标签: capacitor

A few weeks ago, Max the Magnificent considered what was described as The Great Capacitor Plague of the Early 21st Century . This prompted several readers to share various levels of experience with recent capacitor problems. However, a few smudges on printed circuit boards, some smoke, and the occasional flame all pale into insignificance with regard to the experiences of capacitor users in times past.   In the middle and latter part of the 20th century, it was almost a rite-of-passage for anybody making claims on the electronic engineering profession to have sat through a storm of aluminium flakes, paper, fluffy chemicals, and heaven knows what else that rained down following the explosion of an electrolytic capacitor.   What follows is a true story of how an exploding capacitor delayed the development of part of the British computer industry by at least six months. The name of the individual involved has been withheld to avoid embarrassment. We will call him "The Young Man," or TYM for short.   Setting the scene This exploding capacitor story starts long ago in a distant place, when transistors made of germanium were in their infancy and magnetic cores were the memory of choice for the latest embryo electronic computers. The distant place is Croydon, England, now a part of greater London, once famous as the site of the first London airport, perhaps now more celebrated for a ladies' hair style named "The Croydon Face Lift."   The airport had spawned number of light engineering electronics companies, including Philips/Mullard (radios and televisions), ICT/ICL by their earlier name of Powers-Samas (tabulators and computers), Creed (teleprinters), Dictaphone and Muirhead (fax machines), Aeronautical and General Instruments (AGI) (aviation instruments), and many more, all moving by various degrees into electronics.   Now visualize a bespectacled school leaver -- TYM, who is central to this story as a very young man -- who had bicycled around all these companies to find out what they were about. TYM finally ended up knocking on the door of Powers-Samas -- which was to evolve into International Computer Tabulators (ICT), then International Computers Limited (ICL), and eventually Fujitsu -- asking if there were any jobs working on electronic computers. He had an interest in electronics, he'd built a few vacuum tube and germanium transistor radios, and he had also won the Physics prize as his most significant grammar school academic achievement. He was directed to the training school. In those days, large engineering companies in the UK ran training schools, offered formal engineering apprenticeships, and paid for employees to attend college and/or university. Here TYM met a kindly Mr Wood who managed the training school and who listened sympathetically to TYM's electronic engineering dreams and employment requests.   This is the point in our tale where luck steps in and plays its part, because Mr Wood was a mechanical engineer who knew very little about electronics, as was the case with most of the senior and middle management of Powers-Samas / ICT at that time. However, Mr Woods had recently been visited by the head of the physics department at the local technical college, who was recruiting students for his department. For TYM, Mr Woods then uttered some life-changing words: "Physics is halfway between Mechanical Engineering and Electronics, so we will pay you to study Physics with Electronic Engineering as your Applied Physics subject, and we will employ you as a Mechanical Engineering Apprentice." All of this was undertaken on a day(s) release program.   Sometime later... While studying during the following years, TYM was provided with experience in the design and manufacture of all manner of machines involving punched cards and their associated mechanical engineering, while at every opportunity trying to associate himself with any electronics-related developments. As a result of endless pestering, TYM found himself as a know-it-all kid attached to a very small select group of top-class electronic computer design engineers who were developing the first British ferrite core-based computer in secret. Along the way, TYM had some development contact with another secret company computer called the Samastronic, which was a machine constructed by highly-skilled mechanical engineers as their best effort at computer building. TYM also carefully avoided any involvement with the company's first all vacuum tube electronic program controlled computer -- the PCC -- which managed to become a disaster without TYM's help.   The new computer, which was called the FCC (Ferrite Core Computer), had all the architectural features of a modern single-chip processor or microcontroller using discrete components. The program and data memory were magnetic core, some of the magnetic core drivers were vacuum tubes (good old EL34s), while the arithmetic unit was implemented using diode transistor logic (DTL). The diode part of the DTL employed the very fast IN96 point-contact germanium diode -- a device that will feature later in this exploding capacitor story. The high-speed scanning for one co-ordinate of the core memory used a modified version of a nuclear weapon coaxial cable trigger device. The computer programs were loaded into the computer with paper tape or punched cards or a simple keyboard. There may have been a teleprinter as the output device.   As a very junior member of the team, TYM was provided with an excellent opportunity for learning from some of the very best computer designers in the country/world, and to carry out small circuit development tasks. One day, somebody must have looked at TYM's company experience resume and seen that he had been involved in the design and construction (mostly construction) of the power supply for a machine that could photo-electrically read pen or pencil marks on a punch card and punch holes in the correct locations. Thus it was that TYM was given the job of designing one of the many FCC power supplies -- the one for the high voltage required by the vacuum tubes.   