Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG The oven with a mind of its own Dave Thompson I’m probably not the only one among us who finds modern appliances failing relatively quickly so annoying, especially as it always seems to happen just outside the factory warranty. Some often joke that they must put a timer in there! Appliances can be a significant investment, and when compared to the whitegoods of old, they seem to reach their end-of-life very quickly. The term “planned obsolescence” comes to mind. My mother used a well-known branded mixer every weekend for nearly 40 years before it needed replacing. Admittedly, it had a new armature fitted after 30 years (by dad) to keep it going, but the mixer she shelled out good money for, as a replacement, lasted just six years, and was deemed ‘unrepairable’ by the service people. There is a tendency to think that because modern equipment is far more ‘feature rich’ (read: complicated) that it is more likely to fail, but that only explains a fraction of the problem. A lack of replacement parts and the high repair cost, even if you can source the parts, is another bugbear. This time, it’s personal When we bought our current house six years ago, we renovated the kitchen (among other spaces) and installed a Samsung wall oven. We’d had a previous model in our old place for several years and liked it a lot, so it was a natural step to upgrade to the newer version here. It’s a very good oven; it even has a feature so you can put a heat-shield divider in it and cook a roast on the bottom and a cake on top, but why you’d want a chocolate-flavoured lamb roast and a lamb-flavoured chocolate cake is beyond me! The controls were also ‘upgraded’, and not for the better (in my opinion). Of course, we only realised that once we had unpacked and installed it. For example: on the old model, the various touch functions were backlit, so in dimmer light, you could see what you were doing. On the new one, they rely on a dull graphic printed on the glass touch panel, and unless you are in good light, they are practically invisible. Why the designers thought that was an improvement is beyond my pay grade [it sounds like it was designed by accountants – Editor]. One of the most commonly-used controls on our oven is the timer function, which is initiated by pressing a bell icon on the aforementioned touch area, then by tweaking one of the two very modern push-in, pop-out infinity knobs to dial in the desired time. Then you either wait for a few seconds for it to automatically set that time, or press the almost-invisible timer button again. To set it, I usually just fish around on the panel in the general area of the timer touch button until I hear the beep and see the timer display show; I then set the knob to my time and walk away. Of course, turning off the timer alarm when it starts harping on at me is another fishing expedition if the light in the kitchen is not that great. I’m used to it enough now that I usually hit it every time, but I still think it is a bad ‘feature’ and a step backwards. One thing we do like is the oven’s ability to self-clean using what they call a ‘pyrolytic’ system. While this might invoke thoughts of robotic hands moving all about the inside and leaving things sparkling clean, the reality is that it is a far cruder system. What it basically does is lock the door and pump the temperature to a ridiculous degree (har!). This turns anything inside the oven to a fine ash, including any burnt-on grime and Items Covered This Month • • • • An oven with a mind of its own Testing lifeboat sets Repairing a bricked NAS A not so fusey MPPT controller Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz siliconchip.com.au Australia's electronics magazine March 2022  61 grease. Obviously, this can be a problem if you forget to take out the roast, or any cookware that might disintegrate above 400°C. But it does seem to work well enough, and it is easy enough to wipe the ash out afterwards with a damp rag (once the oven has cooled down, of course). The rub is this can only be done around 30 times in the entire life of the oven, as it places a lot of stress on the oven’s components and seals, so we have to mete out cleaning as-needed with an eye on the longevity of the appliance. The soup thickens So, besides these gripes, for the last six years it has been going well. That is, until a few weeks ago when I walked past and the LED display was flashing randomly between the different program settings. I touched the left-hand knob (which controls these things) and it suddenly went quiet again. I was a bit perturbed, but after giving the knob a good back-and-forth tweaking and nothing untoward happening, I thought it must be just the pot or encoder inside getting a bit dirty (I’ll call them pots for simplicity; they’re sealed so I don’t know what mechanism is inside). However, one night we were sitting enthralled in the latest streaming drama on the box when I heard pinging from the kitchen. Once again, the display was going crazy but this time we had sound to go with it. And again, merely touching the settings control stopped the graphics and noise. 