Wow. ah hello, Hi! In a recent video if you haven't seen it, which I'll link in, We uh, repaired an Alec um, heater and a spoiler alert. Uh, it turned out to be a failed X2 uh, cloud class metallized uh poly? put the kettle on uh, filter capacitor or in this particular case, it was used as a dropper resistor for a zener. And by the way, don't mind the voice and music in the background if you can hear it.

I'm just like on hold with a bank. So yeah, it's going to take an hour. So I thought I'd shoot a video. so that was this one here.

and um, I tried to get a replacement one from another uh, power supply and it was a Corsair vs. 650 power supply which had failed. and no, this wasn't the reason now for the fail. Um, but I got the actually the exact same one out of it and both of these had drastically dropped in capacitance.

And that's to be expected if they've been surged overload because that's how metallized, self-healing polypropylene. Uh, probably put the kettle on capacitors actually, um, work. If you put an impulse into them, then they'll get a little energy discharge inside of there which will then have a little plasma arc and it'll just vaporize the metal and then it, uh, it won't short out because that metal is vaporized, it'll actually just form an insulative hole there and it retains its capacitance so it won't short out. But do this too many times and you lose too much square area of your capacitance.

And uh, that's how these self-healing caps work. Anyway, a few people asked, could I actually take these apart, put them under the microscope and see if we can see any damage? Well, um, I don't know is the answer. I guess we'll find out. So before we get Medieval and Sas, we'll just measure it here.

Here's the Uh failed one from the Alec heater at one Kilohertz: 103.5 Farad. So that's less than half of the 220 uh, Nano farad uh, nominal value in there. and the other one is absolutely plummeted. 109 19.2 Nanofarads.

Wow, more than the order of magnitude and the Esr is up to, you know, oh, looks all over the shop. Yeah, the Esr is better at uh, a hundred kilohertz. so I'm just testing this on a good one. This is a Uh, got this from J Car.

This is a sun tan. I took the side off it there and you can see that. Yeah, it's pretty well. the potting compound has gone all the way in there.

Anyway, I forgot to tell you, this is a Utx brand and we do actually have uh, the data sheet. uh for this one that we're going to attempt to tear down. It's in Chinese, but we get the deal. In this particular case, it's a polypropylene jobby.

so they're metal. uh, metallized film. They just call them film. Capacitors is the Uh is the generic Uh name and it's got a metal sputtered layer on in.

this particular case, the polypropylene film. And there we go. So we expect it to be wrapped in there like that. The interesting thing is is that you can actually get different failure characteristics: internal plasma, um, arcing in there.

Um, when they're tightly wrapped or loosely wrapped. So a tighter wrapping actually gives you a different effect in the way that these things fail apparently. So yeah, that's interesting. So let you know.

a smaller one like this, a two to four. They might be the same voltage rating, but a Uh 220 nano farad in this size might act and survive very differently to a the same one. Here, 220 nanofarads in a much larger size because this could have a thicker metal layer and a larger or a less tight wrapping there. so less tight pressure, uh, between the metallized Um surfaces.

So yeah, it can make a difference apparently. Well, there you have it. I just accidentally broke that off, so that answers that. Yep, they are rolled exactly like I suspected.

I don't know why the wrap looks like that. it's got. maybe that's just part of the foil manufacturing process, but you can see that it's got stripes on there, which makes it look like you know it's stacked in this direction. Like this.

it's like a one stack two stacks. but it it doesn't. You can clearly see the wraps in there, so maybe that's the way to do this. Sort of tear the legs off.

Looks like we haven't done too much damage right? So I'm just testing how to get these things apart and then looks like that has a contact which then goes over into the middle. Okay, so let's get Medieval. Oh I can't Nope. I know I'm gonna need a bigger boat.

The potting is too good. The way I did it last time is to kind of like yeah, get in there like that. But yeah, this potting hasn't really taken to the foil, So potentially ah yeah. there we go.

We can't actually get this off. Ah, Bobby Dazzler, you're that all right. So you just have to be gentle getting the end plates off. Although the em damage is not gonna be in the ends, it's going to be in the middle somewhere, so I assume there's nothing but pressure holding them together.

I don't think they'd apply an adhesive when they wrap them. I think they just wrapped them. And Bob's your uncle, All right. Can we see where the wrap starts? Is that is that it? So that's an additional plastic wrap? Maybe we can.

Okay, yeah, we can get under all that. Don't worry, I'm not going to stab myself. I am a professional. Ah, okay.

there we go. Is it starting to look metallized? See if I put a bit of light under that you can see. Ah, it's starting to get more metallized isn't It Looks just getting more and more That's that's interesting. So you've got to remember this is a brand newy that'll pull it apart.

