Fullscreen HTML5Med Res (??MB)HTML4

SERVICEMAN'S LOG (What) were the designers thinking? One problem with being an engineer, a serviceman, or similar is that whenever we see something mechanical or electrical, we immediately (and mostly subconsciously) assess it for a range of criteria, such as what it is made of and the methods used to manufacture it. Or, in my case (especially in my younger, more inquisitive days), how one could take it apart. Basically, we can’t help but be analytical. Let’s face it – everything we use, buy or make has a design element to it, and supposedly some theory sits behind that design. However, I often find myself quietly cursing the stupidity of some of these designs because it appears that no actual thought processes have gone into them. This phenomenon is by no means modern; engineers have been bemoaning poor design for as long as there have been people making things. It might seem more prevalent these days because of the number of YouTube channels dedicated to pulling apart everything from tools and machinery to appliances and cars, often for laughs. But they always seem to be asking the age-old question: what was the designer thinking? I can reel off several examples for you. I grew up with older British-made cars and some of the decisions made during the manufacture of those vehicles begs the same question. To get the engine out of a Morris, Austin, MG or Wolseley 1100/1300 for example, you either have to have double-jointed hands the size of a small child, cut an access hole in the passenger floor pan, or have an impossible-to-source specially-made spanner. I had the motor in and out of my 1300 so many times we cut a hole in the floor (replacing it with a suitable access cover) and used a special spanner Dad made after seeing one at a garage in town. And take the original Mini; iconic though it is, one gets the distinct im62 Silicon Chip pression that they designed and built the car and then discovered they’d left nowhere for the battery, so it went in the boot. There are many other examples and while I know these cars were built down to a price, making life more difficult for the poor sod who inevitably Australia’s electronics magazine Dave Thompson Items Covered This Month • • Learning a painful lesson HP 3585 spectrum analyser repair *Dave Thompson runs PC Anytime in Christchurch, NZ. Website: www.pcanytime.co.nz Email: dave<at>pcanytime.co.nz has to work on it is just not cricket. Editor’s note: making fun of poor British car design is like shooting fish in a barrel. Who hasn’t heard of Lucas Electrics, AKA the Prince of Darkness? They were truly innovative: they invented siliconchip.com.au the intermittent wiper, the first selfdipping high-beam and they also produced a very potent anti-theft device: their ignition systems. It’s difficult to steal a car that won’t start! A design dud in the kitchen A while ago, we purchased a sandwich maker; one of those clam-shell types that closes and bisects the sandwich while cooking it. In my opinion, this device has several design problems. Firstly, the top plastic cover protrudes out over the bottom section at the rear of the maker, ostensibly to cover the hot hinges and keep them away from wayward hands. But the steam produced while cooking is very effectively trapped by this hollow, overhanging moulding. The steam then condenses and drips down onto the bench and power cable. We have to put a folded paper towel or similar beneath it, to avoid pools of greasy water forming under the back feet. While this is no deal-breaker, it clearly isn’t a good design. And that isn’t the only problem, either. The more significant failing is the fact that no matter what brand of bread we buy, none of the slices fit correctly. The bread slices either fit entirely inside the cooking cavity and don’t make contact with the edges, or they end up with one edge sticking out the sides. (Editor’s note: perhaps you are using metric bread, and it was designed for imperial? Try using a British Standard Loaf and see if that fits.) While the smaller slice might seem the way to go, fillings (cheeses in particular) bleed copiously from any seam that isn’t clamped and sealed by the elements and then proceed to leak out all over the rest of the machine and the benchtop. This leads us to another design flaw: the gaps between the elements and the case mouldings. From the first sandwich we made, these cavities fill up with crumbs, cheese, water and anything else that might be cooked in the appliance. After just a few months of weekend-lunch use, this sandwich maker was so filthy and impossible to clean properly that my wife refused to use it anymore. Being the gentleman serviceman, I took it to the workshop and stripped it down to clean it out, and what a lovely job that was! It is generally pretty easy to take siliconchip.com.au apart, except for the vain attempt by the manufacturer to prevent me undoing four, pimple-bottomed Torx-style “safety screws”. But I did have to cut the connectorsecured power cables off to free the bottom plastic case shell, as the power lead clamps to it with flying wires connecting to the elements through a small hole in the case. Australia’s electronics magazine The reason I went that far is because the bottom half was almost completely covered with both greasy and rockhard melted cheese and breadcrumbs. I had to chip some of it off with a screwdriver. I used my heat gun to soften the baked-on gunk on the outside of the case because it couldn’t be removed without otherwise damaging the shiny plastic finish. I would