0
$\begingroup$

the differential equation for the current flowing through a serial capacitor (see for example https://www.allaboutcircuits.com/textbook/direct-current/chpt-13/capacitors-and-calculus/) indicates that one must take a derivative of the signal but in the literature, a serial capacitor (for instance a coupling capacitor) is often considered as a unit that just eliminates a continuous component of the signal. Well. Is it necessary to take a derivative or just get the first moment of the stochastic signal and eliminate it in order to simulate the passing of that signal through the coupling capacitor?

$\endgroup$
3
  • $\begingroup$ I recommend Papoulis' "Circuits and systems: a modern approach". It deals extensively with this kind of problem. $\endgroup$
    – MBaz
    Mar 18 at 15:26
  • $\begingroup$ Are you looking to make a digital DC blocking filter? I suggest a search on that phrase -- "DC blocking filter". $\endgroup$
    – TimWescott
    Mar 18 at 22:42
  • $\begingroup$ No. Sooner I make a coupling capacitor but do not know how to do it. Clearly, if I need to take a derivative then I change the signal dramatically. If I need to apply just a high-pass filter with a cut-off freq standing below "any signal frequency", then I need just eliminate a mean value. $\endgroup$
    – Sergey
    Mar 20 at 11:09
0
$\begingroup$

Is it necessary to take a derivative

Yes

or just get the first moment of the stochastic signal and eliminate

That's just a very crude approximation of taking the derivative. In some cases that's sufficient, but in most it's not.

For DC blocking the capacitor is typically used together with a resistor to form a high pass filter. The approximation is often "good enough" if the corner frequency of the high pass is well below any signal frequency of interest.

$\endgroup$
4
  • $\begingroup$ I get the impression OP is talking about the full impulse response, i.e. transient plus steady-state, not just the cap as a DC blocker. So what OP refers to as the "first moment of the stochastic signal" might be the transient response leading into the steady-state (0) (therefore no frequency involved). In which case considering the cap, alone, is not enough. But only if I'm not mistaken. $\endgroup$ Mar 18 at 18:38
  • $\begingroup$ Let's focus on a coupling capacitor. Can it be seen as a DC blocker only? And of course, there is a resistor (in the form of the input impedance of the stage, for example). Well. Is it advisable, in this case, to simulate a coupling capacitor as a high-pass filter (Butterworth, for example) of order one, if it influences a signal as a high-pass filter (together with the input impedance)? $\endgroup$
    – Sergey
    Mar 18 at 19:52
  • $\begingroup$ Actually it's not necessary to take the derivative; that's just an artifact of how we usually solve such problems. You can integrate the average, and subtract. You can certainly view that as the action of a capacitor in such a circuit. $\endgroup$
    – TimWescott
    Mar 18 at 22:41
  • $\begingroup$ Does integration need to be made on the scale t, where t is the time constant of the high-pass filter (which is formed by the capacitor in question and the input impedance of the stage)? If yes then it is advisable to use a readily available function of Butterworth's high-pass filter. $\endgroup$
    – Sergey
    Mar 20 at 11:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.