Okay, so this one should not have any fails. Finally got through to the bank. By the way. Uh, have a look here.

I put some gloves on. So now you know that's not the metallized layer I expect, right? But there's obviously something I expected. Like a solid, not not to be able to see through it at all. So you could say that it's getting more and more.

Maybe we're not actually into the cap. That's just not what I was expecting at all. Roll and roll and roll in. and wow.

Oh, here we go. This is what we expect. There you go. So several layers in.

Oh whoa. hello. It's like you got like bad toner or something like that. Um, yeah.

you really have to have the light behind here. There's um, stripes? By the way. that's just my, um, adjustable color temperature. Um, studio light and turn it all the way blue.

Look at that. isn't that neat. Ah, here we go. Come on, Is that just another? I don't know.

That's okay. Oh, hang on. Oh that ripped, very thin polypropylene layers. Oh god, it's doing it again.

No, it just keeps just keeps going. keeps tearing like that. Yes, it's getting frustrating. Am I back onto a perfect roll now? Yeah, you can see one metal layer, there's some pull back there, and this is in the construction.

You can see it in the color there. But yeah, they do actually tell you the pullback distance in some data sheet. It has a specific name. I'll spear the gory details.

Yeah, I'm curious to see what happens when it gets to the middle. I presume that there's a some sort of small core it starts on. Unfortunately, unless you start the un rolling process perfectly, then you're gonna come. Agata, haven't peeled off this these side bits properly.

So ah yeah. really ripped it as I've so that really wasn't wasn't the way to go about it, Was it? We're not done yet folks. that's not all folks. Um, no.

we have to keep going. That's one side and then the other. You can see the other side clearly poking through there. So and then and then you can see the polyprop film.

You can see the reflection, the poly pop film on top of that. So they manufacture this as a completely insulated thing. It's almost like powdery, isn't it? It's going to be really interesting to open the, um, dead one and see how it drops that amount of capacitance it must have. Like, I can't imagine how many like huge holes it must have in there.

When you talk about these things, you know, blowing holes in them and self-healing and stuff, it has to be on a small scale. I've still got a ton of wrap left in here. Oh, it's torn again. It's hard to unroll these things without tearing.

You're getting the idea. What's inside one of these puppies? Okay, well. I certainly don't need these gloves. so it all seems very clean.

So that's good stuff going. This doesn't end. It does not end. This is too much fun.

There's an art to unroll in these and I don't have it yet. This is my first one. I'm losing the will to live. Help me Larry, Help me.

You have to be in on the dumpster. Joke on Twitter to know who Larry is. Well, you can't say you're not getting your money's worth in your capacitor. I can't.

Come on. there's there's got to be nothing left in there. There's more layers. No.

I. I think we're done. I can have a go at. uh.

taking apart the failed one. Like surely. if it's lost half of its capacitance, right? you'd expect to see tons of like huge blowholes in this foil. Um, because it's a.

you know, it's a wrap. Like it. It's a square surface area thing. This is going to be intriguing.

I don't know how you can lose half the capacitance in there. Okay, so here's the Utx one that's lost half of its. uh, capacitance. So this is the one from the heater Gonna have a crack.

Literally. There we go. Of course they're all going to have varying amounts of pot in and stuff like that. but there you go.

It's basically the same construction as the other one, so these are different brands. Really adheres to the casing. way better than the other one. This one seems like bonded.

You can see here how they do the end contact and uh, this is called shoopage named after Max Shoop. It's kind of like a pasty mixture kind of stuff that gets put after they put the wrapping on and then the contacts extend all the way to the edge of the film on each side of it. So you've got one that extends all the way over here. one that extends all the way here.

You've seen that before with that overlap gap, uh thing in in the film and then, um, all this pace, which is, uh, it used to be like a tin lead, um kind of just like a solder, uh, paste. But of course Ros compliance. They don't have the lead anymore, so it's like a tin, zinc or something aluminium, uh maybe. And yeah, they mix all this stuff together becomes I believe some sort of like a pasty type thing and you can see the film in here.

but you can see how the end here. This has actually formed all the impression here because it's it's just got inside each like a wrap of the film and it's just it's just peeled off like this. But this is how they do the end contact. There's not just one end contact point because if there was, then you'd have the giant, uh, you know, roll like this and then you'd have huge self-inductance in there.

