Was the early 2000s capacitor plague corporate espionage, or just industry woes?

Video takes a wider look at the busted caps of early 2000s PCs and other gear.

See full article...

 

[mrkite77]

One of the very first capacitor fixes I ever made at work was on a broken ufo-style Airport Base Station... maybe around 2005?

 

[first_citazen]

Thanks, I had this queued up in my Youtube feed for a while, now I'm watching it

 

[Glorified Desktop Support]

Being glorified desktop suppport, I certainly remember that. Had a bunch of HP computers and eventually just pre-emptively declared the problem models as obsolete and replaced them.

 

[Erbium68]

Back in the day before SMPS became the norm and then got miniaturised, large electrolytics were common for power supply smoothing - by large I mean hundreds of microfarads at, say, 400V, or 68000 microfarads at 15V (nobody ever called them 68 millifarads).
I once met the son of an R&D engineer who had been killed by an exploding 68000 microfarad capacitor on test. There was supposed to be a pressure vent in the design but it was concluded it had failed. The capacitors hold very little actual energy - only about 5J as it happens - but when they start to leak due to failure of the separator, heat can build up very fast and the electrolyte boils.
My point such as it is, is that electrolyte isn't that secret, though manufacturing is more likely to have trade secrets. Whatever the cause, the temperature increase as more power went into motherboards seems the most likely explanation. Thermal management in that era wasn't that good.

 

[voojtek]

I replaced hundreds of Dell motherboards with bulging capacitors when I worked as a tech in a school district in the early 2000s. I had janitors freaking out over the little syringes of thermal paste I'd toss after swapping out all the motherboards in a computer lab.

 

[FSM4ever]

I remember our brand new academic Linux cluster made up of 32 beige-box Pentium IIIs stacked on metal shelves started all of the sudden dying circa 2003. Once I opened one of the boxes and saw the snow heads on top of bursted capacitors I remembered some news articles I recently had read. Our supplier, a corner computer shop, had to come over, remove all motherboards which they had to send back to Taiwan to replace the capacitors. Made my PhD longer by a few months…

 

[William Hern]

My parents' G5 iMac suffered from the problem in the autumn of 2007 - even though it was out of warranty, Apple replaced the motherboard for free.

 

[Resolute]

This was the first major disaster I had to deal with in my career. My employer of the time had just finished replacing its ENTIRE fleet of computers and dumb terminals with IBM PCs that all failed. First in a trickle, and then in a flood.

At first, IBM demanded we follow their RMA process for each individual machine. We squawked loudly enough that they finally started just sending us cases of replacement motherboards so we could swap on site and then return the defective ones. And half of those were no good because IBM was rushing replacements so fast that QA went out the window - half the boards we got had a severed lead on a capacitor sitting at the edge of the board.

I logged so many kilometres across western Canada as I went store to store to store and swapped the board on every computer they had. To this day, I still can't stand A&W breakfast sandwiches as I ate so god damn many of them during that time.

 

[AstroPhoto]

Slightly off topic, but it’s cool to see Jon’s video getting some media attention. I’ve been a fan of his channel for a while, so I’m glad he’s getting the exposure.

 

[sword_9mm]

I replaced so many Dell Optiplex machines during those days. Worked at a University and those things were everywhere.

Good times? I actually never thought to look up what was happening or why. Just looked at the board, yep, and swapped out.

 

[egosumliber]

Ah! I had a small flat screen TV from this era, that failed due to burst capacitors. I was able to buy some replacements from Radioshack and get it working again... but never knew why they would have burst in the first place.

 

[Tom Terrific]

Wishing that Jon Y, the force behind Asianometry would have a regular column here on Ars. Really good deep dives into technology, especially semiconductors.

 

[Jim Salter]

Being glorified desktop suppport, I certainly remember that. Had a bunch of HP computers and eventually just pre-emptively declared the problem models as obsolete and replaced them.

I sold a ton of computers in that time frame, and boy howdy I got (inaccurately) blamed for "using cheap parts" frequently during the capacitor plague era... Even as high end Dells failed right alongside my own builds.

 

[motytrah]

Slightly off topic, but it’s cool to see Jon’s video getting some media attention. I’ve been a fan of his channel for a while, so I’m glad he’s getting the exposure.

His videos on semiconductor FABs (of which he has many) are great. Really good at breaking down the topic into east to understand segments without needing a physics degree to understand.

 

[Juventas]

Desktops around this time had rapidly growing power consumption, were terribly inefficient, and had poor cooling. In other words, they were like ovens. I'm certain this was a significant contributing factor of the early failures.