have expected this type of mess after a year’s worth of use by a family of five, but it seems a bit excessive for a few months use by two of us. Of course, we could use a smaller quantity of ingredients, but where’s the delight in an empty toasted sandwich? A quick visit to some big-box stores with a pocket tape-measure confirmed my suspicion that all similar sandwich makers have the same size cooking cavities as ours, or very close to it. Has no sandwich-maker designer ever purchased a loaf of bread at the supermarket? Perhaps there is a theoretical standard slice size, but if so, not many bakers adhere to it. If I designed a sandwich maker, I’d make it so that the majority of everyday bread slices fitted it properly. February 2019  63 Ah, the bad old days Do you have any good servicing stories that you would like to share in The Serviceman column? If so, why not send those stories in to us? We pay for all contributions published but please note that your material must be original. Send your contribution by email to: editor<at>siliconchip.com.au Please be sure to include your full name and address details. someone had knocked my hand away with a bat. When I pulled my hand in front of my face, the top 5mm of my right thumb was hanging by a thread, and there was blood everywhere. Yikes! I didn’t know what I’d done at the time, and can’t recall how I got out, but I do remember walking up to the guys in the workshop by the hangar floor holding my hand out and telling them I thought I might need the nurse! It turns out that some bright spark in a design department somewhere had dreamed up a modification to keep some equipment cool in the bay and a 120mm steel-bladed cooling fan had been fitted to the rear of one of these racks on a previous maintenance visit. I didn’t know it was there, and the powers-that-be saw no need for a safety cover for the fan because it was mounted “out of the way”. Now, modifications like this aren’t just dreamed up by some bloke with a hangover one Monday morning and installed that afternoon. This would have been thought out and put through rigorous checks and balances at the highest levels by the manufacturer, or one of their contractors. They would first see if it was required, then ensure it was implemented to the highest possible standards. This process also involves updating the very comprehensive manuals that go with every aircraft; these huge books document every nut, bolt, panel, cable, washer, rivet, system, component and piece of equipment inside it, along with all the specifications, circuit diagrams and schematics of literally everything on board. Not one of the dozens of people involved in this design process thought about installing a safety screen. Any new addition to an aircraft, regardless of whether it is a plastic bung for the end of a seat frame or a whole new navigation system, is also rigorously tested – destructively and nondestructively – by the aircraft manufacturer and interested third-parties before getting anywhere near a plane. It can take many months, if not years, before the information and parts are eventually available to the endusers for implementation. Something fell short here. Inevitably, I had to take time off work and the injury also caused me to miss my final exams – which I had Australia’s electronics magazine siliconchip.com.au A more severe example of designer negligence harks back to my days as an aircraft engineer. Actually, I was technically still an apprentice at the time, in my final year and just before my final trade exams. For those not familiar with widebodied aircraft, at least back in those days, most of the major electronic components are kept in a special area called the E&E (Electronic and Equipment) bay. Much of an aircraft’s radios, controllers, computers and various other components are packed into this space, mounted in special racks and cabled in with vast looms of wires strung from the farthest reaches of the plane. This room is accessible via a hatch beneath the aircraft (sometimes from inside as well) and it varies in size depending on the aircraft. The E&E bay on a 737 is a lot smaller than that on an A380. On some planes, I’d need a step to boost me up enough to climb in, while taller guys could often work standing on the ground. On the particular aircraft I was working on then, an ageing Boeing 727, one of three then-used as transports by our Air Force (nothing but the oldest and cheapest crates for our boys!), the bay required me to climb a special ladder to access it. As I was of a smaller and skinnier physical size back then, I got all the jobs the bigger guys opted out of. Being an apprentice also had a lot to do with this, and apprentices usually got fun jobs like cleaning bird remains from the engines of aircraft that had suffered bird strikes. Bird strikes are rare (thank goodness!), but when they happened, the task of cleaning the rear section of the engine would be passed down the chain of command until it hit the bottom – an apprentice. As one of the guys said at the time, “manure rolls downhill”; although those are not the exact words that he used. You’d think that a seagull going through all the turbines and vanes spinning at thousands of RPM inside those big jet engines would essentially be vaporised. This is not the case, and it is a particularly unpleasant task to kneel in puddles of aviation fuel in a confined and grimy metal tube scraping crispy bird remains from the sides of the engine. It put me off chicken for years. Thankfully, as I moved up the pecking order (hah!), that job