Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG A brush with disaster Dave Thompson It’s that time of year again; the days are longer, the clocks have changed, and there’s more time to contemplate all those jobs I’ve been putting off doing over the winter. While the lure of the interweb, streaming services and reading back-issues of Silicon Chip magazine keeps me office-chair bound, duty calls – and it’s about time I answered! I have several yard projects lined up; not that I’m really into gardening and landscaping mind you, but they need doing, and there’s nobody else in the frame. So I’m the one who has to do them. You’d be forgiven for thinking there’s nothing electronic-servicemanworthy in this type of work, but you’d be wrong; power tools come under that umbrella! I’m into DIY as much as anyone else, and like most engineers and servicemen (and as elaborated upon previously), I like to use the best tools for the job. However, this can create problems, especially when the tools run into certain ‘minor’ problems, such as no longer working. I recently had to sand some timber in the garden. While most people wouldn’t care that the odd fence joint doesn’t match up, as a former furniture creator and hobbyist luthier, that type of thing annoys me greatly. In an effort to make it as tidy as it can be, I fired up my new-ish Bosch 1/3-sheet sander to straighten up some edges I had cut badly. To my surprise, it didn’t work very well. This was unusual, because it had performed admirably in the past, and I’d used it to do a small amount of work during a house renovation I completed a while back. But now, while it powered on, it seemed to labour terribly and there was a distinct electrical-type smell coming from it. You probably know the smell I’m talking about; it’s a type of ‘Eau de burnt insulation’ scent that indicates that something is not quite right. onto my workshop floor a while back while doing another job. I thought nothing of it at the time, and it worked fine afterwards; or so I remembered. I have dropped power tools before; they are built to be tough (frequent power tool users are generally not renowned for their elegance!), so I didn’t think much about it. Even though my tools are not designed for ‘commercial use’, they are Items Covered This Month • • • • The brush arcing investigation Fluke 77 DMM repair DAB radio screen repair A ‘simple’ SMPS repair *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz Past mistakes come back to haunt me I then recalled dropping this sander siliconchip.com.au Australia’s electronics magazine December 2020  61 very well made, and rightly or wrongly, I’ve used most of my power tools in that role over the years. I am now starting to think that the fall mentioned above might have something to do with the current state of affairs. So anyway, I powered it down and made a quick visual inspection of the tool exterior. The problem is that there is not much to see from the outside. The sanding plate at the bottom seemed free enough to move, but something was really stressing the motor out, and I could see the commutator and brushes arcing like mad through the small plastic cooling vents/grilles on the side. The case around the motor also got unusually warm after just a minute’s use – a classic sign that something is not right. As I don’t have an X-ray machine, the only way to find out for sure what was going on was to open the sander up and take a good look at its guts. Fortunately, there were none of those terrible security fasteners in sight; however, the PK screws holding the two halves together were embedded very deeply down moulded plastic channels. To gain access, I needed a very long-reach and thin-shanked number 2 Philips screwdriver, which in a way is a security feature in itself. Not 62 Silicon Chip many people have such a long-reach driver available, but as I do (several in fact!), it was just a matter of removing the screws and separating the case. Tools not made to be serviced This is where things started going a bit wrong. The two sections of the case didn’t want to ‘let go’. They parted ways by about a centimetre, but would go no further. I couldn’t see what was holding it up either, as there is a lot crammed in there, and due to the narrow gap and limited viewing angle, not much was visible. Australia’s electronics magazine I reluctantly used a pry tool to try to coerce them apart gently. I don’t usually like using tools to separate cases; in my perfect world, they should just come apart when all the fasteners are removed. But in this case (hah), it was fighting me all the way, and I had to use quite a lot of mustard to lever it apart. Eventually, the upper part of the case came away, and I could see everything inside. The first thing I noticed was several bits of what looked like one of those Airfix plastic model kits after the parts had been snipped out of it. Also, I found a coil spring lying in the bottom of the case. I was a bit puzzled at first, but it didn’t take me long to figure out from whence the bits came. At the top end of the motor, near the commutator, is a rectangular grid made from injection-moulded plastic. This shaped part includes the brush housing, and this whole assembly is aligned and located into a corresponding moulded void in each half of the case by a small plastic tongue. This is designed to keep the motor assembly properly centred and the brushes aligned with it. These tongues had broken away on both sides, leaving each of the pieces half-stuck in their slots in the case. siliconchip.com.au Marvellous! This alone pretty much ruined the tool, as without the motor assembly correctly aligned, it