Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN’S LOG Spy games and super-villain gadgets Dave Thompson It’s no secret that I like gadgets and I suspect many electronics enthusiasts do too. In my case, the weirder the gadget, the better, and even more so if the device is something I can make myself. However, some gadgets move beyond weird and into worrying territory, and this time I had a chance to repair one. As a kid, I was always fascinated with Maxwell Smart’s shoe-phone. Not so much that it was a phone, but that it was hidden in a shoe. Throughout the 80s and 90s, I used a couple of space-age Ericofon Cobra phones with the dial on the bottom as my home telephone as an homage to that shoe phone. (I have three in my collection and they all still work!) The fascination extends to spy gadgets in the slightly less silly spy movies: rotating number plates, oil slicks at the push of a button, rockets mounted behind the headlights and so on. Of course, all those things are faked, made for the big (or small) screen, but they are no less cool, and they sparked my lifelong interest in such things. I get the same feeling when I see a wall safe concealed behind a hinged picture frame, a secret door to a passage behind the fireplace, or a bookcase operated by a hidden lever. A long time ago, I messed around making ‘bugs’, tiny radio transmitters that broadcast to a transistor radio. Of course, in the spy movies I grew up on, bugs are tiny little things that can be stuck anywhere with the press of a finger and transmit several kilometres, even without a prominent antenna. In reality, they need to be a little bigger and, even then, can reach out only a few dozen meters, even if they are relatively sophisticated. Someone needed to be nearby with a receiver to pick up the signal, which of course, was also detectable by the bad/good guys with electronic bug sniffers. Spy games indeed! Editor: for some fascinating related stories, see our articles on Cyber Espionage in the September & October 2019 issues at siliconchip.au/Series/337 As time went on, I made more advanced projects, though I would still build any ‘bug’ that appeared in any of the magazines of the era. Some worked OK, some very well, but regardless of performance, I loved experimenting with them. I never used them in any surveillance role, but they gave me hours of fun. This is how I learned; by doing. A little while ago, I wrote about an old-school night-vision device a customer brought in (April 2022; siliconchip.au/ Article/15283). That device was featured in a late-1970s project magazine, and around the same time, another gadget was advertised in those publications as a “pain field generator”. At the time, I was very curious as to what this thing was and how it worked but never looked into it any further. It turned out that a friend of the guy with the night-vision scope had purchased a short-form kit and plans for one of these ‘generators’ years ago and, had tried to put it all together, without much luck. He contacted me after the night-vision thing worked out and wondered if I would like to look into it for him, and perhaps get it working. I sure would! Pain in the wotsit The plans he’d imported included a reproduction of the original magazine project article featuring this device, explaining how it worked and what to expect from it. It turns out that this project (or one very similar) is also featured in one of those ‘Evil Genius’ project books that were popular a few decades ago. On the face of it, it seemed straightforward; it is essentially a high-frequency oscillator that could be manually varied in frequency and modulation to produce some very annoying high-level sounds that could potentially be damaging to humans and animals. I suppose this is the “pain” they are alluding to in the blurb. There are several iterations of the project, from a ‘pocket’ version up to one you could mount on a perimeter fence. The main difference was the output power and the speaker array used. The parts used in the project (or their modern 72  Silicon Chip Australia's electronics magazine siliconchip.com.au Items Covered This Month • • • • Spy games and super-villain gadgets An overloaded Onkyo receiver Intermittent lights in a trailer tow bar Fixing washing machine PCBs 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 equivalents) are still widely available, and I saw no reason not to have a go. However, I was initially a bit cautious because I didn’t know these guys or what they would use something like this for, and I was unsure of the legalities of even owning/ building something like it. A bit of research on the topic showed that