Reforming Capacitors - Everything you wanted to know

[@kpanic23]

This is also an issue with grossly overrated capacitors: If you use for example a 50V capacitor on a 5V power rail, it will initially work as intended, but with the operating voltage being so low the capacitor can't properly reform the oxide layer and will lose its formation over time. It's always a good idea to leave some safety margin for the eventual stray voltage spike, but too high is bad as well.

[@ResonantBytes]

Wouldn't it just slowly approach being a 5V capacitor then with the oxide layer rebuilding at a much lower rate? The large safety margin will be gone after a (long) time, but I can't see why it wouldn't continue to be operational at the lower-than-designated voltage. Maybe I'm missing something, though?

[@projectartichoke]

Likewise, for filter caps, using 10,000 mfd to replace a 10 mfd may not be a good idea even if it seems like more filtering is better. Larger value capacitors have more current inrush when initially powered which can damage other parts of a circuit. It's not usually a problem, but it some power supply circuits it can be a problem.

[@projectartichoke]

@@ResonantBytes Notice that the reforming curve is just that, a curve. Since it's not linear, it can be difficult to predict how using a higher voltage cap will function in the long run. Generally speaking, a lot of time and effort goes into the engineering of most electronic circuits and the engineers often choose specific values and types of components for reasons that may not be immediately apparent. But, it's always a safe bet to replace a component with one of the same material type and values.

[@SianaGearz]

​@@projectartichoke a lot of time and effort? I give you the salt bae meme with capacitors being sprinkled. Capacitors are often specifie based on prior experience and just industry tradition, and if it hasn't caused issues in engineering samples, that's it then, that's what it's gonna be. If this wasn't quite the optimal choice, well what are the odds that the commercial/useful life of device will exceed 5 years? It's going to be fine.

[@ResonantBytes]

@@projectartichoke Agreed, it's hard to predict what happens exactly inside a certain cap if it's not used in a way that was tested. But that doesn't convince me that there will be more problems when using a capacitor with significantly lower voltage than it's designed (and tested) for. In analog circuits that happens all the time e.g., when using an electrolytic cap to block DC from an AC signal. There the maximum voltage might only ever be reached in certain conditions, even only in fault conditions, but still they aren't failing more often than caps connected to a power rail. I'd argue it's even the opposite. Those caps might fail due to other reasons, especially heat and high currents, but if those are the limiting factors, I doubt the original argument about the influence of higher spec caps even more. Yes it's not linear. But it doesn't have to be. It's an equilibrium system. When the oxide layer degrades, current may be able to flow at the current voltage. If it does that will rebuild the oxide layer, reducing the current that can flow at that voltage until an equilibrium is reached where the oxide layer rebuilds as fast as it degrades (or other effects start to have an influence). The oxide layer will be weaker than if the cap had been powered near its designated limit, but any current due to degradation of the oxide layer would have contributed to the regeneration of it. Also presumably less electrolyte will have been used which would actually prolong the life of the capacitor.

[@oliverw.douglas285]

In the Telecom Industry, we are running into this more frequently, as legacy equipment shows its age. Most of our older Voice Switches employee electrolytic capacitors, to add further filtering to our primary office battery of -48 to -53 volt DC. Over time, these filters caps develope leakage currents outside acceptable limits, & thus they are fused accordingly, to reduce catastrophic failures. As part of our testing, we precharge caps with a series light bulb, of a known value. When the light extinguishes, the cap is charged with minimal leakage. However, if the bulb remains illuminated, it's time to replace the cap. It's old-school, but is time tested, & relatively simple to do.

[@lilblackduc7312]

I like that idea. Thanks!

[@zer0b0t]

yeah, visual test rather than using testers, I like it

[@johncasteel1780]

@@zer0b0t You need to do both.

[@swimfan6292]

And the electric company wants to come install a smart meter on my house, HA! jokes on them!

[@RickTheGeek]

3:50 almond

[@joshuawalker7054]

I like how you reversed the footage of you taking a capacitor apart to show how it went together. That was smart!

[@nickwallette6201]

Keep going until the caps are new again!

[@alexplorer]

Now you're thinking 4-dimensionally, Marty!

[@johncasteel1780]

@@nickwallette6201 An old cap will NEVER be "new" again. It may be usable, but that is different.

[@floorpizza8074]

@@johncasteel1780 I don't have much of a sense of humor (just ask my wife), but even *I* could tell it was a joke.

[@klausstock8020]

@@alexplorer I guess he actually used a time machine for that video. Because he accidentally left the packaging in the shelf behind him.

[@marpintado]

The best part of this video for me is that now I understand the reason electrical equipment disconnected for a long time will bust when reconnected again. Thank you!!!

[@1pcfred]

Yup. When weak capacitors are hit with full voltage they go poof and tend to take other components with them when they do.

[@johncasteel1780]

@@1pcfred And many times, the other components are made of unobtainium. Power transformers come to mind.

[@1pcfred]

@@johncasteel1780 I don't see a capacitor taking a transformer with it. Transformers are fairly rugged and usually fuse protected. But sensitive active components could be at risk. Even resistors often masquerade as fuses in circuits. When they go it is often impossible to read what value they were. They'll either burn up or completely vaporize.

[@dell177]

i worked at a power supply manufacturer for about 30 years as a technician and then engineer. Our units supplied anything from 3v at thousands of amps to hundreds of volts at several amps. The older designes used computer grade caps with screw terminals and these became hard to source. I used to setup a group of caps with individual resistors in series with each cap. I'd start at about 25% of the rated voltage and check the srop across each resistor. After they settled forwn I'd increase the voltage to about half the rated voltage and do the same thing. Then I'd go up to 75% and let things settle for an hour or two. before going home for the day I'd go up to the full rated voltage overnight. After each voltage level I'd check the drop across each resistor and anything over 5-10 MA leakage was a failure (these were old can caps). Atfer the last full voltage run I'd put a red dot on the caps that passed. This proved to be a good indicator of cap longevity but anything over 10-15years was very iffy, even if the caps were NOS. Old caps just are not as well sealed as new ones so they can dry out just sitting in a box. cCps in use can last a very long time because they are kept active, I've got tesxt equipment that is 60 years old with 500v electrolytucs that still work just fine because they see use (Heath 0-400v regulated power supply I built in HS).