just wasn’t going to work correctly. Then that familiar serviceman’s muck-up sinking feeling set in as I realised I’d done this damage while prying the case apart. Now I had to scramble to find a solution that would get this almost-new tool back up and running. While potentially show-stopping, this damage didn’t explain why the sander laboured in the first place, so I likely had two problems to solve. Excellent! As if I don’t have enough to do anyway…. Getting on with the job One of the bits lying in the bottom half of the cover was a small plastic pillar, about 1.2cm long and 3mm in diameter. It was smooth on one end and had obviously snapped off on the other. Looking at the remains of the plastic latticework brush housing assembly, I could see another similar but intact pillar holding one of the coil springs used for maintaining brush tension. I could also see where this pillar had broken off the other side of the brush housing, assisted by the tension of the spring. That must have happened in the original fall, and without spring tension keeping that brush in good contact with the commutator, the motor would run erratically, if at all. At least I’d discovered a possible reason for the initial running-rough problem; the good news is that I still had the plastic pillar, the brush coil spring and the various broken pieces of the brush plate assembly. This might be salvageable after all! Pondering plastic permutations The problem with plastics is that not all can be successfully glued together; at least, with the glues I have on hand. Some plastics are too oily, some overly porous and some just too weak once they have been broken from their original, moulded shape. The grids of this brush housing frame measure about 3mm across for the most part, while the plate itself is about 1.5mm thick; why they didn’t make the whole assembly from one solid piece of moulded plastic is a mystery. The saving of perhaps one gram of siliconchip.com.au weight and a cent or two in production costs would make no engineering sense, except perhaps if it were made this way for airflow/cooling purposes. My guess is that like most manufacturers, they wanted to minimise parts cost, maximise profits and build in some obsolescence along the way. By making the internal components so fragile they cannot withstand a drop from standard working height to a floor, or even survive the separation of the case for maintenance, then it makes some sense. I suppose this is also what separates the home/DIY tools from their hardier (and usually much more expensive) commercial cousins. But no amount of theorising would solve my problems. To get this back up and running, I’d have to repair that broken spring holder and the frame it all mounts to. If I had a 3D printer, I could probably sketch it out and print one up, but as I don’t know anything about 3D printing, I have no idea whether a piece I could print at home would be strong enough for the job anyway. I’m no expert, but as dad owned an injection moulding machine for many years, I got familiar with some of the more commonly-used plastics. By the look of it, this was some flavour of glass-filled Nylon. As is typical, there were no markings or part numbers on it, and I couldn’t locate a service manual for the sander either, so I had to guess. In my experience, this stuff usually glues OK with the likes of a good epoxy resin, and as that’s what I had in my glue drawer, that’s what I chose to use. (When all you have is a hammer, everything looks like a nail, and all that.) First, I had to disassemble the remains of the brush holder assembly from the commutator. That meant pulling the bearing off as well, and of course, it was pressed (and likely glued) on very tightly to the armature. It always amazes me how a small job can snowball into something requiring a workshop full of specialised tooling! Fortunately, I’ve been collecting tools for years, so I had what I needed; using an arbour press and various vices and mandrels, I first cracked the glue and then slowly eased the bearing from the shaft without any damage to the armature. If I didn’t have those tools, I’d likely be throwing this sander Australia’s electronics magazine December 2020  63 in the bin and writing off the purchase price of a new one. If anyone needs an excuse to buy more tools, look no further! Feel free to clip out that last paragraph and bring it with you next time you walk out of a hardware store with several hundred dollars in tools you hadn’t planned on purchasing. Anyway, once the brush assembly was free, I pieced it all back together on the bench, CSI: Christchurch style. I only had seven bits to assemble, so it wasn’t exactly rocket surgery, and fortunately, they had all snapped cleanly and fit back together quite nicely. Once I was happy with it, I put a few strips of two-inch (~50mm) masking tape down on my flat melamine workbench and mixed up a swag of longcure epoxy resin (the quicker-setting epoxies aren’t as strong). After painting the ends of each piece with glue, I press-fitted it all together. I held it all as tightly as I could with more masking tape, leaving the gaps clear, and then filled those gaps with what was left of the resin. I left it for two days before attempting to get it off the bench. By then, it was as set as it was going to get, so I removed the tape from the top, then carefully lifted the now-almost-solid brush housing off along with the masking-tape foundation. It came off relatively easily and was dead flat. After removing any remnants of tape, I prepared it for reassembly. One problem left to solve Except, I still had a broken brushspring pillar. The sander wasn’t