Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG If it isn’t one thing, it’s another Dave Thompson Sometimes, even when there are no customers lining up, work comes along anyway. It isn’t always welcome, but when your tools go down, you have to fix them. It doesn’t help that I’m afflicted with the Serviceman’s Curse, so I’m allergic to paying for replacement tools when it’s possible to (uneconomically) fix them! Over the past year, plenty of local businesses have folded; there simply isn’t the customer traffic to keep the doors open any more due to lockdowns and general economic malaise. While our overall revenue has dropped, as you would expect with a lot less work coming in, the silver lining is that I finally have some free time to get onto those little jobs that I’d been putting off. Those of you who live the rock and roll life of a serviceman know that sometimes things don’t go according to plan. An anticipated five-minute job can easily turn into a two-day mission in the flash of a shorted battery connector or a clumsily-placed screwdriver. That sort of thing doesn’t happen to me, of course! But I do hear rumours that it happens to other, less-careful people. The first small job created itself when I went to use my soldering station, and the pencil was still cold 10 minutes after I switched it on. The astute among you will know soldering irons are meant to be hot, so the fact that I could hold on to the wrong end need, of it without being burntyou told me that something was up! The pencil connects to the soldering station using one of those multi-pin screw-on plugs, sometimes called a GX-16 series connector. I removed and re-connected it, and it seemed sound, so I guessed that the pencil’s element had gone open-circuit. Confirming this theory proved to be more difficult than I imagined, mainly because the pencil itself appears to be a moulded unit. Everything is set into it at manufacture, and it cannot be disassembled to reveal the innards. The cable stress reliever at the bottom can be prised out, but the element appears to have no means of being removed, other than by cutting into the pencil’s plastic body. This makes them inexpensive to manufacture, but not great for repairs. I think they expect people to throw away the dead pencil and buy a new one. The problem is that I’ve used this pencil for a while now and having just ‘broken in’ a new tip, it is perfect for the work I do. To bin it without at least trying to repair it you want. would be, well, frankly against my serviceman’s code! So electrical checks would have to be made via the GX connector. I For full details on how to enter, drawing & rules head to: jaycar.com.au/dmax-jaycar Items Covered This Month searched for circuit diagrams online for my model. Once located, my mul• It’s always the other thing timeter confirmed there was no resist• Coin counter repair ance or continuity through the element from any of the pins, let alone the • Alternative security systems designated ones, which told me all I • LED rose garden repair needed to know. It was dead! • Electric fence energiser repair get that thing win that thing SPEND $50 OR MORE FOR A CHANCE TO WIN AN ISUZU D-MAX *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz siliconchip.com.au Borrowing a spare Fortunately, I have a spare pencil. But when I say spare pencil, I mean spare soldering station. While it is very much like my usual one, I originally Australia’s electronics magazine March 2021  61 * purchased it as a desoldering station. My faulty unit has a temperatureadjustable soldering iron on one side and a hot air wand on the other. In contrast, the ‘spare’ station has the same soldering iron setup, with a temperature-adjustable desoldering gun and suction pump arrangement on the other. The desoldering ‘gun’ on this station looked great on paper, but doesn’t work well in practice. The ‘real’ version this unit was copied from might work very well, but this one doesn’t, at least for me. It has one of those pistol-grip style handles, with a hollow, heated tip. Pressing the ‘trigger’ on the handle activates the suction pump, so I can theoretically hoover up the molten solder. Sadly, it is useless; it wouldn’t suck the froth off a lager. Also, the element won’t ‘take’ solder, similar to the metal on some cells; the solder simply doesn’t stick to it. I’ve tried ‘seasoning’ it with various solder pastes to no avail. So trying to heat a solder joint is a lesson in frustration. The bottom line is that I’ve never gotten it to work, which is why it sits in the corner of the workshop. I expected it to work as well as the other station, but it just doesn’t. I’ve long accepted this and have moved on. On the plus side, the integrated soldering iron has almost never been used. I simply unscrewed/unplugged that pencil and installed it on my other station, which immediately resolved my cold-tip issue. However satisfying this quick-fix may have been, I still had a dead pencil, and by extension an incomplete soldering station (useful or not). And as a serviceman, that bothered me. A quick search on AliExpress revealed that a replacement element was available for just a few dollars. Or, I could get a whole new pencil and cable/connector for only a few bucks more than that. Even better (for a tool junkie), a new pencil plus two spare elements could be had for around the same money! You already know which option I went for. Of course I am going to try to fix the broken one. Now I know what you are thinking; did I fire six shots or only five? Oh no, sorry, wrong script. You are thinking that if the pencil is moulded and cannot be disassembled, how will I replace the element? 