Hallo Tim,

In addition to that, what other repliers stated, I sent you this:

If you have a look into a datasheet, you will find, that for
electrolytic capacitors reverse voltage up to 2V can be tolerated. The
reason is, that there's a passivation process, which prevents normal
formation at lower voltages. But if reverse voltage is greater than 2V,
so called 'formation' takes place, producing considerable heat. This
condition must be prohibited!!

Tantal capacitors accept NO static reverse voltages!!!!!! Datasheet
tells, that reverse voltage can be tolerated, if this is limited to less
than 0.15 x Un, BUT ONLY FOR SHORT TIME! But they don't tell, how long
this 'short time' lasts. So, forget this possibility of having reverse
voltage.

By the way, when applying reverse voltage danger comes from heavily
increased current flow. And if this is limited somehow, damage can be
prevented to some amount. I have seen some circuits, which were designed
like this. BUT THIS CANNOT BE RECOMMENDED AT ALL!

In the schematic above, you can see a simple methode to fabricate a
bipolar electrolytic capacitor by the help of two standard electrolytic
capacitors:



Technique is called 'back-to-back'. With this configuration voltages of both polarities can be handled.

At the right of schematic you will see simple circuit, which is often used in analog electronics. At first sight one might think, that for big amplitudes (think of a sinus with amplitude of about 14Vs) also dangerous reverse voltages are applied to the electrolytic capacitor. At least for the negative going part of sinus.
But this is not so. Because RC unit acts as a frequency dependent voltage divider, you will find big negative voltages only for very low frequencies. And if the content of very low frequencies is limited somehow, e.g. by the help of some high pass filter earlier in the signal path, then indeed a polarized electrolytic capacitor can be used like shown in the schematic. I have seen big studio mixing consoles, not containing any bipolar electrolytic capacitors, only circuits like shown in the schematic.

Voltage dropping across electrolytic capacitor can be calalculated by the following formula:

Uc = 1 / SQRT( (2 x pi x f x R x C)**2 + 1) x Uout

In our example at f = 100Hz at capacitor only 0.22Vs will be observed with Uout = 14Vs. For f = 20Hz only 1.11Vs is observed.

If resisitve load is much lower or if lower frequencies with relevant amplitudes are to be expected, capacitance shoud be increased. Other option, of course, is the use of a bipolar type.

Kai