going to work without that being resolved. The problem I had was that no amount of glue (that I had or could fit in the limited space) was going to be able to hold that spring pillar in place. I could try gluing it, but my experience told me that as soon as I tensioned the coil spring, the pillar would just pop off. I might be able to get it stuck onto the plate strongly enough with a decentsized blob of glue, but this would foul the operation of the spring, so I needed something more robust but which still allowed the spring to do its job. I ended up drilling a 2mm hole through the brush plate, precisely in the centre of where the pillar used to stand, and literally bolted in what looked like an old tape machine capstan screw from my parts box. The shaft height and diameter of the ‘screw’ 64 Silicon Chip was almost the same as the broken plastic one, and there was enough thread protruding underneath to fit a decent nut and washer. There is nothing directly below that part of the plate anyway, so I had plenty of room. I put the brush into the holder and slipped the spring onto the screw first before mounting it, and a dab of Loctite on the threads before tightening the nut up should ensure it doesn’t move in a hurry. I pondered whether I should do the same thing to the other pillar as well, but I’d tempted the servicing Gods well enough already. The first acid test was to reassemble the tool and hope everything fit into the case with the ‘mods’ I’d made. It was a bit finicky putting the brush plate back onto the armature and refitting the bearing with it in place, and I was very careful not to put any stress on anything lest it all come crashing down. It went together OK, though. I used an old trick (which no doubt everybody else uses too) to keep the brushes out of the way while I installed the commutator/armature assembly. I pulled the brushes out of their holders about halfway past the springs, then used the tension of the springs to hold them open while I slipped the commutator between them. Once in place, it’s then a simple matter of prodding the brushes inwards a little until the springs snap back behind them. With all that now in one piece, it was time to re-fit it back into one half of the case. This is where it could all come unglued (ha ha!). I was reasonably sure the brush plate would fit, because it was glued back together flat, but you never know until you try. I also knew that if I pushed the now re-glued locating tongues into position, I might not get a second chance at repositioning them. After a bit of give and take, I managed to get the assembly sitting flat into the case. Halfway there! Making sure everything was in the right place, I got the top half of the case and gently positioned it until I was reasonably sure it was in the right place. The next move would make or break the repair. After lining it up, I gradually applied pressure and finagled everything into place; the cable clamp boot, the wiring pushed into the channels, and the case perfectly aligned to the Australia’s electronics magazine bottom half. With a final push, it all went together. I installed the screws, held my breath and plugged it in. It powered up and with a bit of arcing (I’d probably put the brushes back in the wrong holders), it worked a treat. After an hour’s work, it is sanding as well as it ever did. Phew! The job was done! Editor’s note: power tools with brushless motors are becoming more common and are now available at reasonable prices. Besides avoiding this sort of problem (you can’t have a brush spring detach if you don’t have brushes!), they also seem to have better power, less noise and more battery life than their brushed counterparts. I am impressed, and recommend you take a look next time you are tool shopping. Fluke 77 digital multimeter repair J. R. of Tauranga, New Zealand, had to use a fair bit of creativity for this repair, as the replacement parts he needed were not available. We think his solution is ingenious... I’ve had a Series I Fluke 77 DMM since the early 1980s. It has given me excellent service over countless hours. Unfortunately, I managed to connect it across a 2kV supply, and it was no more. Buy another? Flukes are very expensive, and I can probably no longer justify a new one. Throw it out and replace it with a cheap clone? Perhaps the best option, but nostalgia has its place even in the minds of dispassionate, ruthless engineers. So I thought I would see if I could fix it. Fluke multimeters have input protection which many times prevents expensive damage when oopses like this happen. 80-series meters employ the same concept as the 70-series, as do some of the 20-series. The current circuits are properly fused, and the volt/ohm ranges are protected by a combination of series fusible resistors and either a pair of high voltage MOVs or, on early models, a spark gap which arcs over at around 1500V. In either case, the resulting highbut-limited current blows the fusible resistor and open-circuits the input before anything else happens. Mine has the spark gaps. Indeed, that is what had happened. Spark gap E1 had been destroyed, and resistor R1 was split along its length and measured open-circuit. siliconchip.com.au I had repaired another Fluke 77 multimeter with the same fault around 2013, and at the time, the special resistor and spark gap together cost NZ $69.00 from Fluke. So I while expected the fix would be worthwhile, I knew it wouldn’t be cheap. I quickly found out that the genuine resistor from the Fluke NZ repair agent would be over NZ $60 by itself, but the spark gaps were no longer available. A Google search confirmed this is a well-known