"From 2001 to 2004, Nichicon produced defective capacitors ("HM" and "HN" series) that were used by major computer manufacturers. Some sources claimed that these capacitors were either overfilled with electrolyte or were constructed using electrolyte that was prone to leaking, causing premature failure in any equipment using them. This issue was not related to the Taiwanese capacitor plague." -- Wikipedia How convenient! The capacitors on my ASUS P4B are in fact from the "HM" series. Maybe they just bought the incorrectly formulated electrolyte to save cost? Then it absolutely would be related to this dark time of PC hardware!
Interesting, I bought new Nichicon caps from Radio Shack around this time and I found them to be leaky. I don't know which series they were as I long ago disposed of them.
The industrial espionage theory is one thing, the other is the ROHS theory which is that new regulation to cut down on toxic materials is what made capacitors so bad, it's about the same time that lead-free solder was starting to get used and it provided similar headaches.
Some of those failed capacitors were from rip off manufacturers who also marked their capacitors with known good manufacturers. Many of these rip offs make fake Nichicon capacitors.
2010 through 2012, I was a bench tech for a computer recycling company - I lived through "capacitor plague"! "Bad caps" was our number one board rejection report.
The test circuit you put together to demonstrate the behavior of the bad capacitors was brilliant! It was really interesting to see and it becomes very clear why they'd prevent a motherboard from working correctly. I've tried to test bad capacitors before with component testers and multimeters, but they never seem to look all that bad from a numbers perspective. Sticking them in a test circuit like that tells a very different story indeed. Keep up the great work 🙂👍
Thank you! And yes, you're right. I was not expecting to see such a high capacitance on those badly damaged caps. First I was like, "oh, they are better now!" 😂
The Vloss value is defined in a very simple way. Charge a capacitor up to a given voltage, then disconnect the charging source and immediately measure the voltage. The relative difference of these two voltages (in percent) is the Vloss value. PS, several models of transistor testers with the newest Markus' firmware (1.49 and higher) can measure the leakage current directly. Also you can measure it by charging the capacitor continuously with its rated voltage thru a multimeter in mA or μA mode.
Basic capacitor testers calculate time to reach set voltage. So yes, if it leaks away it will cause that time to extend and the reading will be inaccurately high. Your cup analogy was spot on. Really an LCR meter is needed. You've done a good job with what you had.
Looking forward to the repair! :-) Apart from Nichicon, another reputable brand that suffered bad batches was Nippon Chemi-Con. It was around 2004-2005, I don't remember which series was affected.
There is a good explanation for the higher capacitance values for the failing capacitors. You need to look at the 1st year uni physics formula for capacitance. Due to the leaking electrolyte and build up of gas inside the two plates are pushed closer together increasing capacitance but this is also means the ESR climbs until the current jumps across the gap.
I've probably said this before here and definitely have said this in other repair channels. So forgive me if I'm repeating myself. For removing those two leaded components in a difficult ground plane like those failed capacitors you have I've had really good luck using one of those Weller type soldering guns. I'm talking about the kind that has a copper wire loop as the soldering tip and uses a low voltage high current transformer to make the tip heat up. It gets hot enough quickly enough to melt the solder without trying to heat up the entire board, plus with some extra solder "blobbed" onto the pins you can heat both pins at the same time. Basically heat up the pins while gently pulling the part out. Cleaning out the solder filled holes can be done with a regular soldering pencil and a stainless steel pin. Heat up the hole with a soldering iron and push the needle through while the solder is liquid. I've used a pin vice and a drill bit to even drill out the solder holes (you have to be drilling centered in the hole).
that was a nice testing rig! You're luckier than me with removing the caps with wick and no pre-heating... Not sure how you can do that! On my "Livincontrol" repair video I had hot air + solder iron and I wasn't even able to REMOVE the capacitors!
It may depend on the quality of the board. I must say, this P4B feels premium. Maybe the vias have a lot of material where heat can travel and melt through the board. I could not free the holes of the vias that connect to the copper plane though.
Sometimes parts are hard to take off boards. Some of the techniques used are somewhat counterintuitive too. Such as adding solder to remove parts. It doesn't make much sense on the face of it but it's what you have to do. Sometimes you have to add, remove and add more solder. It's a two steps forward one step back process.