62 Silicon Chip Well, I’m glad you asked as it’s a good question! Anything can be disassembled with the right tools and the right attitude. A hammer tap here, or a Stanley-knife blade applied there, or even a junior hacksaw placed just so can achieve amazing results. These apparently-sealed devices can be opened, repaired and glued almost seamlessly back together without anyone (but us!) knowing about it. I used to watch Dad opening moulded-plastic power supplies using a carefully calibrated hammer tap, and when the thing was repaired, he simply glued it back together. That was in the days when such power supplies cost a small fortune and were worth repairing. That obviously isn’t the case today; I literally have cartons full of these supplies that will likely never be used, but the philosophy of the repair still stands, and I thank my Dad for passing that on to me. I wouldn’t usually do this for a paying customer, but to repair my own tools or appliances, I’ll give anything a go. If I ruin the pencil, I’m out a few bucks, but it’s the serviceman’s creed and the principles of repair that compels me to at least try. I’ll let you know how it goes! Some more light work To be realistic; many repairs are simply not worth the cost. Recently, I had an LED ceiling light stop working. It had only been installed (by me) a few years ago and hadn’t had a lot of use. Maybe a few minutes a week of ‘on’ time, if that. These lights are commonly called “UFO” lights because they look like a flying saucer. But they are different than downlights which require cutting large circular holes in the ceiling. These ones come with a fitting that simply replaces the existing battenmounted socket we are all familiar Australia’s electronics magazine with, and the new UFO light slots into place, hiding the socket. That makes retrofitting ceiling lights a breeze. We did our entire house with these, and it was an effortless job to convert all our incandescent lights to LED versions, without a bunch of tools, mess and headaches. And they’ve been great; the light is better, brighter and more economical than our original lamps and fitting them was super-easy. I’d installed a dozen others in the house, and they’ve all done a whole lot more work. Why this one died is likely down to the fact that 10% of these lights will fail in the first few years, and that’s an acceptable failure rate for modern manufacturers. Retailers simply replace the unit and chuck the dead one in a skip. Now, if you are afflicted with the Serviceman’s Curse, you know you can’t just throw something out without at least pulling it apart and looking at it. So I had to take it down and open it up. At the same time, I ordered another one, because even though it is seldom used, we need a light in that spot and at only $11, replacing it is the obvious solution. It also means that I could work on it at my leisure without hearing “get that bloody light working!” The housing popped open without much hassle. Inside is a 200mm diameter PCB with a bunch of surfacemounted LEDs soldered to it. Another small PCB is mounted in a central cutout, containing the LED driver. A sniff with my serviceman’s nose told me something had electrically given up the ghost. The usual suspects My first step was to remove and check the two electrolytic caps that dominate the driver board. I fired up my soldering station and went to try to desolder those two caps, but the pencil was stone cold. I think you already know how that went! After resolving that, I removed the caps and checked them with my trusty Peak Electronics ESR meter. One measured 15W, which is on the high side. The other one was also high, but not as out-of-spec as the first one. I replaced both and reassembled the board into the light. I powered it up with my non-Variacbranded Variac, and the LEDs sprang into life. Success! It is now reinstalled siliconchip.com.au and happily illuminating our spare room. So it was well worth having a go, and when the replacement I ordered finally arrives, I’ll have a spare. I must be cursed It’s never fun when the tools we rely on to do our job don’t work. Last week I fired up my computer to write this article, and my machine wouldn’t boot. The old saying is that a plumber’s pipes are always clogged, and while I’m not sure what that means, I’m pretty sure it applies to me! My main computer is a monster that I built 11 years ago, so