TYM set to work on a design for a power supply that would have a front panel that fitted in a standard 19-inch rack, with most of the circuit components mounted on a board fixed at right angles. The design was conventional by today's standards -- a transformer, a full wave rectifier, and a vacuum tube-based feedback circuit featuring a neon gas tube as the voltage reference source accompanied by a large electrolytic capacitor to deal with current surges. There were also a few ancillary components like fuses, fixed and variable resistors, and a voltmeter.   The design was reviewed by TYM's immediate superior who approved it and reminded him to make sure the capacitor was "formed" before installing it. In those days, electrolytic capacitors required forming before use, especially after any significant time "on the shelf." This forming involved applying a staircase waveform of increasing voltage steps, where each step voltage was maintained for a few hours.   Explosive times When the power supply was completed, it was installed in its position among the others on the rack. The complete computer occupied four 6-foot racks linked together with the memory stacks cantilevered out. All went well for a few days and the power supply did its job perfectly...   ...until one day when, during the quiet of the afternoon, the computer development lab (indeed, the whole building) echoed with the sound of a massive explosion that filled the computer room with a mixture of aluminium confetti, fluffy chemical fallout, and smoke, all followed by what sounded like machine gun fire.   The senior engineer on the project rushed from his office, peered through the smoke and settling dust at the smoking ruins of his prototype FCC, and demanded an explanation, not of TYM (fortunately), but of the senior team leader -- the strict hierarchy of command and responsibility protected TYM. The team leader explained that he thought a large defective electrolytic capacitor in a power supply had exploded. The answer to the next question -- "Why are all the arithmetic unit circuit boards smoking and all the DTL diodes and transistors so much broken glass and bent wire on the floor?" -- explained the machine gun-like sound that had followed the initial explosion. It would appear that the explosion had somehow provided a 400-volt surge to the 5-volt devices on the arithmetic unit board. The machine gun fire sound was caused by the chain reaction of exploding diodes and other components as the 400-volt power surge progressed through the various circuits.   The FCC before and after (or "Behold the power of an electrolytic capacitor") (Source: Ron Neale)   The team was instructed to not let this happen again. A few weeks or a month or so later, all the damaged boards had been rebuilt and replaced, and the power supply had been rebuilt exactly as before. This time the electrolytic capacitor was formed and double checked. The development and the collection of performance data continued until, once again on an otherwise quiet afternoon, the computer room was rocked with a repeat explosion of the same capacitor.   This time some precautions had been taken to prevent much of the earlier chain reaction damage. Even so, some damage to other boards and power supplies occurred. After much by way of raised voices and robust discussion in the chief engineers office (expletives deleted), careful analysis showed the explosion had forced the cantilevered circuit board of TYM's power supply down onto a power supply below it and there was evidence of a massive arc.   Things were now getting serious. Jobs were at risk. Eventually, careful analysis by the full engineering team of the offending power supply and its design tracked-down the culprit. Neon gas tubes used as voltage reference sources for power supplies get very hot, so placing one on a circuit board next to an electrolytic capacitor is not the best design move. Lesson learned.   A few months later, the FCC was back on course. Eventually, in the early 1960s, half a dozen were built and deployed in the market as the ICL/ICT 558 Computer. By that time, the machine was at least six months late and not the most competitive of machines. TYM suspects that by that time the computer may have had a different power supply design, but he could not be sure as he had moved on to the physics lab in the same company and was now pursuing the development of ferroelectric memory based on triglycene sulphate and barium titanate and multi-aperture ferrite memory -- MAD devices.   TYM pursued his career in electronics in various parts of the world peppered with other explosions, some more serious including -- along the way -- some very large ones underground in Nevada. Sometimes, TYM could be found bathed in the blue light of Cherenkov radiation while crawling around atop of the nuclear reactor at the Wright Patterson USAF base pursuing dreams of radiation-hard memory components.   Rumours that TYM's attempts to produce low-cost, thin-film, radiation-hard squib firing switches for the wire-guided Dragon missile delayed its deployment into the US military arsenal by six months are totally without foundation (it was a difficult development for many reasons).   In later life, after a long career in electronics, TYM was often found masquerading as a technical journalist and on the lecture circuit. He could frequently be found at his favorite diner -- the Pinecrest Restaurant at Mason Geary in San Francisco -- or at one of his many other favorite eateries, especially the Moss Beach Distillery at Half Moon Bay, California. Questions as to why these meetings were often held in the company of a much younger female purser/cabin attendant from United Airlines or a female PR executive are dismissed by TYM as merely selective observation. He informs us anybody suggesting anything to the contrary can expect to hear from his legal team.   Ron Neale

