RC circuits: timing, filters, and real behavior.

An RC circuit combines resistance with capacitance. The resistor controls how quickly current can move charge into or out of the capacitor. The capacitor voltage therefore changes smoothly rather than jumping instantly. That simple interaction creates timing and frequency-dependent behavior throughout electronics.

The time constant

The product of resistance and capacitance is the time constant, written as tau. Ohms multiplied by farads produce seconds. A 10 kohm resistor and a 100 uF capacitor therefore have a time constant of one second.

tau = R x C
tau = 10,000 ohm x 0.0001 F
tau = 1 second

One time constant does not mean the capacitor is fully charged. It means the voltage has completed about 63.2 percent of the remaining journey toward its final value. Each further time constant completes 63.2 percent of what is still left.

Elapsed time	Charging toward final value	Remaining during discharge
1 tau	63.2%	36.8%
2 tau	86.5%	13.5%
3 tau	95.0%	5.0%
4 tau	98.2%	1.8%
5 tau	99.3%	0.7%

Charging from one voltage to another

The most useful form of the RC equation handles any initial voltage and any final source voltage:

Vc(t) = Vs + (V0 - Vs) x e^(-t/tau)

Here, Vs is the final source voltage, V0 is the capacitor voltage at the start, and t is elapsed time. If a discharged capacitor charges toward 5 V through a one-second time constant, it reaches about 3.16 V after one second and 4.32 V after two seconds.

This is also why logic thresholds matter. A digital input does not wait for the capacitor to reach the rail; it changes state when the rising or falling voltage crosses its threshold. If that threshold is poorly specified, the timing is poorly specified too.

Low-pass filters

Put a resistor in series with a signal and a capacitor from the output to ground, then take the output across the capacitor. At low frequencies the capacitor has high reactance and the output follows the input. At high frequencies the capacitor provides an easier path to ground, so the output is reduced.

The ideal cutoff frequency is:

fc = 1 / (2 x pi x R x C)

At cutoff, an ideal first-order filter has a voltage magnitude of about 0.707 times the input, or -3.01 dB. Above cutoff, attenuation approaches 20 dB per decade. It is a gradual slope, not a wall.

High-pass filters

Swap the component positions so the capacitor is in series and the resistor goes from output to the return path. Low frequencies are reduced because the capacitor presents high reactance; faster changes pass more readily. The same cutoff equation applies when source and load impedances do not disturb the assumed resistance.

High-pass RC networks are common for AC coupling, removing DC offsets, detecting edges, and shaping pulses. A coupling capacitor and the input resistance of the next stage form an RC filter even when no separate resistor was drawn for that purpose.

Why source and load impedance matter

Textbook formulas assume a zero-impedance source and an infinite- impedance load. Real circuits have neither. Source resistance adds to the chosen series resistor. Load resistance appears in parallel with part of the network. Both can move the cutoff frequency and change passband gain.

For example, a 10 kohm RC filter feeding a 10 kohm load does not see a 10 kohm output resistance. The load materially changes the network. A voltage follower or other buffer can isolate the filter when the next stage would otherwise load it.

PWM smoothing

An RC low-pass can turn a PWM waveform into an approximate DC level. The average is related to duty cycle, but residual ripple depends on PWM frequency, cutoff frequency, load, and waveform amplitude. A much lower cutoff reduces ripple but makes the output respond more slowly to duty-cycle changes.

This is a tradeoff rather than a single correct value. Decide how much ripple the load can tolerate and how quickly the output must settle, then verify both conditions.

ADC inputs and settling

Many ADCs briefly connect an internal sample capacitor to the input. The external source must charge that capacitor accurately during the available acquisition time. A large series resistor or slow RC filter can prevent settling, even if a multimeter shows the expected DC voltage.

Check the converter datasheet for maximum recommended source impedance, acquisition time, and input capacitance. A small local capacitor can provide charge during sampling, but it also changes the filter and must recharge between conversions. A buffer is often the cleanest solution for high-impedance sensors.

Reset delays and debounce

RC networks are often used to delay reset release or soften switch bounce. They can help, but a slowly moving voltage may spend too long in an undefined logic region. A Schmitt-trigger input, reset supervisor, or explicit firmware debounce gives a more predictable transition.

Button bounce is not always symmetrical, and leakage paths can make charge and discharge times different. Treat the RC as part of a threshold system rather than assuming tau itself is the debounce time.

Initial current and stored energy

An uncharged capacitor initially behaves like a low impedance. In the ideal RC model, charging current starts at:

Iinitial = (Vs - V0) / R

The current then decays exponentially. Large capacitors connected through very small resistance can create substantial inrush current. Stored energy is:

E = 0.5 x C x V^2

For small signal filters that energy is tiny. For power-rail capacitors it may be large enough to stress connectors, switches, rectifiers, or discharge paths.

Real component limits

Effect	What it changes
Resistance tolerance	Moves time constant and cutoff frequency.
Capacitance tolerance	Can dominate timing error, especially with electrolytic parts.
Leakage current	Creates a parallel path that matters in high-resistance timing networks.
ESR	Adds loss and modifies high-frequency and pulse behavior.
DC bias	Can reduce the effective capacitance of ceramic capacitors.
Temperature	Changes resistance, capacitance, leakage, and therefore timing.
Input thresholds	Turn a voltage curve into a timing result, often with significant variation.

Choosing R and C

Many R and C combinations produce the same nominal time constant. The choice still matters. Very low resistance wastes current and may load the source. Very high resistance makes leakage, contamination, input bias current, and measurement probes more influential. Very large capacitors may have poor tolerance, leakage, size, and aging; very small capacitors may be dominated by parasitic capacitance.

Start with a resistance that suits source and load constraints, solve for capacitance, choose a nearby standard value, and recalculate the real range using tolerance. For precision or long delays, digital timing is often more stable than a large analog RC network.

Common mistakes

• Calling one time constant fully charged instead of 63.2 percent complete.
• Ignoring source and load impedance when calculating cutoff.
• Using nominal capacitor value as though tolerance, bias, and leakage do not exist.
• Feeding an ADC through too much source resistance for its acquisition time.
• Choosing a PWM filter only for low ripple and then discovering that it responds too slowly.
• Using a slow RC edge on an input that has no hysteresis.
• Forgetting inrush current when charging a large capacitor.

A practical RC checklist

1. Is the circuit being designed for a time, a threshold, or a cutoff frequency?
2. What source resistance and load resistance modify the selected R?
3. What range results from resistor and capacitor tolerance?
4. Do leakage, bias current, ESR, or DC bias matter at these values?
5. How quickly must the output settle after an input change?
6. Can the next stage tolerate a slowly moving voltage?
7. Does startup current or stored energy stress any component?