Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG Life on the ‘bleeding edge’ Dave Thompson When new technology comes along, I prefer to sit back and watch what happens before I take the plunge down that particular rabbit hole. This is a different philosophy than many people I know, including family members, who simply must have the very latest widget, gadget and toy available. Some people seem to need the latest gadgets. Control your home lights, entertainment system and air conditioner with your phone or home PC? Check. Have the latest electric car? Check. Own the latest drone with an 8K stabilised camera? Check. Ask Siri, Cortana or Alexa to order washing-up powder for you? Check. Don’t get me wrong; I’m not averse to these things and usually embrace technology, especially if it makes life easier. The problem with early adoption is that many manufacturers these days forgo 84 Silicon Chip stringent product testing and simply let their customers do it all for them, attempting to resolve any problems that crop up on the fly, in the hope that product sales will cover the costs of finishing the development (or recalling it in worst-case scenarios). Gone are the days of focus groups, mass testing and in-depth trials. The problem is that consumer security and privacy often suffer from this damn-the-torpedoes, seat-of-the-pants approach. Australia’s electronics magazine This is a tried-and-true business strategy, though. Japanese companies have done this for years. As a nontechnical example, say a manufacturer wants to try a different flavour of ice cream. In Japan, they simply make it and release it into the market. If it takes off, they reap the rewards. If no one buys it, they quietly withdraw it and move on to another flavour. siliconchip.com.au Items Covered This Month • Life on the ‘bleeding edge’ • A failed computer that needed • • new capacitors Cheating the (arcade) system An Astor Mickey OZ repair *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz In other markets, manufacturers would trial the flavour, get feedback from different groups and then decide whether to produce and market it. The latter process takes much longer, involves a lot more people and costs a lot more money. This is partly how the Japanese took over the automobile and electronics industries, and it appears that many western businesses have finally figured out how this approach could work for them. The biggest problem, of course, is that we consumers then become the testers for all new products, and as we all know from history, that doesn’t always work out so well. We’ve heard about mobile phones and laptops that catch fire (and now electric cars are doing this as well due to battery manufacturing faults; the Chevrolet Bolt was recently recalled in the USA due to multiple fires). We have clothes dryers that melt and burn the house down (multiple recalls by several manufacturers) and even aircraft that crash because of some unforeseen software glitch. So there’s a lot to be said then about not being an early adopter; many of us tend only to buy products that have been well and truly wrung out, though, in this era, that is becoming increasingly difficult. Early adoption experiences There are exceptions to my selfimposed rule, though; I recall buying my first digital camera, way back in the 90s; a Casio QV-100 [my first digital camera was the very similar Epson PhotoPC – Editor]. It was a marvel of engineering at the time, and as I was soon to be travelling overseas, I was looking forward to taking lots of snaps with it. At the shop, I asked many questions siliconchip.com.au about it, and the older gentleman salesman openly admitted that I probably knew more about it than he did; he couldn’t keep up with all this ‘new’ technology. I purchased the camera anyway, for a staggering amount of money by today’s standards, simply because I needed one and it was available. It was new tech and expensive, but I considered it worthwhile. This camera boasted a resolution (switchable!) of either 320 x 240 or 640 x 480 pixels. This meant that at maximum resolution, it had a megapixel count of, um, zero point three. That was cutting edge at the time, but sadly we don’t see many Casio digital cameras these days. It took good photos as long as I was in full, glaring sunlight and didn’t move a muscle while taking the shot. I still have most of those shots, and while they look a bit washed out, like looking at a 1970s-taken instant polaroid snap, they are all still perfectly viewable and remarkable for the time. The big problem with this camera, aside from the situation-limited snapping opportunities and need for a huge amount of light, was that it ate AA batteries for breakfast. I just couldn’t keep it supplied with power, and with no option for an external supply, my use of it became very limited. Of course, everyone now has a camera on their phone; some models have over-100-megapixel cameras and batteries that last for days even with heavy use; something early adopters of this tech