Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG ELSEC 764 UV Monitor Repair A friend recently handed me a small handheld humidity, temperature and light meter and asked if I could figure out why it wasn’t working. These meters are typically used in art galleries and museums to monitor environmental conditions, ensuring the best possible conditions for exhibited items. The meter is branded ELSEC model 764 UV + Monitor; it appeared to be a nicely made instrument from the UK and in relatively good external condition. However, on opening the battery compartment, I found the normal 2×AA alkaline batteries were absent and that there had been a fair amount of battery leakage at some stage. I set about cleaning up the terminals and inserting two new cells. The euphoria of having solved a relatively easy problem was short-lived, however. The LCD on the meter sprang to life for a second or so, but then immediately went blank. This repeated over and over every few seconds. The information being displayed on the screen in the short time the display was on gave me some confidence that the microprocessor and sensor circuitry were operational, so it was time to take the back off the meter and explore a bit further. On exposing the main PCB, I found that the alkaline battery paste had spread down one side of the board and engulfed a number of SMDs. I used isopropanol on a cotton bud to carefully clean up the dried paste as best as I could, then used a contact cleaner spray to finish up. After re-inserting the batteries, I was happy to see that the meter’s display was now on and displaying believable numbers for temperature, humidity, UV and light levels. Problem solved, I thought. The meter has a clock function that allows it to display the time and date of minimum and maximum measurement excursions. This is very handy for tracking longer-term environmental variations. For example, you can leave the meter running all night in the gallery and then look to see if there were any out-of-bounds measurements during the night that need further investigation. Perhaps the gallery’s air conditioning system had a hiccup at 4am for some reason. The meter’s time and date are entered using the meter’s front keypad, following the steps of a clock setting function in the main menu screen. Unfortunately, the clock did not begin to advance after being set via the menu. On closer inspection, three of the SMDs affected by the battery paste were a Philips I2C 8583T real-time clock (RTC) chip, an adjoining 32.768kHz quartz crystal oscillator and a very small crystal loading capacitor. Although the meter was allowing me to enter new values for date and time, it appeared that the RTC was not subsequently ‘running’. Under 12× magnification of the chip and crystal, I could see the tracks running between them disappeared under both devices. I could also see what looked like more 82 Silicon Chip solidified battery paste around the bodies of the crystal, the RTC chip and the PCB. I didn’t relish the idea of removing the crystal or RTC from the PCB, so I gently poked around the edges with a fine-gauge stainless steel wire and dislodged a bit more of the paste, especially around the RTC oscillator pins. After another spray with contact cleaner and powering up the meter again, the clock was now advancing. I left the unit running for a few hours and again checked to see if things were OK. The clock had lost a significant amount of time, so the oscillator circuit was still being affected, probably by battery paste lurking out of sight under the RTC chip and/or crystal. Alkaline battery paste is, unsurprisingly, chemically alkaline and thus both reactive (corrosive) and conductive. As a last resort prior to physically removing the chip and crystal, I thought I’d try to dissolve any hidden paste using a mild solution of acidic white vinegar. I shielded adjoining parts of the PCB with soft tissue and applied a few drops of white vinegar to the top row of pins on the RTC chip and crystal while holding the PCB vertically. I could see the vinegar emerging from the lower underside of the RTC chip and crystal, so it was obviously making its way under both devices. I repeated this a couple of times while cleaning it with contact spray in between. After a couple of days now, the RTC is keeping good time, and the meter appears to be working as it should. I was able to compare the meter’s readings with another more modern and more recently calibrated ELSEC 765 meter. All the measurements, including humidity, were almost identical after having sat together for an hour or so. Australia's electronics magazine siliconchip.com.au From the default clock settings at power-up, and from what I can glean from the internet, the meter was made circa 2001 and now, hopefully, will provide a few more years of productive service. Best of all, from my perspective at least, is that a nicely crafted and not inexpensive professional