Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG Getting sucked in by a vacuum cleaner A recent vacuum cleaner repair had me asking myself the rhetorical question: How much of an imbecile am I? Sometimes repairs don’t quite go the way they should, and in this case it might not have even needed a repair! In my defence, I’d never worked on one of these particular models before, so it was very much a trial and error process. The housekeeping duties in our home are shared equally between Mrs Serviceman and myself. When anyone asks me for relationship advice (you’d be surprised how many people ask me how I’m able to spend so much time in my workshop without my marriage breaking down), I tell them this: helping out with the housework beats any bouquet of flowers or diamond ring. Nothing says “I love you” more than doing dishes, 68  Silicon Chip doing the vacuuming or cleaning the toilet! My point, as usual an age in coming, is that the other day I was doing the floors with one of our four vacuum cleaners, a battery-powered Bissell Air Ram (if I’m doing the floors, I need the best tools for the job, right?) when the machine suddenly made an alarming and nasty sound before stopping dead. This was accompanied by a very brief, high-pitched whine and the usually green battery-status LEDs suddenly started flashing red. I promptly hit the off switch and recalled the instruction manual (yes, I do read manuals) stating that if something got caught in the workings, the motor would automatically shut down and the LEDs would flash red as a warning. The manual also mentioned that once the jam was cleared and the lights stopped flashing, the device could be restarted. However, while the LEDs did stop flashing, any attempt to switch the cleaner back on resulted in the lights flashing again, indicating something else must be happening. I flipped the cleaner over and checked out the air intake and powered roller brushes but Dave Thompson* Items Covered This Month • • • Vacuum cleaner repair Faulty capacitors in Behringer active PA speaker CHIMEI LCD monitor *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz couldn’t see anything obvious. As nothing else is visible from the outside, the only option was to dismantle the machine in order to get a proper look inside. Most battery-powered vacuum cleaners are underpowered and thus have the suction of an asthmatic mouse. This unit is powered by a 22V Li-Ion battery and has all the moving parts packed into the compact “head” of the cleaner down near the floor. The only thing in the handle is the battery, making the cleaner lightweight, manoeuvrable and very efficient, as it only has to suck the dust and dirt about 50mm into the dust collectors. The only complaint I would have with it is that the two dust reservoirs are small and fill quickly, meaning it has to be emptied frequently for it to perform at its best. I’d obviously sucked something disagreeable into the thing because the noise it made sounded terrible. I actually thought a fan might have come loose or perhaps it had run a bearing. Being a serviceman, this presented no real problem other than the fact I’d never had one of these apart before and so I wasn’t sure exactly what I’d find once I got in there, or even how to get in there! As usual, there was nothing remotely useful or service-manual-ish on the internet. I assumed only dealers and repair agents would be privy to that information. All I could do was grab my trusty screwdriver and set about stripping it down. siliconchip.com.au I took what I like to call the “shotgun” approach to stripping this machine down. That is, I started by undoing every screw I could find, simply because they all appeared to be holding the vacuum cleaner together. There are about two-dozen visible fasteners dotted around the outside of the case and as it is constructed from high-quality plastics, all the screws are classic PK types. However, instead of using straight blade or Phillips-style heads, the screws were all T10-sized Torx-style splined heads. Fortunately, a long time ago I invested in one of those multibit sets that included all these oddball types, along with a decent-sized driver handle. As such, I have yet to encounter a screw I cannot remove. Regular readers will be aware of my feelings towards those horrible antitamper or security type fasteners, however, Torx screws are growing beyond that use and have become quite popular among builders and constructors. One feature of Torx screws I find very useful is that the bits fit tightly into the heads of the screws and hold fast, making one-handed installation a breeze. This also means you can get away with not having to use a magnetictipped screwdriver because once engaged, the screw hangs on to the bit until you physically pull it off. Disassembly is also less stressful as I don’t have to mess around, fishing out screws that have been loosened but have fallen back into the screw cavity. After removing all