Somewhere between the period of 1999 and 2007 a plague swept through the world, devastating lives and businesses. Identified by a scourge of electrolytic capacitors violently exploding or splurging their liquid electrolyte guts all over the PCB, it led to a lot of finger pointing and accusations of stolen electrolyte formulas. In a recent video by [Asianometry] this story is summarized.
The bad electrolyte in the faulty capacitors lacked a suitable depolarizer, which resulted in more gas being produced, ultimately leading to build-up of pressure and the capacitor ultimately failing in a way that could be rather benign if the scored top worked as vent, or violently if not.
Other critical elements in the electrolyte are passivators, to protect the aluminium against the electrolyte’s effects. Although often blamed on a single employee stealing an (incomplete) Rubycon electrolyte formula, the video questions this narrative, as the problem was too widespread.
More likely it coincided with the introduction of low-ESR electrolytic capacitors, along with computers becoming increasingly more power-hungry, and thus stressing the capacitors in a much warmer environment than in the early 1990s. Combine this with the presence of counterfeit capacitors in the market and the truth of what happened to cause the Capacitor Plague probably involves a bit from each column, a narrative that seems to be the general consensus.
“More likely it coincided with the introduction of low-ESR electrolytic capacitors,”
y’know, that really rings true. Most of the culprits I’ve seen WERE the low-ESR types. Most of the other caps on a board are usually fine, and if they’re not they’re usually located right next to a heat-producing component.
Maybe because the caps that were not low-ESR were not stressed that much?
Heh, your post reply made me think back to just when this was happening.
Heat cycles seemed to play a big part in it, and this would have been the first generation of board throwing some serious heat at the (incidentaly low ESR) caps near the CPU heatsinking. Although I did see failures in the presumably audio system electrolytic cap as well. Those normally were in a fairly clear area of the case.
I can’t recall if PSUs saw an unexpected spike in failures. Which you’d expect if the problem was capacitors generally. I don’t remember that being the case and they are spec’d a bit different.
Around that time I worked in a computer shop. PSU failures also were a daily happening, mostly customers explaining in complete shock how much smoke just escaped from their computer.
I must say however that most of those PSUs were of the cheaper type and always developed a strange smell after about a year of usage (probably due to failing caps).
The more expensive power supplies from Antec, Chieftec, or any other more expensive than bargain basement company seemed to last.
I would not expect the low ESR to be incidental. As that will reduce impedance these caps would also have seen the ‘brunt’ of the current draw larger transient swings that come part and parcel with higher TDP’s.
My biggest problem with Caps was in the old Samsung 22″ monitors. We had a ton of these at work at the time and they’re were great 16:10 and people would get two of them. But they would die over time and instead of tossing them, I convinced my boss to give me a couple which I re-capped, then I did a bunch at work as well.
I’d read the corrosion inhibitor was missing in the stolen electrolyte formula, this was the entire problem.
I prefer 1980’s caps – no safety vent and we used to blow them up for fun.
The video here actually tries to evaluate the sources. The conclusion is that this part of the story may have just been repeated hearsay. You don’t have to agree, but it’s interesting to hear it criticized.
Used to use a model train power supply to blow them up. They would go “POP” then uncurl across the room.
Ofcourse in reverse polarity to speed up the process!
My electronic class was a blast… literally. Someone wired a cap to AC power cord, put it behind plexi and plugged it in. Then turned on the breaker for the outlet.
The plastic jacket was embedded into student’s sweater about 20 feet away and the 1/4″ inch plexi had a hole in the middle. They never found the metal housing.
I’ve never in my life heard anyone refer to electrolytic capacitors as E-Caps. It’s jarring and makes the video sound like something an outsider has swotted up on. The report is well researched, but he brings no insider insights.
Elkos. In German, they’re being called Elkos..
We also still use the antique “condenser” term (Kondensator) rather than “capacitor”.
Likewise, we use “coil” (Spule, for spool) or “choke” (Drossel) instead of “inductor”.
I think it’s interesting how different certain things are in different places!
The electric symbols are different sometimes, too.
“Q” often is a mixer transistor/oscillator instead of a plain transistor.
For that, “T” is used. In English schematics, “Q” is just another transistor.
Then, there’s that wire resistor that English schematics use. Very odd in our eyes.
Over here, we use a rectangular box for a generic resistor.
Or let’s take the trace drawing. We don’t use bulges when traces overlap.
We rather draw over them and use dots in case they make connections.
It’s basically same in Czech and Slovak language… tlumivka (dampener/quieter/choke), civka (coil), kondezator. Schematic symbols and namings are also the same as in German too.
in Danish “lytter” short for elektrolytter
In the US we use T for “transformer, and L for “inductor.” We quit doing the humps to show non-connecting lines on a schematic many decades ago, I suppose because draftsmen were taking too much time as schematics got more and more complex as smaller, cheaper parts allowed getting lots more on a board or in a product.
(FWIW, I am not the other “Garth” who has replied in this topic. He’s not using a last name.)
When I was studying for my Radio Amateurs License Many of the old timers used the zig zag line for resistors. But I had been taught the box in electronic class. We had a short discussion on it at the time. They concluded that the symbol had been changed to meet an international standard and they were wrong. They changed to the new symbol going forward to help newcomers.
