Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG Fixing heaters – it’s a gas Dave Thompson I usually only repair electronic and mechanical devices, not gas appliances. But when our heater started acting up in the middle of winter, I thought I’d better look into it. It turned out to be an electrical problem after all, so it was up my alley! It’s almost the middle of winter here in Christchurch, and as is usual for this time of year, the weather is gloomy and cold. Because of the ‘lockdown’, we are spending a lot more time inside than we usually would, and subsequently spending a lot more on keeping the house warm as well! Then again, as we aren’t driving our cars that much, the money saved and extra money spent probably cancel out. I know what you are thinking; LPG isn’t the most efficient way to heat a 46  Silicon Chip home. I agree, but the 6.5kW Masport gas stove (or fire, depending on where you went to school) installed in our lounge was already here when we moved in. And given that we no longer have a reticulated natural gas supply in town, it is bottled gas or nothing. We never actually intended to keep this fire; we knew the people we purchased this house from and spent many nights enjoying dinners here, but because they didn’t use the fire much, we thought it wasn’t much chop. Australia’s electronics magazine We planned to replace the gas fire with a pellet fire (or stove, depending on where you went to school). We’d used a pellet fire in our old home for the previous decade or so and we were very happy with it. While some love and some loathe pellet fires, for efficiency, they’re tough to beat. The fuel is simply compressed sawdust, which is cheap to make and widely available, and emissions are next-to-nothing. The ash pot only needs emptying once every few weeks siliconchip.com.au when the fire is used all day, every day. Our old Canadian-made Evolution 2 pellet stove (which replaced a log burner) could throw out around 10kW, but we only ever used it on the lowest of five heat settings; otherwise, we’d have melted! Before we moved in, I purchased an identical, almost-new Evolution 2 pellet fire salvaged from a quake-damaged home. It was a bargain, and all we’d need to do was swap out the gas fire with this one, although I’d have to get resource consent and a registered installer to do that work. I could do it, of course; but I had to pay a professional to do it, to satisfy the insurance company. But once we moved in, we discovered that the gas fire could produce some decent heat (about 6.5kW worth), so it wasn’t worthwhile to replace it. That spare pellet fire is now taking up valuable bench and power tool space in my workshop, so if anyone is looking for a cheap, good-condition Evolution 2, drop me a line! fires is the noise of the fan and auger motor. On the low setting, the auger runs for about three seconds twice per minute; on high, more often. It isn’t that loud, and we found after the first few nights we no longer heard it, but visitors would often ask what it was. The fan noise is similar to a small fan heater; not too intrusive but certainly audible. Many people think they couldn’t put up with these noises, but it really isn’t that intrusive, and we soon got used to it. Another downside is that a pellet stove needs electricity, so it was initially rendered useless in the quakes, when we had prolonged periods without mains power. However, I soon had it rigged it up to our generator, so we could at least keep warm if the power failed. And that is pretty much it as far as operation goes. Keeping it as dustfree inside as possible, and emptying the ash pan once in a while is about all that is required; plus a flue clean every couple of years. An introductory course on pellet heaters The problems begin For those who don’t know what a pellet stove is, or how they work, they are actually very clever. Most work similarly, regardless of make or model. (Don’t worry, this is leading to a repair story, I promise…) The top part of the machine is a hopper into which pellets are poured. Pellets are available from supermarkets and hardware stores in 10, 15 or 20-kilo bags, with the largest bags being the hardest to carry, but also the best value. Thankfully, since we usually ran our heater on low, it would only burn through about 15kg of pellets each week. A motorised auger system in the bottom of the pellet hopper periodically feeds pellets into a burn pot, usually within a sealed burn chamber in the bottom half of the fire. You can generally see this burn pot through the glass front door of the chamber, and this is where the visible flames sprout from as well, giving that cosy ‘fireplace’ effect. A blower fan spreads the hot air outwards from the fire. Once the fire is alight, the more pellets you feed in, the hotter it burns. Drop in the pellets less-frequently, and the heat output is reduced. Besides having to feed the hungry fire, the other main gripe with pellet siliconchip.com.au As you’d expect, there are lots of moving parts in a pellet heater, and they need to be in good condition to ensure they are operating effectively. The first problem I had with our Evolution stove was a common one: a failed igniter. Usually, to