Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Cheap fixes for the working Serviceman Dave Thompson is currently on a pilgrimage to Eden Park, home of the All Blacks – an arduous journey undertaken by every good New Zealander at least once in their lives. While he fends off feral kiwi bird attacks and practices his haka, we have a selection of reader-contributed stories for this month’s column. R. W., of Hadspen, Tas was asked if he could put a new fuse in his friend’s Bose subwoofer... I queried why and was told that it stopped working; the LED no longer lit up, and the last time this happened, the warranty repair just involved replacing the fuse. I said I’d be happy to take a look, but fuses blow for a reason, and if the cause was not determined, they would continue to blow. I collected the unit, a “Bose Acoustimass 300”. It is a solidly-­made and very heavy subwoofer that connects wirelessly to a TV soundbar. Its sole LED did not light when power was applied, so I assumed it was a simple power supply fault. The base of the unit houses the electronics in a diecast assembly that forms the base of the enclosure, the lower acoustic port and also acts as a heatsink for the amplifier. Only a couple of screws were visible and, on removing them, the plastic cover wouldn’t budge. Checking under the adhesive felt feet revealed another four screws, but the cover was still fixed after removing them. I located another two screws under the adhesive label and had to cut holes in the label to remove them. Clearly, this was the first time it had been opened, so how had the fuse been replaced? Eventually, the circuit board was revealed and, from below, I could see the HV delineation on the circuit board and took care to handle it from the edges in case some capacitors held charge. It was a wise decision, as a whopping HV electrolytic capacitor occupied the centre of the board. A small multi-conductor ribbon cable needed to be pulled from its socket (what it connected to, I cannot imagine), and after the heavy-duty speaker connector was unplugged, the board was free. About half of the board is occupied by switch-mode power supplies (SMPS), about a quarter is what appears to be a Class-D amplifier and filters, and the remainder looked as if it was microprocessor or wireless related (there was a PCB antenna in the corner). I don’t know how that antenna worked, as it was almost entirely within the diecast enclosure. There was a plastic-encapsulated fuse on the board (covered in white in the photo below) rated at 4A, but it was intact, and there was no evidence that it was not original. Following tracks from the mains cable with an ohmmeter revealed nothing unusual, so I applied power. I measured around 340V DC in several places around the switch-mode supply and main electrolytic. A feed from this cap went via an inductor to another smaller HV cap hidden under masses of what looked like hot melt glue. I was unable to find any low-voltage rails. From here, I was in the dark, and tracing was difficult as the board was a multi-layer type. I desperately needed a schematic. After hours of trawling the internet, I had no useful leads and had ignored all the YouTube fix-it videos. I don’t waste time looking at them as they are generally unhelpful. Eventually, with no other way forward, I watched a video from “Jonny Fix” about an Acoustimass 500, one of several that he has fixed, and they all had the same problem as mine. He had my full attention. It appears that one diode is the culprit, and it’s involved with the smaller SMPS right where I had been looking. Why it fails with such regularity is a mystery. Nearby is a 16-pin DIL IC marked ALTAIR05T N02671, which I determined is a switching regulator. The diode in question was a 3A 100V schottky type in an SMD package, located under a glue river (see the photo at lower left). It measured short-circuit. I removed the diode, and it still measured as shorted. The fault in the Bose Acoustimass 300 power supply was hidden under a big blob of glue. Replacing the shorted schottky diode returned it to a functional state. 86 Silicon Chip Australia's electronics magazine siliconchip.com.au Items Covered This Month • A $3 subwoofer fix • Finding replacements in unexpected places • Repairing an Esseti TIG welder • Revitalising a Miniscamp Computer from EA • Fixing an MTM 4400i inverter generator Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz Cartoonist – Louis Decrevel Website: loueee.com I could not believe I had found the culprit so quickly with no schematic. Thank you, Jonny Fix! Now to find a replacement. Some people in the comments section said that a 1N5408 had worked for them, but I wanted to replace it with an equivalent schottky type. I found a dual 100V 20A diode in a TO220 package (MBR20100CT) at Jaycar for under $3. 