Inside an Electrolytic Capacitor

[willawake]

Lytic Caps Unwrapped by PeteS in CA

What's inside an electrolytic capacitor? In principle, a capacitor is two metal plates or pieces of foil, with a insulator (dielectric) between them. While this is, in principle, true of an electrolytic capacitor, there’s more than meets the eye. Literally and metaphorically, I'm going to take an electrolytic capacitor apart, and describe what is found.

The electrolytic capacitors I work with are larger than those usually used on PC MBs. The latter are typically 8mm, 10mm, and 12.5 mm in diameter, while the ones I work with are usually 16mm or 18mm in diameter (though I have taken apart a 12.5mm part using the technique I'm about to describe). First, I use an Exacto knife and needle-nose pliers to slit and peel off the outer polyester sleeve. Then I slip one of the points of a pair of diagonal cutters between the rubber plug on the bottom and the lip of the can, prying up the lip and doing that all the way around the can. Using the diagonal cutters again, I make two or three cuts (~1/16 of an inch) into the lip of the can. This allows me to use the needle-nose pliers to peel down the can at each cut, tearing the can open down past the rubber plug, so that the plug is no longer retained by the can. Now the capacitor element can be removed from the can. While the core will definitely be wet or moist, there won't be a lot of liquid. Holding the lead wires between the index finger of one hand, I very firmly push the rubber plug toward the ends of the leads, and when it comes loose (some effort is needed, especially with Nichicon products), I slide the plug off the leads. In a 16mm height part I unwrapped recently, the height of the rubber plug was about one third of the height of the can. The capacitor core is two strips of aluminum foil (with leads swaged onto the foil) and two strips of paper in alternating layers, rolled up like a jelly roll. The paper retains and diffuses the electrolyte through all the layers of the jelly roll. Tape is used to keep the jelly roll from unwinding. Removing that tape allows unwinding the layers of foil and paper. Being both wet and very thin, the paper is very flimsy.

So, you have the two strips of foil for the plates of the capacitor, or so one might think. What is the dielectric insulator? The paper? The liquid electrolyte? Neither. The paper is just for retaining and diffusing the electrolyte, and the electrolyte is very conductive. The dielectric is actually a thin film of aluminum oxide which is "formed" onto the anode foil. Technically, the cathode of the capacitor is the electrolyte, and the second strip of foil is used to make contact with the electrolyte. It's going to get stranger shortly. If you haven't mangled the two foil strips by now, you may find that one is slightly thicker than the other, and the surface is a very dull gray. That is the anode foil, and the dull finish is important. The other foil is thinner, and may be almost shiny. That is the cathode foil.

At least one question should have occurred to you. If the electrolyte "forms" aluminum oxide on the anode foil, why not on the cathode foil as well? The answer is that it does, but only a little. The anode foil has the oxide layer "formed" on it long before it goes into the capacitor, by a different electrolyte than what is in the finished capacitor. After the foil is "formed", it is slit into strips, wound with the paper and cathode foil, taped, placed into the can, and the electrolyte is added. The anode forming process is controlled to produce an oxide layer of known thickness and breakdown voltage. While the cathode does get "formed" some, once the capacitor is assembled, its oxide layer is very thin. What you have, then is two capacitors in series: the anode foil capacitor, which is slightly higher than the nominal capacitance of the finished part; the cathode foil capacitor, which is 10-100 times the capacitance (the oxide on the cathode foil is extremely thin) of the anode foil capacitor. When you have two capacitors in series, the formula for the net capacitance is 1/(1/C(C) + 1/C(A)). As you'll find out if you plug in sample numbers, with C(C) being > 10 x C(A), the result will be slightly lower than C(A).

