A big job.
So Gary was flushed with success.
Welcome yet again to On-Call, The Register's weekly column in which we take readers' tales of odd jobs in odd places, tart them up and present them to you as a bit of light relief on a Friday. This week, meet “Gary”, who once had a trouble ticket land on his desk stating that “the PC would reset every time the customer flushed …
Sorry to burst your bubble, but the noise cicadas make is called clittering. So Clitterhouse is, essentially, "the house of noisy insects".
(It could also be an old word for clay ... but adobes don't work all that well in the British climate, and don't make for as good a story either.)
Beer. Because what else could follow the above, especially on a Friday.
Over 30 years ago I worked for a company that supplied minicomputer systems for pathology labs in hospitals. I was on-site at one location, where the customer had complained of the system continually restarting itself.
I was staring blankly at the machine, wondering where to start, when there was a "whump!" that was felt rather than heard, and the lights flickered.
"What was that?", I said.
"Oh, just the X-ray department next door".
In an existence before that, we used to have similar problems on our micro development kit. We noticed that they got more frequent around 5pm. We eventually realised that it was due to the lift in the building.
I worked for a radio station and in the studios you try and eliminate external and internal noise. So everything that can be soundproofed is and things that couldn't be were stuck in a rack outside the studio. They had good air conditioning and were generally kept cool which obviously suited the electronics. We had a new station which was going to be mostly recorded shows being broadcast from the building and a new studio was built for the existing station. The old studio was then going to be used for recording the shows on this new station. However when they came to use the old studio they complained that main computer was shutting down on a regular basis. This was used for recording the voice of the DJ into the station playout machine and that had never happened before, it had been rock solid.
It was one of the computers that we couldn't place outside the studio and so required the studio to be empty whilst we looked into the problem. It was a real mystery because the problem was only happening now and we couldn't work out what had changed since the legacy station had moved out. There was now less use of this PC instead of more so why it was suddenly crapping out every so often was head scratcher. It didn't happen when we were checking the studio over even after a period of hours and we looked at everything we thought could be the cause.
It was only when we were watching a show being recorded everything slotted into place. The fan on the PC wasn't silent enough so we'd set it up to switch off when the microphone fader was up (i.e. people could hear you on air / the recording). We realised that the fader stayed up all the time as they were just recording the bits between the songs. So after a period of time the thermal trip switch would cut the power to the PC as protection because it wasn't getting any cooling from the fan. Previoulsy it wasn't a problem because the fader would be down and the fan on during the songs and adverts. Once we'd worked that out we disabled the mic fader link to the fan and just let it run continuously.
Almost 30 years ago I had a CRT monitor on which the picture would, at regular intervals, get twisted and show a definite tilt to one side.
I was on the point of sending it back as faulty when I realised that the hospital's MRI scanner was directly above me, 2 floors away.
When I went to university many years ago, I was able to buy my first PC, a 286 clone. Just, a morning while I was writing and testing some Matlab code, the PC rebooted suddenly (losing a lot of work). And it did again in other days, always when working in the morning. Only after a while I noticed it happened when the cleaning lady was turning on the vacuum cleaner...
Mainframe O/S occasionally crashed and the customer sent us the resulting post mortem dumps. It had the whiff of electrical noise about it - so we arranged to go in when the engineers had their big weekend maintenance session.
While the engineers did their peripheral maintenance tasks we tried a few experiments. We knew from experience that the mainframe's engineer's panel - even when stopped - would show random changes if a severe noise problem existed.
Card reader - no. Card punch - no. Load a tape drive - after a few tries - bingo!.
Very pleased with ourselves we waited until the Chief Engineer was free a few minutes later so we could demonstrate it was a hardware problem. When he came over - the problem wouldn't happen.
So we asked the question - what has changed in the last few minutes? Nothing - except the engineers had lifted a floor tile to install a cable for an additional new tape deck. Put the tile back - and the problem was able to be recreated.
The space under the tiles was jam-packed with cables - one for each tape deck. Each cable had a metal box some distance away from the tape deck. The metal box for a new tape deck installed the previous week was sitting high on all the other decks' cables. It was high enough to just touch the underside of the floor tile above - but not high enough to stop it going home properly.
The false floor was an unusual design. Instead of corner "mushroom" supports - it was a lattice of steel bars. The floor tiles were backed with an unpainted steel sheet. That had connected the lattice to an unpainted area of the tape deck cable box. The lattice was grounded to the building wiring for electrical safety - and the result was the computer room's "clean" earth was being connected to the building's "dirty earth".
Back in my Central Line days with London Underground in the mid 1990s we had an intermittent earth that screwed up the signalling for about a month. As it was at Marble Arch it caused havoc with the service. Always used to happen in the AM peaks.
The whole area was cleared of any scrap metal etc, engineers on site overnight, nothing found. Points on the crossover given a thorough going over etc.
Eventually the cause was found - there had been a broken chair screw (a screw that holds to the sleeper the thing that holds the rail) in the past that had snapped below the surface of the sleeper. A new screw had been wound in on top of the old one - pushing it out of the bottom of the sleeper. Over time, the track bed had settled a little, thus meaning that the old screw, only under the weight of a fully laden train (hence the AM peak), just made contact with the iron tunnel segment. As the old screw was in contact with the new screw, that in contact with the chair, and the chair in contact with the rail - through which the current controlling the signalling runs, Bingo, a lovely hidden intermittent earth.
There was a core-based mainframe - you know those memories made up of very small magnetic cores arranged in grids and flipped by two crossing circuits. Lots of fun problems with those such as young programmers burning up a core by flipping it constantly on/off....
Anyhoo, this story is about the periodic rebooting of said mainframe. For many months on a not-regular basis the primitive system would sigh and give up its ghost. All the field engineers in the kingdom couldn't determine why this was happening. Replaced cores, replaced CPUs,, etc.
Then one night someone was standing next to the memory box and heard this whoosh, followed by the death rattle. Looking up at the tube overhead that emitted the whoosh he asked what was in the tube that would make the noise. It was determined to be one of those fancy pneumatic tubes used for sending priority messages. The passage of the carrier overhead was enough to induce a magnetic field in the cores.
Way back in the dim and distant 80s, I did some work for BP Minerals who had an exploratory base-station in norther Norway, just up from Trondheim. Back then we were using DEC Rainbows to download and process magnetometer data which was then saved to 5 1/4" flooppy disks. The field crew kept reporting that their floppy disks were being corrupted and they were loosing the data. These very early PC-alternatives didn't have a hard drive, just 2 floppies - one was the system drive with a CPM disk and software we ran the other was the data drive, so loosing data meant the magnetometers became unusable as we couldn't clear the data from their pathetically tiny memories.