[@klafong1]

I had some of these boxed new-old-stock low voltage/high capacitance devices that showed dead shorts when checked with a multimeter. I attempted reforming using a low-voltage supply and a series resistor, such that the forming current would be just a few milliamps. Unfortunately, these capacitors were truly shorted, so they resisted all attempts at being resuscitated.

[@dell177]

@@klafong1 Reforming can revive tired caps but there comes a point where they are just not recoverable.

[@mikehartigan2703]

I build and repair guitar amps, mostly tube amps with supply voltages that can many times reach over 500vdc. A very common practice is to use an incandescent light bulb in series with the power cord as a current limiter when powering up an amp for the first time. The light bulb will light up, then dim with current draw... a nice visual reference, like the needle on your analog multimeter. I was taught to bring the voltage up incrementally to "break in" or "form" the new caps... Now I know exactly what's happening!

[@andygozzo72]

a bulb limiter is a very good idea and i would consider it essential when first testing items of unknown state ...

[@TomKappeln]

I do the same with 70's and 80's high end Pioneer amps.

[@andygozzo72]

@@TomKappeln should ALWAYS test any old electonic equipment with some sort of self current limiting devce, NOT rely on a variac, they dont...

[@StitchJones]

Having a background in electronics. This video is very very helpful. You did a heck of job. Well done sir.

[@ForgottenMachines]

Shelby, you have risen to a new level!!! I'm SO impressed by the technical information, backed by your research and experience here. And most valuable of all, you explain the science behind all of this...so I thank you for making this very professional-quality video and publishing it here for folks like me to learn from...thank you so much!!! Fantastic!!!

[@OIE82]

Yes. Very well done!

[@jonganzer7478]

Incredibly helpful. Someone suggested that I reform some new power supply filter capacitors before I install them (they had been sitting more than two years) and I had no idea what he was talking about. After watching your video, I completely understood what I was doing and why I was doing it; was able to set up the circuit, and do the job safely and correctly for all four 33K ohm capacitors. Excellent and entertaining presentation, very clear and understandable. Thank you!

[@TheBackyardChemist]

This man found a way to turn hardened criminal capacitors into reformed members of the board, WITH ONE SIMPLE TRICK!

[@peterlarkin762]

Lol

[@hardtymz2517]

salvation lies within

[@lilblackduc7312]

"It's never too late to mend".

[@Ed_Stuckey]

_WITH ONE SIMPLE TRICK_ MANUFACTURERS DON'T WANT YOU TO KNOW!

[@johncasteel1780]

MAYBE.

[@TimBielawa]

That web interface is pretty friggin slick!

[@messyfool]

I really don’t understand any of this . But I loved this video. You are so well spoken, genuine , and easy to listen to. You are a amazing teacher and a great salesman. Keep up the good work. Your awesome man.

[@Cotronixco]

Are you his awesome man?

[@LagrangePoint0]

ikr?, I have basic knowledge about electronics but english is not my first language so I had to rewind several times to kinda understand some things, it was interesting and entertaining anyway.

[@MrVeryCranky]

One of the best and comprehensive presentations I've seen.

[@geoffmorrison3648]

The self-healing page at 4:49 is about metallised film capacitors. Entirely different from electrolytics.

[@azurplex]

I've got years of analog electronic experience and never knew this was a thing. I'm still learning, cool!

[@ArchetypalCat]

Me too - although I knew that old caps dry out, leak, or explode (it happened to me ;0 ) and sometimes must be replaced even if the gear they're in has been in constant use. In the audio industry, mixing consoles are "recapped" after some time interval. However, some techs go too far and replace *all* of the mixer's electrolytic caps which usually isn't necessary.

[@TomAtkinson]

That was quite an eye opener! Made me realise something about these old hard drive supplies... some of them take 3 spin ups to get going, amazing it works at all it won't die.

[@Sheevlord]

This is the most comprehensive video about capacitor reforming I've seen. Thank you.

[@cdrive5757]

An easier and more elegant method is to use a Adjustable Current Limit Supply. Dial in your max current, set the open circuit voltage about 15% to 20% below the rating of the Cap. After stabilization keep the current limit unchanged but raise the open circuit voltage to the rating of the electrolytic. BTW, you did a nice job describing the idiosyncrasies and reforming of Electrolytic's! It's been over 65 years since first studying them. Wakodahatchee Chris

[@jam99]

I was thinking the same thing. Why not use a constant current regulator/source with a maximum voltage topping out at the rating of the cap?

[@xxwookey]

I just came here to make this comment. Cheers for doing it for me. A bench PSU with voltage and current limits has been standard for a couple of decades now. All 3 of my s/h bench supplies have it.

[@NaokisRC]

For something like the Data general, you can definitely appreciate why reforming would be required. Also great explanation, thank you for sharing!

[@ashwadhwani]

OMG with this talent, you should be explaining policies to pass bills in the parliament

[@OneTheAndOnlyOne]

Offtopic: At the end of the video the dot matrix printer... I haven't seen one in over 20 years now.. I miss the sound of it. LOL! I want one...again.. Ontopic: great video! That's why some old stuff just don't work any more after setting them away for 20+ years...

[@EdgarsLS]

It might be worth mentioning that for non-polar paper capacitors, reforming like this doesn't work, instead the paper dielectric has absorbed moisture and you need to get it out, for example putting the capacitor in molten beeswax above the boiling point of water... or putting it in an oil, and then pulling a vacuum on the capacitor that will boil off the water as well. I've done this and it works

[@Dinnye01]

He *did* mention that this is for one specifoc type of capacitor. I do hope he will do the rest as well.

[@bzuidgeest]

Please just replace it. Paper caps like that will never be really good or safe again. They are at the end of their life or beyond. Want an original look, restuff them.

[@michvod]

Yes, paper capacitors can be theoretically "repaired", by boiling them in hot wax for an hour. That drives the moisture out and rebuilds the insulation layer of the paper dielectric. But this works with larger unsealed ones, that were mostly from 1930s-40s. There are some videos of some people doing that, and the insulation properties increased from like 1M ohm to few G ohms

[@MrCarlsonsLab]

Completely sealed "molded" capacitors (Think of Sprague black body, yellow writing types,) (paper types) are all bad due to the paper inside becoming acidic, not just moisture. Wax body capacitors have both moisture and acidity problems, so you are just lowering one issue, but the other far worse issue still remains. Think of discolored pages in a an old book (yellowing or browning paper) this is whats going on inside all of the old paper caps. They all need replacement, no excuses, just physics and logic.