became a younger apprentice’s problem. Learning a painful lesson for someone else One job I did do a lot, just because I was one of the smallest qualified guys, was inspecting, cleaning and repairing wiring looms inside the wing-tanks on whatever plane needed it. If you think there isn’t enough room in the wing of a plane for a person to work, you are mostly right; it is a very tight and claustrophobic space, especially when decked out with all the lights and breathing gear and carrying special anti-static tools. My foreman always swore that he’d cut a hole in the skin of the wing if I ever got stuck in there, but now I’m not so sure he’d have kept that promise! But I digress; I had work to do in this 727’s E&E bay, adding wires to an existing loom for some new component going in. This required me to contort myself into a flat position on my back while curling around sideways to get right in behind one of the racks so I could crimp and terminate the wiring into a Canon connector mounted on a frame there. It was tight and taxing work and once again, because I was one of the smaller guys, I got assigned to do this type of thing a lot. The problem wasn’t getting into position, but getting out again when my arms were in behind the racks and my feet hanging out the hole in the floor; I didn’t have a lot to push against to make my way out. I reached up to lever myself against the back of the rack and felt a sudden hard whack, as though Servicing Stories Wanted 64 Silicon Chip to sit later. I also had to play guitar at a friend’s wedding that weekend and did so painfully, with the pick taped to my comically-large thumb bandage. My right thumb is still shorter than my left because the base doctor just tore the damaged bit off and threw it in the bin! I recall it hurting, a lot, and it didn’t help when several guys from the technical department came over and had a look around before interviewing me, claiming they never thought – and I quote – that “anyone would be stupid enough to stick their hand in it”. The only silver lining is that the fan was running in the direction that pushed my hand out, rather than simply ate it whole. A mesh screen was eventually fitted and I’ll wager most other fans fitted after that had one too, regardless of where they were situated. I (and others) couldn’t believe this mod had made its way right through the testing and vetting process without someone realising that having a highspeed steel cooling fan unshrouded in a work environment might be a bit of a health and safety issue. I mean, we are talking about an industry that is typically extremely paranoid about every aspect of safety! I’m guessing that the character sitting at a drawing board who originally dreamt this up, along with the people who subsequently signed off on it, have never had a pair of overalls on or wired a rack in an E&E bay. You just can’t beat hands-on experience. Even the grass is mocking me And on another note, I recently purchased a new lawn mower, given it’s the time of year when you can hear the grass growing. It is a well-known, dayglow yellow brand and I assumed they’d know a thing or two about making lawnmowers. However, when I tried it, even set at the lowest blade position, the lawn was still looking uncut. I like a short lawn; not shaved earth, but preferably cut enough so I don’t have to do it twice a week. All my old mowers could cut this short but this one goes no shorter than the industry standard of 25mm. I contemplated modifying the level notches on the side, but even with the deck scraping on the ground, the cut was still far too long. The blade disc is just too high in the deck. I ended siliconchip.com.au up removing the large, central bolt holding the blade disc on and added 10mm of washer shims to the top-hat style mounting plate before bolting it all back on. It now it cuts perfectly, but it begs the question; did the guy who designed this mower ever mow an actual lawn with it? HP 3585 spectrum analyser repair A. L. S., of Turramurra, NSW, has had one thing after another go wrong with his 40-year-old Hewlett-Packard spectrum analyser. But he’s obviously very attached to it since he keeps on fixing it each time. Here is what he had to do to keep it going over the last couple of years... The HP 3585 Spectrum Analyser is a very versatile instrument, covering the frequency range of 20Hz to 40MHz with a 3Hz resolution, making it useful for both audio and RF applications. It also has a built-in tracking generator. Unlike most spectrum analysers (which usually only tolerate tiny RF signals), its input capabilities are really good with a selectable input impedance of 50W, 75W or 1MW and with 42V peak handling, which is ideal for testing medium-power audio amplifiers. I purchased this one locally for a few hundred dollars on eBay; that’s quite a bargain when you consider that it was more than $30,000 brand new in 1978. Australia’s electronics magazine To put that into perspective, it was the same price as a Mercedes Benz 450SL! My unit had a number of faults that occurred over a two year period, so in that respect, it’s probably about as reliable as an old Merc too. First, some intermittent flashover started happening in the CRT EHT section, usually during periods of high humidity, causing the CRT to eventually fail altogether. I noticed that the LEDs on the front panel appeared to still be operating normally and a quick screen grab using a GPIB/USB connection proved that the instrument was still fully functional, apart from the lack of display. Using the display (X & Y) BNC connections