This is why each layer is effectively um, stacked so that, uh yeah, you get low inductance um in these capacitors. So that's how you can lower the manufacturing cost of this. of course by wrapping things. When you have machines that rotate and wrap, it's you know it's very Once you've got that refined it's you know it's very quick and easy to manufacture.

like huge number of plates like this, but this is how they make contact and yeah, that stuff just peels off when you actually break that. So you know this is the uh, polycarbonate. uh, probably put the kettle on film and it just sort of gets in between all the gaps and then the pin is just uh embedded in that as well. and they probably you know, bake it at a temperature and do whatever, um to make it, um, set in there.

But there you have it. That's stupid and that's why it appears to break off very cleanly like this, because like it kind of sort of adheres to the metal uh, layers that are on there. but when you break it apart, it's just quite soft and it's just, well, it looks like it. It's like just totally sheered off.

It's just peeled away from the plastic film and then the metal layer inside there. It's neat. Really neat. process.

I like it. Shoe beach? that's that's too perfect. We've got the same stripe effect. Nothing looks as good on camera as it does with your mark one eyeball.

I think we're gonna have to sacrifice some outer layers here. Needs of the many outweigh the needs of the few. same sort of thing or much different secret sauce between the manufacturers. It's interesting how oh yeah.

there we go. Whoa. Surely that's oh. Is this what? Where is this? It is.

this it. This is very consistent. Look at this. Look at this.

Wow. Ah hello. This is all the way through. This is not my, um, maybe that cut or that cut there that that.

That could have been my knife. but look wow. look at that. Whoa.

Look at that huge hole there. But it's interesting. This strip down like this looks like I don't know. Ink lots right? If you're a psychologist, how do you feel Dave? Tell me about your mother.

What's the first thing you see in these in these ink plots. It's unusual that this pattern is consistent all the way through. What the heck? And look. We got tearing again.

Surely I didn't gouge it that much. Yeah, maybe I dug into it there. Okay, what's all this squiggly stuff in the middle? It's almost like it's split, so we're getting smaller and smaller gouges. That's what we'd expect until they go away.

So that's that's. obviously my, uh, heavy-handed medieval technique. This pattern is still all the way through it. If you want to see the other side looks exactly the same.

Okay, so this one's lost half of its capacitance. This is not what I expected. Like if I expected, you know, I thought I'm sure it will maybe see something. But I I really didn't know what to expect, but it certainly wasn't this.

But also, all of this blotchy stuff here is another thing, right? That like that's another thing entirely. We just did not see that on the good capacitor. But once again, you need like a control capacitor, the same model, the same brand, everything to pull apart. But there's something like that is just weird.

Here's like the blow here's like the holes, you know. But unfortunately we can't get their film apart right? There's nothing visible on the surface, so whatever's happening, it's happening inside the film that this is all blotchy. And there's holes everywhere, right? And thanks to one of my viewers, Elekami Wolf, who saw my live show where I uh, previewed some of this, um, tear down check out his uh channel here. It's um, it's actually really good.

At only 153 subs, it deserves a lot more subs. There's like great tear downs and all sorts of stuff on here so I'll link in down below. He actually had the exact same Utx branded cap and pulled it apart and this is what we get. and this is like more kind of what I expected.

I expected like big, sort of like, you know, blow chunks taken out of it. You know, from like the impulses and stuff, but you can see it's very sort of blotchy in there as well. And here's another shot here. and this is, you know, kind of what I expected.

like big chunks taken out, but he's unsure whether or not it might have been the peeling process or something like that. But yeah, it's anyway. it is very different, even though exact same brand uh and model cap as my one. and I've got that wiggle wiggle yeah through the middle of it.

We don't see any of that here. But yeah, he's unsure if this was like, um, like when he teared the Um film apart and like it just like peeled off blotchy and stuff like that. Anyway, it's very interesting, thank you very much. That is quite different to my one.

It doesn't look like anything's jumped through the, uh, the top and bottom polypropylene layer. Whatever's happening is happening within. You know, how they manufacture, they're real like this. um, you know.

so they make it that that that's your capacitor. But it's all to do with the difference between these two. But why is it all blotchy and and the squiggly lines? It's just all the way through. Did I physically stress it? And it's and it's torn.

It's stretched the polyprop and that's cracked. the metallization was that done due to my heavy hand and this. It's hard to see how it would be or wiggly wiggly yeah like that. And also that doesn't explain the blotches.

So so all I can tell you. this is certainly very different to the control one that Suntan brand. To be. 100 sure though, I would need another Utx brand.

Perfect. I I tore it there and you can I separated the layers. Look at all those blowholes. Wow, Gotcha right? Surely that has to be what happens.