[@Dandan-tg6tj]

I usually carefully reflow capacitors at 110 Celsius (while they are rated for 120 Celsius). That's all I will ever do apart from replacing them.

[@ArchetypalCat]

Just came across this - thanks Mr. @TechTangents ! Some used audio gear I have acquired is in an unknown state, so I'll be doing the variac and light bulb trick because all of them are most likely old enough to have old-school, linear power supplies, not the modern, switching power supplies ("switchers"). Of course, I'll crack them open just to make sure. It is good that you mention that switchers cannot have their voltages slowly ramped up without possible bad outcomes. I have also one piece of my own, pro audio gear that is about 25 years old but was powered on only once for a few minutes to make sure it worked - and then shelved and never used again. It will be very interesting to see what state its caps are in. Wow = @Mr. Carlson himself commented somewhere below - he da E-Man !

[@popquizzz]

For a minute there until you spoke, I was thinking Garth would pop in from behind the camera and I was being transported into another Wayne's World video on Public Access Television.

[@electronicsandewastescrapp7384]

Ha I've got that same radio shack multimeter I've had since I was a kid in the 80's! Learned more about capacitors here than in all my years replacing them. Great work!

[@PhattyMo]

15:46 - Those might be oil-filled caps,in which case..I don't believe reforming is needed,or even possible,since they aren't electrolytic. Side note: Old oil-filled caps were often filled with PCB (Polychlorinated Biphenyl).. If they are leaky,dispose of them properly! That stuff is nasty and hazardous.

[@victorman2227]

Another factor in speed of leakage current slowdown at 15:09 might be different capacitance values as the circuit is basically a high pass filter loaded by resistance of the analog meter. Oh and if you want to automate without multimeter you could consider a source measure unit, they have more current ranges and some can go up to 1000V (but yeah expensive sadly). Nice vid thanks.

[@andygozzo72]

use a filament bulb, they have ptc resistance characteristic... low when cold/low voltage across, rising with applied voltage, very useful as current limiters, semi stabilisers

[@WestEast3259585]

what about that using that funky "CC" mode in a power supply?

[@victorman2227]

@@WestEast3259585 For the typical power supply the current we want (

[@Green_House]

I did hear about this about 55 years ago. Now I understand it. Well done!

[@robsonrobbi1763]

I means you are really old

[@Green_House]

@@robsonrobbi1763 You are obviously from the 'I' generation. Your observation of the obvious is amazing! FYI: When I first became interested in electronics, there was no internet to find out anything. As a teenager, if I wanted to learn anything, I had to walk down the street, catch a bus to the city, walk to the library and hopefully there would be a book addressing the topic I was looking for, and if there was, I would have to read many chapters to find the thing I was looking for - no quick search feature or hyperlinks. 🙂

[@peteleoni9665]

You sir, are a perfect, ideal, clear and concise tech speaker/writer! I am impressed and educated. Bravo # 1 !!!! Perfect speaking voice as well. Bravo #2 !!!

[@sficlassic]

The idea of reforming old capacitors is cool and maintians the original look of the device. I would rather scrap them when you are working on expensive or antique ( vacuum tube ) equipment. But I love how you go into the detail on how to do it.

[@fallwitch]

Very interesting. Thanks for the informative post. I would love to hear what Mr. Carlson would say to this video.

[@dwagner6]

Lol he’d be giving a thumbs down, for sure. Not worth risking your equipment!

[@Morinaka25]

He'd tell him to at least make his capacitor tester, so he knows for sure if the caps will hold up.

[@larryh8072]

I think he would be fine with reforming. He has made a very nice variac and dim bulb limiter for bringing up older tube equipment. He did comment elsewhere that paper capacitors must be replaced. Not only do the “waxies” take on moisture but the paper becomes acidic. These capacitor types need to be replaced.

[@gabrielleeliseo6062]

A great KZbin channel who talks a lot about caps is Mr Carlson's Lab

[@chironbramberger]

I'm so glad you made this video. I've also done a bunch of research on this but you've put it all together very nicely and made ti relevant for people who like restoring vintage computers. Well done - thanks!!!

[@kstlfido]

Well done! We run into this at CHRS all the time. Reformed many; some are just too far gone. Thanks for spreading the knowledge!

[@dolphhandcreme]

2 things to note here: 1) the sinkrate of the current is also greatly depending on the capacity because it is high on discharged capacitors. When they charge, the current goes down. 2) why didn't you use the current limiting function of your psu? It nearly does the same but without bumps. You can even charge lithium ion batteries with such psus, just set maximum voltage and current. It's called CCCV. (Constant current, constant voltage) or IU.

[@JasonOlshefsky]

Came here to ask about #2 ... it seems like a constant-current of 2-5mA and a maximum voltage at the rated working voltage would work as well or better. Is there something important about the stairstep method?

[@davewilson4493]

Using a psu in CC mode was also my first thought. I guess that with an analog psu with a single potentiometer current control, even if using an external meter which gave precise current readings, it might still be extremely tricky setting a very low current. Rather easier if the psu had coarse/fine controls, but with a digital psu which had current control down to the mA level, it should be really easy, as long as the psu did what it was supposed to do at the extreme low end of its current limit range (I'd still probably put a meter in series to make sure).

[@SeanBZA]

Most PSU you find in use are normally reasonable in voltage control, but current control is normally very poor both in terms of loop stability (they oscillate around a set current nicely) and in actual control ability. you need a SMU, generally not a common, or cheap, item in a workshop, to get good repeatable and reliable current setting ability, or a constant adjustable current supply, also not common. So a constant voltage power supply, and a resistor is the thing to do, though with very large capacitors you can use an incandescent lamp, rated for roughly the voltage maximum for the capacitor (so a 120V lamp for a 100V capacitor, and a 230V one for up to 400V, with a 12V and 24V lamp for lower like 15V and 35V) to provide a faster charge current, yet still provide protection against it going suddenly short. Light bulbs around 15W for the higher voltage, and 5-10W for the lower voltages works. So long as the voltage is within 50% of the rated either way it will survive short term overloads, plus visual indication on the higher voltage ranges by it becoming a photoflood lamp.