62 Silicon Chip There was obviously something going on with the pot or something else in the control board. And what’s worse is a few days later, I walked past again and noticed the oven was on. I was pretty sure the wife hadn’t switched it on for anything and I certainly hadn’t, so it was time to do something about this. The biggest problem with wall ovens is that they live in a hole in the wall, and access to the internals is not great until they are removed. I probably could have hefted the thing out myself, but I’m getting too long in the tooth for those sorts of shenanigans, so I asked a friend over for lunch. This was the guy who initially helped us renovate this place, and I Australia's electronics magazine bribed him with a nice meal so I could make use of his muscles. With the retaining screws removed, it was easy enough for us both to manhandle the thing out and onto a piece of carpet I’d put on the kitchen floor. It has a nice long cable, which plugs via the usual wall switch arrangement into the grid and that had been coiled up behind and just came out with it. Once on the floor, it was just a matter of removing a few chassis screws to remove the outer shell and reveal the internals (after making sure the wall switch was off, obviously). This gave me access to remove the panel that held the control PCBs. Everything on this oven is handled by four printed circuit boards mounted internally at the top of the oven cavity; the main PCB that appears to manage all the heavy-duty power switching functions is away from the rest, with the three smaller ones immediately behind the front control panel. It was those three that I was going for first. The manual controls – the two joggling infinity pots mounted on either side of the oven’s control panel – boasted a small circuit board of their own at the rear, and then one larger board housed all the touchscreen buttons and displays that we could see through the front glass bezel. These were mounted directly to the panel. I removed the entire front panel easily after finding the three screws that held it on, unplugging a ribbon connector that runs off to the main board, and sliding it forward out of the chassis. There were many more screws siliconchip.com.au holding the various PCBs to the front panel but once taken out, the boards just lifted clear. My goal here was just to have a good look and possibly squirt some contact cleaner about, or look for and repair obviously dodgy solder joints before calling in someone more specialised in oven repair. It’s the Serviceman’s Curse! Even though well-isolated heat-wise from the rest of the oven, these boards live in an inhospitable environment. Wall ovens (especially) heat right through every time they are used for an extended period, so the effects of constant heating and cooling must have an impact on the boards and various solder and plug/socket joints. There are several fans and ducts to keep the heat away from the electronics, but it still must have an effect, and that’s what I started looking for. Everything was surprisingly clean inside; I was expecting greasy residue and other rubbish, but it looked pretty good. All the boards seemed in good nick. It looked increasingly like this was all a waste of time, and a bit overly ambitious of me to think I could do something with it. I reassembled the boards to the front panel, and we sat back and had a coffee before we put the oven back into the wall. Once the caffeine hit, I considered that one thing I could look at was the mechanical parts of the control pot. On these model ovens, the whole ‘knob’ at the front can be pushed into the panel – clicking into place – to get them out of the way and give that modern, sleek stainless look (and apparently make it easier to keep clean). When needed, a simple push inwards pops the knob out, and it can then be joggled to whatever program is required. In practice, we almost never pop them in and just leave them looking like regular control knobs. Perhaps there was something physically there that was affecting things. This one certainly gets way more use than the one on the right side, but as I was already clutching at straws, I decided to disassemble this left-hand one and have a look at it. Mechanically, it is pretty simple: the mechanism for hiding the knob is similar to many push-on/push-off switches we are already familiar with and it works much the same way. Except this is the whole knob assembly that can move in and out, and a sliding shaft allows it to operate in siliconchip.com.au