something very similar in theory – though much more powerful – is already in-use at a military level by the USA and shipping companies (against pirates, among other possible uses). The Israeli Army developed another version for use in the Middle East. Called LRAD, or Long-Range Acoustical Device, they consist of a large, focused speaker array that can emit sounds at very high levels to make the ‘listener’ very uncomfortable, theoretically without doing any permanent damage. Another similar ‘civilian’ unit is marketed and used worldwide to prevent groups of young people from congregating around shopping malls or other areas where loitering kids could potentially cause problems. Older people cannot usually hear the high-frequency sound these devices emit and are mostly unaffected, but it is apparently quite uncomfortable to younger ears. Of course, the kids soon learned that by wearing earpods or headphones, they could easily defeat the system. Even so, these devices are apparently still deployed in many countries for this very purpose. I suppose it would be like anything else – a stereo system can be used as a sound weapon (ask annoyed neighbours!), so it matters what something will ultimately be used for as to whether it is deemed dangerous or not. I asked the customer, and he said that for him, it was only about curiosity and experimentation – much like my own motivations for wanting to make one. Perfidious perfboard With that dealt with, I looked at what he’d done already. And the answer was “not much”. The ‘kit’ came with some of that older-style perfboard prevalent in the days before DIY PCBs. The project required component leads to be put through holes and then routed underneath and soldered together to create pseudo ‘tracks’, creating a facsimile of a hand-wired printed circuit board. This construction method is fine when done correctly – however, this one wasn’t; it was a bit of a mess. I thought it best to salvage what components I could and replace those I couldn’t. siliconchip.com.au Australia's electronics magazine August 2022  73 happened. Cranking the pots and toggling the sweep switch did nothing. No pain, no gain For example, the project used a couple of 555 timer ICs, and I couldn’t easily extract them from the rats’ nest. It would take way more effort (and potentially do more damage) to try to remove them, so instead, I reached into my parts box for a couple of new ones. I recovered an IRF540 Mosfet, a custom-wound transformer/choke and a couple of other inductors; the rest I just replaced with new components. I used Veroboard for this build. I know it isn’t very popular among some out there, but for something like this, it is (relatively) cheap and easy to work with. The voltages and currents involved are well within the limitations of a construction method like this. I have used the excellent open-source VeeCAD software (veecad.com) in the past for complex layouts on Veroboard, but this would be a reasonably straightforward build, so I just ‘winged it’. If push came to shove, I could easily redo it using the CAD program. Almost all the components mount on one board, with two Motorola piezo tweeters mounted externally in a suitable enclosure. Those tweeters have off-board inductors mounted directly on their terminals. Two pots are mounted into whatever case is used and these, along with a modulation setting and a power switch, are the only controls. The first 555 in the circuit is configured as a free-running oscillator with variable frequency control. In contrast, the second 555 is configured to produce a sweep voltage that modulates the output of the first oscillator. This sweep is controlled by the second pot and can be switched in and out. The output is fed to a Mosfet and then on to the LC network of the tweeter array. The whole thing actually reminds me of the Barking Dog Blaster project from the September 2012 issue of Silicon Chip (siliconchip.au/Article/529); that is far more advanced, but the output section is very similar. I also made one of those ‘blasters’ back in the day and, with it, successfully ‘trained’ a dog a few doors down. It learned that if it barks constantly, some uncomfortable sounds would come its way! I assembled the project and fired it up on the bench. I powered it using a benchtop power supply (it runs from 9-12V DC) and flicked the power switch on. Nothing 74  Silicon Chip Now, it could be that I am so deaf after years of racing model aircraft and playing in bands that I just couldn’t hear it. Still, as I also didn’t experience any of the disorientation, dizziness or headaches they claimed in the promotional material, I was pretty sure it wasn’t