film capacitor endcaps bonding method?

[coppercone2]

How do they physically construct film capacitors? I assume you can get two conductors with a dielectric in between, stagger them a bit and wind it up, but you are left with layers of conductor with physical space  and dielectric between them.

Do they use like conductive glue to make the end cap? Or have some fancy method of rolling everything up and pressing a plate against it? What is the practical method for shorting out all the foil edges? Or do they bond a thin metal strip to the end of the foil so they can interact with that directly when its coiled up so they don't have to deal with a layers of plastic (much easier and reliable to fold, or even just have it short out while winding). Or is it a ton of bond wires? Laser robot that welds on thin conductive strips that protrude out? Micromachined spiral endplate thats crimped on? laser cut the foil so there is conductive tabs that protrude out the side (but then you still have a layer of dielectric material preventing you from crimping/welding them all together on a pile, and I assume the matalization is way too thin to try to dissolve the dielectric away with a solvent wash.

I thought about just rolling it up and hanging it so its partially submerged in a pool of thin conductive epoxy, but it would have to be staggered enough to prevent wicking to short out the plate ends. Figure epoxy won't wick too much though, so like 2mm stagger will be enough. But its kinda high resistance compared to some kind of crimp or something, and I kinda get paranoid about the capillary action...

I can imagine if you just roll up a film, and its real thin, you might have problems trying to short it all out to make a bunch of parallel plates and not a snake with a weird distributed inductance. I know when people make home made foil capacitors they have a single wire that ususally sandwiches the conductor, but this forms the long long inductance/resistance if its a big cap. If you short it out its like a really fine parallel plate capacitor.

[mzzj]

Metal spraying is the most common
https://metallisation.com/applications/metalsprayingofcapacitorends/

[SeanB]

Depends on the construction method used in the capacitor. The ones made with separate foil plates and insulators, they just leave the one side of each plate projecting past the end, and then either spot weld the tabs to it, or just use tabs that are wound in with the foil each side. However, the majority use a metallised foil, with the one edge of the foil not being metal coated, so the construction is to spray a lowish melting point alloy on the ends ( called Schoop metal) to make a contact to the foils, and spot weld or solder the wires to that. That makes a low resistance connection to all the length of the foil, leading to a low ESR and low self inductance of the capacitor.

Almost all film capacitors are the latter, the separate foils are only used on older oil impregnated paper, or on electrolytic capacitors.

[coppercone2]

Hmm, for signals applications like filter networks not designed for heavy interference do you think you can get away with conductive epoxy?

Is there any literature on  Schoop metal spraying? And why do they not use separate foils? cost or performance?

I thought about purchasing a metal spray guy that runs on acetylene for shop use before, but I don't see this working with thin plastic foil.

Can you maybe dip it in a pool of ultra low melting point solder and have it survive for long enough? Like S-Bond solder, 125-130C. Or is metal spray used because the film metalization does not like sticking by conventional means?

[T3sl4co1l]

"Why do they..?"
"Because it's cheaper."

Also, foil isn't self-healing.

If you insist on making your own, just use extended foil and crimp the ends over.

Tim

[coppercone2]

How do you crimp the ends over if there is plastic between the foil layers? I would imagine for a crimp to work you would need to find foil of the same or slightly less thickness then the plastic and roll it up so it makes a sandwich. Or like I said make a micromachined spiral inside of a metal endcap that you slide on so the exposed metal on one side of the foil touches the endcap, which is then crimped on (I guess you wuold need a machine like you use for doing WAGO wire crimps that closes like a stargate iris ? Thats like 100$ used for a small one, I think you would need some kind of multi-axis grinding machine to make the correct geometry for such a large crimp tool for something like a 1 inch capacitor (say custom made 11uF foil 100VDC)

No idea how you would cut something like a 4 micrometer or whatever spiral trench in a aluminum endcap either. All I can think of is laser ablation.

