On Capacitors: Electrolytics

On Capacitors: Electrolytics

Electrolytic capacitors in series with the signal chain.

Electrolytic capacitors are rarely used directly in the signal chain (note the word directly, please — more on that in a little bit). The only time I’ve ever seen them used in this fashion was in certain highly popular OTL amplifier kits. In that case they were being used to couple the speaker to the amplifier while preventing any DC on the output stage to go through the speaker. The use of the electrolytic capacitor was justifiable in this situation, although it was most certainly the weakest link.

First, because an electrolytic capacitor is essentially a “chemical soup,” it has some characteristics vaguely resembling a rechargeable battery — although they (thankfully) do not catch fire like some batteries do…. They also have a certain persistence — called dielectric absorption — which causes them to slowly build back their terminal voltage after being fully discharged, as if they are recharging themselves by magic. This strange phenomenon results because of the lossy electrolyte in them that is used as a dielectric.

The electrolyte in effect absorbs the electrical energy, releasing it much more slowly than the metal plates in the capacitor does. Thus, if an electrolytic capacitor is shorted out, it will be found, a minute or so later, that it has regained some of the terminal voltage. Needless to say, this is not conducive to hi-fi audio!

Second, is the frequency response problem. Electolytics tend to behave badly at both high and low audio frequencies. This again is due to the electrolyte. The electrolyte has a tendency to vibrate at low frequencies, according to one capacitor manufacturer, and this results in a weak bass response. At high frequencies, the electrolyte simply does not let the high frequencies through, which is why most power amplifiers have a small value film capacitor bypassing the electrolytics in the power supply. By using a large value film capacitor in parallel to an electrolytic in any part of the audio circuit, one can bypass much of the critical midrange portion of the audio signal as well as the highs, largely reducing the electrolytic’s inherent negative audio effects.

There are ways to remedy these problems with electrolytics in audio on the component level as well; hence the manufacture of “audio” electorlytics, in which a barrage of different electrolytes and better construction techniques are employed. These do not get rid the problem, per se, but they do mitigate it, and render these capacitors excellent for use in power supplies. And in series with the signal chain, as in the OTL circuit I was discussing? Well, it was certainly an improvement over your basic electrolytic in that role, and I’ll just leave it at that.


Electrolytic capacitors in the cathode circuit.

Now to the discussion of popular circuits boasting “one” or even “no” capacitors in the signal chain, in which they achieve this glorious status by using a multitude of cathode resistors bypassed with electrolytics. This statement of “one” or “no” capacitors in the signal chain is rather misleading. After all, the use of a capacitor in the cathode circuit is in the signal chain, albeit indirectly. This is because its use is to bypass a biasing resistor on the cathode terminal of a vacuum tube. If the capacitor was not there, degenerative feedback would result, causing significant reduction in signal gain. Thus, the capacitor is actually quite critical to the audio going through, and an electrolytic’s inherent “messiness” is not ideal for this use.

Designers of the circuits described above are frequently using several electrolytics per channel. Our tests have shown that it is better to have two coupling capacitors than only one, and a cathode biasing resistor bypassed with an electrolytic. As usual, more marketing hype.

That said, excellent circuits can still be made with high quality electrolytic capacitors in the cathode circuit. But just using them in that function in order to allow one to announce how few coupling capacitors there are is poor practice at best.

We have found that the best use for cathode capacitors is to self-bias a pentode or a triode-connected pentode output stage. Because pentodes are more complex internally, modern ones tend to be much more susceptible to out-of-control biasing episodes. Self bias is the only reliable way to use them, short of the old-fashioned user-unfriendly meter and manual bias adjustment scheme. But enough on electrolytics.

Next up: Coupling capacitors: Ideal Values and Brands. Get ready — we’ll have some mighty interesting discoveries to announce.