But um, yeah. like when I did my last video I I published that. Yeah, it's actually like it forms a little plasma in there when it over arcs because it vaporizes the metal and it looks like it does vaporize the metal and that causes a uh, decay in your capacitance. But it doesn't short out, it just blows these things up like it.

just like the metal is just gone, right? The metal is just vaporized away and there's just uh, pressure in there and it just. I don't know how it escapes, but it's just right. It's vaporized away. and this is the self-healing nature of these capacitors.

The last thing you want, right? Your capacitors? If it shorts out, right? Bingo. Your uh product's going to go banksy right? When they get an impulse like this, it damages them and causes a degradation in the capacitance that is clear as day. Wow, that crack though. like all the way crack all the way.

I don't know. Leave it in the comments if you can explain that. Uh, in fact, I do have access to a capacitor expert. I might ask him.

These are just like Kiama size blowholes. This is insane. But then again, to lose half the capacitance I like. You really have to lose like half your material.

Here's the next victim. This is from the Corsair uh, power supply and exactly the same capacitor. Exactly the same model. Everything's are the same, I believe.

So Apple's Apple's comparison, but this one has lost or over 90 of its capacitance. So the other one we saw, you know, roughly half the material was missing. Will we see like 90 of the metal missing from this one? Let's find out is our layers. Are we getting into the good stuff? The crack again? Look at this.

but the crack is bigger. Maybe that's just tearing from the unwrapping process. It just seems weird that it's like only a small section though. doesn't seem right.

I mean, look, there's hardly anything left in here, but that's oh yeah. So we just separated, separated the two layers there. Maybe I've unrolled the wrong layers, but then you've just got you know, holes like that and what the heck you know? Like this, this is not peeled off somewhere else, right? There's no matching thing on the other side, so it's almost as if this has happened within the film. But yeah, it's got this weird cracking like there's just nothing left in there.

Look, how does this have any capacitance at all? It's got that weird gap all the way down. I can't see how that's a result of me unraveling this. Anyway, I think we've seen enough. I think it's time to.

uh, phone a friend, phone an expert and see what they have to say about this. Okay, well, as it turns out, it looks like I'm wrong. So we have an expert here who's joining us. Ron Demko from Avx.

Thank you Ron. You're going to tell us what I've got wrong because I thought this was a self-healing issue. But you don't think so. Yeah, I think we could.

We could give you some details on on the real event. All right, thank you. So what's your role at Um, Avx? How long have you been there? Well, I just passed 40 years? Yeah. Wow.

Yeah, Well, you know it's funny though, because the company changes so every three or four years we buy a new division. So at one point at time it was just capacitors. Now it's antennas, connectors, uh, modules, a whole bunch of other things. So yeah, right.

I work in a small group in R D and we do engineering support things like that. Got it? But you're mostly. But you're the capacitor guy, right? You're in the capacitor. Yep, that's your specialty.

40 years in 40 years in capacitors? Well, yeah. Different types. Glass, tantalum, electrolytics, ceramics. Tell us your opinion.

What am I seeing here? What do you think this problem is Because you don't think that it is, uh, being destroyed by impulses on the mains? That's right. Yeah, well. I think we we should start going back and looking at the part. So mainly, these parts are non-hermetic They're using a low-cost epoxy, so you know they're very cost-effective easy to use, easy to choose.

So it's a probability that moisture got in in the application and then that moisture was trapped with that non-hermetic case. So what we have? Well, there's also another scenario there could be poor drying in the manufacturing process, and possibly materials where you know water was put in there. That's very unlikely. So I think what we've got is a couple of easy explanations.

Uh, of course, those small holes. Or maybe the big hole that you have up on the upper right there right is that's a result of actually self-healing Okay, so that's the vaporization. As you can expect. Now, when we start getting into the other area, we can explain things kind of easily.

As you roll this thing out from the left to right, or right to left, you'll notice that there's a variance. Some of that variance is due to the difference and the connectivity from the shoopage on the termination to the actual electrode. So at one point in the video you showed a very light uh, background with some speckled Um scenarios and that was just moisture and a high resistance scenario. But what we're seeing here is in the the the crack lines, uh, the the lightning bolt looking thing.

Yeah, that goes right. That almost went through the whole thing. You're having this electro, uh, well. this reaction of the moisture and what's occurring is that you're getting a uh, a moisture effect that's actually causing the electrodes to change.

Uh, they're getting more resistive. They're getting some some migration or movement, oxidations, changing resistances. So those voids are most likely to some extent like a micro healing or a impulse degradation. That's how it explained that.