[@johnwilson839]

@@SeanBZA It's been a while since I used a bench power supply with a variable current limit setting so maybe it is difficult to set the current correctly. I'm still curious though about the original question of would it work fine without the stair step? it's pretty simple to rig up precise(ish) constant current circuits with a couple of transistors and that would be easier for me than getting GPIB set up.

[@SeanBZA]

@@johnwilson839 Does not really need constant current, error versus simple RC will not be much of an issue, plus you can also use the drop across the resistor as a current sense. While some PSU units are great with CC mode, I find a lot of the lower end ones are not exactly great there, especially with a very large capacitive load on them. After all a 5c resistor is cheap, and can act as olfactory warning of high current flow.

[@patmx5]

Great explanation of the reforming process! I generally use my small HP bench supplies (6212A or 6216A) for caps rated up to 100V, or a 6515A for anything higher. All of them operate in either constant voltage (what people generally think of when working with a power supply) or constant current mode. For the 621Xs, short the output and turn up the voltage a bit, then adjust the current limit for somewhere between 5 and 20 mA. For the 6515A, this step isn’t necessary as its output is fixed at about 5 mA. Once the current limit is set, remove the short and set the power supply to the cap's rated voltage. Turn the power supply off and connect the cap, then turn the supply on and it’ll increase the output voltage to maintain the current setting until it reaches the rated voltage, at which point it will change to CV mode as the leakage current decreases. Less babysitting needed that way, and if the cap doesn’t fairly quickly get up to the rated voltage, you know it’s toast. I monitor voltage and current like you do - voltmeter across the cap, and milliammeter in the low side between cap and power supply negative terminal. I came upon a box of big caps I had squirreled away a few weeks back and have been reforming like a madman lately. Most have come back nicely; only a few went into the garbage. In my opinion, at least for larger caps, it’s definitely worth a shot before just blindly replacing them. Certainly those used in older high end test gear (HP, Tek, Fluke etc.) are normally very high quality components that will last a long time if properly cared for.

[@jp.larocque]

I agree that just using a bench supply in constant current mode with a correctly-chosen target voltage should be more straightforward. On one hand, I _love_ the scripting Tech Tangents did for this project; these kind of scenarios are why I always look for computer interfacing options when I shop for test equipment. But I don't see what stepped voltage increases can do for you when the power supply can regulate the current within a chosen target for you. For that matter, no need for a resistor either (it looks like resistors were mentioned in that specification for use with a constant-voltage-only supply). And no need for external voltage or current monitoring: typical bench supplies should give you a voltage and current readout, and at 2-5 mA you don't have to worry about voltage measurement errors due to drop in the leads. KISS. One exception I can think of is if the power supply doesn't offer fine enough current control or current measurement. Cheap supplies might only offer 1 mA resolution (or worse), but it looks like the supply shown in this video has a resolution of 100 μA, so that shouldn't be an issue.

[@algorithminc.8850]

Really great video - great explanation. Seems consistent with my experience with it. Nice to have the video to share with others. Many thanks. Subscribed. Cheers

[@KeritechElectronics]

Hey, thanks for a nice and in-depth explanation! I've known about reforming capacitors ever since I started restoring tube gear, and I sometimes do that both in existing devices and new devices I build using antique components (can electrolytics etc.). Generally I test them before reform attempt. If there's no capacitance, it means that the lead has corroded internally and the capacitor is dead. Otherwise I'll attempt reforming, bringing the cap up on a 300VDC power supply through a current-limiting resistor, by setting the voltage to get ca. 5mA out of it, leave it for a few hours and bring it up again after the current drops below 0.5...1mA, observing the max rated voltage. Looks like I could use a CC/CV high voltage supply for that and for tube circuit prototyping :)

[@BlackEpyon]

With small capacitors, probably not worth the hassle. The big beefy ones though, definitely worth trying if you got the equipment for it. Those guys are NOT cheap, and are often difficult to source!

[@andygozzo72]

it can be worth it if the cap is an old vintage type and you want to preserve originality as much as possible , i've reformed many signal couplers successfuly, in some cases the circuit will self reform them, after a while

[@TomKappeln]

Pioneer amps from the early 80's .... 20.000yF/63V .... oh yes ... they cost if GOOD ones.

[@cajunfid]

A lot of guitar nuts would disagree. They freak out when someone has hacked up a vintage amplifier and replaced the original tone capacitors with newer variants.

[@TomKappeln]

@@cajunfid : In this case : replace the guitar guy ! Drummers are way cooler 🤣😜

[@andygozzo72]

@@cajunfid if these tone caps are only in low level signal part of the circuit, no high dc across them, unlikely to need replacing anyway,

[@DHGxMcFlurry]

I'm just starting out to learn about electronics/circuitry and most channels are hard to understand and keep up with but this is oddly very easy for me to understand. I would love to see more videos like this about basic components or ics, or different logic systems etc explained for noobs like me

[@angrydove4067]

An intelligent video about reforming, not just DO or DON'T because that is how I do it. Thanks Shelby.

[@AySz88]

Thanks for covering everything from the basics - I suspect people from the "bad caps" days will imagine that "reforming" just means the same (euphemistic) "remanufacturing" or "refurbishing" and all the baggage that comes with that era.

[@ResonantBytes]

Thanks for this video! Using your devices regularly, even just for a few hours every year can do so much! As a repair technician I've been doing capacitor reforming from time to time, but mainly using the variac method due to time and equiment constraints. Most of the time when we get a device that has been in storage for a long time, the owner has already connected it to mains and done the damage before they bring it to us. The method in this video appears to be the proper one, but it is also time consuming. It's all about limiting the intensity of the chemical reaction during the self-healing process and therefore especially the temperature. If you can't take the device apart and desolder all caps, it's still much better to have it sit at a lower voltage for some time before connecting it to the full mains. I've been thinking about what to do with switching power supplies. Regulated PSUs with linear voltage regulators usually aren't a problem, because you can usually regulate the voltage down enough so the input voltage to the regulator is so low that it can't output it's designated voltage and thus reduce the load on the capacitors connected to them. I'm going to do some experiments with switching PSUs to get an idea what we can do to reduce the stress when taking back to use. Have you done anything in that direction or have you just accepted having to desolder every large enough cap in switching PSUs?

[@TechTangents]

Switching power supplies are so finicky that you really can't work around it. You have to remove the caps. I'm not sure quite what you mean about linear PSUs either, I don't think you can reduce the voltage into them to vary their output to reduce voltage to any downstream caps. An LM7805 for example, is an LDO that won't output 5V until you get up to about 7V input.

