siliconchip.com.au self to component replacement due to a heavy coating of resin. I searched the web for information relating to the control electronics without success. A large local supplier of Whirlpool spares said that replacement control PCBs were no longer available. I finally located what I thought was a replacement on a UK website; it looked identical but had a different part number and at a price of £260 plus freight, I wasn’t about to chance it being compatible. I decided to modify the fridge instead to control the defrost cycle independently of the fridge electronics. This could be done quite easily with a mechanical defrost timer but I prefer using a programmable controller that gives me more options for controlling the cycle instead of just on/off. I have quite a collection of refrigeration parts in my workshop, including a few such controllers. The replacement seemed a simple task. I would allow the fridge electronics to manage everything other than defrosting. I would program the controller set point to a very low (unachievable) temperature, then I would connect a double-pole relay to the controller’s compressor output terminal. This relay would remain energised at all times except when defrosting. The compressor and the evaporator fan would be wired through the NO contacts of that same relay and therefore the normal operation of the fridge would be controlled by the fridge onboard electronics, which from time to Australia’s electronics magazine time would try to defrost it for a short time with no result. When my controller entered the defrost state, it would open the relay, thus halting the compressor and the evaporator fan and enabling the inbuilt controller defrost relay which was now connected to the element via another two-terminal defrost Klixon I fitted. The original Klixon was left in place due to the mysterious four additional wires; the two wires originally connected to the element were disconnected. The added controller was mounted on the back of the fridge in a position that allowed the display to be easily observed and the buttons accessed for fine-tuning. But when I fired up the fridge, nothing happened other than the display showing the previous set points for the fridge and freezer. It should have entered the alarm mode due to the high cabinet temperature. What was going on? It seemed that the fridge electronics would not initialise after I had made my changes. The only change I had made from its perspective was disconnecting the defrost element from the original Klixon, so I tried reconnected it and the fridge fired up. The control PCB likely performs a diagnostic check at power-up and if the defrost element is open-circuit, it refuses to continue. But hang on a minute, I thought, if it was correctly sensing that the defrost element was OK, how was it not able to drive it? Anyway, I decided that the best course was to try to trick the control board into thinking that the defrost coil was still connected, even when it was not (so that the added control board could drive it instead). I decided to try connecting the coil of a 230V AC-powered relay as a dummy load. Eureka! The fridge powered up and operated normally. I guess the control PCB isn’t that fussy about the actual resistance between those two wires, as long as it isn’t very high; the relay coil has a resistance of around 3.8kW compared to the defrost coil at 80W. But that was enough to fool it and so I left this relay permanently wired up, in the back of the fridge. Everything is now working fine, with my added controller managing the defrost cycle as required. I made a few adjustments over a couple of days to optimise the defrost timing and it’s now working normally. SC March 2020  73