Lighter center section and the darker blotches are the most likely a result. Yes, yeah, those are going to be mostly a result of end termination corona type effects. Something on that order. Um, so it's always difficult to comment.

Maybe what? Fifteen twenty thousand miles away or so? Yeah, but um, and not having dealt with the parts. but I think that's pretty much the scenario. It's definitely a moisture event. So will the moisture like eat away at the metal? Will it like actually corrode the metal? And what what sort of metal are we talking about here? For the metallization layers? Well, in some cases there there's zinc, but this most likely is not the case.

It's usually aluminum. It's probably on the order of 100 angstroms the dielectric thickness itself. So the film itself is probably in the order of maybe 1 to 20 microns something. I like that.

so it's a wonderful part when they work well. and I suppose it worked in a sense. It worked well. This one is only a year old.

See, that's the thing. That's why, Because when when I think of moisture ingress, all I think of is like old school reefer paper capacitors which are famous for. You know, getting moisture ingress, they crack the cases crack, moisture gets in, it is absorbed by the paper and then when you go switch it on, you've got this big conductive path that just goes boom and the magic smoke escapes. Um, I I just didn't expect that this that moisture ingress with film caps after like a year in, you know it's the actual product it came from is only a year old.

Yeah, I can make a couple of guesses on where the Um. moisture came in and that's a big issue. Choosing the right epoxy is is tough and I'm sure these guys build great capacitors It It could happen to anyone. But um, the right material system's the it's a it's a real secret, right? Yeah.

and is that something you can test like after you've constructed them? Yes, that's a very good point. So how do you do that? What we could do? Yeah, we'd put them on a moisture humidity test. Uh, you may be. Maybe it would be good to have 40 degrees C and I don't know.

40 Rh, maybe another 40 degrees C and an 80 percent Rh Brutal Humidity cell and then do an 85 85 cell and what we would look at. We would look at the degradation of capacitance across time and that would tell us a lot of things about uh, what's happening within the device. uh, that would probably go in with Edx and we would do some uh, x-ray, uh, evaluations of of where the material systems have broken down and and constituency of water, etc. Because as you said, some of those small little points in there might be self-healing Is that certain? Those that is, is that the size of hole you'd expect from self-healing as from an impulse thing? Yes, yeah, it's pretty neat too to uh to throw these on a sample scope and and apply i don't know a couple of different signals to them, and you'd actually see some noise occurring when we have self-healing events.

So that'd be a neat though, I don't know noise. What sort of noise are you talking about? Well, you'd actually, you know, apply maybe full voltage at maybe a moderately high temperature and you might actually see some failures punching through and you'd actually see some variation in the applied voltage of the part as this thing Coronas. So when you get those little micro uh healing points, you actually vaporize that electrode material and as it cools it precipitates around that failure site. So if we could blow this up a little bit more with the Sem or something like that, we actually see a little dark halo around that white point of failures.

So this large hole here? I don't think that's a blow. I don't think that's a self-healing blowhole, Would it. Would you get one that large? Well, you could. It depends what hits it.

I think that's maybe more unlikely. Maybe I've torn it. there. Is it possible that I've torn this when I've taken it apart? Yeah, I think he did a good job on the Dpa, but I think we might have approached it a little bit differently.

right? Okay, well, it was my first time. Okay, so when you like the ones that where we tour down here, where's the ceiling point? Is it The potting? Is it when they potted? Is that what they're trying to do is that they're trying to seal it as best they can? Yes, Yes. uh. It would possibly be at the perimeter of the metallization to the uh, encapsulate, right? Okay, and then through the uh.

shoopage on the side. Really, it'd come in through the side. It wouldn't come in through the actual film wrap, would it? Because I would imagine that that's pretty sealed. It's correct.

That's correct, right? It would be at the interface of shoepage or some errors or voids and stupid. Thank you very much Ron. This has been awesome. Um, I was completely wrong.

I just went into this video with the mindset that it's all self-healing and I didn't think about moisture. Who knew, right? And so this is going to be a surprise to a lot of people. I think that surprise, it's no, no. Well, it it introduced something to think about.

Um is like should as electronics designers. Should we avoid? You know if we use them in these sorts of applications? Um, we're like a capacitive dropper. Should we avoid like a no-name brand sourced cap for this reason? Or are they all pretty good? Or I don't know if I could comment about that? They can't comment. Okay, just but I have to say films are a good part to use and it's probably a great application for that.

All right, it's been awesome. Thank you very much Ron.