[@SianaGearz]

​@@TechTangents the LDO will however happily turn into a series of diodes when the necessary input voltage is not reached. So you can gradually raise the voltage through it.

[@ResonantBytes]

@@TechTangents "Linear" in the sense that there is a transistor that does the voltage control and turns any surplus voltage into heat, be it inside a 78xx/79xx or similar or discrete. "Continuous" may be a better term, but "linear" seems to be the one that's most commonly used when talking about regulated PSUs (may be a non-English-native-speaker-thing though). "LDO" means "low drop out [voltage]", which I guess you seem to use for that kind of circuit design? That's actually a special class of those "linear" regulators and the 7805 certainly is *not* one of those. Having had a night's sleep over the matter I do think it's possible to reform the caps of potentially many switching PSUs without removing them first. Most switching PSUs I came across have a rectifier made of schottky diodes on the secondary coil of the transformer (or multiples thereof) which connect to the filter caps. At least one of those power rails has a voltage divider which feeds back to the switching control circuit via an opto-isolator. As long as that's all that's connected internally to those rails (and I think that's the case for most simple switching PSUs), we can safely feed an external voltage into the *output* while the PSU is turned *off*, because the diodes prevent any current back into the transformer and the voltage divider has a high enough resistance to be negligible. Practically you could connect your setup to one of the voltage rails as a whole. Repeating that process for all voltage rails we'd have reformed the secondary side of the PSU. For the primary side things are a bit more complicated. Some PSUs might be OK with a lower input voltage after the secondary side has been reformed, but that's what I want to investigate further: Which kind of circuit can cope with that? For example PSUs which can work with 120V as well as 240V (without a manual switch) can be powered with down to under 100V at times which could be low enough to reform the primary capacitors depending on the circuit. The current through the PSU would make the process you described here impossible of course, but the eyeball-approach using a variac could still be a valueable option here. Worst case scenario you could still desolder the primary caps and reform the secondary ones in circuit with many switching PSUs, which could save a lot of time.

[@ResonantBytes]

@@SianaGearz That. :)

[@eDoc2020]

I figure any switching PSU made to run off of grid power won't be damaged by just a few minutes of running at reduced input voltage. I would just dim bulb it for a bit. Any reforming current will be much higher than ideal but it will let you know if any are excessively leaky.

[@TastyBusiness]

Reminds me of when I ran the process a few years ago, with decent results. Saved a bunch of money on some very large caps, with an 80% success rate. Ultimately other problems with that PSU led to retiring it and installing a modern switchmode option, but it was good practice.

[@Xsiondu]

Used to repair welders and my boss showed me this neat thing where we put a run-out gauge on the side of a capacitor and tested the welder with our load bank. Watching the caps change diameters under use was interesting. It had no diagnostic value as far as i know but it was interesting in a " huh? Well would you look at that" kind of way.

[@thehobe150]

Very good explanation and description of the process and chemical situation (coming from a 35 year EE analog design engineer) but the prices of just replacing the capacitors with newer ones makes it hard to justify the time spent. Mr Carlson ( on Utube) is a master of replacing especially very old capacitors including mostly non-electrolytic capacitors. I think your video is actually very educational and between the lines, give more reasons to replace than reform (especially the unknow ESR factor you mention in the video. The newer switching power supplies actually rely on low ESR type capacitors to limit the internal temperature of modern capacitors.

[@ProtonOne11]

Very well made video. Easy to understand and follow. I would have hoped that you got your hands on a professional RLC meter to actually do a before/after comparison of a reformed capacitor. You don't have to buy one for thousands of dollars, there are places that rent out equipment like that. We even have companies that send us test equipment for free in a "try before you buy" kind of deal, so we can play with the instruments for a few weeks and then decide if we want to buy it, or just give it back with no cost or strings attached.

[@mrgcav]

As an electrical engineer BSMEE I have heard of capacitor reconditioning and reforming. I have worked on and designed Amplifiers and power supplies ( Linear and scan rectified)and Frequency generators and Programmable control hardware for 40 years. In the field it is easier and cheaper just to replace a suspect capacitor. On vintage equipment is another story. If the capacitor is not easily replaceable because its in accessible or because modern capacitors are much smaller for the same value uF or even for ascetics, Then reforming may be an option. Problem with reforming is you do not know the quality the final result and reading the uf capacitance value is no indication. It may last a day or a year. But a new Capacitor is always better than reforming an old one. You post no credentials. But you provided a in depth wel explained explanation of what is happening on the inside of a capacitor. Suggest you do a video on the capacitance of components that are not capacitors, like a Zener diode, a MOV, a step up transformer, A Transistor, An DIP IC socket, a touchscreen and why this is so.

[@jackhreha4907]

This show puts you on my A list. I like how you use the use the junk i bought when i was a kid. God bless Radio Shack and the battery club. And remember Boys and Girls a new cap can have issues that will make you cry. Good luck to you all on reforming your cap world. Take it from me when you do this stuff wear the safety gear. Best Regards Jack.

[@Gracana]

Great video topic, and well explained. I first encountered the concept of electrolytic capacitor reforming in a manual for an industrial variable frequency drive (for speed control of AC motors, if you're not familiar). It was a totally foreign idea to me, and I was surprised to learn that it's necessary due to the way electrolytic capacitors work, as you described in the video. I had no idea at the time, and I still think it's pretty wild.

[@dinkc64]

This is the definitive video on capacitor reformation. Thank you TT!

[@JonTheGeek]

Well researched videos about topics that no one else bothers with is exactly why I'm here. I love this video, and good information to have as an Auto Tech in an ever expanding electronical field.

[@threethousandyen]

Fantastic info. Love the idea of using GPIB to automate. I have that same bench meter and a PSU with GPIB. I’ll be giving this a try in the future.

[@maherf768]

I live in Syria in a remote area, buying components is hard due to rarity (I have to travel long distance to buy in person in bulk) and sometimes unavailability and expensiveness. I use components from old boards and for my projects and repairs and it feels so good! I always test components and 99.9% of the time they work great

[@stamasd8500]

OR, instead of using fancy computer-programmed gear, you could go the analog way. Construct a simple constant current source and use that to put a constant current through the capacitor. Set the voltage source to close to the maximum voltage you want across the capacitor (plus a couple extra volts to account for the drop across the source), and use it to feed the capacitor with a constant current through it. The constant current source will automatically adjust its internal impedance to generate the required current. You only need one JFET and one resistor, or 2 BJTs and one resistor to make a simple constant current source. You do need to fiddle a bit with the value of the resistor initially to get the exact current value you need because that depends on the exact parameters of the JFET or BJTs you're using, but that's a one-time process only.

