I understand that capacitors can have the property of filtering certain frequencies of signal, and I understand they can be put on a line to ground to shunt away high-frequency spikes in order to keep them away from the rest of the circuit.
I've got a few detailed questions for the electronics experts on here...
is this frequency-filtering property constant or is it related to the capacitors charge and discharge cycle?
when the circuit is first powered, is it correct that a decoupling capacitor will draw current until it is charged? after this it should have no effect on the rest of the circuit aside from the intended high-frequency shunting?
is this frequency-filtering property constant or is it related to the capacitors charge and discharge cycle?
It's based on the impedance of the capacitor for the frequencies involved.
Ideally, an electrolytic cap (say 470uF) would look like a short to high frequencies, but it has parasitics that reduce the actual performance vs theoretical. That's why you'll see references saying you should parallel a 0.1uF ceramic with the electrolytic cap or ceramics of a higher value. The ceramic cap has lower inductance than an electrolytic cap at higher frequencies.
Xc=1/(2pif*C)
(This forum doesn't appear to support tex for formatting mathematical expressions...).
when the circuit is first powered, is it correct that a decoupling capacitor will draw current until it is charged? after this it should have no effect on the rest of the circuit aside from the intended high-frequency shunting?
True they need to charge, but you only need to worry about big electrolytics. It's possible to have high inrush currents that may need to be reduced.
is this frequency-filtering property constant or is it related to the capacitors charge and discharge cycle?
Good point. In an ideal capacitor no the charge has no effect.
But on certain types of ceramic capacitors the actual capacitance does depend on the voltage across them. This will normally have little effect on the decoupling abilities but is why ceramic capacitors are avoided by audiophiles in filtering and coupling circuits.
is this frequency-filtering property constant or is it related to the capacitors charge and discharge cycle?
The impedance of any capacitor falls with frequency, more capacitance has less impedance at any given frequency. The amount of charge has nothing to do with this (except for very non-linear capacitors).
when the circuit is first powered, is it correct that a decoupling capacitor will draw current until it is charged? after this it should have no effect on the rest of the circuit aside from the intended high-frequency shunting?
Yes, it charges up - there is thus a current spike unless the supply is current-limited.
MarkT:
The impedance of any capacitor falls with frequency, more capacitance has less impedance at any given frequency. The amount of charge has nothing to do with this (except for very non-linear capacitors).
Electrolytics have particularly high impedance for high frequencies, as they start acting more like an inductor than a capacitor (in part due to the construction of the things, with the two electrodes rolled up around one another). That's why the parallel ceramic, to deal with those frequencies.
Some designs even call for multiple parallel ceramics, like 10 nF and 1 µF, each handing a different frequency best.