instrument was spared from today’s scrap heap. David Worboys, Baulkham Hills, NSW. Rescuing a muddy Toshiba C665 laptop My son came home from town and presented me with two muddy laptops that he’d picked up at the tip shop for $2. The first one was an HP that was wrecked and covered in mud; it was too bad to save. The second was a Toshiba C665 that appeared to be in good condition, apart from all the mud. It was missing the hard drive, RAM and their covers, but it was otherwise complete. I looked inside the back and it looked clean and dry, so I fitted a RAM module, plugged in a charger, held down the F2 key and pressed the power button. The laptop switched on, and the BIOS screen opened. I could see that the screen had water in it and would need to be replaced. I pressed the right arrow key but it did not work. Some other keys also did not work, so it meant that the keyboard also had water in it and it would need to be replaced as well. But the main thing was that the shell was good and the motherboard worked; everything else that was faulty could be replaced. Now to dismantle it. I removed the keyboard and I could see the CMOS battery (CR2032 cell) there, so if it had just needed the cell replaced, that would have been easy. I removed all the screws on the bottom, then split the top and bottom shells apart. I removed the lid and put it aside, then took out the motherboard, as I wanted to check the heatsink and fan, and clean them. The fins on the heatsink were pretty clogged and the fan was caked in dust. I removed the heatsink from the motherboard, cleaned the fins and cleaned off the old heatsink compound. Then I applied new heatsink compound, refitted the heatsink and cleaned the fan. After that, I cleaned the inside of the bottom shell to remove the dust and mud. Next, I refitted the motherboard and CPU fan, then turned my attention to the lid. I removed the four screws from the corners of the lid, pulled off the front screen surround and removed the last two screws. That got the screen loose. I unplugged the connector on the back and removed the two hinges. It was the standard 40-pin connection that many screens use. I’d recently wrecked a non-working HP Pavilion G6 laptop that had the same screen as this Toshiba C665, so I connected the screen and set things up for a test before I started the reassembly process. I used a USB keyboard so that I could go to the BIOS setup screen and check that the replacement screen was good and free of defects. It’s always a good idea to test any replacement parts with the laptop partly assembled. I always do this whenever I upgrade the CPU as well. The screen was good, so I put the laptop back together, replacing the CR2032 cell with a new one as I did. I got a matched pair of two 2GB PC3 RAM modules and fitted them. I had a 320GB hard drive I could install, but I needed siliconchip.com.au Items Covered This Month • A leaky ELSEC UV monitor • Plus a muddy Toshiba laptop • A signal analyser and smoking transformer • Repairing a Sanwa multimeter • A simple fix for a DigiTech radio • Pfaff sewing machine repair 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 A close-up shot of the ELSEC 764 UV monitor’s PCB, showing the SMD ICs that were most affected by battery leakage runoff. Australia's electronics magazine October 2026  83 to find a hard drive bracket. I searched in my box of hard drive brackets, found a suitable one and used it to install the drive. Now I needed a RAM cover and a hard drive cover. I checked my box of covers, hoping that I had suitable ones. I’ve wrecked a lot of laptops over time, and I keep any parts that might be useful. When I was nearly at the bottom of the box, I found a RAM cover, then shortly afterwards, I found a usable hard drive cover. Next, I got out my box of salvaged keyboards, hoping that I could find a good keyboard to replace the faulty one. It didn’t take long to find the right keyboard, which I connected, then I sat it in place to test it before it was finally installed. Now it was time to install Linux. I hadn’t previously installed Debian on a laptop, as I had been using lighter versions on the old laptops that I’d been working on. This laptop was newer, so I thought it would be adequate to run Debian. I downloaded Debian with the KDE Plasma Desktop and burned it to a single-layer DVD. I booted from the DVD and installed Debian. Using Mousepad, a simple text editor similar to Notepad, I tested the keyboard and all the keys worked, so I then reinstalled the top strip. I found that Debian was a bit slow with the i3 processor. I checked the RAM usage and it was only using 2.2GB of RAM out of 4GB for basic tasks, but 8GB would be better for more intensive use. At this stage, I discovered that the battery was not charging, so I checked my box of salvaged batteries. I had two, but one was faulty and the laptop wouldn’t even boot with it in place. The other only held a small amount of