[@sjwhitney]

One thing I would like to see covered and examined is how many manufacturers cut corners and install capacitors in devices that are barely able to handle the rated voltage in their installations. A couple cases in point here that I ran into years ago... I had purchased a plasma TV when they were all the rage. It worked fine for about a year and then I noticed longer and longer start-up times for it. Then one day I hear the dreaded high-pitched "pop" from inside and smell the all too familiar odor of the electrolyte. I set the TV aside and we bought a new LCD one. Everyone seems to understand the power demands of the plasma versus the LCD (and now LED) sets. However, I was not going to give up on the plasma. I had heard of them failing all across the country it seemed!! I was convinced of a simple explanation. Taking apart the plasma TV took longer with the about 800 screws (exaggerated of course) holding the case halves together, I found that it was neatly laid out with separate boards for the various functions, one of them being the power supply. As an aside, these power supplies have NOT radically changed and are still used in all sets! Looking at the board I clearly saw the popped open capacitors and others that were bulging. I looked at the rated voltage of them and quickly decided that the manufacturer had not allowed any leeway as they were right on the actual voltage applied to them. I took down all the ratings and then went to e-Bay and bought a whole new set of them at DOUBLE the rated voltage. Now, here's the most interesting part... The circuit board was actually set up to use larger capacitors than what were installed!!! This was plain to see by the pictorials drawn right on the board and the extra space around the originals. Nothing else on the board had extra space. The new capacitors arrived, and wouldn't you know they fit the actual available space perfectly. That plasma set worked fine afterward, and it is still working like a champ in my living room 20 YEARS later!!! I only wish I had started repairing plasma sets for other people as so many were thrown away for nothing more than $12 in caps and a little labor on my part. My second great instance was the control board for my $1 Jacuzzi hot tub that failed (yes, it cost me $1 for an eight-person hot tub!!). It sort of mirrored the plasma TV in how it gradually failed. I took the board out and saw it looked almost perfect. A couple caps were peaked though but not ruptured. I found the test sequence on a web search and proceeded to measure everything until I could go no further. At that point I sent an e-mail to a person in AZ who could rebuild the boards. I gave him my thorough test results and he was so impressed that he called me! He said that it was going to be an easy fix since I knew my way around and that I wasn't going to believe what I had to replace. The capacitors were my first guess and was right. Again, I ordered double the existing voltage rating and oddly, there was extra space on the board again... Seeing the manufacturing pattern here? After replacement the hot tub was back in operation, and I was relaxing in it until we moved out of that house and sold it before that move. For $3K in case you're wondering! So, the question remains as to why manufacturers are shortchanging the capacitors in everything. Are they simply doing it knowing that they will fail and the average person will run out and buy a replacement device? I smell a rat here....

[@vencdee]

This is the strategy of limited lifetime of device. So after guaranty is over, the device will fail and you must buy a new one... Not only capacitors but insufficient cooling on power transistors and ICs etc.

[@shawnlennon1947]

You’re a good teacher! Thanks 🙏

[@waterflame321]

It was really fun watching you work on this stuff live!

[@AndersNielsenAA]

Nice one! Just what a sceptic like me needed :) A few times you made it sound like a cap is supposed to “have a load” or “be a load” which is a bit confusing, but in this context it does make a little bit of sense, considering we’re talking about ESR. I still have a hard time finding a use case since it’ll still be easier to replace an old one and less risky - time is money and a new one is less likely to blow up 3 months later leaving the neighborhood soaked in electrolytic. Sometimes the original caps might be a big part of the aesthetic though - in that case reforming is your only bet. Always good to have another tool in your arsenal, thanks!

[@gorak9000]

Like he said in the video, it's definitely a viable option for the large electrolytics (like soup can sized or bigger) that used to be in old products - those things are super $$$$ to replace. For your average up to 4700 uF caps you find in modern stuff, most of the reasons they've failed is because they were garbage quality / china off-brand no name garbage to begin with. Best to replace with name brand quality caps - and they cost ~$1 or so anyway, so who cares.

[@phildem414]

Very instructive 🤔 I have a suggestion to automate this with a regulated power supply: set your lab power supply to current limitation mode and set the voltage to the target voltage. The regulation will gradually up the voltage with respect to the current limit. To be tested before reforming the caps of your vo stage Marshall 😏

[@Kenzofeis]

I just wrote a whole rant about that a few minutes ago, now I see you mention it far down in the comments, yes that is the simplest way to do it

[@a10killroy33]

really like your research, though I think I will just replace them whenever needed. About your method, you could also use a high end power supply with current and voltage limiting. if you first set the current limit while shorting the power supply, and than set the voltage limit by opening the outputs, your power supply will automatically ramp up the voltage as the current tends to drop. of course, your power supply need to have fine adjustment for milli/micro amps for the rated capacitors you intend to reform.

[@markm1514]

This is a super interesting discussion, and a great explanation (or part of one) why the 12v ac adapter I bought 10 years ago stopped delivering 12V 2A and became a questionable 11V 250mA supply. If you have a project that calls for lots of coulombs and you have the e-waste, I say go for it. I can envision a few scenarios for it but honestly I would probably just go with new caps for anything serious. I'm not an expert, I don't have all day, and capacitors are cheap. 👍

[@danmanx2]

I was just thinking why not write a program to automate this and bam....there's python. I love your channel dude!

[@hobbified]

If you have a reasonable digital CC-CV bench power supply with 1mA current resolution (like a Rigol or a Miniware) you can do moderate-voltage caps without all the fuss - just set the power supply's voltage limit to the capacitor's rated voltage, the current limit to a few mA, turn it on, and let it sit. If after a few hours the power supply is still in current limited mode (hasn't hit the target voltage yet) then your cap's leakage current is higher than the limit you set, and you will have to use good judgment to decide whether to bump up the current a bit more, or discard the cap as faulty.

[@eDoc2020]

As his HP power supply he is using in the computerized version has that resolution I was wondering why he eidn't just use that.