 

[plectrum]

I don't entirely buy the increasing power demand theory, because a number of failures were in audio equipment that weren't suffering from a Moore's law arms race. As the video suggests, likely more than one factor was at work.

 

[Lavonheim]

I sold a ton of computers in that time frame, and boy howdy I got (inaccurately) blamed for "using cheap parts" frequently during the capacitor plague era... Even as high end Dells failed right alongside my own builds.

I'd like to think the capacitor plague (which I only learned about years later) was the problem with my home-built computers before I had my first post-college job, but the difference between them and those I built after was almost certainly cheap parts. Before, they'd last maybe three years (granted with how fast CPU speeds were increasing there, not a big loss). After.... I'm typing this on the computer I built with my savings from that first job. Still running fine (more or less) after 14 years.

 

[Nekko]

His videos on semiconductor FABs (of which he has many) are great. Really good at breaking down the topic into east to understand segments without needing a physics degree to understand.

I have the degrees and worked in an Intel fab for several years and can confirm that he gets it right almost all the time. Nobody says it better than him when he states "all this stuff just barely works". I subscribe to his Patreon.

 

[Erbium68]

I don't entirely buy the increasing power demand theory, because a number of failures were in audio equipment that weren't suffering from a Moore's law arms race. As the video suggests, likely more than one factor was at work.

You may well be right but local heating can be a problem. We had one unit in an ultrasonic test set that kept failing with a blown capacitor. Eventually we decided to do a more in-depth analysis. It turned out a resistor adjacent to the capacitor was getting very hot indeed. We didn't even bother with a return to manufacturer - just replaced the quarter watt resistor with a one watt one stood a little way off the board, no more trouble.

 

[Hard Thoughts]

Gonna add to the voices here that the channel is absolutely amazing. Totally worth a few hours, both for his tech dives and history lessons.

 

[J.C. Helios]

Actually not one of his stronger videos, because usually he leaves you feeling like you have a satisfactory understanding of the topic at hand, whereas this is more of a debunking of a popular-but-perhaps-simplisitic narrative.

He clearly found it hard to research. If people had gone to jail for stealing trade secrets or selling bad capacitors then there would be more of a paper trail, but instead it seems like everything got swept under the rug?

 

[Jim Salter]

Desktops around this time had rapidly growing power consumption, were terribly inefficient, and had poor cooling. In other words, they were like ovens. I'm certain this was a significant contributing factor of the early failures.

I built systems with 120mm chassis fans and high end power supplies and CPU coolers, exclusively. They ran typically 10-20 degrees cooler ambient than dell workstations, let alone optiplexes.

My systems still failed due to bulging caps at about the same rate as the Dells.

 

[angleiron]

Wishing that Jon Y, the force behind Asianometry would have a regular column here on Ars. Really good deep dives into technology, especially semiconductors.

He did just join the Stratechery bundle, so maybe he's spoken for, but if not, I totally agree. I've learned so much from him.

 

[J.C. Helios]

Just looked at the two motherboards I have here, and it's all solid-state capacitors. I kept expecting him to say that the industry started using more solid-state capacitors, and that was part of the end of the plague. But he didn't bring it up, and I don't know if that's actually true.

 

[adespoton]

Thing is, while there are definitely some issues with those caps, electrolytic caps spilling their guts is nothing new. Macs from 1992-1994 are famous for leaking over everything. There's an entire re-cap industry that has grown up around this, that was in place before the wave of problems started in 2002. This muddies the waters significantly, as it's hard to know if a failure was due to THIS bad cap design or just bad electrolyitic caps in general (design or manufacturing). There's multiple potential points of failure in the "good" designs, and the MTTF is all over the place, even for the ones highlighted in the myth.

And yet I've still got 1980s electrolytic caps that haven't leaked/dried and are working just fine, as well as ones from 2004-ish. It's just a bit of a mess.

 

[Thegs]

What a strange coincidence, Youtube recommended me this same video earlier this week.

I may have taken a risk on the recommendation algorithm because the creator uses a deer avatar, and you can always trust people with cervid avatars.

 

[mekantti]

Veering off into a slightly different topic, it would be great to have some deep dive of the soft touch plastics plague that seemed to quickly follow capacitor plague. Who came up with the "too sticky and gross to use any more" idea for the planned obsolescence of older computer peripherals (as well as numerous other kinds of equipment)?

 

[Iphtashu Fitz]

Being glorified desktop suppport, I certainly remember that. Had a bunch of HP computers and eventually just pre-emptively declared the problem models as obsolete and replaced them.

I wish it was that easy for me. I replaced literally hundreds of Dell motherboards due to this one. Well, I didn't replace them myself. I managed a team of Dell employees that were brought on site to do the grunt work.