the screws in the business end of the cleaner, I could only get a couple of small panels off; siliconchip.com.au one on the left side front and one opposite that on the right. Nothing else would give, no matter how I pushed or prodded it. These two panels provided access and anchor points for the rotating brushes at the front bottom of the cleaner. With the machine running, these brushes would turn briskly and sweep anything in the way rearwards into the path of the suction intake. On most cleaners I’ve seen, these rolling brushes are one-piece, beltdriven devices spanning the head of the cleaner. This model has two shorter rollers, one each on the left and right sides, driven by a centrally-mounted gearbox, like the differential on a car. When the side panels came off, the brushes came off with them, leaving the square metal drive-shafts exposed in the centre. Mounted in the side panels were bronze bushes for the rollers’ axles to run in. They looked very dry which wouldn’t help things but the axles still turned easily in them. While each roller brush had what appeared to be multiple hairs and threads wrapped tightly around it, none of these would have caught or choked the machine to a standstill. It took a good half-hour with a knife and tweezers to remove those threads from the rollers; no doubt after a few hours of use they will be just as bound up again. The bushes appeared to be oil-infused bronze types. I don’t possess a vacuum chamber, so there was no way to re-infuse them properly so I soaked them overnight in a cap of “3-in-1” oil, hoping they’d absorb enough to be lubricated for a while longer at least. Then it was back to stripping the head unit down. I could see the two rear side panels had clips on the bottom but the screws holding them on at the top were buried in behind lots of plastic, which meant the centre assembly would have to come out before I could undo those screws. Based on this, and a couple of other buried screws I could just see down inside if the viewing angle and light was right, I concluded there must be another way in. Perhaps there was something in the moulding that the handle mounts onto that was holding this centre piece in? After popping out the battery in the lower half of the handle, four screws were exposed. Once these were removed, the handle’s cover split apart and with that out of the way, I could see a couple of larger screws below the battery connector assembly that might be connected to something further down inside the body of the cleaner. Holding the handle just so, I had clearance enough to remove the top left screw. Twisting the handle back the other way, I similarly exposed the right-hand side screw and removed it. As I did, something inside the head let go with a loud click and a spring fell out - never a good sign! The centre assembly still didn’t move; it was as if I’d not removed any screws at all! This was becoming frustrating and as I now had to get in to re-fix that spring onto whatever mechanism it had fallen off from, I was past the point of no return. Tricky stuff, these vacuum cleaner repairs! Servicing Stories Wanted Do you have any good servicing stories that you would like to share in The Serviceman column? If so, why not send those stories in to us? 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. May 2017  69 Serr v ice Se ceman’s man’s Log – continued I then concluded there must be fasteners hidden behind the two closelyfitting plastic wheels. All I had to do was figure out how to remove them. There were covers over the entire surface of the wheel and I could see they were held on by two wide plastic clips mounted 180° apart. A thin metal spudger was strong enough to ease the clip off one side and this allowed me to lift that side up. I did the same on the other side and with a bit more persuasion, the hub cap came off. (So what’s a spudger? It is one of those plastic or metal tools one uses to pry open a smartphone or tablet. I have several different types and I have to say they are bloody handy things; I use mine for cleaning fingernails, scraping glue or paint off items and even for prying the wheel covers from Bissell Air Ram vacuum cleaners!) Underneath was a circlip holding the 90mm diameter plastic wheel onto a 12mm metal axle. I dusted off my circlip pliers – the first time I have used these in a very long time – and removed the clip. The wheel lifted off and a bigger bronze bush and steel washer came off along with it; I soaked these bushes along with the others. Sure enough, partially hidden behind the wheel was one very large stepped screw and one smaller screw. The smaller screw, when undone, wouldn’t come out but just wound around with apparently nothing on the other end. The corresponding screw on the other side did the same thing. As you can probably already guess, this made absolutely no difference to the centre assembly’s coming out. (These screws turned out to be holding simple cable clamps that