Your tester must be using a charge-curve instead of discharge curve for the capacity and thus will be fooled by high self discharge/leakage. It should do both. Measure on both halves of the cycle to identify leakage correctly.
I have the same tester. You have to calibrate it every once in a while jumping 1-2.3 together and then you need a 15pf cap. Those good 3300uf caps should be near .05 ohms.
I had a 2002 ePox mobo for the Athlon XP in which every single capacitor turned like that. As a result it would only start up every third or fourth time, and make awful screeching noises when loaded with some heavy applications. So I thought, hey, I might try to replace them, good for a project to learn how to solder on proper boards... and ended up massively failing, taking out a few caps with their vias and putting in none, and the rest just left to stay. Funny enough, this didn't fix the screeching noise, but the computer would now boot up every time rock solid :D
As the ESR increases so does its power dissipation which heats up the caps. A cascade effect happens as they internally heat up, boiling the electrolyte, and rupturing the caps. Another thing to be aware of, electrolyte is caustic. When the bottom seals fail, the electrolyte can damage adjacent components as well as the circuit board itself. It can be very difficult to properly remove the residue from the boards, and must be done, when attempting quality repairs. Great content .. Cheers :)
I remember doing a lot of work replacing these low ESR caps as a result of this back in the day, and remember this episode only too well, If memory serves it was posted on 'The Register' the story was hilarious
Sadly this capacitor plague was the nail in the coffin for A-Bit. Luckily I have a BE6-II which predates 2001. Still runs great to this day. I was working for Dell in 2006, where half of our calls were motherboard replacements for Optiplex GX270, SX270 and GX280. A lot of people screamed at Dell for the issue, but I took the time to educate owners that this was a known across all motherboard manufacturers and System Integrators.
i almost got BE6-II for 8$, but missed an auction. Someone was were happy, cause it was inside whole ATX celeron build. All cost 8$ plus shipping, functional. Not many people noticed, cause on photo, only part motherboard was seen, but I've identified it according BIOS string. Damn, I've missed it by 5 minutes. Could have BE6-II. I have only BE6 , but that board is earlier one, and not so good. It has some issues. Still good board. Just it's not recommended to use ATA-66 IDE connectors. It also hinders overclocking potential. Always wanted to have BE6-II rather.
Considering I have seen the capacitor plague strike all sorts of equipment, even stuff like dash clusters around the 90s with the dodgy brand capacitors. I would say the plague and hunt for saving cents really brought on a lot of failures back in the day.
I should say it's often not that. In the 90s the electronics manufacturers simply believed cap manufacturer promises on unproven high density products which allowed them to engineer slicker and smaller devices. But the stabilising magic ingredients turned out to not be quite magical enough and ended up degrading. So it wasn't even cost saving but trying to make progress happen when it didn't want to be happening yet.
Nichicon is not a dodgy brand. But Nichicon has made some dodgy models. They have brought shame on their family for generations. Rubycons are the best.
During those times, we replaced some capacitors with higher uF and voltages, if same value got out of stock from our supplies, all worked fine as long as it for high frequency and low esr types. We do the same on the power supply, some crt monitors, and other electronic devices that may be using same type of capacitors
I remember having at least one motherboard from that time where the capacitors started bulging and leaking. I think it smelled a bit too. Honestly, I never even thought about replacing the capacitors back then, even though I knew enough about soldering. I'm sure many tons of hardware got tossed around the world due to this problem. What a shame.
I recapped a big screen TV. It was worth doing. It's still running today. A lot of stuff isn't worth recapping depending on how many capacitors you have to buy and replace. A motherboard cap kit can cost half the price of the motherboard and there's no guarantee just capping it will fix the board. As when the caps went they might have took other parts with them.
@@1pcfred Motherboards with failed caps, mostly booted fine (or which noobs tried to fix). But crashed when Windows loaded or bit later. I got P3 500 with bad caps, when I bought about ten new ones , clerk at the store asked if those were for motherboard :) Most hobbyist don't need ten electrolytics and nothing else.