I’ve been reluctant to upgrade it. That’s because it was still going very well, played all the high-performance games I ever wanted to play, and it has always been there for me. For it to fail to boot up one morning was quite devastating. While I’ve always tried to make these servicing stories non-computer-centric (as it is a dull trade), some readers might find it interesting. If a machine doesn’t boot, I usually start by removing everything but the absolute basics to get the motherboard up and running. In this case, that still resulted in no boot. I then started removing and replacing RAM, and suddenly, I had a POST (power-on selftest) screen. I replaced the single stick I’d left installed one-by-one with the three others, and with the third, the machine didn’t boot. Leaving that out, but with the other three sticks installed, the machine booted happily. After 11 years, one stick of RAM had failed, and that broke everything. You just cannot take things for granted in servicing. printer had broken. The bureaucracy she works for had said this 12V device was “too dangerous to use” in this state! That sounds to me that this is a statement from someone with more ego than knowledge. In any case, the solution was simple. It took me an hour or two to design a new cover in OpenSCAD. I was then able to 3D print the cover, and the deadly 12V inside was safely locked away. The next coin counter fault was a problem common to many cheap devices: the front panel membrane switch developed a fault, so the PRINT button no longer worked. Pulling it apart was easy enough, with just three countersunk self-tapping screws holding the upper case in place. I did have to cut one wire tie; this prevented the rotating coin counter mechanism inside catching on the multi-core wire connecting the front panel. If I had designed it, I would have used a reusable clip/U-channel to position the wire safely. But this worked, even if it did make disassembly (and reassembly) a bit more of a hassle. Next, I unscrewed the front panel PCB, unplugged the front panel membrane switch connector and very carefully removed the complete front panel membrane switch assembly. Then, using a thin blade (the knife from a Swiss Army Card works really well), I separated the two halves of the membrane near the PRINT button. Unsurprisingly, one of the conductive tracks was open circuit. To determine if this was the only fault, I ran some leads to the correct pins on the PCB and found that the coin counter beeped when I shorted these two wires together, simulating a press of the PRINT button. Well, I had found the fault, but now I had to figure out how to fix it. The likelihood of getting a replacement front panel was low-to-zero. This was a major problem for the canteen, as the coin counter is used every (work) day and having to count coins manually meant a lot of extra work. Then I had a thought: as only one button was faulty, why couldn’t I just add another switch? Initially, I was just going to drill an extra hole somewhere and use a standard 6mm pushbutton switch; that would work but look a bit ugly. But then I remembered I had some tiny switches that measured 3mm x 6mm and were only 1mm high. I purchased these as spares for repairing car remote controls. After a bit of measuring, it looked like I could shoehorn one of these switches to fit into a hole cut in the plastic case underneath the existing PRINT button, so that’s what I did. For speed and simplicity, I drilled two 3mm holes next to each other and Coin Counter repair G. C., of Salamander Bay, NSW has found (like many others) that it can be easier to replace a cheap failed part with a higher quality alternative than it is to fix the original part. In this case, it was one of those horrible membrane buttons. His solution means that it’s unlikely to fail again… My daughter runs a canteen for a large organisation and even with email ordering, it’s still necessary to count all the cash each day. They use an unbranded coin counter that’s simply labelled “Coin Counter”. I have fixed it previously; it was a totally unnecessary repair, in my opinion. The small cover for the docket siliconchip.com.au The coin counter and small docket printer. Australia’s electronics magazine March 2021  63 filed until the switch fitted inside, then used double-sided tape to re-glue the membrane switch to the case. I pressed the tiny switch into its new home and tweaked its vertical position. When the switch activated reliably when the PRINT button area was pressed, I added a couple of drops of super glue to make it permanent. I then soldered thin (wire-wrap) wires to it, then soldered the other ends to the PCB tracks. Happily, the coin counter is back in service and working perfectly and, even better, it’s externally unmodified – the only difference is that now you can feel a click when the PRINT button is pressed. I would go so far as to say that after this repair, it’s better than new! Alternative security systems R. M. of Scotsdale, WA, found out that there are really cheap security systems, and really expensive security systems, and neither is all that appealing. Luckily, he found a middle ground... Our community shed needed better security as several of the keys had disappeared over the years. As the only member with any electronic knowhow, I was volunteered to search out a suitable replacement. I went to the biggest security shop in the local town and made enquiries. They suggested an RFID system priced at around $3500. As the shed is a small non-profit organisation, I knew that we couldn’t afford that. After some discussion, the committee authorised a budget of $200 and let me loose. 