didn’t have to be removed at all.) I felt sure the larger, stepped and specially-machined screws would, however, be holding this assembly in place; after all, it made sense that the larger screws would be doing the job and besides, I couldn’t see any other screws left to remove! I was confident the centre piece would fall out onto the bench once I took out these screws so I was extra careful to make sure everything was supported as I removed them. Once they were out, everything came apart. Well, when I say everything, I mean the handle mount separated from the head of the cleaner, 70  Silicon Chip leaving it dangling by a couple of wires from the battery enclosure. As for the centre assembly, it remained fixed in place! I just couldn’t see what was holding this darned thing on. At this point, I was starting to feel a bit of the “red mist”descending, so I walked away and spent an hour or so tidying up the workshop; nothing’s worth losing one’s cool over! On my return, refreshed and relaxed, I sat the unit on the bench and just looked at it. I concluded there was only one possible way it could go, and that was straight up. There seemed to be nothing mechanical holding it that I could see, and as no one part could come off before another, there could be no other way. I found a couple of large screwdrivers and found a leverage point on each side that could take a bit of pressure and gently started applying upwards force, testing to see what would give. As I put on a little more pressure, I could feel something starting to shift and with even more pressure applied, the centre assembly slowly worked free of the base unit. Vindicated, I silently heaved a sigh of relief; I really didn’t know where I was going to go if that hadn’t worked! It turns out that the centre assembly is held by just six small screws, meaning I didn’t have to take any of this other stuff apart at all. It was a classic waste of time and effort, due to lack of talent. A large, moulded electrical plug on the bottom of the vacuum unit pressed into a corresponding socket in the base of the cleaner, and this transferred battery power to the motor buried in the vacuum unit. The brushed electric motor, very similar to those used in electric model aircraft, is only 30mm in diameter and 60mm long and powers an 80mm hard-plastic impeller within a clear moulded air duct system. A shaft connected to the armature of the motor drives the two rotating front brushes through a differential system and I could see everything was designed for the most efficient use of the motor’s power. I could also now see what had jammed the impeller; a half-burnt incense stick had been ingested and had hit the fan at just the wrong angle, stopping it and causing the built-in circuit-protection system to activate. I’m glad they included such a system as I’m guessing that 22V Li-Ion battery could deliver some serious juice if put to the test and this would easily burn out the wiring or the motor if not disconnected. Splitting the vacuum assembly apart was a simple matter of removing two siliconchip.com.au Faulty capacitors in Behringer active PA speaker G. D., of Mill Park, in Victoria, managed to find a suitable circuit diagram for a PWM power supply to help him repair a pair of active PA speakers. He writes . . . I was recently asked if I could have look at my mate’s daughter’s speaker systems. The power LED and clip LEDs were flashing briefly but no sound would come out when her guitar was connected. Her diagnosis was that the fuse had failed, so could I help? So two “Behringer Eurolive B115D Active 1000W PA speakers with wireless option and integrated mixer”, both with the same fault, were loaded into the ute to make the journey to the workshop. After removing numerous screws, the power module was lifted clear of the speaker box and once the speaker connections were released, it was laid on the bench. Another six screws needed to be removed and the lid of the aluminium box housing the electronics could then be opened, only to reveal more screws holding the circuit board in place, plus several clamps that held the various active devices to the case which also acted as a heatsink. It was a messy task, with a copious quantity of the heatsink compound making its way to my fingers. The operator’s handbook was with the speakers but it contained no information about the actual construction and definitely no circuit diagram, so I was on my own. A search on the internet revealed a number of other people had experienced the same fault but offered no solution. However, I did find a circuit for a Eurolive B215D that I downloaded; my thinking being that there would be some commonality but once I started to compare the diagram with the actual circuit board, all hope vanished. Nothing seemed to be in common except the brand name. A visual examination of the faulty circuit board showed all solder joints to be good and there were no signs of any distress in any components. The