could only dream about. For better or worse, technology marches on. I still have that camera somewhere, but it is obviously of no use to me other than some sentimental value. So, while being an early adopter has some perks, it isn’t always the best way to go. I recently broke my rule about being cautious in this regard when I had the opportunity to upgrade my home computer. What’s that old saying? A plumber’s pipes are always clogged, a cobbler’s children run around barefoot and a mechanic’s car is always on the verge of breaking down. I’m sure there’s one of these idioms for every profession. The fact is, I last purchased parts for my home computer in 2010, just before the quakes hit and ruined everything. Since then, I’ve installed an SSD or two, but the main components (motherboard, CPU, and RAM) Australia’s electronics magazine were all from that era. It was a monster machine at the time and far more powerful than I needed it to be, but I used well-worn, triedand-true technology when building it. There were faster CPUs and newer tech components, but I chose bits I knew worked with each other, and history has proven that I made the right call. I didn’t even really need to upgrade; the machine was working fine and even played the latest games quite well, so there was no mad panic. I’d been planning on buying something new for a while, though, and when some money became available from my mum’s estate, I took the plunge. This time, against all my instincts, I looked at the latest new tech and assembled a machine based on what was available. This was made a bit more difficult as many of my suppliers have been hit by COVID-19 and the resulting chip shortage that has crippled the likes of Toyota, Tesla and other high-tech manufacturers (and is still in full effect, if not actually getting worse!). Obviously, this is eventually going to roll downhill all the way to me, a tiny micro-business trying to supply computer hardware to my clients, and that is precisely what has happened. These days I’m fortunate if I can get a new Intel or AMD CPU, RAM, or a decent motherboard with which to build machines. And as for graphics cards – fuhgeddaboudit! In many ways, I was painted into a corner as to buying what was available for myself, and this is even more onerous when I try to buy parts for my clients. The cautious among us might ask why I just didn’t wait, but with no end to the shortages in sight and an increasingly turbulent market, I decided to just go for it. I ended up with the very latest Intel ‘prosumer’ (HEDT) CPU (with 16 cores!) and a motherboard to match it. I decided on 32 gigabytes of RAM, just because I could, and one terabyte of the latest M.2 solid-state drive that mounts directly to the motherboard. My excitement knew no bounds as I waited for the parts to arrive; this machine would be bigger and better than anything I’d ever owned before, and I couldn’t get it soon enough! Eventually, the parts arrived – some had to come from out of town, which made the waiting even more fraught. When all the bits were here, I set about October 2021  85 assembling it all. There was nothing that I hadn’t done before a thousand times, so I expected it to go together and just work. How wrong I was! The fact is, my mum probably could have worked out how to put it all together. Computer people seem to like making a big mystery out of the whole thing, likely so they can charge more money, but there is really no big secret to assembling a computer. In the old days, when one had to manually set IRQs and other weird parameters, perhaps it was somewhat more difficult, but today it is a bit like building a Lego set. Everything from putting the bits into a case and installing Windows is so turn-key that anyone who gives it a go would likely succeed. Like with many disciplines, though, the real skill comes when something goes wrong. For example, anyone can plumb in a gas line, but what if it isn’t done correctly? Anyone can wire a three-pin plug, but what if they make an error and swap Active and Neutral? Or leave some strands of wire sitting outside the plug? One of my first electric shocks when I was a kid was because of a single copper strand left caught in the plug body... So, I assembled all the bits and, with the wiring and cabling sitting temporarily away from everything, hit the ‘go’ button. All the fans and the built-in LED lighting fired up. I’m not a fan of all this lighting stuff in computers, but as it was built-in and a controller was supplied with the motherboard, I wired it in anyway. But there was nothing on the monitor, and after about 30 seconds, there were five short, sharp beeps alerting me that something was wrong. Interestingly, most motherboards and computer cases don’t come with a speaker anymore. Older cases had threeinch (75mm) permanent-magnet speakers mounted somewhere, and motherboards usually have speaker output connectors. These days, tiny piezo speakers are more popular, so I added one to this build as I’ve collected several. Most