Why polarized cap in audio circuit? (Beginner question)

Hi,
I've studied a few audio amplifier circuits on the web, and have even got one to work on the breadboard (I'm trying to build a colour organ). But I can't get my head around why a polarised capacitor is specified to decouple the audio input from the amplifying circuit. Why wouldn't a non-polar cap work?
Is it something to do with current leaking across the plates?

You are mixing descriptions; there are bypass capacitors (such as an emitter bypass in a transistor amplifier), and then there are coupling capacitors (such as coupling between stages of amplification, if the DC level at the source is different than at the destination).

If you use a polarized capacitor, then the + sign needs to be connected to the more positive node in the circuit. If you use a non-polarized capacitor, then polarity obviously doesn't matter. Non-polarized capacitors are rarer, bigger, and more expensive than polarized ones.

MikeMl said:

You are mixing descriptions; there are bypass capacitors (such as an emitter bypass in a transistor amplifier), and then there are coupling capacitors (such as coupling between stages of amplification, if the DC level at the source is different than at the destination).

If you use a polarized capacitor, then the + sign needs to be connected to the more positive node in the circuit. If you use a non-polarized capacitor, then polarity obviously doesn't matter. Non-polarized capacitors are rarer, bigger, and more expensive than polarized ones.

Click to expand...

Thanks for the quick reply.
The capacitor I see in many circuits sits between the input source ( a mic, or mp3 player or something) and the base of the amplifying transistor. The +ve side of the cap is connected to the transistor base. Now, as I (obviously don't) understand it, the input source is oscillating +- a small number of millivolts. The base of the transistor is biased to just about 0.7v. As the -ve side of the cap goes negative, the positive side goes correspondingly positive, which sums with the biased base volatge. This causes current to flow and we've got some amplification. But as the signal goes into positive, then the transistor side of the cap goes a little bit negative, so the base of the transistor is now at 0.7V - a few mV. So current doesn't flow between collector and emitter. What I don't understand is what function the polarity of the cap plays. I can see that the positive side of the cap always sees a positive volatge ranging from 0.7 - some mV to 0.7 + some mV, but why wouldn't a non-polar cap (of the same capacitance) do the same job. Is it just because they're more expensive?

The nonpolar cap WOULD do the same job. Have you tried buying some non-polar capacitors in the several uF range?

MikeMl said:

The nonpolar cap WOULD do the same job. Have you tried buying some non-polar capacitors in the several uF range?

Click to expand...

Aha. So it boils down to a question of cost then. Oh well, at least I can stop searching for some magical property related to polarity and move on to the next problem with my project.

Thanks for your time, and have a good evening.

mknott said:

Aha. So it boils down to a question of cost then.

Click to expand...

And size. Non-polar capacitors are generally physically much larger than polar (electrolytic) caps.

Wait a minute.
Today, very few audio circuits use a transistor at the input because it has a low input impedance (which requires a high value electrolytic input coupling capacitor) and has high distortion. Most use an opamp that has a very high input impedance so a cheap and small low value (220nF) film non-polar capacitor can be used. A 220nF/63V film capacitor is smaller than a 10uf/16V electrolytic capacitor.

I know. There are some tiny 10uf/16V capacitors but they dry out then fail in a couple of years. A film capacitor lasts "forever".