lproven | How capacitors REALLY work (Reply)

By way of an update to my earlier post. Mailing list correspondent Doug McNutt sends this correction:

That's a bit of oversimplification in a helpful description aimed at less than engineering folks. The "short charge/discharge cycle" is misleading because it implies electrochemistry.

An aluminum electrolytic capacitor depends on a very thin film of Al₂O₃, aluminum oxide, which is formed by passing a DC current through the metal and an acid, usually sulfuric, which is placed as a paste between the aluminum foil "plates" as they are wound into a cylinder with some paper spacers.

It's the same process that is used to anodize metal drinking cups. The cup is hooked to a wire and immersed in acid. An electric current, with the cup positive, causes a controlled oxidation of the surface. After winding, the plates of a capacitor are similarly formed by applying a DC voltage to anodize them. It's the very thin oxide spacing between the plates that gives rise to the large values of capacitance. In use there is no electrochemistry going on except perhaps that the DC bias must remain in the correct direction to avoid dissolving the thin oxide layer.

The acid remains in the envelope which has to be vented so that the capacitor is not altitude and temperature dependent. Unfortunately that means it can leak. Some capacitors are better than others but the best ones are more expensive. Rated lifetimes for aluminum electrolytics are in the range of a few thousand hours at temperature - much less than (2005 - 1984) years.

Another technique is to form the oxide layer in the pores of a sponge created using the techniques of powder metallurgy. Aluminum is replaced by tantalum and the resulting product can be sealed up. The technology had a long infancy with many engineers insisting that tantalum-free is good. Modern tantalum sponges are much better. They don't leak and they are useful to higher frequencies than aluminum foil devices. They are also a bit more expensive but still cheap enough to be my recommended replacement for SE/30's.

Douglas P. McNutt PhD
The MacNauchtan Laboratory
7255 Suntide Place
Colorado Springs CO 80919-1060
voice 719 593 8192

http://www.macnauchtan.com/
ftp://ftp.macnauchtan.com/