motherboards have what they call POST (power-on self-test) codes programmed into the BIOS. These beep codes tell us what is going on, be it a memory, video, or CPU fault. Without adding a speaker, I’d have no idea what was going on, or why I was getting no video output. So, five beeps. Another problem with this new technology is that when you buy a motherboard, there is very little documentation with it. I recall buying Windows 95 back in the day, and it came with a paperback book on how to use it. Tech manufacturers soon realised that printing hard-copy books was both a money-pit and a profit-losing strategy, and soon stopped doing it. Instead, these days a ‘QR’ code is included so you can go online and download the user manual for any given piece of hardware. I hit the web to find out what a fivebeep code means on my Gigabyte motherboard and soon discovered it means a CPU problem. I re-seated the CPU, ensuring once again I didn’t have any wayward pins or obstructions in the socket, but no matter what I tried, I could not get past the five beeps problem. Great! This is just what I needed. I counted myself fortunate this wasn’t a client’s machine; at least I could sort it out at my leisure because it was for myself. In the end, I had to call the tech support guys at my suppliers. They often hear of common problems or glitches and can advise workarounds or solutions. In this case, they 86 Silicon Chip knew nothing because the motherboard and CPU were so new that no information had filtered down the line. All they could do was offer to take it back and get it working. When paying in the region of two grand for just these parts, having these safety nets in place is often a lifesaver. While I was pretty sure (especially after assembling in the order of thousands of machines over the years) that I hadn’t messed anything up, anything is possible. As I couldn’t get it going, I ended up sending the entire box back to the supplier. They contacted me a few days later and confirmed that the problem remained even though they had tried other CPUs of the same type and other supposedly supported motherboards. There were no advisories and no updated BIOS for the motherboard, even though it was nearly six months old by this time. So all they could do was offer me a different CPU that they eventually confirmed did play nicely with my motherboard. They told me that occasionally the system would give five beeps and not boot in about one out of every twenty attempts, but it always powered up normally on the subsequent try. I was OK with that; all I wanted was something that worked, at least most of the time, for now. They shipped it back to me with the offer that if it was still playing up after six months and any interim BIOS updates, they would replace the board and/or CPU to get it working properly. Again, I was OK with this solution; I empathised with these guys as they try to keep abreast of all the new tech streaming out from manufacturers. I also know all too well that most of this hardware these days is thrown out into the marketplace with the bare minimum of testing. Manufacturers will simply placate consumers from suppliers on down through the chain to me with the next model if it proves to be too flawed to fix. This is the way of the world now, and while I broke my own early-adopter rule and paid the price for it, at least now I have a machine that works. In fact, that is what I am typing this column on. I haven’t had any instances of it not booting yet, but if I do, I will take those suppliers up on their offer to provide a new, more stable platform. At the end of the day, this is all they can do, and indeed is all I can do now as well. In the future, I will be a bit more cautious about buying the very latest thing, especially with a customer’s machine. While they might want it, I will be relating my experiences as a warning that it might not be the best way forward. If this had happened with a customer’s machine, it would have been an embarrassing situation. I’d have had to explain why their brand-new whiz-bang machine doesn’t work correctly and that it would take a few weeks before we could get it resolved. That just makes me look like a cowboy, and I don’t like that one bit. A failed computer that just needed new capacitors A. M., of Blackburn, Vic, was faced with an old, broken computer that nobody wanted to fix. But the problem seemed obvious, and the replacement parts were inexpensive, so why not give it a go... The unit in question is a TECS computer of about 2002 vintage running Windows XP. It started playing up in early 2020, being hesitant to get to the desktop promptly, sometimes going through various problems, screens and repeated restarts before finally getting to the desktop. Australia’s electronics magazine siliconchip.com.au Once there, the machine worked with no faults – it was only the startup that was the problem. Finally, it got to the stage where it would not get to the desktop no matter what. I took it to a computer repair outfit in