@@Tegelane5 failed caps lose the attribute they need to do their jobs. They're on a board for a reason and when they fail they're just not doing that job anymore. They become capacitors with no capacity. That's bad. It is especially bad at times of extreme load. When the other components need those caps to pick up the slack they let everything down.
you can use a domestic hairdryer to thoroughly pre-warm a board and it can make a big difference when you need just a little more heat from a small iron to desolder lead-free components on large pads
I recapped so many boards and graphic cards. Mostly our business systems from Fujitsu-Siemens. They didn´t had Nichicons but OST as manufacturer. Also recapped my mother in laws socketA Gigabyte board. New caps just 5€, new Board 60€... Was a good deal and she was happy.
@@1pcfredjust note down the values and buy from any electronics supplier instead of buying a kit, digikey. Mouser. Farnell or the likes all got em cheap instead of costly premade kits
Haha. Yep. I remember this. We had hundreds of systems sold on contract to an engineering firm. We were replacing caps at a rate of about 4 boards a week for over a year.
I remember when I bought in early 2000, my dream computer with Celeron 466 and Voodoo 3 then I wanted a high-end motherboard. I chose an Abit BE6-II, which cost nightmarish money at the time. Legendary 440BX, SoftMenu III, UDMA66 and great performance. After only a year I noticed a problem with the capacitors. Almost with all of the power section.... They were replaced under warranty, but soon after that I sold the board.
@@simon_1987 Maybe at your place. In our country (Poland) it was very popular because of its good performance and price. As a reminder Price at introduction $169, while Pentium III 450 - $348
I have a whole batch of pc's with AMD Thunderbird 900mhz processors that have these issues on the boards. When it gets bad enough they stop posting. I have yet to replace the caps to see if they start working again. Good to know that you can kind of replace then without a hot air rework station since I don't own one of those yet.
You definitely can. Can even use a cheap DIY style heat gun very carefully to pre-heat a board to make soldering iron more effective, obviously with caution :)
As the capacitor get's internally "damaged" over time, it's capacitance can increase while the voltage rating decreases, that causes them to bulge or burst because they can't handle the circuit voltage anymore.
Yes, the electrolyte missed the protective compounds that prevents internal damage to the aluminum housing. What I've read is that this chemical reaction creates gas that causes the capacitors to swell up.
the best test for a capacitor is in an LC oscillator circuit. The frequency is a function for the capacitance and the Q-factor a function of current leakage.
All electrolytic capacitors fail eventually. Plague caps just failed sooner than they should have. Electrolytic caps are just bad by their very nature. Every electrolytic cap has a rated lifespan though. Which ranges from 2,000 to 70,000 hours. But that rating is at max rated temperature. So it's not like a 2,000 hour cap fails instantly when it is 2,000 hours old. Every degree below max temp nets you more lifespan. It's all somewhat complicated.
The event that precipitated the plague did not occur until 2001. All electrolytic capacitors fail regardless of if they have bad electrolyte in them, or not. So you cannot attribute every failure to the plague. Every electrolytic capacitor ever made will fail someday. That is the nature of the component.
I'm wondering if that increased resistance is causing the capacitors to warm up more. Might explain how the fluid eventually boils over and pops. Without enough fluid in there, it can't really hold a charge very well as a result. If it turned on, I could see all the current CPUs tend to pull heating up the capacitors more, and without their abilitiy to smooth out the waveform, possibly causing the CPU to hang. That was still a very neat demonstration how a damaged capacitor behaves, usually people look for the bulge and decide it's time to go. Some others like to watch them blow up fully.
Hi, also please take of one cap that looks good from the board and test it. Maybe we can find some interesting values. Great video, keep up the good work!
There are two other interesting types of capacitors on the board. When I install the replacements, I will have a look. There are 1000uF 6.3V (Sanyo) and 1200uF 16V (Nichicon, HD series) caps. The HD series was not affected by whatever was the reason of the 3300uF capacitors (HM series).