What’s with the dull appearance of the anode foils, and why is this important. Several factors determine the capacitance of a capacitor: the dialectric constant (an intrinsic characteristic of each dialectric material); the thickness of the dialectric; the area of the plates. The dialectric constant of aluminum oxide is very high and it is relatively thin, so aluminum electrolytics have high capacitance. But manufacturers go a step further. The effective surface area of the anode foil is increased by etching pits into the foil. Since the dielectric is formed chemically, it conforms to all the hills and valleys of the surface. The etching is controlled so that the spacing and depth of the pits are dependent on the capacitor rated voltage - getting maximum surface area without the dielectric filling in the pits.

contributed by our friend PeteS in CA

[willawake]

Pics by Jim Arneson

[willawake]

Pics and comment by Topcat

Havent we all wondered what a dissection of one of these crappy caps would look like? Well I thought I'd find out....

The bad capacitor used for this was a JPCON 2200uF 6.3v low-ESR cap removed from an Abit KA7-100 motherboard.

Before the 'surgery', this capacitor looked like just about all typical bad capacitors, bulging from the top and the canister seal was protruding from the base. I carefully cut the canister open with a razor blade, so not to further damage the insides. Upon pulling out the paper innards, I noticed that the paper dielectric material was severely scorched and completely dry. I then began to unwrap it to gain access to the two metal plates and the paper filling just crumbled and fell apart in my hands. Its brittleness and crumbling reminded me of very old newspaper. When I finally got to the plates, they were slightly melted and disfigured, as the capacitor had obviously gotten extremely hot.



Ok, this next capacitor was a good one to use as a demostration of what it SHOULD look like... It was a perfectly healthy Nichicon 2200uF 16v low-ESR.

You can see the obvious color difference! Everything is completely intact, and was soaking wet with electrolyte! You will note that the paper filler and the metal plates are all in one piece and intact.



Ok, if you've ever wondered, now you'll know...

Pics and comment by Topcat

[willawake]

Comparison of Panasonic FJ and FC by BluBlur

[willawake]

Chhsi / Lelon and Gloria / GSC by Willawake

[PeteS in CA]

Unwrapping fried lytics is not pretty. Basically, the electrolyte is conducting the ripple current, an amp or two. If it is losing electrolyte, less of it is still trying to do the same work. It gets hotter as it deteriorates, accellerating the deterioration. The paper basically separates the two foils and retains and distributes the electrolyte throughout the core. It usually is pretty flimsy, so heating it to >>110C is not going to strengthen the paper, . On that one cap where you said the metal flag attaching the lead to the foil had been damaged, that sounds like really poor production processes. The connection of the foil to the metal flags is accomplished by the little swages. These are mechanical contacts, not soldered/welded; multiple swages on each flag makes for multiple contact points. Since these are conducting the ripple current, if they are done poorly or the foil gets cracked/torn, the contact area will get hot and become a point of failure.

[Maxxarcade]

I'd like to see inside one of the big old soup can sized caps, but I'm too chicken to dig into one. Isn't that stuff toxic to get on your hands?

Still amazes me how much smaller caps are these days when compared to their older counterparts for the same value and voltage rating. Does anyone know how exactly they would make something so much smaller have the same amount of capacitance? I'm assuming there's more than one reason.

[Super Nade]

The most common reason would be to use a higher k dielectric. The capacitance of any cap is proportional to k, the dielectric constant, so with higher k, you can decrease dimensions. However, high k caps have larger dielectric absorption, which implies the charging-discharging curve are "slow start" and there are transient fields even after steady stay is reached. This is not very good in high speed applications.

[PeteS in CA]

I'm not so sure whether the K of the dielectric has changed - isn't it still aluminum oxide? OTOH, the process of etching the foil, especially the anode foil, has certainly gotten more sophisticated and better controlled. This increases the surface area of the foil.

[arneson]

Caps in my opinion have not really shrunk that much.
The operating votages are lower and you don't see the combo's so much anymore.
One of my favorites would be a 10/120/200 mf at 200wv all in one can twisted bolted and soldered into the chassis. One section would be blown and the other two would follow.

[PeteS in CA]

With a 5U4GB rectifier?

[arneson]

Full Wave.
No, with FET's and power management controller.

[arneson]

Pete, Why didn't those tubes have a fan and thermal paste mounted on top?

I used the tv to heat my room back then.

[denyu]

MMmm maybe not that hot...