Electrical Engineering › Section G · Capacitors › Topic G.3
Charging and discharging transients
A capacitor cannot change its voltage instantly, so when a switch closes in an R-C circuit, voltages and currents glide to their new values along exponential curves. One number, the time constant RC, tells you how long that takes.
The transient phase
In the steady state a charged capacitor carries no dc current and acts as an open circuit. But between switching something and reaching that steady state, there is a transient phase in which the capacitor's voltage changes and current flows. Transients set how fast a camera flash recharges, how long a power-on reset holds a microcontroller in reset, how a switch debouncer ignores contact bounce, and the timing of many oscillators, including the familiar 555 timer chip, which times by charging a capacitor through resistors. This topic covers a single capacitor with resistors and a dc source.
The charging phase
Close a switch at \(t = 0\) connecting a source \(E\), a resistor \(R\) and an uncharged capacitor \(C\) in series. At first \(v_C = 0\), so the whole of \(E\) appears across \(R\) and the current jumps to \(E/R\): an uncharged capacitor behaves like a short circuit at the instant of switching. As charge builds, \(v_C\) rises, the voltage left for \(R\) shrinks, and the current falls. The result is
with \(v_R = Ee^{-t/\tau}\), so that \(v_R + v_C = E\) at every instant, as KVL requires.
The time constant
The quantity \(\tau = RC\) has units of seconds (ohms times farads), and it sets the pace of the whole transient. Because every quantity follows the same exponential, the fractions are universal:
By convention the transient is treated as over after \(5\tau\). Each further time constant closes 63.2% of whatever gap remains, so in theory the capacitor never quite arrives.
Worked example: charging
A 12 V source charges an uncharged 22 µF capacitor through 47 kΩ. 1. Time constant: \(\tau = 47 \times 10^3 \times 22 \times 10^{-6} = 1.03\,\text{s}\), so charging is effectively complete after \(5\tau = 5.17\,\text{s}\). 2. Initial current: \(12/47\,\text{k}\Omega = 0.255\,\text{mA}\). 3. At \(t = 2.0\,\text{s}\): \(v_C = 12(1 - e^{-2.0/1.034}) = 10.3\,\text{V}\), and \(i_C = 0.255e^{-1.93} = 0.0369\,\text{mA}\). 4. When does \(v_C\) reach 9.0 V? Solve \(e^{-t/\tau} = 3/12\): \(t = \tau\ln 4 = 1.43\,\text{s}\). 5. Check: 9.0 V is 75% of E, which lies between the 63.2% at \(\tau\) and 86.5% at \(2\tau\), and 1.43 s lies between 1.03 s and 2.07 s.
The discharge phase
Now remove the source and let a capacitor charged to \(V_0\) discharge through a resistor \(R\). The capacitor acts as the source, so the current flows out of its + plate, the opposite way to the charging current. Both the voltage and the size of the current decay:
with \(\tau = RC\) for the discharge path, which need not equal the charging resistance. For example, the 22 µF capacitor above, charged to 12 V and discharged through 100 kΩ, has \(\tau = 2.2\,\text{s}\); after 5.0 s it holds \(12e^{-5.0/2.2} = 1.24\,\text{V}\). After one time constant 36.8% of the voltage remains.
Initial values and the general equation
A capacitor need not start uncharged. Because its voltage cannot jump, its value just after switching equals its value just before. If it starts at \(V_i\) and heads for a final value \(V_f\),
This one equation covers charging (\(V_i = 0\)), discharge (\(V_f = 0\)) and every case in between. To find the time at which \(v_C\) reaches a value \(v\), rearrange:
Find \(V_f\) by treating the capacitor as an open circuit in the final steady state.
Thevenin equivalent: τ = R_Th C
When the capacitor sits in a larger network, remove it and find the Thevenin equivalent seen from its terminals, as in Thevenin's theorem. The capacitor then charges as if it were in a simple series circuit: its final voltage is \(E_{Th}\) and the time constant is
The initial current, just after switching, is \((E_{Th} - V_i)/R_{Th}\). This is the standard method for any single-capacitor circuit, however many resistors surround it.