get the fire going, you just push a button. It starts the auger motor and an internal fan. The auger drops pellets into the burn pot. When they have built up into a small pile, the igniter, which protrudes slightly into the burn pot, glows red-hot and sets the pellets burning. It’s helped along by the calibrated airflow in the chamber. When the chamber temperature rises to a set level, the main blower fan kicks in, and it’s away. This usually takes about 10 minutes or so, but after a few years, it took increasingly longer, and eventually failed to ignite altogether. This wasn’t as disastrous as you’d think, because I could easily start the fire by opening the door, manually igniting a small number of metho-soaked pellets in the burn pot and then closing the door; the stove would then be going almost instantly. However, this took away some of the convenience, so I looked into replacing that igniter. I ended up getting the supplier to send out their maintenance guy who Australia’s electronics magazine Items Covered This Month • • • • Fixing a pellet heater Upgrading a Labtech Q1590 frequency counter Asus monitor repair LG TV power board repair *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz replaced it, telling me that poor design meant that as long as the stove was ‘on’, the igniter was powered and glowing red hot. This makes little sense, as once the thing was alight, it didn’t need any other ignition source and all this did was considerably shorten the life of the igniter. It went again after another two years and that time I replaced it, at considerably less cost. When it failed again two years after that, I left it as-is and simply used the metho starting method. After about 10 years, the auger motor bearings failed, and that made a really nasty noise. Fortunately, they are standard bearings and easily replaced, but it goes to show that the more complex a system, the more breakdownprone it becomes. All this influenced our decision not to replace the Masport gas fire/stove. For one, it is relatively cheap to run (compared to electricity) and as we use the gas for cooking as well, it makes no real sense to replace it. Even the fire is sick of the lockdown So it was a bit ironic that barely a few weeks into the lockdown, the fire would periodically go out. I never saw it going out; I just noticed that while the built-in fan was still running, there was no fire on the fake logs. Re-lighting it was also difficult. At first, I thought that the 45kg bottle was empty and needed swapping (I use a manual switching system, so I know when one of the bottles needs replacing). Usually, all I have to do is open the tap on the fresh bottle and flick the gas switch over, and all is well. But this time, I could see the go/no-go indicator in the gas line was still showing green, so the tank wasn’t empty after all. August 2020  47 I went back inside and tried to ignite the fire to no avail. To start it, I push and turn the main gas valve to ‘light’ and hold it down while I press the piezo igniter. Typically, it takes a few strikes to light the pilot lamp, and after about five seconds I can let up on the valve and the pilot stays alight while I hear gas enter the burn chamber, beneath the fake rocks. A few ‘WOOFS’ later and the thing is going. We rarely have to turn this one up either, with level one or two sufficient to warm our space. However, this time I needed to keep the valve pressed much longer, and even then the pilot barely lit up. Once going, we needed to run it on level five just to keep it alight. Something was obviously wrong… While I know a little about a lot of things, I know next to nothing about how gas fires work. But a quick internet search gave me all the information I needed, as well as an excellent service manual for the appliance. At least that allowed me to investigate what could be wrong. I know one thing though; messing around with gas and fittings is something that absolutely should be left to 48  Silicon Chip the professionals. I can still clearly recall sitting at my workshop desk a few months back, and feeling/hearing the massive bang as a house about four kilometres from me literally blew to bits because of a gas fire leak. Lesson learned! Gas is not to be trifled with. Editor’s note: as detailed in this column in the past, just because you get a professional to do the job doesn’t necessarily mean that you will get a good result. Our newly-built house had a recurring gas leak (as did our neighbours, in the other half of the duplex). That was despite it being checked and approved by the relevant authorities! The good news is that as there was some electronics involved, I had a legitimate reason to at least have a poke around. These fires are actually very clever; all gas fires must have a fail-safe system that shuts off the flow if either the pilot light goes out or the main gas valve is opened without lighting the fire. This stops the room filling up with gas and suffocating anyone, or converting the home into a bomb. The gas-flow system is controlled by a solenoid which is held open (and Australia’s electronics magazine thus allows gas to flow) only while a flame heats a thermocouple (or thermopile). If the pilot