One question.... I tried a lot of different ways for removing old capacitors. I always add new solder (63/37) and flux, just like you did. I tried sucking the solder with manual pumps, with Hakko gun, wiggling like you did... that's the one I'm using at the moment: I leave a blob of solder and wiggle the capacitors always. I used the Hakkon glun to finish the job and keep the hole through clean for the next capacitor. One thing that happens sometimes, specially on the negative/grounded side: when I try to solder new capacitors, the solder doesn't adhere to the pads. It is not a matter of temperature, quality of solder or flux - I guess something happened to the pads. I normally get it fixed by bridging it to a nearby negative or ground ( I always check with a multimeter to ensure they are connected by design). I still have to find out why that happens (a mistake from my side) and learn how to avoid it - I'm sure one day there will be no easy way out, like no near place to jumper it. I never saw anyone reporting such issue, I wonder how common that is and how people deal with that
I haven't encountered such an issue yet. Maybe it is linked to a large ground copper area. It has quite the effect on soldering. But it could also be a corroded pad. Cleaning the pad with fresh solder, wick and flux may help to make the solder attach better
A year ago I had a case with a Samsung 17" LCD monitor not turning on (only slightly blinking with its power LED). The problem was in six 820 μF low-ESR capacitors, all six leaked and popped. Five of them were completely open (not recognized by the transistor tester at all!!!), one had approx. 100 μF, high ESR and a lot of leakage. Replacing all six fixed the issue without any additional repair.
Ive got one of these P4Bs - with a celeron 2ghz - it was a scrap find - initially it was dead as a dodo - i replaced the main power caps first - and heard it say someting to me - i thought what the hell was that ? then it went dead again . I put it away for another day - then I ended recapping the rest of the board with solid caps (the green caps actually didnt look bad ) but she came alive - and now is exited to say "system is posting into the operating system". Its also a very handy board as a test rig for graphics cards - as she will say "vga card not detected" during a post
De-soldering motherboard components can be a real headache. Factor in the lead free solder many motherboard manufactures use & that makes the process even more tedious. Preheating the board & adding some 60/40 solder to the pads helps to release & then re-solder the caps. Even then it is still nerve racking when you have to use such a blunt & very hot soldering iron tip so close to the often very tiny traces on the back of computer motherboards.
I think I faced some issue measuring the resistance. If I recall correctly, my multimeter was constantly jumping between different units (Ohm, kOhm, mOhm). Unfortunately, I no longer have the capacitors to measure them.
I have the same board, and mine worked mostly fine with faulty capacitors, but was not very stable, especially if you put a GPU in the AGP slot and load intensive games. Swapping them out fixed my stability issues.
I wonder if you are overloading the VRM with those caps. The caps I tested leaked a lot of current. I could imagine that the power regulators get extremely hot because of the caps. Mine were too far gone, otherwise, I would have measured power draw at the outlet as well as temperatures on the regulators.
2x of my OG Xbox have also suffered from Nichicon's own capacitor plague. Surprisingly they both still work despite some having split open completely. And 1 of them I only opened due to it not staying turned off (that was the super capacitor having leaked and shorted the power button). It was the same for a Pentium 4 540J PC that I'd been given as well. It however had a replacement PSU at some point (I know because 2x SATA HDD, yet only 1x SATA connector) and said PSU had a scorch mark behind a bad solder joint. So I didn't even bother attempting to repair the motherboard (reference Intel LGA775).
I picked up an Asus P4B266-C that had some bad capacitors, and I replaced them with some that I had on hand. It works perfectly with the 1.7 GHz P4 that it came with, but when I tried to use a faster CPU I ran into an issue I've never before encountered. Even with a 2.2 GHz 400 FSB CPU (the fastest one officially supported, but I also tried a 2.5 GHz 533 FSB) the board will POST, but lock up soon after. The faster the CPU is, the more quickly it locks up, but it works perfectly with the 1.7 chip that it came with. I assumed that some component on the board is overheating, but I bought an infrared thermometer to try and troubleshoot it, but it was no help. I updated the BIOS, tried different memory, different video cards, but I wasn't able to resolve the issue, so it sits in storage.
Thanks for this fun video. Yeah anything from this era is junk and should be replaced. Funny how people say Nichicon caps are good when they are not. Panasonic ones seem to be the best, but in things like the Turbo Express handheld all the Panasonic SMD caps go bad. Those were from the early 90s though. Funny how caps from the 80s were so much more durable. A lot of them where made in the USA back then though. I would love to see you work on a Coleco Adam computer some day.
recapped many boards during that tme, the bad brands where mostly: G-Luxon, OST, GSC, Lelon, Fuhjyyu (inside Antec power supply's) Nichicon HM/HN, Rubycon MFZ, Chhsi.