We thought that maybe the floppies had been corrupted by the airport x-ray machine so we put them in a metal box and wrapped them in tin foil... didn't work... So I, as the programmer of the data processing software, got sent out for a week to work out what was going wrong. This was April and the snow was still on ground and I got to rid a skiddo to the log cabin in the middle of nowhere on the edge of the arctic circle to spend a week in a log cabin with a PC and 2 Geophysicists. On the first day they went out surveying and I sat with the PC checking everything I could think of. There was an electric wall heater on a thermostat/timer and I noticed that whenever it switched on the lights flickered a bit and the fan on the PC changed pitch... Lo and behold disk errors started to appear... Quick trip back to Trondheim to pick up a rectifier and bingo no more disk errors...
From that day on it became part of the standard field kit...
I believe it's pond side slang for a device that corrects brown outs. I did have a couple that came over with some kit from the states many years ago, but I never used them because... 240V. Huge great transformer it looked like, must have weighed 15kg, with a couple of black boxes slapped on the ends, one of which contained a diode pack that looked like it could have rectified the entire National Grid and a capacitor of lethal dimensions.
Words for technological devices mutate over there. I had so many customers come into Radio Shack wanting new batteries for their convertor. Now a convertor is a cable TV box, but they call the remote control for that a convertor. And there are a dozen other examples of inaccuracy in the common parlance for things electrical.
Ha ha ha. Brown outs.
No. A rectifier is a collection of diodes in these parts, same as in Blighty. Not sure what you are talking about, but it sounds vaguely like an inverter. Why it should correct for brownouts is anybody's guess.
We have 240V here, in nearly every home. Yes, most small appliances run on 120V ... However, most homes have two legs of a generator's output, called L1 and L2. Each provide 120V when you connect across the neutral. But if you connect L1 and L2, you get 240V ... Heavy electricity users, like clothes driers, water heaters, ovens, air conditioners, some pumps, etc. use 240V. Yes, that's simplified. Intentionally.
Never heard of a set top box called a converter. It's a set top box.
Never heard of a TV remote called a converter, either. It's a remote (control).
That device sounds suspiciously like a ferroresonant power conditioner... Dirty great lump of transformer with capacitors and 'magical' connections that prevents the changes in line voltage from reaching the connected equipment (for a second or two) by storing power in the caps/inductors. They act also like a very short term UPS (1 second max, istr)
Damn good things, but weigh a ton (sometimes literally). Makes a noise like someone twanging a plastic ruler on the edge of the desk when you switch it on.
I can see them being called 'rectifiers' because they rectify power problems?
Also they'd not be transferrable between left and rightpondia, because of the different frequency, they're resonant at one, and only one frequency.
Also: (rightpondian here for clarification). Cable tv set top boxes are called converters, officially. The old Jerrold general instrument boxes were labelled as such. More of a technical term that a consumer one though.
"Also they'd not be transferrable between left and rightpondia, because of the different frequency, they're resonant at one, and only one frequency."
A lot of them could be tuned by changing caps around.
They were scary things, using saturated cores and dissipated a _lot_ of heat (useful in the tropics if you needed to keep your printer paper dry)
Yes, "That device sounds suspiciously like a ferroresonant power conditioner." I found one of these in the HVPSU of an Austrian designed TV transmitter/transverter that I installed in the fairly peaceful and very beautiful KSA/Yemeni Asir border area in the '80s.
the ferroresonant device was 50Hz, but the previously installed TV repeater station genset was 60Hz, I was able to fudge it once I had worked out why the tetrode EHT fuses were popping! (for some reason I ate mostly parmesan cheese whilst doing this maintenance as the local market of horribly beweaponed ghat growing tribes sold whole round cheeses) The UHF driver amplifier was also multimoding as a second fault, and had to be filled with a lot of aluminium foil inside a plastic bag, to try and dampen down the feedback/gain/sprogs, all this whilst trying to breathe at 3500 metres asl, on top of Jabel Fayfa.
Never heard of a set top box called a converter. It's a set top box.
It appears to be a holdover from when VHF-only TV sets had to have a downconverter box added when UHF transmissions were introduced. And for viewing DVB-x on a conventional analog telly you'd also need a (rather convoluted) D/A converter.
from when VHF-only TV sets had to have a downconverter box added when UHF transmissions were introduced.
Converters long predate that . The original ones were added to single-channel TVs so that they could receive Band III ITV as well as the original Band I BBC. My grandparents had one.
Strange. I grew up in Palo Alto, we were very late to the CableTV party and made do with UHF and VHF broadcast TV, with a simple aerial feeding the built in analog tuner. Might be why I honestly don't remember anybody calling the set top box a converter ... It probably doesn't help that I never really saw much use for DearOldTelly to begin with. Far too much noise, not nearly enough signal.
Thanks for the input, guys. Have one on me,
A "Converter" is the shortened name of what was usually called a "Cable Converter", a box that the cable TV companies used to convert their signalling down to VHF to be fed into an analog TV set, usually on channel 3 or 4. When the US switched to digital TV broadcast, the term "Digital TV Converter" was also the standard term used to convert the digital broadcast to analog for older TV sets.
Don't believe me? Go to walmart's web site and type in each term.....
"Pardon my ignorance, but what is a rectifier supposed to be when it's not a collection of diodes, and if that's what it is, how does it prevent a major load from overwhelming an inadequate power supply?"
Its 35 years ago, I can't remember exactly what it was called now - it prevented power variations trashing the floppies is the most important thing, so yes it "rectified" the problem :)
I found a fairly new Braun toaster in a skip, it worked but the lever would not stay down. I discovered that the narrow handle of a teaspoon fitted into the slot and so I used it in manual mode for some months. Strangely, after moving house the toaster worked normally again for a while.
I managed to remove the weird* headed screws and revealed the problem. The clever Germans had fitted a switch that detected if the crumb tray was fully inserted, but the plastic catch was weak and if the tray moved out by a couple of millimetres the toaster stopped working again. I soldered a link across the switch and hey presto it was restored to toasting glory!
I suppose that nowadays the toaster would have to be connected to the internet before it functioned :(
*Why do they fit these stupid anti-tamper screws?
I'm all for recycling, but I'm not sure I'd want to scavenge a food preparation device out of a skip... For example, I once through out a perfectly functioning toaster at uni because I came home to discover mice stealing the crumbs from within the toaster, and on closer inspection found evidence of piss and droppings in there.
Enjoy your toast though :-D
* anti-tamper screws to keep warranty costs down. Normal screws, a person who could just about wire a plug would probably have a go, make the problem worse, put it back together, claim under warranty denying all knowledge. Anyone savvy enough to own a triwing screwdriver would do the same thing, but less likely to make it worse..
"Anyone savvy enough to own a triwing screwdriver would do the same thing, but less likely to make it worse.."
A neighbour asked me to look at her Honeywell Tower fan that had stopped working. They had used a combination of Phillips-head screws and small size Torx screws. The latter screwdriver tips tend to only come in full Torx packs - not in the usual domestic couple of larger sizes.