[@SeanBZA]

@@eDoc2020 Because not everybody has such a nice power supply, and spending close to $50k, to test a $5 capacitor, is not exactly good economic sense.

[@eDoc2020]

@@SeanBZA But he used the GPIB which _requires_ an expensive system. If he just used the current limit a sub $100 PSU would work just fine.

[@popquizzz]

A cheap high voltage power supply can sometimes be reclaimed from an old welding machine and some industrial varistors.

[@smudgeone]

I deal with AC variable speed motor drives and occasionally have to reform the capacitors in them. Manufacturer recommends to do it every 2 years if they've been powered off. They are usually 450V capacitors in a series/parallel setup. I use an 120V AC variac, a 600V:120V transformer in reverse and connect to the AC input of the system. The drives have a pre-charge diode and resistor for normal starts and a thyristor bridge on the 3ph power in. I can bring the voltage from 0 to 600Vac (0 to about 850Vdc) using a standard 120V wall power. I have also done a similar manual setup with a 600V rectifier and a rather large resistor (I think it was 10ohm, like 100W) to make my own DC supply for models that did not have a rectifier circuit.

[@Nerven86]

What a great video! From Greece, greetings!

[@juanferreira5931]

A clear and concise explanation. Keep them coming.

[@rogerwalter2500]

This is the video I was looking for, thanks buddy and KZbin suggestions

[@Bishka100]

09:45 Ha,ha. I have one of those multimeters (I have a blue one) I got it from Tandy in 1984 and I still use it.

[@mikemater]

One of my fav youtube channels man. Keep it up. Also wish i could grow hair and copy that flow.

[@miscbits6399]

2:40 - apart from the issues of reforming a cap, this small snippet demonstates why older caps have very high inductive reactance Newer designs offset the foils so the ends can be folded over and the contact lead touches ALL the coiled up foil on its side of the insulator @kpanic: it's a tradeoff of voltage spikes. I spent a lot of time in the 1980s stripping out and replacing failed 16V caps on 12V rails (plastic cases, they usually _split_) with 50V ones. These were in landmobile base station receivers worth about $1k a pop so it was worthwhile. We ended up replacing ALL caps on sight if we found older devices in them as there were too many instance of replacing the failed ones only, to have other caps fail in-service several months/years later (with a 50-200 mile roundtrip to replace the things including a slog up a mounatin goat track)

[@cannaroe1213]

wow i have no idea who you are, and i watch a lot of youtube, but I think i watched a game video from you before. you are really really really good at youtube, education, tech... honestly bro idk why you're only at 163k subs, you should be over 200k and probably over 400k.

[@christ2290]

Reforming capacitors is like buying underwear and socks at a garage sale.....

[@realraven2000]

8:25 failing the possibility to slowly raise voltage, would it be valid to to bursts of power? E.g. swithc on device for 2mins, power down,then wait 1 min, switch on for 5mins, power down, then switch on for 15 mins?

[@vencdee]

Probably not... the capacitors will explode or so immediately... if not then maybe it could help, but you never know the currents and proper times...

[@stco2426]

Really interesting. I knew what a capacitor is and that they sometimes fail but that was about it. I now know a bit more and what to look out for on a failed capacitor. Interesting to learn about reforming and when to try it v replace. Many thanks for taking the time to make this and share your knowledge.

[@thecorruptedbit5585]

I watched the Data General videos, and a few days later there's a whole explanatory video about the mysterious "reforming" you were planning to do. Amazing!

[@deltavee2]

Interesting plus the Compaq hat is cool and the dot-matrix printer snapped me back a substantial number of years. Thumbs up for educational.

[@CrassSpektakel]

I studied Electrical engineering and even I didn't knew about reforming capacitors. Though to my excuse, I was mostly into writing software, designing digital circuity and radio/em stuff. VERY NICE!

[@HyperVectra]

NO EXCUSE!!! :P

[@ReneKnuvers74rk]

Isn’t beneficial to slowly rotate the cap for some time over two axes? To distribute the remaining electrolyte? We see very similar issues in really large DC (2500V or up) capacitors that are used in DC locomotives.

[@Guitarhero1000]

Really all you need is to rig up an incandescent light bulb socket to the power in. The light bulb will feedback any over wattage and protect anything you are working on

[@1pcfred]

A bulb limiter may protect your house from burning down but I wouldn't guarantee no component failures. In the time it takes a lightbulb to flash a lot can happen in the world of electronics. Some of it rather irreversible too.

[@Guitarhero1000]

@@1pcfred true. But it's better than nothing I guess

[@1pcfred]

@@Guitarhero1000 current limiting can certainly avert major disasters. But as far as stopping damage entirely goes I wouldn't bet the farm on that. It does not take much current to take out sensitive circuit elements.

[@ArchetypalCat]

@@1pcfred that's why I use a bulb and a variac together ...

[@1pcfred]

@@ArchetypalCat bringing equipment up slowly can save it. The bulb is a good idea to monitor the current. Exceeding current is what burns stuff out.

[@r3v0lv3rz]

Thanks, I didn't know that you could repair caps like this until your video.

[@xantronix]

Thank you very kindly for everything you do, Mr. Shelby Technology Tangents.

[@billysgeo]

Thank you! This video gave me a true nerdgasm! Haven’t had one of those in a good while! 14:10

[@aaronsmith593]

Also for decades every tech gave Panasonic capacitors in particular the best rating for accuracy, power handling, and longevity.

[@glennlopez6772]

Thanks for the video! How should I handle my Yahama Midi keyboard and its original power adapter. The equipment has not been used for a long time. Hope some one replies! Using a light bulb as someone commented is nice but should I tackle the adapter alone without the keyboard plugged in! What next? Thanks!

[@Bob_Lob_Law]

You could also just use the current limiting function of your power supply, and skip the whole resistor and adjusting voltage debacle. It might be a little challenging to set it that low, but even with a knob power supply, it could still be done.

[@RexxSchneider]

A constant current source that supplies current in the range of hundreds of microamps to a few milliamps is simple to build. For voltages up to 40V, an LM334 plus one resistor will do the job. For voltages up to 100V, use a cheap IRF530 or IRF540 with an adjustable source resistor to set the current. For higher voltages, there are high-voltage mosfets like the IRF820 that handle 500V, but obviously cost more and require more care when working with such high voltages.