flame goes out, the thermocouple cools, its output voltage drops and the solenoid closes, stopping the gas flow. It’s simple and highly effective, as long as all the components in the system are working. So based on the symptoms, I could at least start to troubleshoot this problem without having to take any gas lines apart. The first possibility was a blocked or partially blocked gas line. If the blockage was further up the line, towards the bottles and the fittings, I wouldn’t be able to do anything without a gasfitter’s ticket. However, we use the same system for cooking, and our gas hob rings all burned at full noise, so it was unlikely to be a problem with the lines, at least to the junction where the fire and gas hob feeds split off – which is situated handily right behind the heater. That meant that it was unlikely that the gas lines to the heater were blocked. But we could have simply had a blocked pilot light, and that assembly is readily accessible after removing the escutcheon and one glass panel from the front of the heater. Once exposed, I used a bent piece of copper wire that just fit into the pilotlight jet to clear any potential blockage. It felt clear, and a quick puff with one of my rubber-bulb circuit-board dusters ensured that it was clear of obstructions. The pilot light on this fire has three flame paths: one towards the bare-copper igniter wire, one to the thermocouple, positioned opposite the igniter wire, and one into the main part of the fire. I used a pipe-cleaner soaked in white spirit to clean these out, and as they came out remarkably clean, that was likely not the fault. I could hold the gas valve down and turn the main gas input tap at the back of the fire on and off, and could hear a decent gas flow through the system. So I doubted that it was a flow problem. The next thing to check was the thermocouple. These are a known consumable, and replacements are widely and cheaply available. After removing the rear access panel, I could see where the thermocouple connected into the main valve. This is a plumbing-type fitting that is easily removed/undone with an open-ended spanner. siliconchip.com.au Once free, the copper tube-like electric lead can be unfurled to bring the connection out so I could get my multimeter probes onto it. The tube is grounded, while the internal wire is the ‘hot’ lead (LOL!). With the meter set to volts, I played the flame of my small gas torch over the thermocouple tip, where the pilot flame usually hits it. I measured just on 9mV. According to the book, I should read at least 15mV, so this was a potential (haha!) problem; 9mV may be barely enough to hold the solenoid in. A replacement M9x1 thermocouple was only $39 including delivery, so it made sense to replace it and see what happened. I tested the new one when it arrived, and got a reading of 16mV. Fitting it was as easy as loosening a retaining nut and bolt, removing the old one and threading the new one in. I re-connected it to the main valve, and the fire’s been going perfectly for a month now, so I think that’s job done! Labtech Q1590 frequency counter upgrade C. K., of Croydon, Vic, went a bit beyond the usual remit in this column of making something that’s broken work again. Instead, he took an older test instrument that was functional but a bit inaccurate, and modernised it so that it is super-accurate. As you can imagine, it took a bit of doing... What can we do with test equipment, years old but still functional, The interior view of the Q1590 frequency counter. There is an oscillator module inside the centre metal enclosure. that is well off the pace in regards to accuracy and stability? This was my dilemma when I tried to calibrate a Labtech Q1590 multi-function counter. I bought it probably in 1989, and it has never failed me. But these days, digital communication technology requires extreme frequency accuracy. Only a few Hertz out, and digital messages cannot be decoded. I tried to calibrate the counter using a GPS-disciplined source of 10MHz, but the readout was about 150Hz too low. Taking the case off revealed an oscillator module inside a metal case, with two trimcaps which can be accessed through holes in the top (shown above). One trimmer is for a 10MHz crystal, and the other is for a 3.906250MHz crystal, the purpose of which is not clear to me. As the readout was low, the 10MHz oscillator frequency was too high. But adjusting the trimmer still did not give me a correct reading. Adding an 18pF capacitor across it helped, but the adjustment was difficult and tended to jump. By replacing the trimmer capacitor and the parallel fixed capacitor with new ones, and with very careful adjustment, I could get to within about 3Hz. But I wasn’t satisfied with that. Fig.1: the small circuit designed to utilise a cheap TCXO found online as a replacement oscillator module in the Q1590. siliconchip.com.au Australia’s electronics magazine August 2020  49 The original oscillator module shown without the metal cover (left) and with the oven and NPN transistor removed (right) Also, on turning the counter on, the reading started about 50Hz low and after a couple of minutes overshot by about 8Hz, then over several hours, it gradually