Melbourne. They found a large number of capacitors on the motherboard that were bulging, and suggested that this was the cause of the problem. But they were not willing to fix it. As I had nothing to lose, I opened the computer and found seven bulging capacitors scattered over the motherboard. Carefully taking photos and notes of which cables went where, I took the motherboard out and examined both sides. It seemed to be only a double-sided PCB, which gave me a chance. I carefully noted the polarity of the faulty capacitors, even though the printed overlay on the board indicated this. It was tricky getting the old caps out of the plated through-holes, but with care and a hot iron and some solder wick, I got them all out. Suitable direct replacements are difficult to obtain at a reasonable cost. All were 6.3V devices, mainly 1000µF, but one 3300µF. I got replacements with higher voltage ratings that were physically larger than the originals, but there was enough room to fit them. After two evenings’ work, I had replaced the defective capacitors and reassembled the computer. Upon powering it up, I had to answer a few silly configuration questions; then it went to the desktop right away. All the programs seemed to work, but the big test was a restart. I shut it down, restarted, and it went straight to the desktop. I consider that a victory. Three months later, it is still going well. After this, I made sure to save all my critical files to an external hard drive. Cheating the (arcade) system M. F., of Wyongah, NSW was reading the Arcade Pong article (June 2021; siliconchip.com.au/Article/14884) and was reminded of a repair he was involved in some time ago... After arriving in Australia in 1988, my family and I initially settled in Newcastle, NSW – a beautiful place. My first job was as a service technician for A. Hankin & Co, a Newcastle/Sydney company that had quite a few Arcade Game centres. They also had a manufacturing facility based in Darby St, where they made their own arcade machine cabinets from timber. One of my main tasks was troubleshooting problems that the field techs couldn’t fix, and one such situation arose on a Monday morning. I was called into the boss’ office as soon as I arrived, and was told to get to Pelican airport ASAP as I was on the early morning flight to Sydney. Upon my arrival, I was picked up by one of the Sydney techs and taken to their main arcade showroom in George Street. When we got there, the lady in charge showed me a gadget and said that she had confiscated it from some kids over the weekend. It was a gas igniter and gave one heck of a spark. I tried it on a couple of machines, and each gave 99 credits (the maximum amount). If the kids had used it sensibly, they would never have been caught! I found that it didn’t work on every machine, but mainly on ones with a credit board manufactured in-house. The board just accepted pulses from a coin mechanism siliconchip.com.au Australia’s electronics magazine October 2021  87 This Astor Mickey OZ was inherited in a state of large dust accumulation. However, on the bright side the underside of the chassis was fairly clean. The underside of the chassis had some of its electrolytic capacitors replaced, and a new power cord had to be installed. The restored radio can be seen in the adjacent photo. 88 Silicon Chip Australia’s electronics magazine and issued a credit. It could be used with mechanical and electronic coin counters. I went back to my trusty workshop and grabbed such a board. Sure enough, every time I hit it with a zap from the igniter, 99 credits would come up. Like the Pong machines described by Dr Hugo Holden, I used an antenna to ‘catch’ the zap. I managed to fix it by running some wire around the PCB, close to the edge. I left one end open, and connected the other to the Reset input. The wire was simply glued to the original boards until subsequent batches had it embedded as a track. Now if the board was zapped, it simply reset and sat waiting for a real credit pulse. Game over! A tale of an Astor Mickey radio When C. K., of Mooroolbark, Vic saw a vintage radio sitting unused, he couldn’t help himself. He offered to get it working again, and succeeded in that endeavour... We have a favourite restaurant in the Dandenong Ranges, east of Melbourne. It was there that I saw an old radio used as a decoration. Looking into the back, I saw that it was in a bad way, absolutely choked with decades of accumulated dust. I asked the owner what she knew about it. Apparently, it had been in the family for a considerable time. I suggested that I might be able to restore it for her. She agreed to this and said next time we are in, she will give us a free meal! As you can see from the photos, the dust accumulation was unlike anything I had ever seen. It took quite a lot of work with the vacuum cleaner before it could be handled. Surprisingly, under the chassis was quite clean. I discovered that this was a 1934 Astor Mickey which had been featured in a Silicon Chip article