problem, with lots of people asking where to get them and none showing up anywhere, even on eBay. It seems like when you blow up the protection, you now have to throw away the meter! I then thought about converting the input circuit to use the two MOVs that the later models use instead. However, I found from Fluke that the MOVs were NZ $55 each, so without any labour cost, a repair would be nearly NZ $200 once freight and packaging were added. Anyway, I couldn’t fit the MOVs without butchering the PCB. Even third-party MOVs on eBay ostensibly meeting the Fluke specification were also scarce and expensive. It is essential to use the correct fusible resistor for R1, because it limits the energy in a fault and then opens, so preventing real damage and perhaps even injury. Any old 2W resistor looks pretty much like the real thing, and I found quite a few sellers on eBay offering “Fluke R1 resistors” or “fusible resistors for Fluke” for a few dollars each. But on close examination, none of the resistors being offered seemed to be anything but ordinary 2W metalfilm resistors which the sellers could have been buying for a few cents. The manufacturers do not state the fusing characteristics of most resistors at all, and they are typically designed for 300V whereas the correct resistor has a short-time withstand of 1000V. Before giving up, I had one last look on Google and ran across a chap who repaired meters, and who seemed pretty authoritative. While he had no solution for the spark gaps or MOVs, he had done the homework and found that at the time of posting (2015), one type of resistor was still being made by TT/IR which was fusible and had the right short-term voltage rating. He stated that it had been used by Fluke in the past, although they had superseded it with an upgraded version, siliconchip.com.au An old Series I Fluke 77 DMM from the 1980s. The destroyed spark gap E1 and opencircuit resistor R1 adjacent. The input protection section of the Fluke multimeter circuit. Australia’s electronics magazine December 2020  65 New rectangular spark gaps were made using copper wire and polyester resin. The spark gaps removed from the moulds. A hobby mill was used to cut the 0.008in slits for the spark gaps. A new resistor was fitted for R1 and the spark gap placed as well. The spark gaps were tested with a high-potential (hi-pot) insulation tester. They were consistent and arced between 1550-1650V. 66 Silicon Chip Australia’s electronics magazine he could not locate it from third parties. I found the manufacturer’s specs and confirmed what he said. Even better, I found they were still available in 2020 from Mouser. So I had a source of safe resistors but no MOVs or spark gaps. The Fluke 77 has two spark gaps, but only one was damaged. I had a close look at the clean one and found the air gap to be just under 0.2mm (actually 0.007 inches, ie, 7 mils). I found a website which had a credible relationship between air gap length and strike voltage for small air gaps (and no it’s not 30kV/cm!) and tried a few sums. A gap of 0.007in predicted a voltage just a little under the 1500V specified by Fluke, so given things lined up, it looked as if there was probably nothing magic about the gap, meaning it might be worth trying to make one. I could buy a 0.008in (0.2mm) slitting saw. Using the website formula, I found that the arc voltage wasn’t very sensitive to the gap; the predicted voltage for a 0.008in gap was a bit over 1500V. So I ordered a saw and the Mouser resistors, which are TT/ IR SPH1001J 1kW 2W wirewound fusible types. I made the gaps out of copper wire and polyester (fibreglass) resin that I had in the shed. The new gaps are rectangular: 8mm wide, 4.5mm thick and 10mm high and more-or-less fit into the space occupied by the original oval commercial ones. I machined a few simple mould shapes into a polyethylene chopping board. I thought the resin would release easily from that plastic, but I sprayed the holes with Teflon garage door dry lubricant to make sure. I used Blu-tack to hold the wires in the right place and poured in the resin. Once set, I removed them from the moulds (easy) and eventually got rid of the Blu-tack (hard). I won’t use Blutack for this sort of job again. I then put them into the little hobby mill and cut the 0.008in slits. I found the gaps were a bit bigger than 0.008in – they were actually around 0.009in. This was probably due to run-out on the saw arbour, or the saw itself. But the web formula indicated they would still meet the spec. I then tested them with a hi-pot insulation tester. They were pretty consistent, arcing at between 1550V and 1650V. The hi-pot has current limiting, which I set to 20µA, so the gaps were siliconchip.com.au not damaged or altered by the test. Since Fluke’s specification for the original spark gaps is 1500V ±20% (1200-1800V), the home-made ones are satisfactory. I tested the remaining good original spark gap, and it arced over at 1500V. I then fitted a new resistor for R1 and one of the home-brew spark gaps and reassembled the meter. When compared with an HP bench voltmeter, the Fluke 77 is as good as ever. No, I don’t propose to deliberately test the over-voltage failure mode! All in, the exercise cost me about $70 NZ, of which FedEx got a fair chunk for shipping $5 worth of resistors from Mouser. The repair (apart from my labour) was therefore economic; nostalgia has its place it seems. I now have nine spare resistors and six extra spark gaps, a mould plate and a slitting saw in case I do it again to my meter or come across someone else who needs the same fix. Radio LCD screen repair G. McD., of Jindalee, Qld had some spare time over