board comprise three sections: the mains input to a rectifier via a common mode input filter, a PWM controller to derive the DC output voltages and an amplifier section that has two class-D amplifiers, one for the bass speaker and the second for the tweeter. So now what? A discussion with the workshop owner determined that the PWM chip (NCP1271) or one of its associated components was the most likely problem. A packet of 10 NCP1271 devices could be had for five dollars (including postage), so I placed an order and while waiting I noted down all the active component numbers and went looking for their datasheets. When the PWM chips arrived I swapped it but the fault persisted. I now concentrated my attention to long-ish screws and cutting through some clear tape sealing each top and bottom join. Once done, the stick fell out easily and to rub some salt in, if I had known what this looked like before this happened, I could probably have extracted it from outside using a pair of long-nosed pliers without taking out a single screw. Imbecile indeed! I reassembled the impeller assembly, replacing the cut tape and plugged everything together in order to test run the motor. After rigging up the battery, I pushed the button only to find the LEDs still flashing red. Concerned that the jam had burnt something out, I made sure the impeller was free to turn. I dragged one of my bench power supplies out and dialled in about 12V and set the current limit to about half (2A) before connecting the motor directly to it. siliconchip.com.au the NCP1271 datasheet and noted the example circuit shown on page 18. It matched what I was seeing on the Behringer circuit board. A note in the “operating description” section of the datasheet, under fault conditions, detailed a requirement for a 130ms time to allow a feedback signal to be received, or else a fault condition will be recognised and the PWM will not start. The example circuit shows two 100µF capacitors across pin 6 (VCC) but in the Behringer circuit they were 47µF and although they showed no signs of distress, they measured less than 20µF, with the worst being only 5µF. I had some 50µF caps handy and replaced them and with the power applied, the circuit responded in the correct manner. So I began the task of reassembly, trying to keep contact with that sticky white stuff to a minimum. Once the box was assembled with power on and a microphone was connected, a healthy amplified sound was produced. Since the second unit had the same fault the repair took only a few minutes, plus the hour and a bit getting it apart and back together. The example circuit diagram in the NCP1271 datasheet that is similar to the Behringer circuit board. The motor powered up fine, which was a relief, but why then was the protection circuit still activated? The only visible component (I assumed the rest of the electronics were up in the handle behind the LEDs) was a 1N400x series diode across the motor terminals, which I assumed to be a snubber diode to limit back-EMF from the motor. A quick in-circuit measurement with my Peak semiconductor checker showed the diode to be a dead short. May 2017  71 Serr v ice Se ceman’s man’s Log – continued No problem; I have a box full of these and I soon had it replaced. This time at switch on, the LEDs showed two greens out of four, indicating the battery was down to about half capacity. Even so the motor spun up at an alarming rate. Now all I had to do was reassemble all the bits I’d unnecessarily removed, including the two cable clamps and the spring-loaded detent for the handle assembly, which is where the spring sprang from during disassembly. The silver lining is that I now know a lot more about this device, so if I ever need to repair it again, I’ll be prepared. Job done! By the way, there is a good 3D look at the cleaner in question at: www. bissell.co.nz/air-ram Three different faults in a CHIMEI LCD monitor Sometimes you have to go back three times before the repair sticks, as A. C., from Sunnyvale in New Zealand experienced during a long saga with a CHIMEI CMV T38D LCD monitor. I recently acquired a 20-inch LCD monitor with a fault description that sounded like it could be due to bad capacitors: “takes several tries to turn on and is getting worse”. The prospect of a cheap 20-inch LCD was rather enticing and since I had repaired other equipment before simply by replacing dead electrolytic capacitors, I figured this would be just as easy. How wrong I was! This particular monitor is delight- fully easy to disassemble. A plastic shroud covers the stand hinge and mount, held in place by some plastic clips and this just pops off with little effort, by pulling on it from its bottom edge. This reveals the stand mount, held on with four screws in a standard 50mm VESA arrangement. There are just three screws left for the back cover, which comes off almost as easily (watch out for two clips in slots at the bottom – use a flat screwdriver). In