20A is overkill, but it was the closest 100V schottky I could find. I used only one of the diodes within, and had to use some hot-melt glue to affix the body to a nearby capacitor for rigidity. I left the metal tab open to the air. On powering the unit, nothing happened, and the LED remained dark. I was about to utter a comment when the LED started flashing orange to indicate that pairing was in progress. When I re-introduced it to the soundbar, it performed faultlessly. As for the mystery of the fuse replacement, it appears the retailer simply replaced the whole subwoofer, and the fuse excuse was a furphy. Lucking into replacement parts T. B., of Kogarah, NSW found a rare part he needed in the most unlikely of places... I began my aircraft electronics career as an apprentice with The Flying Kangaroo in the early 1960s. We had the best electronics test equipment, including beautifully made (hand-wired, I assume) Tektronix scopes. They were valve-powered, unlike the Boeing 707’s all-solid-state electronics. So I did not see much valve technology for repair. Fast forward 40 or so years, and in retirement, I was persuaded to join the HARS Aviation Museum at Albion Park, south of Sydney, resuming my trade as a volunteer. Early on, I spotted one of the aforementioned Tektronix ‘scopes gathering dust in the back of the hangar. I couldn’t let it go unused, so I set it up on the workbench and removed the cover. It looked so good, but a valve was missing, with “6DQ6A” helpfully stencilled next to the socket. My valve experience was limited to building a Radio, TV & Hobbies “Playmaster” amplifier, and it had none of those types! Nearby, there was a lead with a top-cap fitting, so I needed to do more research and start looking for one. Where, though? Fast forward to the following Saturday, ie, two days later. I needed to buy a desk for my home computer; a second-hand one would be good enough, so I called into siliconchip.com.au a nearby used furniture store. They showed me to their desks, but none were suitable. As closing time was near, I headed for the exit. My eye caught a glass-fronted display case at the end of an aisle with some interesting bits and pieces piled up. Worth a peep, I thought. Right on top of the pile was a Mullard valve box, with the end visible, reading 6DQ6A! Surely, it wasn’t possible, but it was. I couldn’t believe it; I picked it up and headed for the checkout. “Price for one valve, please?” “$5 for you.” Done deal. “Probably from a deceased estate,” I was told. Next Saturday could not come fast enough. I fitted it to the ‘scope and was greeted by a nice sharp trace on the screen. A joy to my eyes! It was truly a chance in a million to find such a prize in a used furniture shop five minutes before closing time. Scratching the itch to service a welder R. H., of Waverton, NSW discovered, as many others do, that a seemingly serious fault can result from a single failed component. That’s why it’s often worth trying to fix failed equipment... It was about 1998 when I purchased my Esseti Inverter TIG welder. The cost was then $1700; quite expensive! Although I did not realise it then, this welder was quite advanced; it featured hot start, arc force and anti-sticking, but no high-frequency ignition (HFI) – that would have been an optional extra at $450, too expensive! The arc is initiated by ‘scratch start’, requiring constant practice to gain confidence to achieve this method. It also contaminates the tungsten. I constantly toyed with the idea of adding HFI, as lately, many HFI boards and simmer coils have become cheaply available on the internet. Many YouTube sites explain this but leave out essential details of exactly how to connect these items to the welder. I tried, but the result was that my welder failed. What to do? Buy a new TIG welder, with HFI now considerably cheaper, or have a go at repairing it? After some thought, I decided to take a shot! Australia's electronics magazine May 2024  87 The manual that came with the welder was in Italian, with some attempt at English translation and, of course, no circuit diagram. All seemed good at the front panel. The readouts and the gas solenoid worked. But no 90V at the TIG torch. The main board had two spade terminals, marked with a plus sign, adjacent to two L7912CV 12V regulators. Both measured 12V. I next tested the four RURG8060 rectifier diodes and found that one had failed. All the other components looked OK! I hadn’t tested the two G4ON60B3 N-channel IGBTs yet, but I placed orders for four RURG8060 diodes and two G4ON60B3 IGBTs to be on the safe side. Once received, I replaced all four RURG8060 diodes, checked it over and powered it up. With