The only user maintainable part was a cover designed to be safely removed so that the fan blades could be vacuumed. It was held in place by one set screw - a small Torx!
I had a light fitting from Phillips and apart from the usual BS about including a "natural light bulb" which was in reality the standard "premium" appallingly not-very-cheap-but-nasty yellow phillips CF unit that took the obligatory 5 minutes to get to the stable colour but the unit had a bloody star-hex bolt to hold the inner compartment with the bulb in place. The kind of thing allen key that you only have one of, as in the one that came with this single lighting unit that for "reasons" required it and would therefore inevitably lose it and be unable to find it when the cheap'n'nasty philipps bulb failed or I chose to replace it with something better. I can't think of any good reason for this bolt and the two square metres of hieroglyphical generic warning text didn't indicate why either so I replaced it with a spare cross head bolt that I had lying around...
Strangely, the negative reviews about this unit all centred around the appalling bundled CF bulb and the moronic star-hex bolt that was used.
I have a small fitted plastic case containing 154 bits that can be used to remove nearly every "tamper proof" screw known to man. It cost me all of $19.95 from Ace Hardware about 20 years ago. They are not the best quality tools on the planet, but they have done the job for me all these years. And wonder of wonders, I haven't lost one yet ...
Anyway, all you "remote job site" folks out there would do worse than to keep such a kit (and the required 1/4 inch driver) in your traveling gear. I also throw in a complete 1/4 inch ratchet set, just because I can. Has saved me I don't know how many hundreds of hours over the decades.
Yes, 1000 times this.
I work as an electrician, but even when NOT working, i keep such a set in the car for the inevitable 'our microwave quit, you know stuff about this...' calls.
I posit the reason you don't lose them is because they come out ONLY when needed, and the neat plastic container lets you know instantly if one is missing. Mine came from a DIY stall on a local market, cost about the same but in UK pounds (about 14.99 if i recall) and also has no bits missing :).
+1 also for the ratchet set.
New server delivered and installed on site. This was back in the days of 1/2" tape drives & Torx screws being new and almost unknown. The tape drive had a transit lock secured by Torx. A couple of days later the vendors engineers (plural - where do you get service like that these days?) rolled up to remove the lock and were a bit miffed to discover that I had a screwdriver set with Torx bits in it.
For some random tamper proof fixings that trick with the elastic band and an Alan key can work - if you can find something to jam the thing holding said fixing safely against while you apply the enormous pressure required to force the key and band firmly into the available space. Use one on a 1/4 inch driver and not those right angled ones that can pop out the back of your hand alien style!
....Screwdrivers tend to have legs...
How true. I have never skimped on the quality of tools. A cheap screwdriver will operate for 1/20th of the time of one costing only 4 times as much.
I once visited a customer 100 miles from our office and was happy to be reunited with a screwdriver with my initials carved into the handle. I'd never been there before, but another chap from my office had, and had to supply some of those items to the right as compensation.
"And wonder of wonders, I haven't lost one yet ..."
When you know you _really_ need it, you tend not to let anyone borrow them and you also tend to make sure they get put away afterwards.
I've lost a lot of tools to loans over the years (sometimes loans by a flatmate, they were a bit miffed when I expected them to replace, pointing out that they did NOT ask my permission to loan MY tools out, even if it was to a good friend of mine). The policy now is that before you borrow, you give me a "bond", in cash, equivalent to the cost of replacing the tool. If the tool is one of a set then it's the cost of an equivalent set. If you damage or lose one, I have the $ to replace it. If you lose one of a set, I replace the set and you get to keep the rest of the old set. If the set comes back OK, you get the money back. Oh, "all there but not in their proper place" counts as "damage", as does "comes back dirty". Yup, don't clean the tools then you can keep them; I'll keep your money and buy me a nice shiny new set.
Where it's a tool that someone gave to me, especially the few I have from stuff my parents gave me, the replacement cost is your life, which you must give up willingly before you can borrow the tool - the tool is irreplaceable (sure I can buy a new 14mm spanner, but I can never again have a spanner that Mom brought me on my 16th birthday shortly after I first started tinkering with motorbikes). And me using said tools on your vehicles comes at a mere $standard_hourly_rate which depends on who you are, what I'm wanting to do with that time if I wasn't helping you, how good you are at making coffee and so on.
t cost me all of $19.95 from Ace Hardware about 20 years ago. [..] Anyway, all you "remote job site" folks out there would do worse than to keep such a kit (and the required 1/4 inch driver) in your traveling gear.
Yup. Years back I built up a tool kit for when I was on the road, stocked with the best cheap tools I could find. When the boss tried to berate me for it for the umpteenth time (he had a go every time we were on a job together) I pointed out that 1) I'd had the odd tool/bit needed when his kit didn't, 2) my entire kit cost less than his multimeter (a $NZ600+ Fluke with all the trimmings, whereas my $20 one did just about everything his did, though I didn't have some of the special probes (some I only ever saw in their case in the few years I worked there) and 3) if someone stole my kit I'd be able to replace the bulk of it for the cost of a few days lunches, and I'd get the rest next week, whereas if his went walkies it was an insurance claim, and a long wait for the supplier to travel down (or ship them down), shipping costs alone enough to cover a decent set of screwdrivers, sockets, and oddbits such as Jake described.
Anyone savvy enough to own a triwing screwdriver would do the same thing, but less likely to make it worse..
TriWing, Robertson, Bristol multi-spline, Torq-Set, ASSY, Torx sockets, XZN, snake-eyes, Pentalobe ... And if none of those fit it's time to bring out the Stanley Fubar XXL.
Why do they fit these stupid anti-tamper screws?
So people won't try to fix stuff, and will have to replace it.
SWMBO bought a "juicer" (an electric version of what my mum used to call a "lemon squeezer") for about £60, to make her morning orange juice (out of actual oranges, fancy that! Anyone would think she'd never heard of cartons).
After just enough time for the warranty to expire, the clever mechanism whereby the motor starts automatically when the half-orange is pressed down onto the clear plastic thing that it apparently called a "ream cap" ceased to work. I had to drill through three plastic screw covers to get to the three screws that held the top on the motorized base to expose the internal workings. The repair involved replacing a simple microswitch with a spare I got from Maplin for a couple of quid (which would probably have been 5p if I'd gone somewhere else and bought a hundred). A satisfying saving of about £58, though.
Strangely, the microswitch is supposed to be good for 5 million operations ... that's an awful lot of orange juice!
I may know that link. Or one like it. Across the Bristol Channel wasn't it?
The interesting thing is working out why it fails on a tide-driven schedule, as there's obviously nothing obstructing the line of sight. It wasn't passing ships, as they were not regular enough.
I'll leave that as a reader (or maybe radio-engineer) exercise.
Something similar years ago in the 80's\90's, new (first one in the house) colour TV, would lose colour information on a house near a estuary, the aerial installer had aimed at the transmitter & got a colour picture "satisfied job done".