I am not sure about you, but I tend to put quite a lot of faith in the quality of electronic components. Call me "hopelessly optimistic" if you will, but I sort of expect them to perform the tasks for which they were intended (so long as their environmental specifications and ratings are met, you understand).   Thus, I was a tad surprised when my chum, Rick Curl, sent me an email pointing me toward this Tech Trivia video from the guys and gals at Arrow Electronics.   So, just what was the great capacitor plague of the early 21st century? Was it perhaps a scourge of robotic parasites that grew in computers and infected users via their mice and keyboards?   Well, perhaps not quite than bad, but it was pretty disastrous at the time. For reasons described in the video, humongous numbers of capacitors manufactured between 1999 and 2003 started to fail in systems around 2002 to 2005. The end result involved losses of hundreds of millions of dollars and a lot of very unhappy end-users.   I must admit that I was a tad disheartened after watching this video, but then something happened to brighten my day. Rick sent me another link, this one pointing to the Megaprocessor.com website saying: Well, it ain't mechanical, and it doesn't have vacuum tubes, but this project is still pretty amazing!"   The reason this cheered me up was that I was able to respond: "Ha! I finally got there first!"   Are you aware of any other interesting "component plagues" or weird and wonderful home-grown processors? If so, please take a moment to share them with the rest of us, because I love this sort of stuff.

My chum Rick Curl went to see me in my office about a week ago. We had both decided that we desperately needed some face-to-face time to talk about the current Doctor Who (our feelings are mixed at the moment), and also to try to wrap our brains around the implications of the last episode of the current season, which aired a couple of weeks ago (I'm still reeling from the multi-layered surprises).   Speaking of surprises, Rick presented me with one of his old high-school science projects. (My understanding is that he's having a clear out -- getting this out of his house will make his wife happy, and I think Rick thought that having it in my office would guarantee that it would be lovingly displayed.)     This is a bit of a monster, being about three feet wide. It features an old telephone selector along with a bunch of relays. These relays are a mixed bag -- some switch really quickly to handle the pulses from a telephone dial (which used to be plugged into a connector, but which has become lost in the mists of time), while others have delays to handle the gaps between digits. The whole thing is something of a wonder -- especially for a junior at high school -- I'm not surprised to hear that Rick won first prize for this project.   During his clear out, Rick had also discovered a stack of around 1,000 brochures he'd had printed ages ago. These list the Ten Commandments of Electronics , along with Rick's company's name and contact details. Rick thought I might be interested in these, so he brought a few for me to play with.     These are really rather clever -- in addition to being humorous, they also offer some very useful advice, as follows: Beware thee of the lightning that lurketh in an undischarged capacitor lest it cause thee to be bounced on thy buttocks in a most ungainly manner. Cause thou the switch that supplies large quantities of juice to be opened and thusly tagged so thy days may be long and fruitful upon this earthly vale of tears. Prove unto thyself that all circuits that radiateth and upon which thou worketh are grounded lest they lift thee to high-frequency potential and cause thee to radiateth also. Take care that thou useth the proper method when thou taketh the measure of high-voltage circuits in order that thou doth not incinerate both thee and thy meter; for verily, though thou hath no account number and canst be easily replaced, thy meter doth have such a number and -- as a consequence -- its loss will bringeth much woe unto the supply department. Tarry thou not amongst those who engage in intentional electric shocks, for they are surely nonbelievers and are not long for this world. Take care that thou tampereth not with interlocks and other safety devices, for this will incur the wrath of thy seniors and bringeth the fury of the safety officer down upon thy head and shoulders. Worketh thou not on energized equipment, for -- if thou doeth -- thy buddies will surely be purchasing beers for thy widow and consoling her in ways not generally acceptable to thee. Verily I say unto thee, never service high-voltage equipment alone, for electric cooking is a slothful process and thou might sizzle in thine own fat for hours before thy Maker sees fit to drag thee unto His fold. Trifle thou not with radioactive tubes and substances lest thou commence to glow in the dark like a lightning bug and thy wife be frustrated nightly and have no further use for thee (excepting thy wage). Commitest to thine memory the words of the prophets, which are written in the instruction manuals, which giveth the straight dope, and which guideth thee such that thou will not maketh any mistakes.   The funny thing is that Rick says he no longer remembers where he discovered these commandments or who wrote them in the first place, but when he recently performed a search on the Internet, he discovered that several people had attributed them to him.   Now, looking at the above commandments, are there any additional ones you think we should have? If so, please post them as comments below.