Worked example: a network with a starting voltage
A 20 V source, − to ground, feeds node a through 10 kΩ. 15 kΩ joins a to ground, and 4.7 kΩ runs from a to node b. A 10 µF capacitor from b to ground already holds 2.0 V, + at b, when the source is switched on at \(t = 0\). 1. Thevenin at b: \(E_{Th} = 20 \times \frac{15}{25} = 12\,\text{V}\) and \(R_{Th} = 4.7 + 10 \parallel 15 = 10.7\,\text{k}\Omega\). 2. \(\tau = 10.7\,\text{k}\Omega \times 10\,\mu\text{F} = 0.107\,\text{s}\). 3. \(v_C = 12 - 10e^{-t/0.107}\) volts, starting at 2.0 V. 4. Initial current: \((12 - 2.0)/10.7\,\text{k}\Omega = 0.935\,\text{mA}\). 5. Time to reach 9.0 V: \(t = 0.107\ln\frac{2 - 12}{9 - 12} = 0.129\,\text{s}\). 6. Check: 0.129 s is a little over one time constant, and 9.0 V closes 70% of the 10 V gap, a little over 63.2%.
The current i_C = C dv/dt
Since \(q = Cv\), the capacitor current is the rate at which its charge changes:
The current depends on how fast the voltage changes, not on the voltage itself. A steady voltage, however large, gives zero current; a fast change gives a large current, and an instant jump would need infinite current, which is why \(v_C\) is continuous. If a 0.47 µF capacitor's voltage rises steadily from 0 to 5.0 V in 1.0 ms, the current is \(0.47\,\mu\text{F} \times 5000\,\text{V/s} = 2.35\,\text{mA}\) throughout. Differentiating the charging equation gives back \(i_C = (E/R)e^{-t/\tau}\).
Common mistakes to avoid
First, unit slips in \(\tau = RC\): kilohms times microfarads gives milliseconds, so 10 kΩ and 10 µF give 0.1 s, not 100 s. Second, assuming the capacitor voltage jumps when the switch closes; it is the current that jumps. Third, using the charging resistance for the discharge when the paths differ. Fourth, ignoring a starting voltage and using \(E(1 - e^{-t/\tau})\) anyway. Fifth, using one resistor in the network instead of \(R_{Th}\) seen by the capacitor. Sixth, forgetting that discharge current flows the opposite way to charging current.
Key takeaways
Charging: \(v_C = E(1 - e^{-t/\tau})\) and \(i_C = (E/R)e^{-t/\tau}\), with \(\tau = RC\). After one time constant the voltage reaches 63.2% of its final value, and after \(5\tau\) the transient is treated as over. Discharge: \(v_C = V_0e^{-t/\tau}\), with 36.8% left after \(\tau\). In general \(v_C = V_f + (V_i - V_f)e^{-t/\tau}\), because capacitor voltage cannot jump. In a network, use \(E_{Th}\) and \(\tau = R_{Th}C\). The current is \(i_C = C\,dv_C/dt\), zero at steady dc.
Flashcards
Flashcard
The time constant of an R-C circuit, in seconds
Test your understanding
Question 1 of 5
A 33 kΩ resistor charges a 100 µF capacitor from a dc source. What is the time constant, and roughly when is charging complete?
Question 2 of 5
An uncharged capacitor charges from a 10 V source through a resistor. What is its voltage after two time constants?
Question 3 of 5
A capacitor already holds 3.0 V when it starts charging towards a final value of 15 V. What is its voltage one time constant later?
Question 4 of 5
The voltage across a 10 µF capacitor is rising at a steady 2.0 V per millisecond. What current flows into it?
Question 5 of 5
A capacitor is connected to a network of several resistors and one dc source. Which resistance sets its time constant?