As the tide washed in\out the signal level dropped off & the TV just about kept enough of the signal to display in glorious mono.
Aiming at a different transmitter solved the issue.
A microwave beam has a bit of spread. When the tide is high some of the edge of the spread is bounced back to the receiver - when the path length of this bit is some multiple of half wavelength.s it weakens the signal. If you imagine a wave trough as a concave mirror you can see how it could actually focus enough out of phase signal to cause it to drop out.
Hope I tell this right - it's not really my area: My father in law told me of something similar happening in the Missouri countryside, as heavy mist formed in the morning. The mist would rise and cut off communication. Their solution was something along the lines of having multiple transmitter/receivers at different heights on the towers. I believe the communications had to do with controls of a remote power plant. Probably Tomsauk, because he was one of the engineers involved in the design/build. Worth googling - it was very odd to drive to the top of a mountain and look at the vast expanse of water in the reservoir.
"It turned out to be tides, the link ran over a stretch of sea and high tides were enough to disrupt it."
In that case whoever engineered it should have been sacked. Tides are a known issue on microwave links over water and easily solved with a receive diversity antenna (it won't just have been high tides)
We experienced the same problem on a link that had been functioning flawlessly for years. The link would drop for a short period and re-establish. Everyone was perplexed by the issue and no one could work out what it was until I was sent up to the roof to look at the dish, turned around and looked towards our other site some 15 Km away and saw the newly built crane swing into the line of sight. Management initiated frantic discussions with the owner of the new building and the crane was raised about 5 metres to get above the line of sight. The cable swinging through the microwave link was not enough to break the link but the structure of the crane was. Problem solved.
"and saw the newly built crane swing into the line of sight."
I'm kind of surprised about this. We unofficially tested the fade resistance of some of our 5 foot dish links by hovering a Huey in front of them. Didn't skip a beat.
The fesnel spot over a 30km link is suprisingly large even 10 metres from the dish.
A land registry database was discovered to be losing access to some owner records. The extent of the problem only showed up when weekly housekeeping from transaction archive tapes was done. The database tidy would complain that a "new name" transaction failed because the name already existed.
The local application programmer had decided that the very common family names in that culture gave a data record key that required too much searching by software.
Instead a name index was created which assigned an incrementing number to each new name as a substitute key for the data records. The search to get the number from the name index was automated by a clever use of self-modifying disk search commands. The disk controller either returned the number - or a "not found".
If it was "not found" then the terminal operator was allowed to create a new index record. Unfortunately if the "not found" was a spurious "miss" - then the new index entry overwrote the old one - thus orphaning all data records using the name's previous number.
At that time support queries to headquarters were sent by expensive commercial telex - so had to be succinct. The reply came back "not a known problem".
A small test program was written to reproduce the problem - which proved to be very intermittent. It had good days and bad days. At the record level it became obvious that it was misreading a significant dummy record key of hex "all ones" that terminated a track's data.
Eventually an engineer and myself had the whole machine to play with over a weekend. The engineer checked the voltage tolerances on the disk controllers - all were within the mandated 10% tolerance. He even changed boards speculatively. It was then discovered that the fault could become solid. Adjusting the voltage to within 5% fixed it - a hardware fault that could be inherent in any machine.
A triumphant telex to headquarters received the reply "Oh - we know about that problem".
"Instead a name index was created which assigned an incrementing number to each new name as a substitute key for the data records. The search to get the number from the name index was automated by a clever use of self-modifying disk search commands. The disk controller either returned the number - or a "not found"."
One of the more clever tricks in this kind of thing is to increment by a prime number and set at least 2 keys based on parts of the name with different prime numbers assigned. The advantage was that you were less likely to have fewer increments before you found a blank spot (or the record you wanted) which was important in the days of floppy disk seeks.
Ah the days of having to have a large namespace (electronics parts catalogues) accessed on a small dataspace (floppy disks) - we revelled in the sheer luxury of a 5MB drive when it became available.
Way back, I had an Amiga 1200 with an intermittently fautly mouse. I couldn't work out what caused it to stop working, the switches weren't the silly silver plated ones, the ball (remember those) was clean, as were the rollers.
I began to notice the thing stopped working always at a certain time of day and gradually realized the problem was caused by sunlight shining through a nearby window into a gap in the mouse casing. This was sending one of the internal sensors daft. Solution: A piece of paper taped over the gap.
I could have pulled down the blind but I liked the view over Rivelin Valley from my window.
A friend occasionally caught her PC scrolling intermittently when she wasn't at the desk.
Her optical mouse was often illuminated by sunshine through a window - but only scrolled when hit by the oscillating dappled shade caused by wind in a tree outside. When she was holding the mouse her hand blocked the effect.
Had the reverse of that. In early Nmos design we had a circuit that had been tested to pump a current into the chip substrate to reduce leakage. It was used on many test designs and worked perfectly and was chosen to be used in some chips for a serious demo. The cct board of chips worked fine and was boxed up for the demo. And promptly stopped working, The cct board was removed for testing and worked flawlessly. Rinse and repeat. Then the light bulb moment and indeed a lightbulb was put in the demo box to provide the photons needed to make the pump work until the demo was over and then a frantic redesign of the pump so it could work in the dark like the rest of us.
Solution: A piece of paper taped over the gap.
Had that as well. Moved a couple of computer manuals from one end of the desk to the other, so that they were between the mouse pad and the window!
(I think it was a passing cloud that made me figure out what it was. The mouse was driving me nuts, when it started working fine, stopped, started again.. I got up to grab the old one with the faulty wire but-still-worked-better-if-you-held-it-just-right, saw the bright sunlight across the mouse pad for a moment before it faded as another cloud passed, and the lights went on so-to-speak)
My company used to install account top-up machines, which worked with a cashless system for catering, etc.
One of the machines was mysteriously loading more money than it should, intermittently.
TL:DR it was installed next to a lift, and the electric field from the motors was triggering data pulses in a parallel ribbon cable that connected the coin counter to the control board. Those data pulses happened to mimic the signal for a £2 coin.
Solution: better shielding or use more upmarket components.
Had a similar experience while living on school campus - every time someone used the bathroom the computer would restart. No carpets but the bathroom didn't have a window.
After about 3 months several possible causes were dismissed - only the bathroom light remained. A fluorescent tube paired with outdated wiring will do that. The solution was spending 70 bucks on a really fancy extension cort with interference supression filter and surge protector. Of course that didn't help with the frequent power outages because someone plugged a hair dryer in...
I had a similar case with a friend with intermittent broadband faults.
It was around Christmas, so my first thought was her tree lights - but it turned out she had also replaced an electronic starter in the flourescent lights in her kitchen, and that was kicking out enough interference to knock her router off-line. A quick MW radio test confirmed my suspicions - nothing much in the way of noise from the tree lights, but turn on the kitchen lights and there was a massive burst of squealing from the radio.