charge and it would not charge above 0%, but it would run the laptop long enough to unplug the charger for a few seconds. This was an interesting exercise in saving a piece of junk and turning it into a useful laptop once again. The total cost was under $10, including the $1 my son paid for the laptop itself. I may purchase a new battery in the future, but they are around $40. Maybe a good one will show up in a dead laptop at some point. Bruce Pierson, Dundathu, Qld. Recently, I switched it on and shortly later noticed a tiny wisp of smoke starting to emerge from the ventilation holes! I freaked out and immediately pulled out the IEC connector from the mains, even though the display looked perfect. I thought it must have been burning up in the region of the IEC mains socket. This was a problem I had come across previously in an Agilent 3458A multimeter. In that case, the capacitors and inductors that make up the mains filter in the socket burnt up. The filter box contained cooling oil, which leaked out and allowed the device to overheat and smoke up the whole room! With that instrument, there was no part number listed for the filter assembly, so I replaced it with a Jaycar MS4003 and the instrument worked fine. Back to the 35670A; it is a very depressing moment when smoke appears from an instrument worth thousands of dollars, so I put it away for a later repair. The day arrived; it was time to disassemble the device to confirm that the IEC mains filter was faulty. Four Allen-head machine screws on the rear released the cover to expose the internals. I undid a few more screws, then was able to extract the switch-mode power supply module. I removed the multi-pin connector and placed the power supply on the bench for closer inspection. I held the supply horizontally the whole time because I did not want to have any oil spill all over the works. The smoking analyser This Agilent 35670a dynamic signal analyser is identical to the instrument I reviewed in the June 2012 issue (siliconchip.au/Article/652). I often use it for setting up microphones for testing loudspeakers, infrasound detectors and many other projects. It can analyse audio signals from 290mHz to 102kHz, including frequency response and distortion measurements, but I don’t use it every day. 84 Silicon Chip The very muddy Toshiba C665 laptop that Bruce Pierson needed to clean and refurbish. Australia's electronics magazine siliconchip.com.au The power supply module (left) for Allan Linton-Smith’s Agilent 35670a dynamic signal analyser (right). The device had smoke coming out of it when turned on, which ended up being caused by the floppy drive connector. This SMPS was a separate box totally enclosed by a cage, which I partly removed to inspect the components. The IEC filter looked perfect and there was not a trace of oil... so where had the smoke come from? I put the power supply back in its place, but left the cage open. I switched the instrument back on and watched with a torch to see where the smoke was coming from. It came from a switching transformer – making it seem like it would be an expensive repair! I searched online and found a power supply from a Singapore-­based dealer I know, but it was very different, so I went back online to look for used 35670As. Just about all of those available either had no power supply or were sold as “not working – for parts only”. Possibly because of the same fault. The repair manual refers to the fact that there could be an overload downstream of the power supply, so I removed the connector and then switched it on as before. Bingo – no smoke! All I had to do now to fix it was to work out which module had a problem. I removed the power supply connectors from each module one by one, switching it on at each step to observe if some smoke appeared. I was not too confident that I could disable each module because most of the connectors were hidden amongst a stack of three large circuit boards. On the top of the instrument, exterior to the main circuit boards, there is a floppy drive with a multipin connector that was easily disconnected, so I began there. I switched on the power and waited for a minute. There was absolutely no smoke. What luck! The floppy drive must have jammed and put an enormous load on the power supply feeding it. I left it to soak for a couple of hours, but everything remained normal. siliconchip.com.au I use the floppy drive to save configurations, to avoid the many button presses that would be required to set the instrument up from scratch each time I want to make a measurement. The floppy drive jammed and refused to eject the disk some months ago, so on a subsequent start-up, the electric motor must have been acting like a short circuit, overheating the power supply. Because the power was shut off quickly, the 800414B transformer (shown photographed at upper left) seems to have survived, but a replacement is on my wish list. Fortunately, I had