working. Time to dig a bit deeper. The first thing I did was check my layout and make sure I’d cut all the tracks that needed cutting, ensuring I hadn’t cut any that didn’t! Tracing through the circuit diagram and comparing it to my layout, it all looked fine to me. The next easiest thing to check was the Mosfet. The IRF530/IRF540 used in the project needs a heatsink fitted, but even though I hadn’t added one yet, the component wasn’t getting warm. I didn’t bother checking it; I have many types with similar specs in my parts bins that would work in this circuit (N-Channel, 100V 30A 100W in TO-220). I found a suitable alternative and soldered it in. The only other thing was the choke, which the customer wound himself. I know from my own experience of winding transformers and inductors that they can be tricky things to get right. Since the bobbin and E-cores the guy used came as part of the kit, I knew they were at least the correct types. As it has only 50 turns of “#24 magnet wire” on the bobbin (24 gauge, or 0.51mm diameter enamelled copper wire), it was easy enough to strip it off and rewind it myself using nice new wire. Fortunately, using one of the winding jigs I’ve made up over the years made this a simple, though laborious, task. One anomaly I did spot redoing this choke was that the original plans called for three six-thou (0.006-inch or 0.15mm) shims to be placed between the two E cores. No such shims fell out when I pulled the cores apart, so I created some from plastic and tacked them to the prongs of one of the cores with superglue before putting it together. According to the component description, it should measure 1.5mH (millihenries), and my Peak LCR meter tested it as 1.71mH, which was close enough for me. After taping everything up, I scraped the enamel from the two flying leads and soldered it back into the board. After another quick check-over, I powered it on once again. This time I could hear noise from the tweeters. It almost sounded like white or pink noise until I started messing around with the frequency and modulation controls; then, all hell broke loose! This thing was loud! I powered it off, closed the workshop doors and put a pair of earmuffs on. I also buried the tweeters under a couple of folded-up drop-cloths. I tried it again and ran it at a reasonable level, noting the current draw and onboard temperatures. It was almost unbearable at the audible (for me) end of the range. At the upper end, all I could hear was that slight hiss, but I could feel a kind of pressure in my skull, a very odd and uncomfortable sensation. This sound pressure level is still likely to cause hearing damage if I was exposed to it for long enough, even though I couldn’t hear the actual output. Australia's electronics magazine siliconchip.com.au The Mosfet was now starting to get warm, and as I knew the unit was going to work, I powered it off and set about prepping to mount it in a case of some sort. As I was fixing a heatsink to the Mosfet, there was a knock on the workshop door; it was our nearest neighbour wondering if our alarm was going crazy. I apologised and assured him everything was fine and that it was just a project I was working on. I also apologised in advance, explaining I would have to test this thing again once I’d built it into an enclosure. I told him I would try to keep any noise to a minimum (if that was really possible)! I found a plastic Jiffy box that would accommodate the circuit board, though it would require the usual drilling and chopping around to fit all the stuff into it, and I’d still have to find some way to mount the tweeters. Looking around my workshop, my eye settled on an old set of computer speakers under the bench. These were reasonably large, with timber backs and sides and a moulded plastic front. I reckoned the whole shebang would fit into one of them, and the pots and switches could poke out of the back side – this way, they could be manipulated with the speakers pointing the other way! There was plenty of room, and all I’d need to do was remove the plastic front (held on by four screws) and the old drivers with it, and replace it with one made from Thinline MDF. The tweeter holes were easy enough to mark out and cut in the timber, and with a quick sand and a spray with matte black paint, it looked like a bought one. I used the original mounting holes to fix the tweeter array to the rest of the cabinet. The project was designed to be portable and run on batteries; the customer was not keen on this and asked if it could be mains-powered; he’d only be playing with it around his home anyway. I dug out a 12V 1A ‘wall wart’ type power supply from