I have NO IDEA how to determine the required crimp pressures and stuff on that kind of complicated mess. I can see the plastic membrane deforming and sliding on the front like butter trying to push the crimp off

Can solidworks plastics do it? I have the plastics addon.

[coppercone2]

This is the process I am imagining


It sounds like you have a better way, can you elaborate on the pure crimp procedure?

Are you saying they sell foil thats plastic foil plastic foil with an exposed foil edge? Won't the multilayer bonding cause materials stress and weird electrical properties? I thought the plastic should be at the lowest possible tension, bonding multiple layers or making them thicker will increase the stress when bent I thought? I guess you would need to find a resin that has the same temperature coefficient as the plastic too, maybe just put it in a box with compliant mounting instead of resining it, and use braid instead of wires for connection to PCB, if its sensitive enough, I don'[t care to put it in artillery or space

Also maybe you should solder it before crimping, not sure what leads to higher integrity


Does anyone see one of those fancy 'sleek' zip ties doing the business of a crimp in this situation? I mean the ones that kind of seal parallel rather then perpendicular. Especially metal-zip ties.

Zip ties of this style have minimum 'cusp force' as I like to think of it

https://www.ebay.com/itm/PANDUIT-MLT4H-LP316-STAINLESS-STEEL-CABLE-TIES-14-3-LONG-50PK/302829185429?epid=1401500704&hash=item468206a995:g:EowAAOSwzOxUXSxu

https://www.cabletiesunlimited.com/panduit-14-50lb-pan-alumr-aluminum-cable-ties-50-bag-part-mlt4h-lpal.html?fee=1&fep=1808&gclid=EAIaIQobChMIzda3iM6I3QIVxpyzCh2b5AiYEAQYCiABEgL2bvD_BwE

The regular type will have a kind of right angle form that will lead to uneven stress, but the styles listed should have fairly even tension, especially if you lubricate the interface between the zip tie and the capacitor with some graphite or something, and I think the steel ones are fully analog, not some kind of plastic mechanical DAC with little ribs like the plastic ones, I think the metal zip ties are completely smooth and the thing literary digs into them so you can get any tension you want as the force is induced by microgalling (can you do the insane strenght of steel zip ties in that form factory with only nondeformative friction?), I think. You would need to find real small metal ones though.

Maybe plastic would offer better temperature tracking to the foil though. They also have rubber zip ties that will fully conform to the surface you are tensioning so long its hard enough, but I don't know if their mechanical stability is similar to that of foil. Timing belts are pretty good so maybe? They would offer very even pressure around the circumference.

Other ideas?

Another idea I have is to basically use a helical spring that is wound on the right bobbin and hardened correctly to tension the whole thing up without any kind of deformation, just untension the spring, slide it over your assembly and let it self tighten, should be stiff as fuck and extremely uniform without any kind of crazy metal deformation engineering shit, you could do any dimension spring so long you buy a set of all the different kinds of steel pin machinists use (or machine pin standards kit, similar to gauge block kit) and a few different gauges of spring steel wire, so it would be friendly for rapid prototyping without any kind of special dies or crimp tools. I don't see anything getting the uniform tension a decently made helical spring would provide. I am thinking I would need to be a mechanical engineer to design a proper deforming crimp. Maybe you can use alumium or copper alloys or something too to make springs that are weak, since they only tension a soft plastic/copper interface, so they will be easier to engineer then sspring steel springs

Unless you get a fully exposed metal surface without any plastic your not gonna cold weld anything either I think, because the plastic will just squeeze out unless its stiffened with some kind of hard media, and I would consider this a shady crimp, I am thinking this whole capcitor/metal interface has to be like a leaf spring deal.

Can you have a stable metal-metal-plastic-metal crimp interface?

I thought at best you can do plastic-metal-metal-metal-plastic, where the plastic is only temporarily tensioned and not tensioned for long periods of time (wont it start to creep? or is a permament cold weld made at this point between the inital metal metal interface that is stable even if the force of the last metal bit is decreased because the plastic yielded?)