64 Silicon Chip Australia’s electronics magazine I tried sourcing a cheap (~$50) RFID unit from overseas via eBay, but shortly after I hooked it up and got it working, it failed. So I had to send it back and get a refund. During wanderings through Google, I had come across a more elaborate (and more expensive) four-door controller from Jaycar. After the “fleabay” controller failed, Jaycar (bless ’em) put it on special and dropped the price by a good $50. The committee agreed, and we were soon in possession of a nice sturdy box of tricks that actually worked. That left the actual door strike. Electro-mechanical striker latches require the door to be sturdy and close-fitting. But our big metal shed ain’t all that flash. The door (square steel frame, steel sheet) swings on one of the portal trusses. And when the wind blows or the sun shines, there is a perceptible movement of a few millimetres between door and frame – enough to make the standard latch system unreliable. However, at the back of my farm there’s an old ute with two solid door locks. I nabbed one which is now doing excellent service on our shed. With a 12V actuator to pull the release lever, it has enough slop to handle the geometry changes, and plenty of strength to hold the door shut. For our little installation, the supplied software is overkill. It’s designed to control many doors of many departments and keep records of all the workers’ movements. It is mind-boggling in complexity, and the instruction man- siliconchip.com.au ual is a masterpiece of confusion and poor translation. For example, on the circuit layout, it shows two pins labelled “J9: Joint of closing door by force”. I contacted Jaycar to ask what this meant, and eventually, the answer filtered back: it is a disable input. Close the circuit (joint of), and all doors would stay locked (closing door by force). It makes me wonder if they created the manual using machine translation! I spent many hours decoding the manual and experimenting with the software. The process of registering each user is vital, and there are two ways one can do it. One is a bulk entry method, and the other, more detailed, allows individual entry. I chose this way as we wanted to enter our members’ details one at a time, but I could not get it to work. I thought it must be my fault, so I summoned the local PC expert. He went straight to the “bulk entry” system, and it worked! Don’t bother with that other way, he said, it’s no good. So, finally, we have a working secure entry setup. The total cost was around $370 with the backup battery and trickle charger. Now if a member leaves, or doesn’t cough up the yearly subscription, we can simply click a button and forbid entry. He’d have to resort to removing a sheet of corrugated iron with a screwdriver to get in! LED rose garden light repair B. P., of Dundathu, Qld has some unusually ornate solar garden lights, so siliconchip.com.au when they started to fail, it was worthwhile taking the time to fix them... Some time ago, my wife was given a white LED rose garden light. We noticed on the packet that there were also yellow and pink LED roses, so after finding out that a nearby discount shop sold these, we got a yellow one and a pink one too. Later, my wife received another two LED roses, pink and yellow. However, this new yellow rose only had a single flower, whereas the others had two or three flowers each. We noticed that in the mornings, it was only the single yellow rose that was still lit; the other four roses were no longer lit due to having more LEDs (and presumably exhausting the battery charge faster). After a while, we noticed that both the white rose and the original yellow rose no longer lit at night. I had a look at the yellow rose to see what the problem was. After removing the four screws from the bottom of the small box containing the solar panel and battery, I could see that water had leaked into the box, causing the positive battery connection to become rusty. The wire had also broken off it and the battery terminal was rusty as well. I re-sealed the wires coming from the solar panel properly with hot melt glue, then cleaned up the battery terminal. It was so badly rusted that I decided to clean it and coat it with solder while I was soldering the wire back on. I also tested the battery and found that it was still OK. Australia’s electronics magazine With the battery refitted, the rose still