the Christmas break and decided to spend some of it repairing the faulty LCD screen on his wife’s radio. Here is what happened… My wife bought herself a brand new DGTECH BC76183 DAB/FM digital radio soon after digital radio became available to listeners in the Brisbane metropolitan area. It served her well on a daily basis until the LCD screen suddenly went blank. The idea of binning it and purchasing another had crossed our minds. But first, I decided to have a closer look at it to see if I could repair the screen as she was otherwise happy with the radio. Before pulling it apart, I ran some quick tests to try to narrow down the likely cause of the fault. I switched on the radio and sure enough, the sounds of ABC Jazz came through as impressively as it did when the radio was new. Then I reached for my torch and shone it directly at the screen. As if by magic, I could once again read the name of the station as well as details of the tune being broadcast. This indicated to me that there was nothing wrong with the LCD screen itself or the wiring connecting it to the PCB. The fault lay with the LCD backlighting. It was now time to explore the innards of the radio. After turning it off and unplugging the mains lead, I unsiliconchip.com.au did two screws on the back cover as well as the two marked screws hidden beneath rubber pads on the underside of the enclosure. The back cover could then be carefully prised open, after gently pushing the earphone socket clear with a small screwdriver. This exposed the inside of the radio, but the rear cover remained connected to the main body by the speaker wires. I unplugged this and three other wire harnesses from their sockets, allowing the two halves of the enclosure to be separated. I now had access to the PCB on which the LCD screen was mounted. But as is usually the case, I couldn’t get to it as the screen was on the underside of the board; I would need to remove the PCB. The PCB was held in place by four white plastic retaining clips which needed to be swung clear. Next, two of those that were marked needed to be unscrewed as they provided additional stability to the pushbutton arrangement. Then four screws on each corner of the PCB were removed and set aside. After pulling off the volume control knob, the PCB came clear and turning it upside-down exposed the LCD screen. I unclipped the white plastic cover that butted up to the right-hand side of the screen; this housed the LED that I suspected to be the problem. To verify this, I reconnected the four wire harnesses, plugged the power cord into the mains supply and turned the radio on. Once again, the radio came alive but the screen remained defiantly blank. I turned on my DMM, which had been set to read 20V DC, and probed the two leads of the unlit LED; I obtained a reading of 2.65V. Now that I had verified there was voltage across the LED, I assumed that the fault lay with the LED. Not having a spare white LED in my spares storage, I paid a quick visit to Jaycar and purchased the closest I could find to the original, Cat ZD0192 for $1.65 each. On arriving back at the workbench, I discovered that the replacement LED was longer than the original and wouldn’t fit into the space provided (the original LED was flat-topped and not domed as the one I had just bought). This was verified after I had removed the offending part, cleaned up the two through-holes with solder wick and tried the replacement for a fit. Australia’s electronics magazine So it was out with my trusty modeller’s knife, with which I began shaving away at the inside of the housing until there was just enough space for the replacement part to squeeze in. The new LED was soldered into place, the leads trimmed and the lid of the housing clipped back into place. All that was left for me to do then was to reassemble the radio, taking care not to crimp any of the wire looms during fitting, and run a “smoke test”. To my delight, the screen lit up immediately upon switching on, with everything operating as it should. My wife can now look forward to many happy years of service from this excellent radio. A ‘simple’ SMPS repair R. S. of Fig Tree Pocket, Qld, has a lot of experience repairing switchmode power supplies. This turned out to be one of his simpler repairs, although not necessarily the easiest to diagnose... I had a problem with a Bosch 30V 0.5A battery charging plugpack for a cordless vacuum cleaner. It stopped producing any output, so the vacuum cleaner battery was not charging. This charger uses an On Bright Electronics OB2358 IC in an 8-pin DIL package. This IC has an inbuilt 600V FET, which connects directly to the primary of the flyback transformer. The OB2358 was not starting up, and therefore not generating its own supply voltage, via an extra winding on the flyback transformer. For some reason, there is a surfacemount zener diode on the board from the feedback pin 3 of the OB2358 to ground pin 8. This was leaking enough current to hold pin 3 low and prevent the circuit working. Removing the zener diode got the plug pack working again. I cannot see any reason for the zener diode; it is not shown in the typical application circuit in the IC data sheet. Editor’s note: probably to protect the IC from damage if the feedback mechanism stops working for some reason or the output is externally pulled high. Luckily, this plugpack can be split open without damage. Usually, they are glued together so well that the case breaks when you try to open them. One other note: the feedback circuit on the low voltage side uses a 6-pin surface mount IC marked OD=28X instead of a TL431. Can anyone identify this part? SC December 2020  67