crept to within 3Hz. Both crystals are wrapped in a piece of copper that is heated by an NPN power transistor (shown above). A thermistor glued to the copper sheath provides feedback so that it maintains a more-or-less constant temperature. But apparently, the temperature still was not stable enough. Since I hate to throw things out, I decided to come up with an improved oscillator design using a TCXO (Temperature Compensated Crystal Oscillator). These can be expensive, but I found a 10MHz model on AliExpress for less than $20. That seemed suspiciously cheap, but I decided to take a punt anyway. The 3.906250MHz crystal was a problem – I couldn’t find a TCXO at that frequency. So I decided to use a DDS (Direct Digital Synthesis) chip like the Analog Devices AD9850 (as described in the September 2017 issue; siliconchip.com.au/ Article/10805). Modules using this chip are available cheaply on eBay and elsewhere, but there was not enough space in the Frequency Counter to fit such a module and associated micro. Fortunately, I had a couple of the bare chips in my stock of parts. Virtually any microcontroller can be used to load the tuning word into this chip, and as I have heaps of Atmel AVRs on hand, I decided to use an ATtiny2313. The circuit I came up with is shown on the previous page. If it looks familiar, that might be because it’s quite similar to my Circuit Notebook entry on pages 96-97 of January 2020 (siliconchip.com.au/Article/12231). But that circuit used an Arduino and an AD9850-based module, compared to the more basic approach taken here. There is a small problem in that when the AD9850 has a 10MHz input frequency, the 3.90625MHz we want at the output is a bit too close to the The fixed frequency counter with new oscillator module – the repair cost totalled less than $50. 50 SILICON CHIP Australia’s electronics magazine siliconchip.com.au The finished replacement module was made using a custom PCB. Nyquist frequency of 5MHz, resulting in a very distorted output with many spurs that could read as false edges. I solved this with a tuned circuit that cleans up the waveform, based on transformer T1 plus one fixed and one variable capacitor. I managed to locate the original counter schematic and discovered that the 3.90625MHz crystal is driven by one stage of a 74HC04 hex inverter, operating as an oscillator amplifier. So I just had to remove the old crystal and feed the output of the AD9850 chip into pin 13 of the 74HC04. All that the ATtiny2313 does is load five bytes into the AD9850 to set up the correct output frequency. The tuning word is 0x64000000 (hex). Obviously, there is some magic power-of-two relationship between the two frequencies to get such a simple number (and this hints at why a seemingly odd frequency was chosen). Designing the PCB was a bit tricky, as there is not much room available, so I mostly used surface-mount components. Some resistors and capacitors are on the underside of the board. It was fortunate that the pins connecting the oscillator board to the motherboard were at 5.08mm (0.2in) centres. I used socket strips on the motherboard and matching pins on the oscillator board so that it became a plug-in module. Once the custom PCB arrived, I loaded the components and plugged the module in (shown at left). Holding my breath, I connected the 10MHz reference to the input. And up came 10000000 – spot on! I decided to leave it running several hours, in which time there was the rare jump to 10000001, but only for one count period. I was quite surprised and pleased that the cheap TCXO is so accurate. There is a sticky label on the oscillator which gives access to an adjustsiliconchip.com.au ment, but I am rather glad that I did not have to fiddle with that. I still don’t know what the 3.96250MHz frequency is used for. I believe it has to do with the 100MHz to 1GHz range of the counter. Having spent considerable time on this repair/ upgrade, I did not feel inclined to do a full analysis of the original design. Has the exercise been worth it? Not if I count the (unpaid) hours I spent on it. As I already had most of the components, I spent less than $50 in total. But the satisfaction of extending the life of an otherwise useless instrument certainly made it worthwhile. Asus monitor repair Poor, innocent bugs are often unfairly targeted as the cause of electronics misbehaving. But in the case of one particular monitor, B. P., of Dundathu, Qld, found the culprit to be of the reptilian variety instead... We’ve been using an Asus computer monitor in our camper as a TV, with it connected to a personal video recorder (PVR). Recently, my wife told me that the monitor was dead. I found that she was right, so I had to take it apart to see if it could be repaired. Often something that is totally dead is easier to repair than something that partly works; I was hoping that would be the case here. Opening the monitor up proved to be quite tricky. Computer monitors, in general, don’t seem to be built with repair in mind, as they are clipped and Australia’s electronics magazine not screwed together. So it’s often difficult to get them apart without damage. The usual way of opening them is to pull the front plastic surround away