my bins and simply added a socket to the back of the unit to match the plug on the supply. That should be ample. The finished device looked pretty good. I once again packed bunched-up material in front of the tweeters, put on earmuffs and switched it on. And again, I was greeted with a lot of noise, and after playing around with the controls found I could get some hugely annoying sounds out of it. I could see it would be very disorientating if someone were suddenly exposed to it. With earmuffs off, it was literally unbearable to be around, with even my teeth feeling as if they were vibrating when the sweep was set just right. Nasty! The guy was very happy with it and looked forward to his ‘experiments’. I don’t think his neighbours will be that happy, though! I almost feel sorry for the real-life pirates at the receiving end of the ‘big daddy’ LRAD devices. Almost. Those guys are not quite as affable as Jack Sparrow, and the AK-47s they carry are a bit more menacing than a single-shot flintlock pistol... An overloaded Onkyo receiver R. S., of Fig Tree Pocket, Qld repairs a wide variety of devices. This time it’s an old Onkyo TX-SR506 7.1 AV Receiver which would have cost a pretty penny new. Here’s what he found... The Onkyo receiver would not switch on, indicating an overload. It has seven amplifiers, and one of them had shorted output transistors. Q6053 and Q6063 (visible in this section of the circuit diagram overleaf) had failed short-circuit. Replacing the output transistors with new ones, protected by 100W 5W resistors in the collector circuits (to siliconchip.com.au Australia's electronics magazine August 2022  75 NAAF-941 U01 AMPLIFIER PC BOARD R6093 0.22 (1/4W) Q6013 -0.6V P6083 ID+ ID- R6103 0.22(3W) Q6063 LIST R6173 VPRO 47K R6183 Q6043 2SA1930 -1.1V 33K R6163 SPSL D6013 220K C6043 +47/50 IDLING CHECK Q6073 2SC2240 C6053 103J R6143 22K VOLH D6003 47K D6003, 6013 : KDS4148U R5193 10 (1/4W) -52.5V A close-up of the power amplifier section of the Onkyo TX-SR506 receiver circuit. Q6004, 6014 : 2SC1740S-S R5184 SR IPRO Q6053 LIST R6193 Q6033 2SC5171 R6083 0.22 (1/4W) Q6003 + C5053 47/50 R6043 2K -0.4V 3.3K 22K R5133 C5093 101K Q5043 2SC2229-Y R5173 LIST +52.5V R6153 12K 470 -0.3V +0.6V R6073 LIST R5203 +1.0V 3.9K 5.6K R6013 R6003 R6033 Q5033 2SA949Y R5163 LIST 2.2 (1/4W) IDLING ADJ. C5113 + 22/100 Q5053 LIST R5063 R5073 1.2K 100K C5023 + 10/50 R5103 D5003 MTZJ5.6B -51.5V R5233 120K R6023 +1.1V R5033 120K C5083 040D C5103 + 22/100 Q5013 R5053 4.7K -46.5V 10 (1/4W) C5043 + 220/25 -0.65V 470 56K 330 NC R5083 R5043 0V 2.2K R5013 R5023 221K C5003 1K R5113 Q5003 Q5003, 5013 : 2SC2240 R5003 C5013 47/50 1K + +50.5V 100K R5093 Surround Left ch +51.5V +52.5V R6026 0 +1.1V R6164 33K R6165 33K D6015 220K R6195 Q6005 + C5055 47/50 Q6015 siliconchip.com.au Q6006, 6016 : 2SC1740S-S 36 Y R5166 LIST 1K 10 (1/4W) D6014 220K C6054 103J C6055 103J R6194 R6154 12K C6044 +47/50 C6045 +47/50 R6155 12K R6074 LIST Q6014 R6075 LIST Q6004 + C5054 47/50 470 3.9K 5.6K 3.3K 470 22K 22K R6015 R6005 R6035 R6055 C5095 101K Q5045 2SC2229-Y R5175 LIST 3.9K 5.6K R5204 R5134 22K 22K R6054 R5205 R5135 C5094 101K C5085 040D Q5035 2SA949Y R5165 LIST R5174 LIST Q5044 2SC2229-Y C5114 + 22/100 470 C5105 + 22/100 C5045 + 220/25 C5115 + 22/100 470 3.3K R6014 R6004 R6034 Q5034 2SA949Y R5164 LIST C5044 + 220/25 Q5054 LIST Q5015 R5186 +51.5V R5116 Sorround Back NC R5085 R5065 R5075 1.2K Q5055 LIST 56K 2.2K NC R5084 R5064 R5074 1.2K R5115 R5015 R5045 Q5005 100K R5095 1K 100K C5024 + 10/50 R5104 D5004 MTZJ5.6B D5005 MTZJ5.6B C5025 + 10/50 R5105 100K Australia's electronics magazine  Silicon Chip SBR C5084 040D Q5014 Q5004 R5014 56K 2.2K R5044 C5004 221K SBL 76 C5104 + 22/100 1K 100K R5094 R5114 R6024 (1/4W) the supplies), did not work. The output of the 10 amplifier cosmetic panelling of the boot. This picks up the various 2.2 (1/4W) Q6034 +1.1V Sorround Right ch 2SC5171 went to the positive rail immediately. The driver transis- signals (brakes, turn indicators, taillights and soIPRO on) and R6084 Q6054 tors, Q6033 and Q6043, were also +50.5V shorted, as was the Vbe drives the trailer lights putting any load on the 0.22without +1.0V LIST (1/4W) multiplier Q6013, and 100W resistor R6073 (between the car’s internal electronics. Otherwise, the trailer connec+0.6V Q6074 R6144 2SC2240 22K driver emitters) was burned and open-circuit. I replaced tion could interfere with, say, flasher timing or blown Q5004, 5014 : 2SC2240 all of these. globe detectors. R5034 IDLING C5014 R5004PNP The transistorR5024 Q5033 controlling the driver120K transis- -0.3VI took my car to the trailer dealer and tow bar installaCHECK 47/50 1K 0V SPSR + tors seemed OK, but 330 I have had trouble before with tran- tion experts, which involved a round trip of about 200km. P6084 R6104 -0.65V R5234 sistors in this part of the circuit, as they can be leaky. So After spending a day wandering around shopping centres, 0.22(3W) ID+ 120K R5054 IDI replaced it as well. I returned to be shown everything working correctly. 