If the caps look undamaged, I like to keep my fingers on them when I turn on boards from this era for the first time. If they get hot quickly, then I know they are leaky and will pop soon.
I am no expert to answer the pros and cons or if it is even possible. However, the maximum capacitance of ceramic capacitors are a lot lower than electrolytic capacitors. I think the largest I have seen is about 100uF - the P4B capacitors I replace are 3300uF.
I have some motherboards from that time that died prematurelly due to this plague.... brands like Epox or Abit were more likely to suffer from that. I don't recall having a board from Gigabyte or Asus with this problem, I guess I was luck. One thing though: you are showing Nichicon capacitors going bad on this board. Nichicon is a japanese capacitor OEM and one of the top ones, just like Rubycon, Panasonic to name a few. I don't recall japanese OEMs having problems like this. I normally found the "sick" capacitors to be chinese brands... I don't recall the brand, just the color: they were green normally. You should replace ALL OF THEM, not just the bulged ones. Even the ones that look normal can be very off their nominal values. Great topic
Nichicon also had issues that were apparently not related to the capacitor plague. See my pinned comment. This ASUS motherboard has the capacitor series which is known to be faulty.
There was a particularly popular Asus mATX nforce 2 board that suffered from it, me and my friends group bought them and watched them drop flies over the next few years.
@@bitsundboltsAsus motherboard are in general good in resisting electrolytes to buldge. I've noticed, it happens far less than on other motherboards. The worst is PC Chips and ECS, of course. (also other ripoffs like Syntax USA , Matsonic) Gigabyte is somehow in middle, but MSI is usally worse, than Gigabyte. 3 boards from MSI in my collection, started to buldged capacitors only stored in box (i've rechecked them after 2 years). Intel motherboards have also in general good resistance to this problem.
My Dell computer at work years ago was buggy and unstable. I popped off the side panel to clean out the dust, and saw several bulging leaking capacitors. Opened a ticket with IT, they didn't care, refused to replace it.
could you demonstrate the difference between a big and a small capacitor both with same C and V specs? i need a 470uF 25V 10x20 for a power supply but all local electronics stores don't have this size and instead give me whatever they have. i thought capacitor size didn't matter but now the power supply won't turn on with this 8x16 capacitor. i kept looking for caps but all i found was 8x16 and another smaller one, both with the same uF and V specs.
I have seen capacitors with the same rating, but different sizes. I wonder why they exist. I'd imagine something about density etc, but I don't know. Unfortunately, I don't have capacitors with the same parameters, but different sizes. Could there be another reason why the power supply doesn't turn on? Be careful with high voltage mains power supplies! I opened once a power supply. The PCB, solder work, everything looked awful. Just poor quality and I decided it's not worth the trouble!
@@bitsundbolts its from a 2007 samsung monitor. the capacitor has those specs i typed above. maybe if i solder the capacitors in series it might work? i'll google that later.
I added a link to the lights I got in the video description for reference (it is an affiliate link): But you could also search for 6W 6500K Industry Microscope LED Gooseneck Light. I got lights that I can attach to the microscope arm via a clamp.
Beautiful explanation! I have SO MANY boards and other electronics with bad caps, I need to find a local supplier of good branded caps to avoid buying chunks of them online. What are the best brands? (found in a forum) Chemicon, Nichicon, Panasonic, Samxon, Sanyo and Rubycon? And how do I check if I need ESR caps or not?
Japanese manufactured caps are generally considered decent. Although every brand has released bad series of capacitors. Like these Nichicons in this video. Nichicon is a top rated maker. But this particular series they made was bad. So you do have to research each series you consider using. You need to use low ESR caps in higher frequency applications. It has to do with internal heating. Heat is what kills caps. Electrolyte is a fluid and heat dries the stuff up.
@@bitsundbolts even the poorest capacitors should run for a while. Although I have seen new duds. Those Aliexpress variety pack caps. Which I imagine are mostly rejected batches.
there's no best brand but only good, okay, and bad. among brands you mentioned, Samxon can be okay or bad depending which model you use. try to avoid their lower end model. the others are good but again, it's a matter of model choice. it's always better to determine which model is the most suitable for your specific application. do note, it's best to avoid placing solid polymer capacitor on standard electrolytic capacitor place. not every circuit would appreciate very low ESR caps. for solid caps, Taiwanese brand is good to use (CapXon, APAC, what else?)