Replacing the faulty starter fixed that pretty much instantly.
I don't remember where, but it was in a column on troubleshooting odd problems. The only differences I can spot are that in the story I read, replacing the weak power supply in the PC, not the well pump, was the resolution.
Of course this could certainly have happened more than once. Either way, a fun problem to diagnose.
Uprating the PSU is not necessarily a good solution. The problem would have returned in a few years when the source created bigger power spikes and the capacitors in the PC PSU have degraded. If possible, try to find the root cause of power supply problems.
"Gary" in the original piece here did a good job at spotting the problem.
The proper solution is NOT to replace the pump. Those things are spendy. The proper solution is to rebuild the pump. They are built to be rebuilt. I have one Sears 1.5 horsepower jet pump that is older than I am. It is on it's 6th rebuild ... and due for another one.
Depends. If it was a submersible pump, you replace it. You may never see it again.
I just had one replaced for $1800 and it was only a 84' hole with the pump at 60' (I have a 3' water table) - original had been in there 16 years. The deeper you have to go, prices increase quite a bit.
Edit: I don't get why the pump went on at every flush - this system not have a pressure bladder tank to stop that from happening?
"I don't get why the pump went on at every flush - this system not have a pressure bladder tank to stop that from happening?"
Modern bladder tanks are tiny. Older ones tend to have collapsed bladders (or someone's helpfully bled all the air out, thinking it was an airlock in the system)
I'm having to explain this to SWMBO about the "professional" water installation in outer bumfuckistan and why it's a bad thing that the pump keeps turning on then off, then on, then off, then on then off each time someone flushes a toilet or runs the shower.
"Modern bladder tanks are tiny."
Only if the system is designed & approved by idiots
Have to agree, especially as the article indicated this was a farm environment. I've seen 200L+ (about 50gal I believe) amounts needed to be drawn before the pump has to engage, and a few places even more - especially if the pump is the only one on the farm. The idea is to have the pump start, run for a long while, then stop - which is much better for mechanical stuff than start.stop.start.stop.start.stop...
failed to boot, couldn't find anything wrong with it. Cold started it, discharged the PSU completely. Checked the voltages on the PSU. All good. Took it into the physics lab at school to look for a clock pulse with the oscilloscope. That was all there... still no joy. So I decided to pull the CPU out and look at the signals there, swap in a fresh CPU if needs be. Touched the top of the chip and got a small kick in the end of my finger. Tried again, and the machine came up straight away.
Took it back to my friend's house, and he plugged it back in and set it up again.
Next day, the same issue, no boot. So I went over on the way to school - he was on his HAM radio set when I called. The aerial lead ran right under the BBC, and there was a little bit more of a SWR reading than I would have expected. On closer inspection, the aerial cable was heavily kinked right where the computer had been sitting on it, and I don't know how, but a charge was building up on the surface of the CPU and jamming it.
Solution? A new lead for the aerial, a small sliver of kitchen foil glued to the top of the CPU with a lead running off to the motherboard ground, and a chicken wire Faraday cage around the BBC, monitor and diskette drive, just for good measure.
In my early 20's I was involved with sending data over HF/VHF for the Army Cadet Force National Net.
Using a BBC Master and a PRC 320 man pack radio to send at a ridiculously slow speed to the central station at Blanford Camp.
The main problem was that we had the antenna tuned really well and running parallel to the building, so when we were transmitting, it made the landline phones in the building almost unusable and caused some problems with the BBC Master until we put a Faraday cage around the whole unit, including the CRT.
'Gary now developed a hypothesis he described as follows: “When the toilet was flushed, the pump had to kick in to fill the tank and that when the motor kicked in, it had such a current draw because the brushes were probably worn out.” That current draw was what dimmed the lights and therefore disrupted power supply just enough to also trip the computer.'
I remember (back in my radical youth - back in the days when labour were completely unelectable because they were elf by a pensioner and supported leavign Europe, woudl probably unilaterally disarm and soak the rich) during the miner's strike that someone tried to start a campaign for everyone supporting the strike to flush their toilets at 6pm in the belief that this would cause a sudden surge in water usage trigger lots of pumps to switch on causes a power surge which would cause the national grid fail. Somehow I don't recall it working!
There is an interesting story about an old US prison where the inmates protested against their appalling conditions by flushing all toilets in the multistory building topdown on cue of the chimes of a local church which they all could hear. The resulting damage to the sewage system forced the authorities to move all prisoners to a better location.
Ahhhh.... brings back memories on my time in a large IT company (name not mentioned to avoid lawsuits etc). We were facing a number of mysterious crashes in a cluster of hospitals around the Midlands. Patient systems running on top of Oracle 7 (remember that anyone?) were crashing and forcing DB admins to go into horrible recovery mode. Turns out an HW engineer who had not been properly trained would turn up to apply some minor fix to a running server, and simply press the big yellow button to switch it off, apply the fix, press the big yellow button to switch it back on, and leave before the howls of anguish reached his ears.
" Turns out an HW engineer who had not been properly trained would turn up to apply some minor fix to a running server, and simply press the big yellow"
What sort of DBA would allow anyone, H/W engineer or other, anywhere near a running server without escorting them? Some people just don't have the requisite levels of paranoia.
Had a report form a site that all their mice had stopped working, Attended site after ordering a boot full of mice from stores. Checked the existing mice, no problems on the desk. However none of them worked on the mousemats. It turned out that they had replaced all the mouse mats with company branded mats, however the colour that they had choosen (red) ment that none of the mice worked
I came across one similar when I came across the first introduction of optical mice as an upgrade to the ball and wheel mouse predecessors. Turns out that the optical mice didn't like the glossy surfaces of the standard issue branded mouse mats which worked fine with ball mice. It was considered ironic that the workers that had the best working mice were either those with the dirtiest environments, as in their previous mice failed all the time due to the dirt or those that didn't toe the company line and had alternative mouse mats.