at the back of the instrument, I was able to obtain a very good working display on a 20MHz analog oscilloscope set to X/Y mode, as shown in the accompanying photo. This only works with an old-fashioned CRO though; most digital scopes are just not fast enough to do this! I used this arrangement for some time while I waited for a surplus CRT and EHT supply to arrive which I had ordered from the internet. When they finally came, I first had to turn the instrument upside-down, which was difficult because it weighs 40kg! I removed the HT supply box to take a closer look and to figure out how to change the CRT and it did not look easy. There was a fair old February 2019  65 Circuit diagram for the motherboard of the HP 3585 from the second service manual (A41; page 192), which can be found here: www.ko4bb.com/getsimple/index. php?id=manuals&dir=HP_Agilent 66 Silicon Chip Australia’s electronics magazine amount of accumulated dust around the EHT leads and components. This was probably the cause of the sparks; after a long spell of high humidity, the dust absorbed enough water to become conductive. I cleaned everything out, including the interior of the EHT box, using rags and methylated spirits and then dried it all out. Then I noticed a small perspex cover which concealed a 1A fuse and a quick check revealed it had failed. I replaced it, thinking that it would instantly blow again, but magically the whole instrument lit up and the CRT display was bright and normal. The instrument then functioned well for a year. Then one day, it refused to start up properly. All the LEDs on the front panel lit up but nothing else happened. Switching the instrument off and on a couple of times brought it to life and that is how it operated for a few months until finally, nothing happened at switch-on except for the front panel lighting up. I left it for some time to gather cobwebs because I had another HP 3585a which worked properly, except its knobs, which meant that I could not control the tracking generator amplitude. Eventually, I decided to tackle the fault by exchanging a few spare low voltage regulator PCBs that I had obtained from the internet. There are five of these in each instrument, so I changed them one by one but that didn’t cure the fault. The service manual, which runs to three volumes and 884 pages, points out that one faulty board may be capable of damaging another so that the end result might be several damaged boards and this worried me a bit. The output voltages from these boards measured correctly and that plus the fact that the five green status LEDs were all lit, suggested that the power supply was functioning correctly. Again, I put the instrument aside, fearing that the fault was too elusive, and went back to the older working instrument. Some time later, I decided that I had to throw it away or fix it. I decided to substitute the PCBs from the working unit and take a risk that I might end up with damaged boards and have two dud instruments. There are about 24 boards in total, siliconchip.com.au An analog oscilloscope was used temporarily as a replacement screen. The inside of the spectrum analyser is packed with boards (24 in total). The HP 3585 motherboard shown above, corresponding to the circuit diagram at left. The 80-pin CPU (U13) is located under the large ribbed heatsink. siliconchip.com.au Australia’s electronics magazine many of which are difficult to access, so I just started at one end and replaced the suspect boards with the good ones. Rather than swap in one board, check if it works and then swap it back, I decided to move all the known-good boards into the unit one by one, in case two or more boards were faulty. Then suddenly, after I replaced the central processor board, everything came to life again. As it happened, I had a spare processor board on hand, amongst a set of boards that I had previously purchased from eBay for spares. It was horribly dirty but the unit worked with it in place, so I had two working instruments at last! But that is not the end of the story. Ironically, just as I was finishing this story, a similar fault occurred and the instrument would not start up, even though some LEDs were flashing normally on the front panel. It came back to life when I substituted the processor board (03585-66541) from “old faithful”. Apparently, I now have two dud processor boards and there are none to be had on the internet. The usual parts sellers told me that they had sold all their processor boards. I was back to square one. I tried to repair the faulty boards referring to the excellent repair manual; I replaced all the electrolytic capacitors (and the other capacitors for good measure) but nothing worked. The voltages around the main ICs all measure good but there are 23 of them so replacing them would be a massive undertaking, even if they were available. And then there is the main processor chip which is the size of my wallet and definitely not a standard item! I found one internet seller that had three dud HP 3585 analysers for sale. He was asking $800-1,000 each for shipping but the photo of the instrument showed a very dim glimmer of a trace which indicated a healthy processor board, so I asked him if he would accept half price and pull out the desired PCB and post it to me via USPS for about $40. After about 20 emails back and forth, we finally came to an agreement on price and the board arrived safely and it actually worked. The reason for the failure of these boards remain a mystery but I have kept the two dud boards for future spares. SC February 2019  67