the welder switched to stick, up came the 90V DC. Wonderful! However, when I switched to TIG, the output terminals only measured 12V. Why? I rechecked everything and could not find a reason. Then, when I was lying awake at night, the idea came to me that the 12V was a sensing voltage, and when the arc was struck, up would come the 90V and many amps. So I tried that, and it all went well. As for the HFI addition, I will leave it until I find out more about how it’s supposed to be installed. The moral of this story is: don’t throw it on the scrap heap – have a go. You never know your luck. Miniscamp Microcomputer rejuvenation J. W., of Hillarys, WA decided to try out a computer he built around 46 years ago. Computers were pretty simple back then, so there wasn’t much to go wrong... A few weeks ago, I decided it was time to clean out the workshop. I came across the first computer I ever owned in the back of a cupboard: a Miniscamp microcomputer. The Miniscamp was a project published in Electronics Australia in April 1977. I built it around 1978 and modified it over the next few years to include a serial interface and more memory (ROM & RAM). I decided it would be worth the effort to get it running again after over 40 years of gathering dust. The Miniscamp used a National Semiconductor SC/MP microprocessor and, in its original form, had 256 bytes (that’s right, bytes) of RAM, using two 2112 RAM ICs. 88 Silicon Chip Australia's electronics magazine siliconchip.com.au Input and output were in the form of eight LEDs, eight data switches and 10 address switches. Programs were entered by switching to DMA (direct memory access) mode, setting the address switches to the required address, say 0000h, and setting the data switches to an instruction, say 08h (no operation/NOP). The deposit switch was then pressed, and the instruction would be saved at the selected address. So the first address would now hold the NOP instruction. Entering a large program was a lot of work and concentration. The Miniscamp was basic, but a good learning tool for microprocessors. The first thing to do was to get some information about the design and instruction set. I found an archived copy of the original article and many of the following articles, so I had the circuit diagram and some sample code. I then located the National Semiconductor SC/MP programming and assembly manual, so I was ready to go. I removed the cover and was surprised to find the circuit board in perfect condition, just like it was in 1978. After connecting power, I found the +5V and -12V supplies to be good. Now was the time to see if the hardware was working. I had upgraded the original kit with a ROM containing National Semiconductor’s KitBug ROM (512 bytes), more RAM and a serial interface. I checked the original circuit to find the serial output pin, connected my CRO and pressed the reset button. I was greeted with a stream of pulses from the Flag 0 pin. I had a USB-to-serial interface on hand, so I used a few transistors to isolate the PC and Miniscamp from each other. I ran PuTTY (a serial terminal program) but had to set the baud rate; I thought I had set it to 500 baud all those years ago. I was spot on, and after pressing the reset button, I was greeted with a hyphen as a prompt. The manual indicated that there were only three commands recognised: T for displaying memory, M for modifying memory and G for running a program. I remembered upgrading the original KitBug ROM to add the ability to save (S) and load (L) programs from cassette tape and set breakpoints (B) in a program. I could now enter a program by typing commands instead of setting switches. The original article had a sample program that displayed a binary counter on the LEDs but, upon siliconchip.com.au Australia's electronics magazine May 2024  89 running it, the LEDs did not flash. After some investigating, I realised that to add the extended the RAM, I had to disconnect the LEDs as they took up a 256-byte bank of memory. After a bit more investigation, I disconnected one bank of RAM and got the LEDs working again. I then tried to use the breakpoint command, but it did not work. After printing out the program listing of the ROM, which I had extended to 1kB, I saw what I had done to get the breakpoint feature working. I then set about writing some code of my own. Assembling the code by hand was a bit of a job, so I investigated and found an assembler on the internet to do the job. I was then satisfied that it was all working correctly. Now I just need some tasks for my 40-year-old computer! MTM 4400i inverter generator repair G. C., of Toormina, NSW has an MTM 4400i inverter generator that has served him well during blackouts and remote work over a 10-year period... For