Water flow is often used to illustrate basic electronics in the form of analogies. In this blog, after learning the hard way, I will attempt to explain basic water flow in terms of electronics.   It started one morning with a wet carpet just outside the laundry room and the water heater. The water was warm, and the puddle started at the base of the water heater which had no drip pan. We assumed that the water heater had sprung a leak, which is a typical mode of their behavior.   While I had done some copper pipe soldering many years ago, I felt that installing a new water heater was probably beyond my basic plumbing skills, so we had the heater replaced by professionals. They showed up at 9:00 a.m. and had the job done by noon, as my wife reported while I was at work. My draining the tank ahead of time might have helped.   But then about 3:00 p.m. my wife called back to report that the carpet was still getting soaked even after sucking up the water with the carpet cleaner. And the puddle was hot, so this was not a residual puddle from water still hidden under the wall. Uh-oh -- sounds like a leaky pipe inside the wall.   The heater hot water outlet entered the laundry room wall about six feet above floor level, and I assumed it went straight down from there. I had two choices: I could drain and move the new water heater and cut through the wall behind it, or I could investigate from the other side of the wall (which happened to be the rear inside of a kitchen cabinet). Due to laziness, I chose the latter to start with.   After measuring from the laundry room side of the wall I used a three-inch hole saw to cut out the rear of the kitchen cabinet and through the drywall halfway from floor level, and luckily got centered over the pipe the first time. It was wet. Then I drilled another 3-inch diameter circular cutout two feet higher -- the pipe was dry.   Bingo! I now knew approximately the location of the leak -- about two feet above the concrete floor. It was very fortunate that the leak was not below the concrete slab! After cutting out a larger section of wall and peeling back all the thermal insulation materials, I saw what is in the picture below. So I set up a small funnel and short drainage hose into a large cooking pot to collect the leakage while pondering (no pun intended) what to do:     The sliced-open tin can was planned as a heat shield for my brand-new propane torch to solder the pinhole closed. Yeah, right.   I shut off the cold water intake valve to the water heater, opened all the household hot water faucets, and waited for the above stream to subside. And waited... and waited... and waited.   After many fruitless hours I finally concluded that maybe the shower and kitchen faucets that combined the hot and cold feeds into a single outlet might be allowing the still-pressurized cold feed to leak over into the hot system. Something like this with a "leaky" lower diode, consider the incoming cold water pressure as akin to a positive DC voltage:     My hypothesis was later confirmed by a friendly, long-experienced, retired plumber at the local home improvement center. He referred to it as "cross-feed," similar to what we know as "cross-talk." Ask these guys, they have a lot of good tips for novices. And it did work -- shutting off the outdoor curbside main cold water feed stopped the flow from the pin hole a few minutes later.   But now the situation was a lot more serious -- lack of hot water is an annoyance (cold showers, etc.), but NO water is a magnitude more of a problem. We take for granted how easy it is to flush a toilet, and it's not always mellow yellow. (Note to renovators -- always make sure your new shower/bath/kitchen faucets have two individual hot/cold controls.)   No big deal, I thought, just solder the pinhole closed once the water stops leaking out. Well, we fluxed and tried over and over again. The solder simply would not flow. Having done a lot of electronics soldering, I recognized that even with the torch going full blast the pipe was not getting hot enough. And judging from the steam hissing from the pinhole, there was still standing water in the pipe at the pinhole level, and being at the lowest elevation it had nowhere else to go.   