@n.shiina8798 I wish I would know more about electronics to avoid mistakes like replacing standard electrolytes with polymer may not be a good choice. Well, I did order solid polymer capacitors. We shall see if those will work on that board. I thought polymer caps are superior in every aspect. That's a case where "too good" becomes bad?
This was a really cool video!! Would you like to share the schematics of that test curcuit? Would like to replicate that myself, and need more than whats shown in the video :P
I have prepared a schematic and uploaded it to my Patreon account - available to everyone. There is also a link in the video description above the chapters. Let me know if you manage to recreate the circuit with the diagram!
It's crazy but it wouldn't surprise me if the board would still POST with the bad capacitors. Somebody brought in a PC which was controlling an industrial system which was _starting__ to act up. It had clearly bulged capacitors and many of the ones I took out weren't recognized at all by my M-Tester. If I hadn't seen it boot with my own eyes I probably wouldn't have believed it. When the capacitors on the output of the buck converter get weak it usually just increases the ripple voltage and as you know from overclocking usually there is a decent margin on the voltages before things crash.
I was surprised as well because I heard many times that they are one of the best. There is a Wikipedia article about the exact series which were affected. On my P4B board, there is the HM series which I removed in this video. There are also 1200uF 16V Nichicon (HD series) which look like new.
@@bitsundboltsthe worst nightmare is to replace those big capacitors on some 2000s and 2001s motherboards... like MSI K7T Pro etc... there are giant capacitors instead of common one. They are also too high... motherboard is harder to store, and also, they can move easily and can be loose. Luckily, manufacturers realised, the better is to use middle ones, in bigger quantity, because then, it's easier to replace them, and after quite short time, they've stopped used big capacitors on motherboards.
I wish I kept those capacitors to verify this. But yes, I think you're right. The difference is though that they still have storage capabilities for energy - which just drained very quickly. Resistors don't have that. If I place a resistor in my test circuit where the capacitors go, I believe the LED would immediately dim (not turn off and then fade in to medium brightness).
The yellow board transistor tester is probably the worst variant of the transistor tester. It lacks the TL431 reference IC and instead uses a zener diode reference. Plus it's just kinda crappy. You want one with a rotary encoder. They're just nicer. Chinese clone transistor tester variants are a minefield to navigate. I have 3 and one is the yellow board model. I got it when I knew less about them. It doesn't look like my two favorites are available anymore. I got them both years ago now. Looks like 2015. I assembled them both from kits.
I think I am at that stage when you knew less about them. I hope I can upgrade my testing equipment over time when I learn more and when it makes sense.
If capacitors are leaking (vloss) they test with a higher value, it's the test method that does this. (There is a good video about this at mr. carlson's lab page)
@@ralphwuu theres a number of factors involved with component specs but everything else being equal a larger capacitor can dissipate more heat than a smaller one can. It just has more surface area. Size matters.
The capacitor plague is still ongoing, and will go on until people stop using these types of components in electronics altogether. No matter how well they are produced, they come with an expiry date.
Rubycon caps can last for 40 years or more under normal conditions. The 30,000 hours listed in the datasheets for the extreme long life versions are actually torture tests. The capacitor tier list goes Rubycon -> Panasonic ---------------------------------------->everyone else. Rubycon and Panasonic are in a class by themselves.
@@atomicskull6405 And how many percent of electronics in use today use these capacitors? The capacitor plague is still ongoing. For instance, I have a large box of power bricks I tested this summer, in 2023, and 30%-40% of them no longer worked to spec, and most of that was due to failing capacitors. I repair all my own electronics, and still today electrolytic caps is the main cause of failure, even in more expensive gear. Consumer electronics today isn't built to last. Designers know this and take conscious decisions to save cost and NOT use the most expensive brands of caps. People generally don't notice if they swap phones and laptops every year, but I take pride in using my stuff until it breaks.
What's funny is that I brought this capacitor failure up to tomshardware, anandtech back when it happened with the abit bp6, and they both shouted me down saying i didn't know what i was talking about. they both can suck it.