 

[jforbes]

Being glorified desktop suppport, I certainly remember that. Had a bunch of HP computers and eventually just pre-emptively declared the problem models as obsolete and replaced them.

Worked for a school district late 2000's and we had some awful mini HPs in a couple of the buildings. They really didn't like being unplugged for long, which went well with the classrooms being emptied into the halls for cleaning and re-waxing over the summer. Was probably 1 in 4 that popped the power supply when we got to them right before the new school year started. I learned to plug them into the strip, then the strip into the wall so I was well back when they started sparking and smoking. HP eventually just shipped us some absurd number of replacement parts, which lasted until the life cycle was up. Horrible machines too, was almost impossible to do the replacement without damaging the case.

 

[Meathim]

A couple of months ago I replaced some caps on an old Asus socket A motherboard (A7V iirc. Still didn't boot, sadly, so further troubleshooting is necessary.

 

[electric_mayhem]

I never had issues with motherboards failing back in the 2000s on the gaming PCs that I built. But I recall that I did have regular issues with the power-supplies frying capacitors. You would be chugging along fine and then you hear the loud POP, the whole PC immediately dies, and then the lovely smell of fried electronics wafts out of the tower. Ah, those were the days...

 

[Fatesrider]

Boards and computers bought from Abit, HP, IBM, and, infamously, Dell, among others, suffered these faulty capacitors and were handled with recalls, repairs, or, sometimes, silence.

This is one of those things that happened back while I was running an on-site computer service business, but oddly, (or perhaps not so oddly), I didn't encounter it very much. Most of my clients had bespoke computers and didn't have the kind of warranties that would replace a whole system if part of it damaged other parts.

Since this was MOSTLY an issue in bulk manufacturers of PC's, which didn't generally provide boards for the hobbyist/up-scale boutique computer market, it kind of flew under my radar. In reading the story, I recall the reports in the news, but didn't encounter it much because most of those computers weren't off the shelf.

I can think of a few cases where I was called to fix a computer, and asked what the model was, and when they got it. If it was one of the troubled models, I always recommended a bespoke build to replace it, which allowed more control over the quality and sources of the components, and eliminated the headaches of crapware that at least three of the four companies above were infamous for adding to their offerings (I don't recall dealing with any Abit computers back then). Fortunately, in most cases, the data from the old hard drive was usually fully recoverable.

Nice to find out why those bad computers decided to self-fry, and a very interesting backstory on that, though.

 

[Mad Klingon]

I remember those days. Had a personal MB get the problem. Got it replaced after one of those class action lawsuits that normally result in a $5 coupon, but in this case got me a referb MB of a newer style. Worked for an agency with a building full of Dells. Dell finally sent us a reserve supply of MB and PS and as we replaced dead ones, the reserve was made good when we reported the Asset Tag #'s repaired. Good deal on both sides as it saved Dell hundreds of service trips under warranty and we had the dying PCs back in service in less then an hour.

Another gizmo that went bad was a model of 3Com 10/100 eight port switch. You could tell one near EOL because it would start a soft hissing noise. They were several years old when the hissing started so not a quick fail.

 

[cre8tiv369]

I was doing I.T. at a university and I replaced well over 200 failed motherboard from Dell Optiplex PCs… The GX270s all failed around the same time. I had sporadic issues with other models of Dell PCs and a few servers, but nothing from Apple nor any other brand. I clocked a lot of overtime with my anti static mat and grounding strap… That was such a PITA… While it only took 15 minutes to do a swap, there was a lot of time burned traveling to all the different places on campus where I had to swap MBs. Of course it had to be done on a priority basis, so there was a lot of back and forth and late nights to pacify egos. I never want to see another POS Dell for the rest of my life.

 

[MrTom]

I was working at a computer repair shop around that time. We had something called "eMachines special". It was a new motherboard and power supply. Because when the power supply died from bad caps, it took out the motherboard too.

 

[kre]

Asianometey is a channel many Ars readers would enjoy. Highly recommend.

 

[einnocent]

It’s great to see Asianometry get some love on Ars. He covers a lot of interesting and esoteric topics with a solid amount of research and humility.

 

[arvedui]

I remember around 2004 the Gateway computer my parents bought us for Christmas in 2000 started to have intermittent issues booting. We couldn't figure it out and my dad got pretty pissed and blamed me for it because of "all my downloading on the Limewires". I ended up opening the computer and sure enough 2 of the capacitors had leaked all over the motherboard. Unfortunately I lacked the skills at the time to replace them and we ended up getting a crappy new Dell Precision desktop.