I tried snaking some 1/4-inch icemaker tubing as a siphon down the hot water pipe that by now had been disconnected from the heater (thankfully the new heater had been installed with flexible pipe with threaded attachments instead of the original soldered piping) and sucking on it to remove the standing water (gross!). It seemed to work -- for a few minutes. Then the water level rose back up to the pinhole, I could still see a small trickle leaking out again. In electronic terms, here is what was happening:     Both the hot and cold systems had water stored in the higher-elevation pipes, represented by the capacitors. Discharging at the pinhole let the remaining stored water drain back to the level of the pinhole through the "leaky-diode faucets" and slowly escape. This would take forever to drain out through the tiny pinhole, or to be sucked out through the mouth-operated siphon, which I wasn't looking forward to doing again.   The only way to speed up the discharge process was to bite the bullet and remove the water heater again, cut into the drywall behind it, and with a brand-new, mini-hacksaw made for this purpose cut three inches of the bad section out of the pipe so that a six-inch "repair coupling" could be soldered in. This allowed most of the standing water to discharge quickly, but only down to the level of the lower cut-off pipe. Now the situation looked like this:   The capacitor voltage (water level) cannot discharge below the forward bias voltage of the silicon diode. A residual charge will remain since the electrons have nowhere to go. The standing water in the lower pipe also had nowhere to go, and the pipes still could not be soldered.   To discharge a capacitor, it must have a current path. A negative current source can discharge it. In this water case, a wet/dry shop vac coupled to the cut-off pipe through a duct-taped "impedance matching network" to seal the air flow for maximum suction transfer, along with opened water faucets throughout the house, did the job nicely. With no more standing water I was finally able to solder the new section of pipe in.   Thank goodness for duct tape and shop-vacs!   In retrospect, had I thought of using the wet/dry vac first, I could have simply attached it to the hot water pipe removed from the top of the water heater and applied the suction there. Would have saved me a lot of work, and I would have been able to solder the pinhole through the rear wall of the kitchen cabinet instead of moving the heater and cutting into the drywall.   Next time I will know better. Maybe this experience will help another neophyte with a similar plumbing emergency.   Glen Chenier is an old-timer who grew up with vacuum tubes and therefore has been amazed and fascinated by the many advances in the electronics industry since.

Water flow is usually used to explain basic electronics in the form of analogies. In this blog, after learning the hard way, I will attempt to explain basic water flow in terms of electronics.   It started one morning with a wet carpet just outside the laundry room and the water heater. The water was warm, and the puddle started at the base of the water heater which had no drip pan. We assumed that the water heater had sprung a leak, which is a typical mode of their behavior.   While I had done some copper pipe soldering many years ago, I felt that installing a new water heater was probably beyond my basic plumbing skills, so we had the heater replaced by professionals. They showed up at 9:00 a.m. and had the job done by noon, as my wife reported while I was at work. My draining the tank ahead of time might have helped.   But then about 3:00 p.m. my wife called back to report that the carpet was still getting soaked even after sucking up the water with the carpet cleaner. And the puddle was hot, so this was not a residual puddle from water still hidden under the wall. Uh-oh -- sounds like a leaky pipe inside the wall.   The heater hot water outlet entered the laundry room wall about six feet above floor level, and I assumed it went straight down from there. I had two choices: I could drain and move the new water heater and cut through the wall behind it, or I could investigate from the other side of the wall (which happened to be the rear inside of a kitchen cabinet). Due to laziness, I chose the latter to start with.   After measuring from the laundry room side of the wall I used a three-inch hole saw to cut out the rear of the kitchen cabinet and through the drywall halfway from floor level, and luckily got centered over the pipe the first time. It was wet. Then I drilled another 3-inch diameter circular cutout two feet higher -- the pipe was dry.   Bingo! I now knew approximately the location of the leak -- about two feet above the concrete floor. It was very fortunate that the leak was not below the concrete slab! After cutting out a larger section of wall and peeling back all the thermal insulation materials, I saw what is in the picture below. So I set up a small funnel and short drainage hose into a large cooking pot to collect the leakage while pondering (no pun intended) what to do:     The sliced-open tin can was planned as a heat shield for my brand-new propane torch to solder the pinhole closed. Yeah, right.   