didn’t work. I measured the voltage on the battery terminals with the solar panel lit and got a reading of 2.18V, so the solar panel was charging the battery, but some other fault was preventing the LED from turning on at night. I removed the circuit board and checked the switch, which was still functioning correctly. The circuit board looked clean, with no corrosion. I then realised that I had a spare circuit board from one of our garden lights that had been run over and smashed by a courier, so I decided to use that to get the LED rose working again. This board did not have a switch, but I didn’t think that really mattered. I disconnected the original circuit board and wired up the replacement circuit board, which was quite easy, as all the connection points on both PCBs were marked S+, S-, B+ and B-. The wires from the LEDs were too short, so I just used some scrap wire to extend them. When I tested the rose, it still didn’t work, but I noticed that the YX8018 IC was bent over, so I straightened it and checked the bottom of the circuit board. I found that the solder joints were cracked, so I re-soldered them, but the rest of the PCB was OK. Now the rose worked, so I sealed up the hole where the switch used to be mounted with hot melt glue. With the yellow rose now reassembled and working, I took a look at the white one. I found that the same water ingress problem had affected this light. After re-sealing the wires from the solar panel, I found that I could use the battery connector from the smashed garden light to replace the rusted one as it was the same size. Once cleaned and reassembled, the white rose now also worked again. The next morning, I was outside before dawn, and I was amazed to see that the yellow rose that I had just repaired was still lit brightly. I put this down to the more complex circuit in the garden light PCB that I had transplanted, which is apparently more efficient than the simple circuit in the original LED rose. As can be seen in the photos, the ‘basic’ rose PCB consists of just one YX8018 IC and an inductor, whereas the better garden light PCB has both of these plus a diode and capacitor. March 2021  65 At left is one of the ‘basic’ LED rose PCBs which had started to rust, while to its right is a superior garden light PCB. This makes quite a difference to the efficiency of the circuit. I have also repaired several other garden lights. Two had bad solder joints on the LEDs, and I also had to replace the RGB sequencing LEDs in several lights when they malfunctioned (the blue elements failed). I found some seven colour LEDs on eBay, which sequence red-green-bluewhite-green-pink-warm white-repeat, as the three-colour RGB LEDs were harder to find and more expensive. I’ve also replaced the 150mA Liion cells in all our garden lights with new “1000mAh” (probably actually 400mAh) cells, so they now last all night, as long as they have sufficient sun exposure during the day. Even though these lights are not expensive, it was still worth repairing them, as it saved some money and saved them from landfill. Electric fence energiser repair K. G. of One Tree Hill, SA, has repaired quite a few electric fence energisers over the years, but this one posed some unique challenges... I’ve written about the repair of electric fence energisers before (July 2015; siliconchip.com.au/Article/8707); every now and again, I get one from the local fodder store owner to see if I can fix it. These devices apply a short, high voltage pulse to a bare galvanised iron wire running along the fence, supported on insulators. They are used to control the movement of stock. Most operate from the mains, but some models are powered from 12V DC, generally with a solar panel to keep the battery charged. The Earth side of the energiser output is connected to three or more Earth stakes spaced out by a few metres, as recommended by the manufacturer. The pulses have a typical duration of about 30µs and the region of 5-8kV. The pulse repetition rate is about 1.3 seconds. The main difference between units is the energy in each pulse (measured in Joules). This Gallagher model MBX1500 was made in New Zealand and is the largest I have come across yet. Its pulse energy is 3J, and it is suitable for fence runs as long as 94km! A touch on the fence wire would be excruciating, but not particularly dangerous. Testing the unit on the bench brought some low-level intermittent clicking instead of the regular pronounced clicks every 1.3 seconds. This model is capable of being run from the mains as well as 12V DC; there is 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 car electronics. 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. 