We did the testing for optical transceivers in a seperate building and every so often they'd all start to fail at once. Wasn't a problem when they were tested in the main cleanroom. The culprit? Mobile phones, you'd be able to tell someone was about to receive a text message as the failures were almost exactly 20 seconds before someone got a message. Wasn't an issue in the cleanroom as mobile phones were banned within the testing area due to the sensitivity of the equipment. I recall one piece of kit in there had a 1 metre taped area around it and a big warning sign not to cross the taped line if you had a pace-maker
I was once showing two VIPs round where I worked at the time and had been told to show them everything which included the racks of equipment in the server room, engineering spaces etc. So I was all in full tour guide mode and at my most enthusiastic when we reached the first security door. I said to them and said you either need to leave your phones here (a small shelf was provided) or turn them off. This was also written on a large yellow sign on the door but I made a point of checking that they understood. One of them was happy to comply and switched hers off immediately. The other was incensed by this and pointed out that I was just being difficult. I said that I wasn't "Just being difficult" I was following the correct procedures and If he didn't like it he could wait outside. He was very unhappy about this muttered about being a VIP and I indicated that the phone on the wall could be used to contact whichever manager had set the tour up. He called them and then after a brief discussion rather red faced apologised to me took his phone out and switched it off in full view.
we had a malware experiment happening in the doubly shielded indoor tent, based on OpenBTS, and was a sort-of IMSI-catcher with an amusing network name.
er. . . somehow the 10 milliwhats of GSM leaked out an' pranked one of the VIP's phones, who just happened to be the president of the national data protection authority, I think our explanations & grovelling lasted about a week
Back in the days when our company had Blackberries, when a mass meeting was called, the first announcement over the PA in the room was to have everyone turn off their Blackberries. Not for the disruption, but because the audio gear would pick up the poorly-shielded CPUs with several hundred of these devices 'broadcasting' in the meeting room.
Had a similar problem, but in this case the system, a VAX 11/725, went down hard around 17:30, once or twice a week. Obvious suspect: cleaners. Considered improbable, because the system was in a recess, the power strip and the wall socket were behind the system requiring some serious gymnastics to reach them, and there was a power socket in plain view on the wall to one side of the recess. The console printer just showed the usual logging chatter, cutting off suddenly, then the power-up sequence the next morning as someone came in finding the system being down.
Initial diagnostics pointed to airflow problems, and over the next few weeks all of the associated components and wiring were replaced, with finally the entire system being carted off to Repair, to be stripped down, cleaned and reassembled. With any even slightly suspicious parts and harnesses being replaced again. A temporary replacement was installed which ran fine for the entire time it was in use. The original system, which must have been the cleanest 11/725 in EMEA, also ran flawlessly under test. In the meantime the mains supply at the customer had been monitored for spikes, brownouts and noise, as there was a manufacturing hall next to the office, with packing machinery being constructed and tested. That didn't yield anything either.
Moving the system back to the customer, same problem again.
So a colleague decided to go and see what the hell was going on; unfortunately I was off for a training course. He sat there waiting for the cleaners to come in and do their thing. One plugged in the vacuum, in the appropriate socket. The other went to take the waste bag from the document shredder, tied it closed and put it down. Right in front of the air intake for the system. Which duly experienced an airflow problem and switched off.
The temporary replacement system, although being technically the same (an 11/730), had a different enclosure and didn't mind plastic bags being put right against it.
"Right in front of the air intake for the system. Which duly experienced an airflow problem and switched off."
There's a similar story about one of the very old (washing machine size) hard drives and a cleaner who discovered that if you put the dustpan under the front ledge it would be automatically emptied for him.
Of course _WHY_ you would let cleaners into your data centre/server room is beyond me. We certainly don't and I've made a point of banning them from any technical areas containing racking. Techs don't kick up much dust and if they're tasked with cleaning up they make sure than things don't get broken.
"Of course _WHY_ you would let cleaners into your data centre/server room is beyond me."
The "why" is easy. It's their job to clean the place, floor to ceiling, board room to bog, watering plants, replacing dead light bulbs & emptying the trash in their wake. The modern world wouldn't run without janitorial staff. Extending this to include the labs that evolved into computer centers in the 1950s wasn't even thought about, it just happened.
Janitorial staff having the keys to the entire kingdom (as it were) was the norm until we in the glass room started putting our collective foot down in the late 1970s/early 1980s. It wasn't until the late 1980s that it became uncommon. By the late 1990s it was as rare as hen's teeth. The last time I witnessed a janitor coming unannounced into a data center "in the wee hours" at a place I was consulting for was 2005 ...
Of course _WHY_ you would let cleaners into your data centre/server room is beyond me.
In this case, the system was in one of the office areas. Hence the presence of a paper shredder close to it.
 The 11/725 was essentially an 11/730 with an 'office-friendly' enclosure. Subsequent office-friendly enclosures, like for the mVAX 2 had a more office-friendly airflow layout.
My personal favourite and I used it as an interview question for years after.
Back in 2000 l worked as IT bod in a typical little gov.uk place, lovely staff and a joy to support.
Got a call from top floor, big meeting, projector not focusing, tried everything, help!
Sensing panic we sent up our pfy, five minutes later HELP ! Big guest just arrived, projector still bad.
Sent our junior bofh, no joy, big guest now grumbling.
Grabbed spare projector and ran upstairs, then removed red hot lens cover and quietly left the room.
OK, so mid 90's I was doing onsite repair for homes and business' for a big computer retailer in Dublin.
I had a call to go and replace a floppy drive. Went out and tested the drive, it worked, but the customer said it was intermittent, so I replaced the drive anyway, tested it, it worked fine. Took the replaced unit back to store, plugged it in to a test machine and it worked fine, so boxed it, labeled it as used working and sent it back to the parts dept.
A week later I got a call (different customer) with the same problem. Went out with another replacement drive and again the customer said it was intermittent. Tested the drive and it indeed didn't work but it wasn't that the drive didn't work it was that the BIOS on boot didn't detect the drive.
So, replaced the drive unit and it wouldn't see the new drive. MB failure then maybe? Had one of those for the same machine out in the car, so brought that in and started unplugging everything to disassemble and had a brief idea to test again with nothing but the monitor, keyboard and mouse attached.
Drive worked fine. WTF?
Plugged the printer and scanner back in. Rebooted, drive didn't get detected by the BIOS.
Now I start thinking about the other customer. What make and model was his machine? Pulled out my Nokia and phoned Customer Care line to get the details of the other customers machine. At this point the current customer walks in to the room and freaks the fuck out and runs out of the house. Another WTF moment. Got the details and find that both customers have the same all in package of computer, printer and scanner.
OK now, for some reason, I decided to power up the scanner before booting the PC.. lo and behold, drive works, reboot with scanner powered off (but still connected), drive doesn't work. Suddenly it all becomes clear.
If this particular machine has its packaged scanner connected but not powered up on boot the BIOS would not find the FDD.
Found the customer outside the house and explained the problem and to just keep the scanner on when booting while we worked on a better fix (how this wasn't found in testing the package before it went to market I don't know).
He apologised for running out of the house when he seen me on the phone and explained that he had a phobia of phones which also explained why I couldn't contact him that morning to let him know when I'd be arriving. That was another WTF?
Way back in the early 80's I worked in the support centre for a large British Computer manufacturer.
I got a call from a site saying that the machine wouldn't load, and that they hadn't changed anything.
Although they did get a printer dump, it was particularly useless, as the only error it contained was KEVM_LAST_USER_OF_GLOBAL_EVENT, which meant that the main VM had failed, of course any stack frame revealing the cause of the error would have long been overwritten.