retrospect, I wish all monitors were as easy to open. The overall structural design is basic but quite clever. The stand is attached to the back of a rigid metal cover which protects the circuitry and in turn, screws onto the back of the LCD panel assembly. The plastic case and frame are actually all clipped onto and held up by the panel assembly. I removed the metal circuitry cover plus the threaded hex bolts for the VGA and DVI input connectors and was greeted by the familiar sight of electrolytic capacitors bulging and leaking throughout the power supply. Ah-ha, I thought. I removed the lot, except for the primary filter capacitor (these generally last far longer). As I went, I noted down their capacitance, voltage and reference designators, as well as the brands and series in a spreadsheet. It’s also usually quite important to note down the diameter and height of the capacitors, as in a lot of equipment, space is at a premium and not Some of the electrolytic capacitors had begun leaking onto the power supply PCB. However, this wasn’t the only fault that was found in this particular monitor. 72  Silicon Chip all replacements will be the same size. Having all the data also means you don’t mix the values up, and makes ordering new capacitors easy, as well as a future reference which (hopefully) doesn’t require opening the equipment again. In my monitor, all the blown capacitors were CapXon brand, although there were a couple of Taicons in the PSU as well. Interestingly enough, the Taicons both looked and tested OK on my ESR meter, while even the CapXons which looked physically fine tested just as poorly as their bulging and leaking companions. This just goes to show that for the same thermal conditions and age, some brands of capacitor just cannot stand the heat. Next, I looked up the datasheets for the capacitors I had removed. They were standard low-ESR types. Replacements should have the same ESR or lower – not too low, as significantly lowering ESR can affect circuit operation, especially in a switchmode power supply (SMPS). The Ripple Current Rating (RCR) is like voltage – choose the same, or higher. Make sure the datasheets both specify ESR at the same frequency. Low ESR type capacitors typically specify it for 100kHz, while general purpose capacitors specify it at 60Hz, or not at all. Some quick work with element14’s parametric search and I soon had suitable replacements lined up (all high quality Japanese brands – Panasonic/ Nichicon etc). Upon receiving the new capacitors, I soldered them in and it was time to test the monitor. I plugged it in, turned it on, and was instantly greeted by a nice crisp image which stayed on the screen. It was then pressed into service as my primary computer display. But a mere three months later, more trouble emerged from the otherwise pixel-perfect paradise. This time all the control buttons stopped working, except the auto-adjust button. I immediately jumped to horrible conclusions about blown inputs on the main control chip (as one does), though as in all other respects the monitor was working just fine. Once again I disassembled the monitor and found that the buttons reside on a separate board, connected by a flat-flex ribbon cable. Unplugging this and running continuity checks on the siliconchip.com.au button board showed that none of the switches were faulty. This meant the fault had to be on the scalar board somewhere. The input handling for the buttons is a pretty simple affair. Each button is pulled up to the +3.3V VCPU rail via a 10kW resistor and inductor in series, bypassed with a small capacitor to ground. The output signal is tapped off between the resistor and inductor, then fed to the input of the main processor, so there’s not much which can go wrong. I started checking voltages at the buttons, and discovered that the auto-adjust button had a much lower voltage (0.86V) across it than all the rest (3.3V). My first guess was that the resistor had gone high in value but the resistors and capacitors in question are all part of two four-way SMD arrays. I managed to remove the RP1 resistor array with a flood-and-wipe method, tested it as OK, and managed to eventually get it back on the board without completely destroying the pads. I didn’t like the idea of trying to remove anything else, so I started probing around some more instead. It soon became apparent that the auto-adjust button line was also showing a low resistance to ground and this didn’t change even with the button board disconnected. Clearly, something was shorted to GND, either the debounce capacitor or the processor input itself. Given the difficulty of working on the resistor array, I didn’t want to attempt removing the capacitor array, as I could see myself lifting pads. Besides, even if I had a safe and easy way to replace it, I didn’t know its value. I saw no sense in risking damage. The monitor still worked, and I didn’t really need to use the buttons anyway, so I