4.7K IDLING This time the amplifier worked, and the output centred So I drove the hundred or so kilometres home and ADJ. R6094 close to 0V. Maybe what happened was that Q5033 leaked R6044 thought, this is pretty0.22 good; I’ll just have a look at the (1/4W) Q6064 2K -0.6V enough current to destroy Q6013, then the voltage between lights before putting the trailerLIST away. Oh, dear! Some of -46.5V -0.4V R6174 the bases of the driver transistors rose high enough to the lights didn’t work or were intermittent. A phone call VPRO 47K the destroy them, and the output transistors as well. to the dealer later, we decided that job had VOLH to go back R6184 The circuit is unusual as it has two Vbe multipliers, with to be fixed. D6004 47K Q6044 -1.1V Q6013 in contact with driver Q6063 and Q6003 in contact After a fair bit2SA1930 of fiddling with multimeters and test D6004, : KDS4148U with the output transistor heatsink. I have not seen lamps, the mechanic decided that the 6014 electronic control R5194this 10 (1/4W) -52.5V -51.5V before; perhaps it offers better quiescent current stability. unit must have a faulty ground (connection to the chasI recently discovered that the quiescent current set- sis). Rather than pull the boot lining out again, he decided ting trimmer pot R6043 pin goes run a separate wire from the+52.5V trailer plug to the vehicle Q6005,to 6015 : 2SC1740S-S +51.5V was faulty. The wiper R5185 open-circuit as the control is adjusted. This would stop chassis. Lo and behold, everything worked again. Problem 10 (1/4W) R6025 Q6035 +1.1V Q6013 from working correctly and possibly destroy the solved. Oh yeah!2SC5171 IPRO Sorround Back 2.2 R6085trip home, I decided to check the output and driver transistors. This +50.5V small, low-cost part has After the 100km return (1/4W) Q6055 +1.0V Left ch 0.22 (1/4W) why; ILIST caused a lot of trouble. lights again. I don’t know guess IQ6075 didn’t have much +0.6V R6145 2SC2240 confidence in the system. Of course, it isn’t necessary to 22K Intermittent lights in trailer tow bar Q5005, 5015 : 2SC2240 tell you what I found, is it? R5035 IDLING R. G. B., of Ararat, Vic had that frustrating experience that I had to fix it myself. Like the 120K R5005 C5015 -0.3VThat’s when I decided CHECK 47/50 R5025 1K 0V of taking +something to the so-called experts, and it still Serviceman, I found that the quality of the wiringSPSBL was atroP6085 R5235 -0.65V you need to take R6105 comes back broken. 330 Sometimes these cious. I have a license to test and tag, so I did have some 120K ID+ 0.22(3W) R5055 things into your own hands, and doing so saved him quite idea of what I was dealing with. It was very difficult to get ID4.7K a bit of hassle... at anything to check if the various circuits were intact, but IDLING R6095that all was well with the wiring. I just finished reading the December 2021 Serviceman’s ADJ. eventually, I determined 0.22 R6045 (1/4W) story about his problems with trailer wiring (siliconchip. 2K The lights were sealed LEDs,Q6065 so that had to be taken on LIST -0.6V -46.5V au/Article/15141). Apparently, the demon affecting trailer -0.4V faith; they worked before, and R6175 it was unlikely they had VPRO wiring has international relatives. The Serviceman could suddenly failed. 47K R6185 VOLH have been talking about the wiring on a new trailer I bought So, I phoned the dealer again and47K told them there was D6005 Q6045 recently. still a problem. For some reason, I was now talking to an 2SA1930 -1.1V When a tow bar is fitted to a modern vehicle, some older man whom I doubted was D6005, a qualified electrician or 6015 : KDS4148U R5195 sort of electronic device a mechanic. There was a bit of silence on the other end, 10 (1/4W)the -51.5V is installed, hidden behind -52.5V C5005 221K P6902A Q6003, 6013 : 2SC1740S-S R5183 +51.5V R6053 SL WHITE 22K 902B 2.2 +52.5V Q6036 2SC5171 IPRO A selection of our best selling soldering irons and accessories at great Jaycar value! 