The processor they were using had a speaker which clicked on every CALL instruction resulting in various levels of noise mixed with gently descending tones. I asked them to reload the machine, and to put the telephone next to the speaker so I could hear the processor rebooting.
After hearing the load I gave them a list of load options, and the machine loaded. They thought I was a genius.
All I did was to notice that the load had almost completed before it failed, and I told them to switch off the last few load options.
Turned out to be a failure in the floating point unit, which caused the VM to crash, but the only option to use floating point in the load sequence was a hardware monitoring stats package.
I live in a farmhouse in France which is at the end of a long line of overhead three phase. When the kettle goes on, the power drops from around 225V to about 190V. The water pump (2kW) kicking in makes the lights flicker. Shall I go on? Let's just say sometimes the LED light in my lamp starts acting weird due to the lack of power...
My Pi doesn't reboot, neither does my PC. Nor the router. Nor the NAS. Nor.........
Yes, modern switchmode PSU's are pretty good at dealing with any voltage you throw at them, including wildly varying mains. Even the cheap disposable ones.
When my last UPS died more than a decade ago PC PSUs had improved enough to handle the brownouts and the short breaks I'd bought it for by themselves.
" modern switchmode PSU's are pretty good at dealing with any voltage you throw at them, including wildly varying mains."
"That depends" - there are 85-260V ones and 220V or 115V fixed voltage ones which are far less tolerant of "fun" shit on the input.
The fixed voltage ones are a few dollars cheaper. Guess what virtually every computer in outer Bumfuckistan (where the nominal 230V has been known to drop as low as 95V for prolonged periods) is fitted with?
Unsurprisingly, the most popular electrical items to be sold there are power delay timers (10 seconds to 10 minutes after the power comes on, it's finally switched back on to the appliance, to ensure it's on solidly on and not going to drop out again - except the collective cross-area *thunk* of the 10 minute timers switching in large loads like ACs and fridges usually causes another brownout....) and "Voltage regulators" - which are basically autotransformers with self-setting motor-controlled output in their original form and yes, that's as dangerous as it sounds. More modern ones tend to use Triac-selectable transformer taps which is still dangerous. Of course the collective effect of all those devices sucking more and more current as the voltage decreases (P=IV remember) means that brownouts get worse and occasionally house supply wiring catches fire (as does the distribution system from time to time).
Do your homework. If you plan on owning said pump for any length of time, get one that is rebuildable. Might cost a little more at the outset, but over the long haul it'll save you a lot of money. Here in the US, Grainger's is your friend ... I also use Palo Alto Electric Motor. PAEM will do the pump rebuild for you, but they are equally happy to sell you the parts to fix it yourself. Both do mail order, and are recommended. Not an owner or an employee, just a happy customer of both for well over 40 years.
at a relative's. Nice hot summer days, his network (multiple routers and hubs, lots of devices), set up a couple years before, would intermittently drop out, or simply lose packets for a few seconds. Was driving us all nuts.
I started plugging my laptop into various hubs/routers and running hour-long ping sessions to other machines. Finally found two connections between hubs that were the issue. Pulled out the network plugs and took a REALLY good look.
Turns out, in each case, he had made the cables himself and one wire hadn't quite gotten all the way to the end of the plug. When it got hot enough (one of these was in an attic area!), the dodgy connection between the pin and the conductor would shift just enough to drop out. Recrimped two plugs, and everything's been fine since.
POS terminal at a garden centre would lose power and reboot itself without warning. We determined it happened every time they used their 20W VHF radio which was nearby.
Turned out the very long unshielded power cable used to cross the yard from the socket to the terminal was exactly the same length as the frequency used and something? happened when a message was sent enough to send the Bird RF meter crazy when we connected it.
Ended up digging a channel in the floor for a pipe and running the cable through it which worked a treat.
That reminded me.
Mate told me the disk drive wasn't working, so I asked him to check both cables were properly plugged in. "Yes." He said "I can see the light is on so it must be trying to do something."
I'd seen this so often I just told him to unplug the 'wide' cable, turn it over and plug it in again, to be greeted by "Oh, I didn't know it mattered".
A lot of cheap plugs had no bump, just the slots on the other side. The sockets on the Beebs only looked for the bump.
Hardwood floor. Not so hard, it would bend under the weight of a single person passing on the corridor. Wall sockets were loose. Walking by that terminal would cause the voltage stabiliser to kick between TAPs and cause juuuust enough of an upset on the power supply after A SINGLE FAN was added to it, forcing a reboot.
Wall sockets retightened, 10 years-old voltage stabiliser dumped, cleaning lady instructed not to sweep under desks by the wall sockets so they wouldn't get loose again after pulling the cables with a mop, problem sorted.
There is a scene in the movie 'Annie Hall' where Woody Allen corrects a pedantic-but-wrong neighbor in line at the movies by reaching out of frame and pulling Marshall McLuhan in to tell the man off.
A very long time ago, when this was still relevant, I was on a help line talking to a woman about operating her Internet Gopher server (I SAID this was a very long time ago.)
I told this woman to do something with Gopher. (I no longer recall exactly what.)
She told me, "You can't do that with Gopher."
I assured her that she could.
She insisted I was wrong.
I assured her that this was completely possible.
She became very upset, cited her bona fides as a system administrator and computer expert, and told me very certainly that she could NOT do the thing I described with Gopher, because Gopher was not designed to do that.
"Ma'am," I said politely, "I'm one of the people who created Internet Gopher. My name is on the RFC, number 1436. Additionally I personally wrote the Unix Gopher client we're discussing, and I absolutely guarantee that my software can do what I am telling you it can do."
She reluctantly agreed to give it a try.
I still run a gopher server. It has one user ... my Great Aunt (101 years young). She is using it to publish her life's story. She started the project in 1992 (before RFC-1436), when I discovered the then early gopher & WAIS (I was at SAIL) and decided to show her what computers were useful for. I'd have moved her over to the Web years ago, but she's resistant to change and quite happy with gopher. I kind of suspect that the project is one of the things that keeps the old girl going ...
Years ago (early 80s) when I was a service tech, one of my customers was the local Electricity Company. They had a system, that worked fine, but a few days a month, it would crash at random - fine for the rest of the month. This was getting annoying, so I got hold of a Dranetz - a very expensive, very accurate line votage analyzer. Sure enough, we could see a mains glitch that was causing the crash. We told their engineers, who insisted that they had a clean line to the computer room from the main switchboard. I doubted that, and insisted that we chase the wires. we did, and lo and behold we found a junction - follow the branch at the junction and it heads into the women's bathroom. What? Those were the days, when only two of the staff (in that building) were women. Keep following the wires, and what do we find - i the Women's bathroom was a Sanitary Napkin Burner - nowadays companies com and take away such things for external disposal, but this place had what was basically a toaster on steroids to burn the items to powder, although these things are still used in India and a few other places.
So, it just so happened that the two women in the building had mentrual cycles that were in close sync, so the power spikes were consistently happening in specific times every month.
