Hi Everyone. I was choosing whether to use a Non-polarized cap instead of a standard (audio-grade) electrolytic capacitor. I was wondering if an electrolytic cap needs to have a bias voltage GREATER than the the AC peaks of the signal it is passing? Scenario: You want to block a DC offset say of 0.5V but the signal might reach peaks of 1V. In this set up I see the capacitor is subjected to minus 0.5V on the negative swings of the signal. Question does this matter? OR should I use a non-polarized cap? I've seen lots of audio circuits (mainly on the signal outputs of op-amps) that this might be occurring yet ordinary caps are used.
The cap needs to stay positive with regards to DC (actually it could even go to something like minus 1-1,5 V though that is not advised). But for AC it can always "swing" negatively, within it's voltage ratings. For example, if the electrolytic is rated at 25V then it can also swing to minus 25 Volts peak AC voltage.
That having said, a non-polar cap may have lower distortions for AC signals, especially at low frequencies where there might be a voltage drop across the cap.
Thanks very much MRupp. You have cleared this dilemma with me.....Interesting what you say as regards non-polar caps having less distortion as all the superfi caps are polarised types and the non-polars made by the same manufacturer (Say Nichicon) seem to be a step down in audio quality in their advertised quality rankings...albeit just a little.
The majority of caps are used for power supply filtering, etc,, where polar caps are used. Only when you are using electrolytic caps as DC blocking or coupling caps for audio signals distortion may become an issue. Here non-polar caps distort less than polar caps. You can overdimension polar coupling caps (say a factor of 10) to reduce distortion, or use a non-polar cap right awy since they usually do not cost much more. Subjectively, it is not evcen clear whether you can hear those differences, so try for yourself.
P.S. There is literature on cap distortion, Douglas Self , C Bateman, Bob Cordell, etc..
Thanks again MRupp. As regards using extra large values when audio coupling...I often wondered why large values were used, say 10uF, when in a particular high impedance circuit, 1uF or less will do as far as the low end roll off is concerned. I'll read the write up by Mr Self, as I have a copy to hand. Thanks again.
Depends on the signal amplitude at low frequencies, i.e. signal content below corner frequency. At corner freq. capacitor sees about 70% of the signal amplitude. For example let fc = 10Hz and the signal is 1Vp-p. For a fsig=10Hz, voltage on the capacitor will deviate +- 0.35V from its DC (no signal) level. Deviation will be much lower at higher frequencies e.g. +-10mV @ 1kHz.
So if the corner frequency is well below the low-freq content, capacitor will never 'see' inverse voltage. In practical terms fc = 1 - 5Hz is sufficiently low to avoid this kind of problem.
Hello sonicles,
I need to correct myself: my previous statement that a polar cap can swing to it's negative rated AC voltage is off course complete and utter nonsense, and would most likely destroy the cap.
As an input cap or a DC blocking cap in an amplifier, the impedance of the polar cap is usually much lower than that of the resistor network it is connected to, so the voltage drop across the capacitor is actually pretty low compared to the signal voltage, but in other locations, e.g. the output stage of an amplifier (Zen) that is no longer the case.
So bottom line is that you should probably calculate the expected voltage drop, and if you have more than say 1 Volt voltage drop, there may be a risk of cap failure, at least in the long run. Considering this as well as the higher distortion, you are better off with a non-polar cap for coupling, or you require a sufficiently high DC bias voltage, e.g. for an output cap in an amplifier.
I need to correct myself: my previous statement that a polar cap can swing to it's negative rated AC voltage is off course complete and utter nonsense, and would most likely destroy the cap.
As an input cap or a DC blocking cap in an amplifier, the impedance of the polar cap is usually much lower than that of the resistor network it is connected to, so the voltage drop across the capacitor is actually pretty low compared to the signal voltage, but in other locations, e.g. the output stage of an amplifier (Zen) that is no longer the case.
So bottom line is that you should probably calculate the expected voltage drop, and if you have more than say 1 Volt voltage drop, there may be a risk of cap failure, at least in the long run. Considering this as well as the higher distortion, you are better off with a non-polar cap for coupling, or you require a sufficiently high DC bias voltage, e.g. for an output cap in an amplifier.
Thanks everyone for the extra information, all very interesting.. re: MRrupp. Thanks for the clarification. This is the scenario I was worried about...using a polar cap at low bias voltage. As you say, using a non-polar cap is the safest bet...but most times, space permitting, I would choose an MKP film type. Thanks again.
the absolute voltage on the cap doesn't matter.What counts is the difference seen between the two electrodes. For coupling caps, it is normallly a non-issue because the cap translates the AC component to DC voltage between two nodes sufficiently apart.
When the quiescent voltages on both sides are close, and the AC voltage across the cap is large enough, it can become reversed, which is to be avoided, but it means you use the cap as a capacitor, for a filter or something else, and not for coupling where it simply acts as an AC short.
Using E-caps in such roles is unadvisable, even if it is done with bipolar caps in cheap Xover circuits.
Even if the DC bias is sufficiently high to avoid reversion, a high AC voltage remains problematic, because E-caps are severely imperfect regarding all sorts of losses, which can have a non-linear component
As said above, it's the terminal voltages that matter. To avoid distortion and response errors you'd always use a cap large enough not to have a significant audio voltage across it, so it's almost never an issue. If you're using the cap as a filter, that's another story, but you'd probably not use an electrolytic there if you want predictable and long term stability of the rolloff.