either the closed or open position. While it can be turned awkwardly closed, it is designed to be operated in the open position. The motion felt smooth, but a little different to the right-hand knob – and in any service situation, it is always good to have a second working component to compare to the suspect one. These knobs also sit in the line of fire; that is, when the oven is up to temperature or something is cooking in there, opening the door exposes the exterior bottom section of the control panel to waves of intense heat and potentially other fumes, steams and smoke – especially the way I tend to cook things. I thought that the control panel could do with a good clean, so I pulled the knobs off, used a spanner to undo the shaft nuts, pulled the pots, removed the display circuit boards and disassembled the whole shebang down to metal parts. It’s all stainless and glass, so I threw it all through a quick wash in the dishwasher. While that was processing, I used isopropyl alcohol and soft rags to wipe everything else down and clear the grime out of the nooks and crannies (which are naturally created by these pop-in knobs). I also cleaned the pots and whatever other contacts I could see with contact cleaner while it was all out, and then Australia's electronics magazine once the metalwork was finished, I put it all back together. I had to bribe my friend with another lunch so he could help me wrangle the thing back into the wall; it all went without a hitch. It hasn’t faulted again … yet, but I fully expect that what I did didn’t do much, and this run won’t last long. If I do need to buy a new control or board for it, apparently there are parts available to order – cost unknown – but it irks me that something this ‘young’ would fault at all, considering the purchase price. Time will tell. Leftovers Another trying job through the workshop recently was an amplifier module – one built by a friend from a design from the ‘80s or ‘90s. It looked like one that I’d seen featured in the likes of Electronics Australia, or perhaps even Silicon Chip, but I couldn’t find a matching project for it in those archives. It had never worked. It came with its own power supply in a case, so I isolated the amplifier board and tested the PSU first; as per the owner’s comments, it did indeed work. The line and output fuses were all good, but after rigging it up on the bench and connecting it to my workshop speakers, there was just no signal getting from the input to the output. March 2022  63 All the soldering looked pretty good, and the owner said he had the original documentation that came with the project somewhere if required. Still, I couldn’t see any obvious component misplacements due to the board overlay or anything else really obvious. I’d have to dig deeper. The output transistors are a good place to start because sometimes one or more can just give up if things don’t go well, so I pulled them one by one and tested them out. All were fine. There were also a couple of homewound inductors on the circuit board, and I wondered if the problem could lie with them. I also pulled them from the board and went to measure them, but I couldn’t get a reading on my LCR meter. And I noticed that the solder that had been holding the inductors to the boards was also just falling off the legs as I tried to get test leads on. I soon clicked to the problem; the enamel coating had not been stripped off the wire sufficiently before the inductors were soldered into the board. How he’d even managed to get any solder to stick was beyond me. I’m sure the documentation that came with the kit clearly stated he had to strip the coating off the copper wire, but I guess he either skipped that part or thought he’d done it using just the soldering iron. This is not the way to do it; that enamel coating is quite tough and while I’ve seen people burn it off – with varying success – on smaller wires with lighters or those wee gas torches, it is far better to do it the old-fashioned way with a craft knife and manual labour. I cleaned off the wire ends, used some liquid solder flux and tinned them properly before reapplying them back into the board. I had a second check for any other dry joints but found none, so I reassembled everything back together and applied power. Now I got a good signal through the amp and, after a few minor tweaks, it was ready to go back to a grateful owner. It just shows that the smallest bad joint can cause an entire project to be a paperweight. Testing lifeboat sets R. C., of Mooroopna, Vic had a frustrating day trying to find a working emergency radio for a lifeboat. It seems that they were not well designed... Back in the 1970s, I was a Commonwealth of Australia Marine Radio Surveyor. The purpose of radio surveyors is to make sure that all the radio equipment onboard ships, from small to as big