Another customer, was the state electricity distribution centre. I was installing a large Minicomputer there, and asked, where I should plug it in. The engineer pointed me at a socket, and being meticulous, I metered it first. Nothing. I told him that there wasn't any power to that outlet - he walked over to a massive switchboard, and threw a switch - and there was a "boomph" from behind the panel, and the place went totally dark.
Electricity Authority - (different to last) - their grid control computers were out in the countryside, and running of a 48V DC line, with a big bank of batteries. . One of their own trucks hit a powerline, that was feeding that facility, and knocked out the grid supply to it. The batteries hadn't been checked for a while, and the 48V Dc, was suddenly 4V DC. This sent the grid management into a tizz, spiking various lines and so on. Interestingly, the only computer to suffer, was the main commercial system at the head office of the computer company. PDP11/70 running RSTS/E with RM03 drives. Well, the System Disk, picked up the power spike, and from the autopsy appeared to do two things, it shorted out the servo head, so didn't know where it's heads were - when that happens it does the safest thing and retract the heads - but it did that which such force that the ferrite rod that was acting as a velocity transducer, shattered. So, it wasn't sure how fast it was moving, so the electronic sent the heads in the opposite direction - until they reached the stop at the far end of the rails (which should never be reached, because the servo would slow before then, but remember it was dead, so it would run right to the end, hit the stop (which had a microswitch, and change directions). Repeat and Rinse... Well, the disk drive started to move itself across the computer room floor.
As I recall, the list price for an RM03 (67MB) was about $23,000 - the immediate suggestion was to trash the drive, and put in a new one. For some reason the IT Manager insisted on getting that one rebuilt - of course, the innards were shot, and I think the rebuild price was about $18,000.
While I am on a roll. Was installing a system at a new Department of Defense site - it was a SCIF (Sensitive Compartmented Information Facility) - basically a large faraday cage. Being in the initial setup, it was still being wired etc. I was setting up the computers, and in comes an electrician to do some lighting or something. He has a tool bag, and on the top is his transistor radio playing the local FM station. He closes the door behind him, puts his bag down, and notices that his radio isn't working. Thumps it a few times, I am heading out to get something, which coincides with him hitting his radio, and he thinks he has fixed it (nope, the RF shielding around it has a gap in it now. ) I come back, he gets up to do something outside, and closes the door again. Swears at his radio. I explain SCIF's and Faraday cages to this electrician...
Farm Related - I had a customer that was a vet science research lab.
1 They had a punch car reader, that allowed card decks to be uploaded to a remote mainframe - One, and only one, of the scientists complained about it misreading - so I came out, spent hours, doing a full overhaul, running thousands of test cards through it, watching the signals on oscilloscope, etc. It was perfect - a week later - same problem, same person. I check it out again. Looks great. Sure enough, a week later, same problem (SAME PERSON), so I turn up, and ask the Manager, can I talk to the guy logging the failure reports - sure, he calls on the walkie talkie - I sit gazing out the window, and in rides a guy on a horse, ties up the horse, and comes in. Introduces himself, and I say "so you have been having problems with punch cards not reading - can we try them now" - he reachs around to the back pocket of his jeans, (that have been sitting between him and horse), and pulls out the deck of cards... I explain that they are very intolerant of moisture, bending etc.
Same organization had a remote cattle research facility - they had an old PDP8E that monitored the vitals of cattle. They had a big fibreglass shell, that was remotely lowered (it was in a barn, and the switches were in the computer room), and it would have sensors that worked out temperature, breathing rate, etc of the animal.
Well, they were getting weird readings so blamed the computer. I pulled it apart, tested everything. looked good, they tried again. Same problem - So, I said "can we have a look at the other end" - we go to the barn, someone lowers the shell over a cow, and it seems agitated. So, I ask for it to be raised, look up into the wiring at the top, and there is the answer - a very large snake coiled in the wiring.
The IT guy in an aluminum mill gets a call late one night. The large, high-resolution color CRT on the system monitoring the hot roll line had failed. The on-site maintenance tech had swapped out every other CRT he could get to with the one on the hot roll line but none of them would work. He was calling the IT guy because the last CRT in the plant was locked in his office.
So at 3am, he wonders into the plant and goes into his office to unhook his CRT. The two of them move the CRT out of his office to the line and proceed to drag it the ½ mile down to the control room. The maintenance guy says, “Wait a minute, I found an easy way to move the CRTs.” He walks over to the 25 ton crane with the magnetic coupler on it and before the IT guy can stop him, he swings the magnet over to the CRT and locks on. Up goes the CRT hanging by the 25 ton magnet and off he goes with it.
There was absolutely no way to ever degauss that monitor.
My best diagnosis was about 15-20 years ago - i was working for a bank in Ireland. we were developing a new backend authorisation system for retail point of sale devices. They were dialling into banks of analog modems and then generating an X25 payload into the backend system for authorisation.
We had a nagging 1-2% transaction failure rate in testing which was totally unpredictable and driving me bananas.
After spending endless hours running live captures trying to capture the interrmitent failures in flight, I eventually found that the X25 packet length for the failures was different (shorter) than all the successes. Then going back form that we established that transaction only failed if the credit card number contained a particular value (card number ended with 00 or something like that).
So it was a software bug in the originating POS terminal software and nothing to do with the backend sytem we were testing. If the card number met a certain criteria then the POS terminal did not append a CRC value that was a required part of the transaction payload. Which resulted in a payload that was 4 bytes too short and rejected by the backend system.
Serious needle in haystack stuff.
Last July, I worked at home with a senior lecturer producing a course document on using IT&C in research. After a long day he left for home needing to just do a final read through and to email the final copy to me.
At 2330 I got a very panicked colleague who reported that the laptop had died and wouldn't start. It was plugged in but nothing. It was fine during the day and had worked perfectly for an hour but died whilst he got a brew.
I told him I would call in in the morning to see what I could do, even if it meant removing the HDD into a caddy. Next morning I headed off to see him with 10 year old Grandson in tow. Whilst I got the low dow on what happened and before the kettle boiled, my grandson walked up to the machine stared at it then pushed the powerlead into the charger fully and bingo. The look of contempt on my grandsons face was priceless as he added, "now can I get a lift to rugby practice, I'm gonna be seriously late!"
My colleague taught electrical/electronic engineering for 31 years previously having been in industry. At 67 he had comtemplated retirement but loved his work.
However, a few weeks later whilst in my home office again, he watched in amazement as my 8 year granddaughter came in apologising for sending her homework to the wrong printer (from her home) whilst bemoaning the fact that the 'HPLIP set up on the laptop was only allowing single side printing again.'
He retires next week. Happy birthday and don't worry, let the kids will deal with it. Get back to finishing your book on thermionics before your knowledge is lost.
Biting the hand that feeds IT © 1998–2020