from the screen carefully, making sure not to damage the screen in the process. I’ve opened up quite a few monitors over time, but this one proved to be a lot more difficult than most of the others I’ve worked on. Still, I eventually got it open. I then sat the monitor face-down on a towel and lifted the back off. I could then see why the monitor had stopped working; there was a blown-up gecko at the side of the metal housing. I could see that the gecko had been burnt by high voltage electricity. Despite that, it had clearly crawled some distance from where it had been zapped. After removing the gecko, I proceeded to disconnect the cables necessary to turn the metal housing over so I could access the circuit boards on the other side. While doing this, I found a dead cockroach in the corner of the video board. I’ve previously had a computer power supply blown up by a roach, but this time, the culprit was the gecko. There was a considerable carbon deposit between the two tracks where the gecko had come in contact. I’ve seen other devices where tracks have been shorted by some wildlife, but this is the first time I’ve seen this carbon between the tracks. I would need to rectify this before I looked into what else might have been destroyed. August 2020  51 I started by scraping all the carbon out of the burnt section of the PCB until it was back to clean fibreglass. This was to ensure that it would not arc when voltage was applied. Next, I touched up the corner of the blown-off pad with solder, although this may not have been entirely necessary. I checked the fuse next, and it had blown. That was potentially a good sign, but it didn’t rule out damage to other components. I searched for a replacement fuse, but as this fuse was a leaded type that was soldered to the PCB, I was unable to find a suitable replacement. I then thought of fitting fuse clips, so that I could use a regular fuse, but I didn’t have any clips of a suitable size. I wondered what junk circuit board I might have that I could salvage some smaller fuse clips from, and I located an old CRT TV board that I hadn’t yet stripped of components. It had suitable clips and even a fuse with the correct rating, so I removed them from the board and considered how I could fit them to the monitor PCB. Because the replacement fuse was shorter than the original fuse, I decided to re-use one of the original fuse pads and fit the other clip to a section of the PCB with no tracks. I drilled 1/16in holes for the clips, fitted them and bent the pins over, then soldered the first one to the pad. It was then just a matter of soldering a wire from the pins on the other clip to the original track. This would save me effort in future if the fuse ever blew again. It’s always hit and miss replacing a blown fuse, as it might just blow again the instant that power is applied, or perhaps it wouldn’t blow but something else would. So before applying power, I decided The dead gecko and the damage done shown below. 52  Silicon Chip Australia’s electronics magazine to make some further checks. I checked the bridge rectifier, and it tested good. I then checked across the power terminals and as there was no short circuit apparent, I decided to apply power to see what would happen. I plugged in a power cable and turned the monitor over, then pressed the power button and the monitor came to life. That was a good sign; it appears that the fuse had done its job in protecting the circuit from the killer gecko. I just had to reassemble the monitor and put it back into service. It has been working well since the repair and I’m hoping for no more wildlife invasions. Unfortunately, there are large openings in the monitor for ventilation, so that is still a possibility. This was another successful repair at no cost, which was a win-win situation. It saved the monitor from landfill and avoided us having to find a replacement. It’s worth having a go at repairing devices, but always remember that electricity kills, so proceed with extreme care. LG TV power board repair R. S., of Fig Tree Pocket, Qld found some damaged parts on a TV power supply PCB and replaced them. But it seems that the damage was more widespread than he thought... This LG TV power supply board had a strange fault on the 5V standby supply. It uses a 3B0365 IC (IC500) with an internal high-voltage FET, which shorted out. When this and the 1.2W currentsense resistor (which I found to be open circuit) were replaced, the 5V supply would still not power up. The circuit (siliconchip.com.au/ link/ab3a and siliconchip.com.au/ link/ab3b) shows that the auxiliary supply generated for the 3B0365 comes from an extra winding on the transformer. This also supplies two other integrated circuits on the board, via a transistor controlled by a powerup signal from the main board. This transistor was shorted out, as was the L6599 IC (IC100), so the auxiliary supply for the 3B0365 was being shunted, stopping it from working. Once these additional faulty parts were replaced, the board sprang to life. What I am not sure about is whether the L6599 failed first and damaged the other components, or whether the 3B0365 failed first and caused the other problems. SC siliconchip.com.au