reassembled and continued using it. Unfortunately, the poor thing died completely a few months later, simply shutting down without warning and refusing to power up again; not even the power LED worked. This time I was not sure where to start, disheartened by the fact that the power LED is driven by the scalar board, and I felt as if my fears about the processor failing were confirmed. But I eventually got around to it, and for the third time, had it open on the workbench. The first thing to do was siliconchip.com.au figure out which board the fault lay on. A dead scalar board could explain the lack of a power LED but so too could a dead power supply. I removed the PSU and started with a visual inspection. There had been no noise when the monitor shut off, so I did not expect a blown switching transistor or such but I carefully eyeballed all the power semiconductors anyway. Nothing was obvious; no burnt parts or bad solder joints. I put the PSU back in the monitor and firmly screwed it back in, as I didn’t want the possibility of a mainspowered board scooting around the workbench while trying to test it. I first measured the voltage across the mains filter capacitor, and found it correct and steady at around 340V DC, so at least I knew the fuse and capacitor etc were OK. I went on to the secondary side. Despite having no schematic, the voltage rails were at least labelled, although they were supplied to the scalar board by a right-angled dual-row 0.1-inch pitch pin header, which was not easy to probe with a multimeter. I got creative. This involved plugging an old floppy drive cable onto the header, which basically broke out the connections to a convenient socket. I was then able to clip one multimeter probe to chassis ground, follow the connections to the other end of the cable, poke a short piece of wire into each socket position in turn, and measure the voltages there. I found that the +12V rail seemed OK but what was supposed to be a +5V rail was bouncing up and down around 2.4V. It certainly seemed as if the power supply was bad, but I didn’t want to assume anything straight away. I know some SMPSs do not run correctly without a load, and I wanted to be sure the scalar board still worked anyway. I tried the reverse approach, taking a standard ATX computer PSU and connected it to the scalar board. Upon powering it up, I was pleasantly greeted by a green power LED on the monitor, and a “No Signal” message on the screen. The scalar board was clearly still working, and this proved the PSU was at fault. Since I was getting something out of the PSU, it seemed then that the primary side was fine, and thus I focused my search on the secondary side. I decided to check all the output rectifiers first. These often fail open- circuit, shorted, or leaky, so they’re a good place to start. Some quick in-circuit testing revealed that D101 was a dead short and this was obviously putting the PSU into a protection shutdown-andretry loop, hence the fluctuating +5V rail. I’m glad the PSU controller was smart enough to do this – some supplies simply blow up when faced with a short on the output. D101 is an SB20200FCT dualschottky rectifier in a TO-220 package and the easily-obtained MBR20100CT from Jaycar was a suitable replacement, although I had to add an insulating thermal pad and washer as the original rectifier was an ITO-220AB insulated variant. With the new rectifier, the PSU sprang back to life with all rails steady and correct. Of course, while the monitor was now powering on again, the buttons were still not functioning. I decided to revisit that fault, armed with better tools, including a hot air rework station. I also got lucky with a schematic, by searching the PCB code (A190A2-HS1) on Google and discovered the same scalar board (and probably PSU) are also used in a Viewsonic VA1912w-1/ VA1912wb-1, for which I found the service manual easily. The shorted capacitor array, CP7, was listed as a 100pF 50V 0603*4 part. (As I later found out, this package is referred to as 0612, and is the same physical size as a 1206 component. For array components, it seems the dimensions are simply written swapped). The magic of hot air and tweezers made short work of removing the old array, and a quick test proved one of the capacitors in the array was indeed shorted. I was also able to confirm that the other three were about 100pF, as per the schematic. Another order later and I soon had some new capacitor arrays ready and waiting. After cleaning up the pads with solder wick and alcohol, I used tweezers to dab some tiny spots of solder paste onto them, before placing a new capacitor array on top and re-flowing the whole lot with hot air. I must say, it’s a marvellous thing to watch solder paste melt effortlessly before one’s eyes, instead of struggling with an oversized iron and solder wire. But the upshot of this long and arduous story? The monitor and all its buttons have been working ever SC since! May 2017  73