already ordered a brand-new floppy drive, which I will use to replace the damaged one. So it looks like in this case, a little smoking has not resulted in any fatal health problems. Allan Linton-Smith, Turramurra, NSW. Sanwa U-50 DN multimeter repair Back in 1972, I bought a Sanwa U-50 DN multimeter that served me well for over 50 years. Its use declined slightly with the arrival of digital meters, but I always found the old analog meter to have its place in peaking adjustments and monitoring pulse-width modulated signals. Some months back, the old meter developed a loose-­ feeling selector switch and, upon opening the case, I found a small steel ball (of the sort found in ball bearings) rolling around freely. What had happened over the years is that the hole in the spring holding the ball against the detent had worn larger and allowed the ball to escape. The instrument was constructed with the point-to-point wiring scheme used in vintage radios, making it extremely difficult to access anything near the selector switch. I decided that one ball was enough for normal use and simply reassembled the meter in its case. A week later, the other detent ball dropped out the same way! I checked the internet for balls of a slightly bigger diameter, but the next size up was too large. Not to be thwarted, I turned a scrap of nylon rod to form a hemisphere on one Australia's electronics magazine October 2026  85 end, with a short stub and a retaining ring on the other end. This took quite some time to fit due to the poor access to the switch, but I finally got it in. The meter worked perfectly, but several weeks later, I discovered that the AC volts range no longer worked. This range was rarely used, so it may have escaped my attention for quite some time. Opening up the meter again, I found a tiny copper-oxide rectifier that had clearly gone short circuit and was almost certainly not available after so long. I mentioned this to a dear friend of mine who kindly sent me a selection of germanium diodes and transistors to try. A pair of OA5s actually worked extremely well and did not visibly alter the calibration when compared to a digital meter. A month or so later, the meter stopped working on all ranges. I first thought the meter movement had gone open circuit. Applying a few tens of microamperes directly to the movement resulted in a significant deflection, so it looked like the fault should be repairable. Examining the works showed that the selector arm was not lined up with the contacts; in fact, it was halfway between the contacts! The whole switch assembly was quite loose and held in place by only the wiring. It took quite some time to locate a couple of screws that secured the switch assembly to the case because the heads of the screws were largely hidden under the closely packed resistors. With the selector arm correctly positioned, these screws were tightened, but still no meter movement. Tracing out the circuit showed three resistors, one of which might be open circuit, but each tested extremely close to the marked value. Every continuity test failed to reveal the problem, but finally, after turning the meter over many times, I saw a momentary near-full-scale deflection on the Ohms range which, at that moment, had a short applied. I had previously tugged at every solder joint without finding any problems, but there had to be some intermittent connection somewhere. It turned out to be the solder joint on the negative input. A glob of solder had somehow solidified into a tiny hook under the connection lug. It looked like a perfect solder joint from the top, and it resisted pulling, so had escaped further scrutiny. Both the lug and the wire were well-tinned, so it only took a brief touch with a hot iron to finally make a firm connection. Net cost: several days’ labour. Graham Lill, Lindisfarne, Tas. DigiTech DAB+ radio repair Sometimes the cause of a fault can be puzzling, but trivial and obvious in retrospect. A vanishing breed of hifi buffs still goes for ‘component audio’, made up of amplifiers, tuners, CD players, turntables and even preamplifiers. In this case, the offending item was a combined FM/DAB+ radio tuner made by DigiTech (Model AR1753). The problem was that the on/off switch no longer responded to pressure. The switch was mounted on a small circuit board attached to the front panel by four small plastic columns. My plan was to push aside the little projections from the columns and prise off the board, desolder and replace the switch, then replace the board. However, on opening the radio, it became apparent that two of the tabs were already disengaged from the board, such that the push button on the front panel (arrowed in the photo) could no longer reach the switch on the board. Snapping both tabs back in place was all that was needed to restore the switch action. In the photograph, I have restored the left-hand tab to its correct position but kept the right-hand