25W Soldering Iron TS1465 $14.95 Build, repair or service with our Soldering Solutions. We stock a GREAT RANGE of gas and electric soldering irons, solder, service aids and workbench essentials. ESD Safe Tweezer Set TH1760 $19.95 Solder Flux NS3070 $17.95 Precision Angled Cutters TH1897 $19.95 1.5 to 3mm Desolder Braid NS3026-NS3028 $5.95EA 0.71mm & 1mm Solder NS3001-NS3096 FROM $2.95 240V Fume Extractor TS1580 $74.95 PCB Holder with LED Magnifier TH1987 $24.95 48W Soldering Station TS1564 $119 160pc Heatshrink Pac k WH5524 $24.95 Shop at Jaycar for soldering essentials: • Battery, gas and electric soldering irons & stations • Wide range of solder • Desoldering braid & tools Explore our great range of soldering gear, in stock on our website, or at over 110 stores or 130 resellers nationwide. • Soldering iron stands, cleaners & PCB holders • Heatshrink tubing • Tools & service aids jaycar.com.au/soldering 1800 022 888 As the PCB had been significantly damaged along with the relay, the new relay had to be connected point-to-point using solder on the underside of the board, with the top of the board coated with epoxy resin. then he came back on and said check the plug. By now, I was nearly at the point of rudeness, but fortunately, I only thought to myself, “of course I checked the wiring to the plug, you fool”. After mildly telling him I had done so, he said, “No, check the plug itself”. When I looked at this device from a mechanical viewpoint, it consisted of several brass holes into which brass pins were inserted. The pins were slit lengthwise, allowing them to make a springy contact with the holes. Inserting a screwdriver into each pin and spreading the contacts slightly fixed the problem. I still check the lights each time I use the trailer, but so far, this simple repair has lasted over two years. Fixing washing machine PCBs N. B., of Taylors Lakes, Vic runs a laundry repair business, so he sees a lot of broken washing machines. Here are some repairs he’s undertaken lately... The first one is an obvious fault, but at first glance, it looks like a write-off. There was a giant scorch mark and significant damage to the PCB around the relay pin that connects to the mains-potential “FS1” spade lug. Replacing the relay and repairing the board was the challenge. The relay switched mains to a high-current resistive heating element to maintain the desired wash temperature. So the repair insulation had to be good, and the resistance had to be low. After removing the relay, I cleaned the soot off the damaged area on both the top and bottom sides of 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. 78  Silicon Chip using an old toothbrush and PCB cleaning spray. I had to ventilate the room well while doing this. After I scraped through and removed the charred area, I filled the hole left with epoxy resin from both sides. When it was set, I marked and re-drilled the hole. Then I soldered in a new relay. The hole I had drilled was large enough to feed a crimp pin through it and onto the relay pin and flow solder through to the relay pin. I had enough clearance between the pin and the PCB, so there would be no problems with arc-over (especially as it’s a non-inductive load). I soldered two strands from a 2mm2 mains cable between the spade connector and the extended pin of the relay, snipping off the excess pin length. A quick insulation test between the spade terminals told me the job would be reliable. After testing it under load, I sprayed plumber’s clear rubber pipe sealant on both sides to seal the deal. On another similar PCB, I had a problem with the door sense circuit not recognising that the door was closed. The circuit for this is a simple series circuit comprising a mains source, a dropping resistor, a 1000V 1A diode, an optocoupler and the return Neutral. The diode tested open-circuit between the diode side of the resistor and the opto-coupler input pin. Still, it seemed OK when I tested the diode directly in both directions. I removed the surface-mount diode, and there was the remnant of the solder mask under one of the pads (the coating applied to PCBs to prevent solder from sticking in unwanted areas and forming solder bridges). I removed the coating and re-soldered the diode, and it tested OK. I then found that the opto-coupler internal LED was shorted, so I had to replace that too. I’ve also come across PCBs with breakdowns in high voltage areas, where white streaks can be observed running between components. This is high-voltage arcing in the intermediate layers of the board. The cure is to drill a hole wider than the arcing track but, of course, not through any internal or external traces on the PCB unless you bridge them out by another path. The sides of the hole and the edges of the PCB can then be sealed with lacquer. Moisture ingress into the PCB can cause this and can also affect layer capacitance, affecting the performance of tracks carrying high-speed digital signals. SC Australia's electronics magazine siliconchip.com.au