written by Rodney Champness (March 2004; siliconchip.com.au/Article/3438); that article included the full schematic. Editor’s note: we will publish an updated article to the Astor Mickey OZ very soon, so we’re refraining from publishing another circuit until then. This design shows how far back the standard superheterodyne design goes. The first valve, a 6A7, is a pentagrid converter that multiplies the signal from the antenna with the local oscillator. The resulting 455kHz siliconchip.com.au difference signal is amplified by variable mu pentode 6D6. There are two intermediate frequency (IF) transformers, L7/L8 and L9/L10. These are tuned by trimmer capacitors in the sides of the IF transformer cans, accessible from the back. A 6B7 dual-diode-pentode valve provides envelope detection and audio amplification. The filtered negative DC voltage from the diodes also provides automatic gain control to V1 and V2. Finally, the type 43 pentode drives the speaker through a transformer. In 1934, there were no permanent magnet speakers. Instead, it uses electromagnet L14 which drops almost half of the voltage from the 25Z5 rectifier. The centre tap of the transformer is connected to the electromagnet coil, which has a resistance of 1.875kW. This puts the centre tap at about -122V, from which the grid bias is obtained for V4 via the resistive divider of R17 and R18. Provision is made for an external electromagnetic speaker via a four-pin socket on the back. There is a rather heavy-duty switch accessible from the side to switch between the internal and external speakers. There was no power cord with the radio, but two pins were sticking out of the middle of the chassis in the back designed for some kind of plug. I replaced this with a small plate made out of 1.6mm aluminium and fitted it with a cable clamp to hold a three-core mains cable. The radio had two large 8μF electrolytic capacitors that I would not trust, so I immediately replaced these with the only large, high voltage electros siliconchip.com.au that I had, which were 100μF/350VW! Checking for any obvious shorts, I carefully applied power. There were no signs of distress; all the valve filaments lit up, and on turning up the volume, I heard some noises coming from the speaker. Attaching an outside antenna to the red wire out the back, I could actually pick up some stations. I was amazed that after all this time, all the 86-year-old valves were still working. So what else needed doing? There were a couple of capacitors with high voltages across them, so I replaced them with modern ones. It was also a time before ferrites and iron dust cores, so all the coils are air wound. This meant that the only adjustments for the aerial and oscillator coils were with capacitor trimmers. As it turned out, they were not far off, but the IF trimmers needed considerable tweaking to get them to 455kHz. Rodney Champness mentioned that this set suffered from overheating. With the rectifier and output valve side by side, the heat discolours the top of the cabinet. The 25Z5 filament is 25V at 300mA. This is a 7.5W heat source, to which would be added the inefficiency of the rectifier and transformer losses. I cut off the filament wire and soldered in two 1N4004 diodes. This gives a slightly higher DC voltage, but is justified in this case. As for the filter capacitors, the 100μF units were a bit over-the-top, so I replaced them with smaller ones, rated at 47μF/350V. One problem with electromagnet speakers was residual Australia’s electronics magazine hum, but with these capacitors, there is none. There is no power switch. I thought of replacing the volume control with one that included a switch, but there is no room. Likewise, there is no dial lamp, as there is no dial as such. There was probably more work in restoring the cabinet than the electronics. It had a few cracks and the walnut veneer had a few missing bits. I filled these with an appropriate wood filler that turned out to be too light, so I had to darken it. The inside of the cabinet was raw timber, so I sprayed it flat black. A couple of coats of satin varnish improved the appearance. Finally, there were some holes in the speaker cloth, so I replaced this with “vintage” speaker cloth I got on eBay, which was a reasonable match to the original. Once everything was tuned up, I measured the performance, and it is certainly not brilliant. The Melbourne stations come through fairly well with an outside antenna, but using a signal generator, it needs about 50μV for an acceptable signal-to-noise ratio. Does that matter? Not really; the radio is unlikely to be used as such, but will continue to be a decoration at the restaurant. That’s assuming the owner will keep it – this model can fetch up to $1000 on eBay! I went for lunch to the restaurant with the restored radio, and the owner was really pleased with it. She followed it up by saying she has several other old radios that I might like to look at. Who says there is no such thing as a free lunch? SC October 2021  89