66 Silicon Chip Australia’s electronics magazine a compartment in the rear to hold a 12V, 7Ah SLA battery in addition to the mains cable and plug. The two halves of the case come apart after removing six deeply recessed selftapping screws. Fortunately, they are normal Philips head types rather than one of the many ‘security’ screws used these days. Inside was an offline switchmode power supply (SMPS). This, or the 12V battery, fed into a DC-DC converter stepping the low voltage DC up to several hundred volts to charge the main capacitor. Then there was the circuitry associated with the pulse transformer. There was also a second PCB containing four LCDs plus a 40pin PIC microcontroller. In its working condition, the unit displayed the output pulse voltage and energy, plus two other parameters which didn’t seem to apply to this unit. So it was quite a complex unit with many SMD components down to M2012 (0805) size. When fault-finding electric fence energisers, I first test the large pulse capacitor(s). There are two in this unit, one 30µF and the other 6µF, both rated at 1200V. After ensuring they were discharged, I measured their capacitance with my Peak Component Analyser. The larger one was 13µF and the other 3µF, both well down on their original values. This is quite common with these components, but such a drop should only reduce the energy of the pulses, not stop the unit working altogether. Despite this, I ordered some replacements from a local Adelaide firm and carried on with my testing. The other component to test early siliconchip.com.au is the main SCR which discharges the capacitor into the primary of the pulse transformer. This appeared to be OK as the voltage across the larger pulse capacitor was a couple of hundred volts and fluctuating with the intermittent clicking. I then measured the voltage out of the offline SMPS which I found to be only 4-10V with fluctuations, instead of a steady 12-15V. I fed 12V DC from my bench supply to the SMPS output with the mains input disconnected. The energiser started working normally, with believable values on the LCD screens. Looking at the SMPS more closely, I noticed a large electrolytic capacitor with a small bulge in the top. It was the 100µF 40V filter capacitor on the output of the mains bridge rectifier. I removed the capacitor and tested it on my Electronics Australia ESR Meter. The reading was about double the typical value shown on the front panel of the meter; not a show-stopper but it needed to be replaced. As I didn’t think this was the main cause of the problem, I refitted the old one to continue my search for the real culprit while I ordered a new capacitor. It wasn’t available from the local outlets as it was a low-profile device 30mm in diameter and only 20mm high, so it took a few days to arrive from Perth. siliconchip.com.au The primary active device in the SMPS was a power Mosfet with an integrated control circuit in a TO-220 package, riveted to a small heatsink. It’s coded TOP227Y (TOPSwitch-II) and is made by Power Integrations Inc. This makes for a much-reduced component count as it contains the oscillator, PWM modulator, voltage reference plus all the protection circuits and the power switching Mosfet. The circuit used in the energiser turned out to be very similar to an application circuit shown in the data sheet (shown below). As the TOPSwitch device had so much of the circuitry in it, I decided replacing it would be a worthwhile punt. The only stock I could find was from Digi-key. I ordered two of the devices as the postage cost from the USA was more than the device itself. I can’t fault the delivery time; I placed the order online on a Tuesday, and it arrived at my local PO on the following Monday. Unfortunately, replacing the TOPSwitch device didn’t help my problem except to eliminate it as the source of the fault. Referring to the figure, I measured R3 and C5 on the control input of the TOPSwitch. R3 was fine, but C5 was rather low in capacitance, about 30µF rather than 47µF. I didn’t pay much attention to that at the time. Australia’s electronics magazine One point of difference between the MBX1500 circuit and that from the data sheet is the use of a TL431 voltage reference rather than the zener diode shown. I replaced the TL431 and also the optocoupler connected to it, but neither changed the result. I also tested the low-voltage DC rectifier (D2) and filter capacitor (C2), the latter for both capacitance and ESR. Both were OK. I was by then wondering where to turn next. I read the data sheet more carefully and realised the part played by the cap on the control input was quite important. So I removed the 47µF capacitor and measured its ESR. The typical value is 1-2W, but this one was so high that it was above the maximum reading of the ESR meter, which is 100W. In other words, it was virtually open-circuit. I lost no time in replacing C5 with a new one, which had a 1W ESR reading. The power supply came good immediately, and the energiser started its regular clicking with 7.6kV pulse voltage showing on the LCD. The capacitor was a through-hole type mounted between the PCB and the heatsink. The elevated temperature of this position probably hastened its demise. I put the three PCBs back in the housing and screwed the cover on, then left the unit to run overnight as a soak test which it passed with no SC problem. March 2021  67