those readers not familiar with an inverter-type generator, they have the advantage of producing a pure sinewave output with the correct voltage and a stable frequency. An alternator in this machine supplies three-phase power at varying frequencies depending on the engine RPM. This three-phase supply is rectified into DC and fed to the inverter, where it is converted to the required 50Hz AC output. Another advantage of this system is that the engine can idle slowly on light loads and rev up to meet the demand if the load increases. That is achieved by a stepper motor, which operates the throttle based on the detected load. 90 Silicon Chip Standard generators are hopeless in this regard, as their frequency and voltage can vary all over the shop, and they need to run at approximately 3000 RPM for a two-pole machine to get a 50Hz output. That means a lot more noise. This generator is rated at 4.4kVA and was made in China. I purchased it on eBay for about $500. We recently had a blackout, and the generator ran for about three hours when it stopped producing power. I checked the generator; it was still running but no LEDs were lit on the front panel. There was no output voltage coming from the machine. I stopped the generator to let it cool down so I could look at it the next day. I then began disassembling it to access the internal workings. The machine looked pristine inside. I was impressed by how well-made everything was and how easy it was to take apart. It used all regular Phillips screws, standard metric bolts and no breakable plastic clips. I had to pinch myself to see if I was dreaming! I began by looking for any burnt connections or loose/broken connections. Nothing showed up in that regard. Thankfully, I have a wiring diagram (shown opposite). I disconnected all the plugs from the inverter board to avoid getting false readings or damaging any electronics while testing. I measured the resistance of the three-phase star-­ connected winding labelled “Main windings” in the diagram. All three windings measured 0.7W to the common star point of the windings. All resistance readings should be equal when testing three-phase machines. Sometimes, this information is available from the manufacturer’s website, but I was on my own in this case. I used a Megger set on 500V to test between these Australia's electronics magazine siliconchip.com.au windings and the machine’s frame. It showed above 100MW, so the windings were OK. Next, I measured the resistance of the exciter winding labelled “Control winding” in the diagram. This showed a reading of only 0.2W, which concerned me, but I also checked with an inductance meter and measured about 16µH. That wasn’t totally convincing but at least it proved that the winding wasn’t shorted. This winding was also over 100MW to the frame on the Megger test. At this stage, it looked like the inverter board was the culprit. Unfortunately, it is potted in epoxy resin and seems quite complicated. I may try to de-pot it at a later date. I jumped on the web and was horrified to find no leads about MTM generators or where to get parts. Thoroughly disgusted, I gave up for the day. The next day, I tried another web search and stumbled on a company called Generator Guru (www.generatorguru. com). This was my saviour. They specialise in saving Chinese generators from being thrown away. I got onto their site and searched for my brand and model number. Up came the inverter board and all the other spares they have for this generator. I was impressed. The sad news was that the inverter board would cost $525 with free shipping. That was more than the cost of the generator itself! I thought about it and remembered that I bought it 10 years ago, so considering inflation, it wasn’t all that bad. I searched the web for other generators of similar capacity, which cost at least $1200. I was also loathe to write this machine off. I decided to bite the bullet and ordered the inverter board, plus the ignition controller board, as a spare. Their service was brilliant, with regular tracking updates about the shipping from Sussex in the UK. The parcel arrived in about five days, which was also impressive. When I had a spare moment, I fitted the board, which has six connectors that are all different, so you can’t muck it up. I left some covers off and fired up the generator. All was well with a 10A kettle plugged in as a test load. I refitted the covers and gave it another test run to be sure. I am pleased to have saved the generator from scrap and highly recommend Generator Guru if you are chasing parts. They also have some repair videos and advice SC columns. The wiring diagram for the MTM 4400i inverter generator. siliconchip.com.au Australia's electronics magazine May 2024  91