as they get, was in good order to maintain communications in an emergency. For example, if a ship starts sinking, its emergency communications equipment must function properly. This is a story about a string of faulty lifeboat sets. I was called to the Port of Melbourne to test the equipment on board a tug that was to sail from Melbourne to Sydney. It required a onetrip authorisation certificate, indicating that all the radio equipment was fully functional. Other surveyors dealt with other aspects of the ship like the hull, machinery etc. I tested all the radio equipment, and the only item left to test was the lifeboat set. These operated on 500kHz MCW, 2182kHz AM voice and 8364kHz. They were powered using pedals or hand-cranks, and when stowed, they would float. I tested the set, a Clifford and Snell RN610, and found it was not working as the wave change switch was damaged. As there was a small fleet of tugs in Melbourne, I asked if there was another lifeboat set, and they replied yes, and obtained it. I tested this one, and it also failed with a faulty transistor. I asked if there were any other sets, so another was obtained, and it also failed. They got another one, and it also failed, each with a different fault. Having had four sets in a row fail with different faults, you begin to wonder if you are doing something wrong. Things were getting desperate, as the tug could not sail unless I gave the 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. 64 Silicon Chip Australia's electronics magazine all-clear that all radio gear was working correctly. The cost of preventing a ship from sailing is high; it was into the thousands even then. You did everything you could to make sure a vessel was not held up. I asked if they had a Solas III lifeboat set, and they replied that they did. That one worked! I had always found this type of set worked well. The faulty sets were repaired later by shore service as the tugs were telephony-­only ships, with no dedicated radio officer. The Clifford and Snell RN610 was a compact set – too compact, as the top cover was shallow, and it was difficult to get headphones etc in the top cover above the operating controls of the set. This meant that the large wave change switch often had pressure on it, damaging the switch and causing other problems. After I moved on from the surveying work, the following surveyor tested the ability of one of the sets to float (they are meant to). It didn’t and continued to the bottom of the Yarra River. That caused a stir. A significant percentage of the troubles with these sets was because the top waterproof cover of the set was so small/shallow that you had difficulty getting the two sections together to make a watertight seal. If you did get it to seal, was the wave change switch damaged in the process? I often wondered how this particular model set obtained authorisation/ approval under the Safety of Life At Sea (SOLAS) Convention for use on ships, considering how often they were found to be faulty. The case of the bricked NAS K. R., of Auckland, New Zealand knows the saying, “if you don’t have backups, you will be sorry”. He didn’t want to be sorry, so he set up a backup system, but it broke and then he had to fix it back up... About ten years ago, I realised that our home PC had become the de facto family photo album as digital camera images replaced film. Worrying stories of people losing these family memories because of hard disk failure were becoming commonplace. The ever-­ increasing pixel count and decreasing price of digital cameras added to the problem as our media storage needs increased exponentially. Additionally, more and more businesses are emailing invoices and siliconchip.com.au statements, so the PC is also becoming our bill filing system. Then there is that gigantic email archive. With 100GB, including 70GB of photos and growing fast, it was time to put a backup system in place. I was satisfied to add another HDD inside the PC and copy the files across for a while. But what if the PC suffered a serious power incident that fried everything inside it, or it was stolen? My next step was to connect an external USB HDD to run a backup each month so that the backup data was in two places, but these backups ended up being done at somewhat irregular intervals, often three months or more. I decided that we needed a separate appliance that could be kept out of sight in a more secure location, like a Network Attached Storage device (NAS). I selected a Taurus brand enclosure with a gigabit Ethernet interface and fitted two 1TB HDDs running in RAID 1 (RAID = Redundant Array of Inexpensive Disks). RAID 1 mode allows one of the two drives to fail without losing any data by ‘mirroring’ the data across both drives. I was now able to schedule backups of the important data each day, and once a week, I backed up