tab as I found it. I suspect that this type of problem is not rare. I surmise that many users apply excessive pressure to pushbuttons. How many switches have been replaced when the problem was mechanical and not electrical? James Goding, Princes Hill, Vic. Pfaff 7550 sewing machine repair This fault is something I have never encountered before or since on any equipment. The story starts over 25 years ago. Sewing is a very serious hobby for my wife; she loved her all-electronic 7550 machine, but she wanted to upgrade to the later 7570 model because it had extra features. A friend asked if she could buy the old machine, paying it off monthly; she got it for a good price. A few years later, she moved to Queensland, and some years later, we visited her. My wife asked if she was still doing much sewing, and she said sadly she was back on her mechanical machine – the Pfaff was consigned to the garage. It had developed an intermittent fault where the The PCB shown here just needed to be pushed further into the locking tabs so that the pushbutton (marked with a white arrow) can properly contact the board. Servicing Stories Wanted Do you have any good servicing stories that you would like to share in The Serviceman column in SILICON CHIP? If so, why not send those stories in to us? It doesn’t matter what the story is about as long as it’s in some way related to the electronics or electrical industries, to computers or even to cars and similar. 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. 86 Silicon Chip Australia's electronics magazine siliconchip.com.au stitching became erratic and slow, eventually coming to a halt with garbled images on the display. The fault could develop after a few minutes, or it could go for more than half an hour. She got three separate repair quotes, all around the $1000 mark (I think one was about $1400) – quite a lot for the time. She couldn’t afford the repairs, all involving main board replacement. We felt a bit obligated, saddling her with a dud, so I said I would take it home and do my best – no guarantees. I knew nothing about sewing machines and would be chasing an intermittent fault without a schematic, but still, it’s just electronics. How hard could it be (famous last words)? I decided to allocate a couple of hours each day to familiarising myself with the board (easily removed through the base), eyeballing it for anything suspicious, looking for bulging caps or discolouration. I checked the voltages, which turned out to be rock solid with no ripple, even when the fault manifested itself – so the power supply was eliminated as a source of the fault. It looked all the world like a heat-related problem, but nothing looked, felt or smelt like it was under any heat stress. I even replaced some electrolytic capacitors and a transistor array, more out of hope than conviction. I got all excited when it went for more than half an hour, but no, the fault returned. I had managed to ascertain that the heart of the system was an Intel processor, but the real-world interface (controlling stepping motors, the display, keyboard etc) was through a large ASIC chip. It looked like, and probably was, a dedicated processor. Since I was getting nowhere with a meter, I decided to check around all the terminals of the ASIC while it was working and see if I could spot anything different when it failed. The chip ran on an internal clock that was presented at one of the terminals. It produced a nice square wave, but I noticed that the negative part of the signal was a little high (about 1V). Still, I didn’t place much significance on it at the time. However, when the fault started manifesting, the clock signal’s negative excursion was now above 2V and slowly rising until everything stopped. I didn’t record any notes at the time, so I can’t recall how high the voltage drifted. Now I was faced with a dilemma. Even in the unlikely event that I could source such a unique component, I would almost certainly destroy the board trying to replace it. Why not try a Hail Mary fix? I ruled out a resistor to ground (it might ruin a high-­ impedance signal). I decided on a low-value ceramic disc capacitor to ground (I can’t remember the value now) and it worked! The signal’s negative extent remained at about 0.5V with no drift. I put it back together and powered it up for two weeks straight, and asked my wife to test it every day, then shipped it off to Queensland. Our friend was recently visiting, and she said it has never missed a beat since. I hadn’t given much thought to it since the fix, but then I read the remarks of one of your contributors in the July 2023 Serviceman’s Log column that a number of people were having main board problems. I would be interested to know if any of your other readers have had a similar problem, or if we just got one unlucky dud. SC Frank Murray, Downer, ACT. siliconchip.com.au Australia's electronics magazine October 2026  87