Capacitor Behaviour and What It Means on a PCB
A capacitor is the simplest component in the circuit and the one whose behaviour is most often taken for granted. It stores charge, it resists a change in voltage, and it passes a changing signal while blocking a steady one. Every use of a capacitor on a board follows from those three properties.
This article reviews what capacitance is, how a capacitor charges and discharges, what its reactance means in practice, and how the electrical behaviour translates into the layout decisions a board requires.
What Capacitance Is
Two conductors separated by an insulator form a capacitor. When a potential is applied, charge accumulates on each conductor, and the amount of charge that a given potential can hold is the capacitance of the arrangement. It is measured in farads, a unit so large that most practical components are rated in microfarads, nanofarads or picofarads.
The quantity depends on the area of the conductors, the distance between them and the properties of the material between them. That is why a capacitor can be built as a component with two plates and a dielectric, and also why two copper layers separated by a thin laminate behave as a capacitor whether or not anyone intended them to.
Charging, Discharging and the Time Constant
A capacitor charged through a resistor does not reach its final voltage instantly. The current is largest at the first instant and falls as the voltage rises, so the voltage approaches the supply along an exponential curve rather than a straight line. The speed of that approach is set by the product of the resistance and the capacitance.
That product is the time constant, and it is the natural unit of any discussion about a capacitor in a real circuit. Discharging follows the same curve in reverse. Increasing the resistance lengthens the charge time by reducing the current; increasing the capacitance lengthens it because more charge has to be moved.
Why the Voltage Cannot Change Instantly
Because charge takes time to move, the voltage across a capacitor cannot jump. This is the property that makes the component useful: it absorbs the transient that the rest of the circuit cannot, and it holds a supply steady while a load changes.
It is also the property that makes a capacitor dangerous after the power is removed. The charge is still there, and energy stored in a large capacitor at a high voltage will discharge through whatever completes the path. That is why a bulk capacitor on a supply is treated as live until it has been discharged deliberately.
Reactance: Passing AC and Blocking DC
With a steady voltage applied, a capacitor charges once and then carries no current, so a direct voltage does not appear on the far side of it. With a changing voltage, the capacitor charges and discharges continuously, and the effect is a current that appears to flow through it. The opposition it presents falls as the frequency rises.
That frequency-dependent behaviour is described as reactance, and it is what makes a capacitor a coupling element for signals and a short circuit for noise. A capacitor placed between a supply and a ground, close to a device, presents a low impedance to the high-frequency current that the device draws, so the current is supplied locally instead of being drawn through a long trace. The planning of a power distribution network is largely a question of where those capacitors go.
<img src="https://www.gopcba.com/wp-content/uploads/2024/09/55387056_2.webp" alt="Decoupling capacitors placed beside an IC on a PCB” />
Coupling, Decoupling and Bypass
The same component serves different purposes depending on where it is placed. In series with a signal it is a coupling capacitor, passing the alternating part of the signal while blocking the direct voltage that the two stages would otherwise share. That is how a signal is passed between stages whose bias points are different: the direct voltage is blocked, the signal is not, and each stage keeps its own operating point.
In parallel with a supply it is a bypass or decoupling capacitor. Its role is to supply the transient current that a device demands faster than the supply can deliver. The distinction in names describes the placement: a bypass capacitor diverts high-frequency current away from the rest of the network, and a decoupling capacitor keeps one device from disturbing its neighbours.
Voltage Rating and Dielectric Choice
Every capacitor has a maximum voltage, set by the dielectric strength of the material between the plates. Exceed it and the insulation breaks down, which turns the component into a short circuit and, depending on the energy available, into a small explosion.
The dielectric also decides how the capacitance behaves. Ceramic types are small and cheap but the capacitance varies with temperature and applied voltage, which is why the construction of the part is specified alongside its nominal value. Aluminium electrolytic and tantalum parts offer large capacitance in a small package and introduce a polarity that has to be respected, and the resistance in series with the capacitor is a real quantity that limits its effectiveness at high frequency.
What This Means for Layout
The electrical behaviour decides the layout more than the value does. A decoupling capacitor is only effective if the loop it forms with the device and the plane is small, so its placement and the vias that connect it matter as much as the capacitance. At the frequencies involved, a centimetre of trace is a significant impedance.
Where a large capacitance is needed, it is usually spread across several parts of different values rather than concentrated in one component, because a single part has a self-resonance above which it stops behaving like a capacitor. The same reasoning explains why a switching supply places its input and output capacitors immediately beside the switching loop, a point illustrated in the layout of a switching converter.
Reading a Capacitor on a Schematic
On a schematic, a capacitor appears as a value and a package, and it is easy to treat it as a component that either works or does not. In practice the useful question is what the part has to do at the frequency of interest, and whether the circuit around it allows that.
That is also where the mixed nature of most boards shows itself. An analogue stage and a digital stage both use capacitors, for different reasons, and the way their supply and return paths are arranged decides whether the two keep out of each other’s way, as the mixed-signal layout guidelines describe.
FAQ
Why is a large capacitor not a substitute for several small ones? Because a real capacitor stops behaving like a capacitor above its self-resonant frequency, where the package inductance dominates. Several small parts cover a wider band than one large one.
Does a decoupling capacitor have to be as close as possible? It has to be close enough that the loop it forms with the device is small compared with the wavelength of the highest frequency involved. On a fast design, that distance is short.
What happens if the voltage rating is only just adequate? The part works until the supply exceeds its rating, and the failure is a short circuit. A margin is required, and the margin should account for transients rather than only for the nominal rail.