The zig-zag symbol for resistors is the ANSI standard. The A in “ANSI” stands for “American” though, and perhaps Europe and other areas have differing standards. I think the zig-zag makes the schematic easier to read and find things at a glance.
Same in Dutch: elco, elektrolytische condensator.
I’ve never liked the use of boxes to represent resistors, because boxes can be used to represent anything: transformers, diodes, delay lines, and whole subsystems are common uses.
Exactly!
I’m a big fan of Asianometry and he normally does a great job explaining some very complicated topics, typically around silicon chip fabrication processes. But the insistence on calling them “E-Caps” got my hackles up and made me question the validity of the rest of the video.
He said that he called them e-caps simply because, for some reason, he didn’t like to have to repeat the word “electrolytic.” If English isn’t his native language, he still does an excellent job of it.
Slightly off-topic, but what peeves my grits is musicians who insist on saying “pup” instead of “pickup” or “tranny” for “transmission” which is confined to car fanciers. No one says “I sent a coded tranny to my associate.” And English speakers who say “VW” when “Volkswagen” is faster. Yes I know in German VW is shorter.
HaD owners should really drop daily article quota and introduce “quality” metric instead. If I wanted to watch five too long and totally boring videos created by content creators I’d just go to YouTube, search “electronics hobby” or “computers hobby” and filter by “long, recent videos”.
How about you just skip over the articles you’re not interested in and leave them for those of us who are?
HaD isn’t written to accommodate your specific interest in every article, nor mine.
A world with positive feedback only would quickly turn into chaos.
And a world with negative feedback only would quickly turn to shit. Your point?
Agreed. I watched the whole video and thought it was interesting. I couldn’t care less about the endless string of “Kristina keyboards” posts, and as you suggest, I just skip them.
I haven’t seen a single one of those in the last ~3 years. Thanks to this uBlock rule:
Oh – the humanity. Brutal.
Meh have you seen the slop everywhere lately? This is quality now
Super capacitors had the same problem. I, and many others that own the Davis Vantage Pro 2 weather station have had to replace leaky super caps in the ISS (Integrated Sensor Suite) or outdoor sensor array. The ISS has a solar panel that provides daytime power and charges the super cap. The super cap (10F) provides power for overnight up to a couple of days of low sunlight. There is a lithium battery that kicks in should the super cap run down. When the super cap fails the weather station would continue running until the lithium battery died and then would start only transmitting data when the solar panel received sufficient sunlight then stop transmitting at sundown. Replacing the super cap was fairly easy except if the ISS station was roof or pole mounted and it was winter time. In spite of that minor past inconvenience the Davis VPRo 2 is highly accurate, has shorter data transmission times (2.5 secs) and has the longest transmission distance compared to far cheaper “so called professional” stations found on various discount sites.
I wish I thought ahead of time to save every bad cap from the great plague that I have pulled and replaced. I want to know how big a jar they could fill up by now. It’s certainly been plenty.
Real life needs a stats page full of counters for various trivial things
We had a bad batch of capacitors at work about that time, or a little earlier. It seems like it was late 90’s, but I could be wrong. These however were not exploding or rupturing, only shorting, and it was only one value, one size, one voltage, 220µF 16V 6x11mm. They began going down after only a few years of service, maybe less, in something that went into private aircraft and did not get used many hours per year. After seeing a bunch of these making up nearly all our repairs, I decided that anytime a unit came in, we would just replace all those capacitors, regardless of what it came in for.
(FWIW, I’m not the same Garth who replied above without a last name.)
Garth Bock here.
RoHS is the reason. At least that is what i think.
RoHS introduced new requirements (being “green”), while at the same time prohibited proven, reliable solutions. So designers and manufacturers focus was shifted away from things like reliability, failsafe, fault tolerance, etc. As a result we got additional E-waste because everything failes much earlier.
In the video, at 13’40, [Asianometry] says that electolyte becomes too alcaline, so this means electrolyte is basic, implying that leaks should be neutralized with some acidic solution.
But in a thread on EEVblog someone says it is acidic (https://www.eevblog.com/forum/beginners/what-chemical-to-clean-the-pcb-after-leaking-capacitors/msg1207699/#msg1207699) and that one should use a basic solution to neutralize leaks (but a post later on in the same thread recommand to use white vinegar…).
So what is the truth?
Beside neutralizing leaked electolyte, what is the best way to remove corrosion it left?
Bodge wire, generally. Replace the copper if it’s bad enough. Although I imagine if it’s more than surface on a many layer board it’s probably stuffed.
This seems like something a strip of PH test paper could solve. But I think what you actually care about is not neutralizing the spill, but removing it from the PCB entirely. If it dissolves in an acid, great! Do that and then wash the excess acid away with water.
Ah, the memories… I retired my old AMD K6-2 around this time, then bought a new PC with money I earned replacing motherboard capacitors. :P
i am lucky, i’ve only once had capacitor failure. i mean, i assume there are some dead devices i’ve run into with bad capacitors but only one mattered enough to look into it. around 2009, my 2003 athlon became unreliable. i looked at the motherboard and the capacitors in its voltage regulator were swollen. so the fun part, i simply underclocked it and it was steady for a year or so until i decided to buy another pc
also reminds me of my intel core 2 duo i replaced it with, eventually the fan died and again i simply underclocked it and ran it without a fan for a while
performance sure places a lot of demands on hardware