the operating system for good measure. Even when compressed to about 100GB, my 150GB of data took about two hours to transfer over gigabit. All was well for several years, and I had to use the backups when upgrading hard drives in the PC and even fully recovered the operating system after the motherboard failed and was replaced under warranty. Annoying as this failure was, it was a pleasant change to have the failure occur just within the warranty period instead of shortly after it expired. Just as the proof of a pudding is in the eating, the proof of the backup is when it is restored. So I was well pleased when it worked as designed. Then, the scheduled backups started intermittently failing for no apparent reason. Power cycling the Taurus NAS would usually fix it, a clue I ignored completely. When the failures became annoyingly frequent, I searched the Taurus NAS website for updated firmware as my first step in troubleshooting. I downloaded the latest version of firmware and then double-checked the process to load it. It was a good thing that I checked because I had downloaded the wrong version. After finding the correct firmware, I started the upgrade, which went just as expected. Once completed, I navigated to the NAS web interface to find a slightly different menu, and there was no longer an option to set the RAID mode. A frantic check revealed I had managed to delete the correct firmware and uploaded the incorrect firmware instead, designed for a cheaper version with a single HDD! Oh well, no matter, simply download the correct firmware again and reload, right? Wrong! The latest firmware has version check software built-in, and it helpfully refused to allow me to load what it now thought was the wrong firmware. Where was this version checking feature when I needed it? Google helped me find the original earlier version of firmware, but it would not allow that to load either; the computer still said “No”. The NAS worked fine; it just would not recognise the second HDD for RAID operation. Every cloud has a silver lining, and with the NAS opened up on the Silvertone Electronics sells a range of Signal Hound spectrum analysers from 4.4GHz up to 24GHz. There's even a 43GHz analyser coming soon! « This 4.4GHz spectrum analyser is yours from just $1677.50 This product and even more can be purchased from Silvertone's Online Store https://silvertoneelectronics.com/shop/ ► UAV & Communications Specialists 1/21 Nagle Street Wagga Wagga NSW 2650 Phone: (02) 6931 8252 https://silvertoneelectronics.com/ contact<at>silvertone.com.au Spike RF analysis software included for FREE with every Signal Hound analyser Silvertone is a reseller of these brands BitScope siliconchip.com.au Australia's electronics magazine March 2022  65 “workbench” (dining table), I realised that the intermittent failure I was trying to fix was actually caused by the HDDs not spinning up every time. A multimeter revealed the 5V rail was spot on, but the 12V rail was about 9V – the power supply was failing. This was four years after I bought it, but I called in at the supplier Digizone and, bless their after-sales service, they replaced the power supply without question or evidence of purchase. When I explained the firmware mess I was in, they said it would need to be returned to China to be re-flashed. I got the feeling their tech had made a similar firmware error at some stage; he was extremely helpful and even offered me a loan NAS if I needed it to recover any data from my HDDs. Now that I had resolved the intermittent problem, I was more determined than ever to fix the firmware myself. I found a reference to the firmware version inside the “imageinfo.ini” file. Using Notepad, I cunningly changed this reference to the correct firmware description and tried again to upload the RAID capable firmware. Not so simple apparently; again, the computer said “No”. Back to Google, and I found a set of instructions on how to ‘unbrick’ the Taurus NAS by connecting a serial data cable to some solder pads on the circuit board. I soon soldered a donor serial data cable in place (Rx, Tx, Power, GND) and installed PuTTy on my PC. You have to admire the determined person who reverse engineered these factory connections. I set the serial port to 19200 baud but cold-booting the NAS with the serial cable connected generated a meaningless stream of green ASCII characters in the PuTTy window, like a sequence from the movie “The Matrix”. It looked like a voltage compatibility problem, so back to Google, where I found an amazingly simple 5V TTL to RS232 converter that used just two FETs and two resistors. A trip to Jaycar, and $3 later, I had it built on a breadboard and tried again. Success! The now perfectly-readable boot sequence could be interrupted with an old-­ fashioned Ctrl-C to present a Linux boot loader menu. Using the TFTP option (Trivial File Transfer Protocol) and tftp32 freeware software on the PC, I was finally able to upload the correct version firmware files, and the NAS rebooted, as good as new. Learning how to use PuTTy and tftp32 was an exercise in itself, but the reward of fixing the NAS was huge, especially when I caused the problem in the first place. I put in place a truly paranoid backup methodology of daily backups to a separate HDD inside the PC, weekly backups to the Taurus NAS and, just for good measure each month or so, I back up to a portable 1TB HDD that I leave at the office. I can backup gigabytes of data to the cloud, but I still use the NAS just in case. A not so fusey MPPT controller S. L., of Whitfield, Cairns had been ‘gifted’ a dead MPPT controller. Despite it giving him the cold shoulder, he went on his way to having it work again... Several companies make almost identical blue coloured MPPT controllers: Victron, Fangpusun and HanFong, to name some popular ones. Recently, I was given a dead Fangpusun MPPT 100/50 unit to experiment with. Editor’s note: the others are ‘clones’ of the Victron units; our experience is that Victron make quality devices, and we bet that the clones will last nowhere near as long. These units are not designed to be repaired. They have an epoxy solution in the lid when it is attached during manufacture. This epoxy binds to the top of several electronic components, making it impossible to remove the lid without destroying components. (“No serviceable parts inside” – yeah, right. It’s full of serviceable parts; you just can’t get to them!). This faulty unit had no output, and the documentation talks about a non-user-replaceable fuse that protects against reverse-polarity battery connection. On a good working unit, a multimeter will show a reverse polarity diode across the battery connection. On this faulty unit, the battery connection tested open-­circuit, indicating the battery had been connected incorrectly, and the reverse polarity diode had done its job by blowing the fuse. Because there is no way to access this fuse, the unit is deemed a throw-away item. After some checking, I carefully cut away a small part A small part of the MPPT controller’s lid was cut away to access a normally “non-user-replaceable” fuse. This fuse was open-circuit; it was bridged and the unit now uses an external battery fuse. 66 Silicon Chip Australia's electronics magazine siliconchip.com.au of the lid with a Dremel to expose the fuse. It was open-­ circuit as expected, and rated at 250V AC 80A for this 50A controller. I bridged the fuse with solder and wire, and the unit is now fully functional. It now requires an external battery fuse for protection. These units automatically detect a 12V or 24V battery and are further configured with a plug-in switch that allows eight different operational modes to be set, to cater for different battery types. These settings are in the user manual, available on the internet. The absorption, equalise and float settings are different for each mode. This programming switch plugs into a 4-pin RS232 connector on the unit. The RS232 pinouts are 1: GND, 2: RX, 3: TX and 4: Vcc (5V). A cable can be easily made up for communications with a laptop. An RS232-to-USB TTL serial converter module is required. The TX, RX and GND pins from the serial module can then be wired straight to the MPPT controller port. To test the cable, use PuTTy (a serial communications program) by selecting the “Serial” button, changing the “Serial Line” to the COM port number assigned to the device, changing the “Speed” to 19200 and clicking “Open”. You should get about 20 lines of data, repeating every second, showing the status of the MPPT controller. To experiment further with this Fangpusun unit, I downloaded the free “VictronConnect” software from www. victronenergy.com/support-and-downloads/software After installing the software and connecting the homemade interface cable, the “VictronConnect” software talked perfectly to the Fangpusun unit. The first thing it did was advise that the firmware needed to be updated, which I allowed, and the latest Victron firmware was loaded into the Fangpusun unit. The unit then rebooted, and I had access to a lot of historical data, and could modify many settings. Now the absorption, float and equalise settings are all individually adjustable, as well as things like the solar and battery maximum current. The unit retains the last 30 working days of data in history. This is not the last 30 calendar days, but the last 30 days when it was producing power. The only thing left for me to do was glue the piece of the lid that I cut out back on, and I had a fully functional SC MPPT controller. After dealing with the fuse, the unit was tested with the VictronConnect software. siliconchip.com.au Australia's electronics magazine March 2022  67