[@BartKus]

Yeah! I don't get why any of that software was necessary. PS: I miss your Law Blog.

[@SuperBrainAK]

@@RexxSchneider I have done similar with simply a transistor and a constant voltage set on the base, using a resistor from the emitter to ground the voltage drop across it blocks current flowing through the base reducing the current through the collector. Super simple but might not be as simple as a dedicated IC, but it is super cheap!

[@RexxSchneider]

@@SuperBrainAK I agree that using a bipolar transistor with a constant voltage on the base and an emitter resistor is also an easy way to create a constant current source, and it would have been my first choice 40 years ago. However, the cost of a bipolar transistor plus a zener diode and an extra resistor is not much different from the cost of a JFET or an LM334. For use to supply the sort of constant current needed in the video, there's not much to choose between them, but as soon as you need temperature stability or are working from low voltage rails, the LM334 wins hands-down. The other option is a current mirror, but you need two very well-matched transistors for that circuit to work well.

[@2tallB]

I wondered about that - you could just set your PS current very low, but I didn't try it, might not be so easy.

[@danielpas368]

Does setting in constant voltage and pulsing a higher current help build the oxide layer? I would think ideally you want to get it on a constant current supply at your chosen rated maximum with a voltage cut off limit set, I would think this would achieve a faster oxidation.

[@joutoob9]

@tech tangents @11:50 That Micronta Multi-Meter, I have that one, passed down by my dad long ago but the red and black test leads have disintegrated. Any idea where I can get replacement wires? I'm sure Radio Shack would have had them but it's gone now. Maybe someone else watching this has had to replace the wires on their same model and can point me to a current item somewhere to order?

[@AmalgmousProxy]

This is wealth of information. Thanks!

[@wesley00042]

Perfect timing. I have a SWTPC 6800 which has a minibar Coke can-sized electrolytic that needs reforming.

[@1pcfred]

In the simple method you just connect a capacitor to a variable PSU through a resistor with a meter across the resistor to monitor the voltage drop. Begin at the lowest voltage and note the meter reading. When the meter reaches zero increase the PSU voltage in a step wise fashion and again watch the meter. When it drops again increase the voltage again. Keep going in steps until you've reached the capacitor's rated voltage and then hold it there until the meter's again dropped and then hold it there. It is recommended to form for one minute for each month out of service. I generally call it good after an hour myself.

[@dynamiceq7215]

The metal can cap shown in 15:36 is commonly used as a motor run capacitor in HVAC equipment. They are very inexpensive in small uf sizes.

[@sjftech]

So many people missing the point here. There are certain caps that you either can't buy anymore or that can be prohibitively expensive. If you can reform them, you can potentially save a device from becoming ewaste.

[@1pcfred]

I reform capacitors purely for the Mr. Wizard voodoo aspect of it. Quite frankly I don't need any capacitors. As they are neither food, clothing nor are they shelter.

[@ArchetypalCat]

@@1pcfred I'd say you got the point on the most fundamental level, with an extra point for humor reforming.

[@napalmholocaust9093]

You have a scholarly voice despite looking like a roadie for Warrant.

[@kuro68000]

Could you use a current limited power supply for this? Just set it to 2mA limit and voltage to the rated voltage of the cap. Would need to be a decent PSU that can reliably regulate down that low, but they aren't expensive now.

[@owenvogelgesang7314]

That was my question. This seems really complicated just to have a current limited power supply

[@83hjf]

yes this is too overkill. all he needs to do is a current limited psu and feed each DC rail individually. "regenerating" all caps at once.

[@tarstarkusz]

The BEST option is to buy an old (very old, tube era like 1970s and older) capacitor tester. They can put a lot of voltage on the cap and measure the leakage and capacitance with a cool magic eye tube. Most go to 600V in steps, like 10, 20, 50,100, 200, 450, 600. A lot of people are using them for this purpose. There are millions of them out there, but it may itself need some restoration work. But it is very easy to reform caps with them.

[@kuro68000]

@@83hjf It would be hard to regulate the current if you were powering up the whole device. Maybe on the mains side of certain PSUs it's possible.

[@kuro68000]

@@owenvogelgesang7314 well it's safer, in that a current limited PSU might put out full voltage for some fraction of a second before current limiting kicks in, but you can overcome that by manually ramping the voltage up at the start. I'm guessing the resistor method was with older analogue power supplies in mind, and so you can see the current reduce to know that it's working. If the current doesn't fall/voltage keep increasing you need to stop because the cap is probably bad.

[@2tallB]

Excellent video! I like that ammeter method.

[@hermitxIII]

I've seen shango066 reform capacitors in old TVs many times with a variac and light bulb, so I had a vague idea of how it works. Never expected to find a tutorial.

[@hellboystein2926]

Heared that for very big commercial ones, so the example that I remember was a capacitor that was big as a coffee cup and used for a ~220V AC-motor inverter, but for the smaler standard electrolyte capacitors i always got told that they 'dry out' and in the same company where i saw the big one that has to be reformed after some weeks of not using it, they changed ALL Elkos as a standard procedure when refurbishing electronic modules.

[@SockyNoob]

Before I even watch, I can only imagine this being especially useful on large, high quality caps. If I see a no-name failing, I'd replace regardless.

[@nonsuch]

I'm sure you probably thought of this or someone else mentioned it. It's very cheap and easy to make a HVPS using a AC transformer from an old tube amplifier, etc that you've rectified with a few diodes and a cap or two to lessen the ripple. Of course, the end result won't be an automated process like you have now but, you'd at least be able to reform caps rated 500+ volts when needed. Yes, this either requires a Variac to adjust the AC or if you're inclined, you can create a simple circuit to adjust the DC side instead eliminating the need for a Variac. Since I don't reform caps very often (hardly ever), I just use the Variac to adjust the AC voltage to intern adjust the DC voltage. There's many videos on here of "reformers" people came up with. The majority of people who will want to reform some caps don't need an expensive setup like you have. Buying a lower voltage transformer and making a voltage doubler/tripler circuit will save you some money even further but, used transformers aren't expensive and can often be had for free. I just use one I took off of an old Hammond organ amplifier that's 600 volts. Honestly though, I only like to reform "Can" type caps with multiple capacitances and rare ones you can't really buy anymore. Otherwise, I just buy new ones that are generally much smaller and have better temperature ratings.