I shut off the cold water intake valve to the water heater, opened all the household hot water faucets, and waited for the above stream to subside. And waited... and waited... and waited.   After many fruitless hours I finally concluded that maybe the shower and kitchen faucets that combined the hot and cold feeds into a single outlet might be allowing the still-pressurized cold feed to leak over into the hot system. Something like this with a "leaky" lower diode, consider the incoming cold water pressure as akin to a positive DC voltage:     My hypothesis was later confirmed by a friendly, long-experienced, retired plumber at the local home improvement center. He referred to it as "cross-feed," similar to what we know as "cross-talk." Ask these guys, they have a lot of good tips for novices. And it did work -- shutting off the outdoor curbside main cold water feed stopped the flow from the pin hole a few minutes later.   But now the situation was a lot more serious -- lack of hot water is an annoyance (cold showers, etc.), but NO water is a magnitude more of a problem. We take for granted how easy it is to flush a toilet, and it's not always mellow yellow. (Note to renovators -- always make sure your new shower/bath/kitchen faucets have two individual hot/cold controls.)   No big deal, I thought, just solder the pinhole closed once the water stops leaking out. Well, we fluxed and tried over and over again. The solder simply would not flow. Having done a lot of electronics soldering, I recognized that even with the torch going full blast the pipe was not getting hot enough. And judging from the steam hissing from the pinhole, there was still standing water in the pipe at the pinhole level, and being at the lowest elevation it had nowhere else to go.   I tried snaking some 1/4-inch icemaker tubing as a siphon down the hot water pipe that by now had been disconnected from the heater (thankfully the new heater had been installed with flexible pipe with threaded attachments instead of the original soldered piping) and sucking on it to remove the standing water (gross!). It seemed to work -- for a few minutes. Then the water level rose back up to the pinhole, I could still see a small trickle leaking out again. In electronic terms, here is what was happening:     Both the hot and cold systems had water stored in the higher-elevation pipes, represented by the capacitors. Discharging at the pinhole let the remaining stored water drain back to the level of the pinhole through the "leaky-diode faucets" and slowly escape. This would take forever to drain out through the tiny pinhole, or to be sucked out through the mouth-operated siphon, which I wasn't looking forward to doing again.   The only way to speed up the discharge process was to bite the bullet and remove the water heater again, cut into the drywall behind it, and with a brand-new, mini-hacksaw made for this purpose cut three inches of the bad section out of the pipe so that a six-inch "repair coupling" could be soldered in. This allowed most of the standing water to discharge quickly, but only down to the level of the lower cut-off pipe. Now the situation looked like this:   The capacitor voltage (water level) cannot discharge below the forward bias voltage of the silicon diode. A residual charge will remain since the electrons have nowhere to go. The standing water in the lower pipe also had nowhere to go, and the pipes still could not be soldered.   To discharge a capacitor, it must have a current path. A negative current source can discharge it. In this water case, a wet/dry shop vac coupled to the cut-off pipe through a duct-taped "impedance matching network" to seal the air flow for maximum suction transfer, along with opened water faucets throughout the house, did the job nicely. With no more standing water I was finally able to solder the new section of pipe in.   Thank goodness for duct tape and shop-vacs!   In retrospect, had I thought of using the wet/dry vac first, I could have simply attached it to the hot water pipe removed from the top of the water heater and applied the suction there. Would have saved me a lot of work, and I would have been able to solder the pinhole through the rear wall of the kitchen cabinet instead of moving the heater and cutting into the drywall.   Next time I will know better. Maybe this experience will help another neophyte with a similar plumbing emergency.   Glen Chenier is an old-timer who grew up with vacuum tubes and therefore has been amazed and fascinated by the many advances in the electronics industry since.