The direct answer to how a capacitor draws current is defined by the capacitor current equation: i = C × (dv/dt). Unlike a resistor, where current is strictly proportional to the instantaneous voltage across it, a capacitor's current is proportional to the rate of change of that voltage. If the voltage across a capacitor is perfectly stable (dv = 0), the current is exactly zero, regardless of whether that voltage is 5V or 500V. It is only when the voltage changes that current flows, as the capacitor either absorbs or releases charge to its dielectric field.
The Capacitor Current Equation: Symbols and Rearranged Forms
To use this formula effectively on the bench, you must understand the exact physical quantities each variable represents. The standard differential form is i = C(dv/dt). For practical engineering where the rate of change is linear over a short interval, we write this as i = C(ΔV/Δt).
| Symbol | Parameter | Standard Unit | Practical Bench Notes |
|---|---|---|---|
| i | Instantaneous Current | Amperes (A) | Can be positive (charging) or negative (discharging). |
| C | Capacitance | Farads (F) | Assumed constant in the ideal formula; varies with DC bias and temperature in reality. |
| dv (or ΔV) | Change in Voltage | Volts (V) | The difference between final and initial voltage across the capacitor terminals. |
| dt (or Δt) | Change in Time | Seconds (s) | The duration over which the voltage change occurs. Smaller dt yields massive current spikes. |
Rearranged Forms for Design and Debugging
Depending on what you are trying to solve for, you will need to rearrange the formula. Keep these variants in your mental toolkit:
- Solving for Capacitance (Sizing a filter): C = (i × dt) / dv. Use this when you know your maximum allowable voltage ripple (dv) and the time between charging pulses (dt).
- Solving for Voltage Change (Calculating ripple/droop): dv = (i × dt) / C. Use this to find out how much your supply voltage will sag when a load draws current (i) for a specific duration (dt).
- Solving for Time Interval (Hold-up time): dt = (C × dv) / i. Use this to calculate how long a capacitor bank can sustain a load before the voltage drops below a critical threshold.
Assumptions, Boundaries, and the Unit Traps That Break It
When the Formula Applies (and When It Doesn't)
The equation i = C(dv/dt) assumes an ideal capacitor. This means it assumes capacitance (C) is a fixed constant, and it ignores parasitic elements. In the real world, every capacitor has Equivalent Series Resistance (ESR) and Equivalent Series Inductance (ESL).
For slow voltage changes (like a 60Hz AC line or a slow power supply ramp), the ideal formula is highly accurate. However, for nanosecond-scale digital edges, ESL dominates, and the capacitor behaves more like an inductor. Furthermore, if you are using Multi-Layer Ceramic Capacitors (MLCCs) like X7R or X5R dielectrics, the capacitance value (C) is not constant. A 10μF MLCC might drop to 2μF when a 24V DC bias is applied. If you use the nominal 10μF value in your equation for a high-voltage circuit, your calculated current and ripple will be entirely wrong.
The Unit Traps That Cause Bench Failures
The most common reason hobbyists and junior engineers get nonsensical results from this formula is failing to convert prefixes to base SI units. The formula only works with Farads, Volts, and Seconds.
• Microfarads (μF) must be multiplied by 10-6 (e.g., 100μF = 0.0001 F).
• Nanofarads (nF) must be multiplied by 10-9.
• Milliseconds (ms) must be multiplied by 10-3.
• Microseconds (μs) must be multiplied by 10-6.
What a Realistic Answer Magnitude Looks Like
Context is everything. If you are calculating current for an I2C pull-up or an op-amp feedback loop, a realistic answer is in the microamp (μA) to low milliamp (mA) range. If you are calculating inrush current for a bulk power supply filter, a realistic answer is tens or hundreds of Amperes. If your math yields 0.0004A for a motor driver inrush, you likely forgot to convert milliseconds to seconds.
Worked Problem 1: Signal Filtering and Realistic Magnitudes
Scenario: You are designing a digital communication line with a fast edge rate. You place a 100nF MLCC decoupling capacitor near a microcontroller pin to filter high-frequency noise. The digital signal transitions from 0V to 3.3V with a slew rate (dv/dt) of 2V/μs. What is the peak current drawn by this capacitor during the edge transition?
- Identify and convert variables to base units:
C = 100nF = 100 × 10-9 F
dv = 3.3V (but we are given the rate directly, so we use the slew rate)
dv/dt = 2V / 1μs = 2V / (1 × 10-6 s) = 2 × 106 V/s - Apply the formula:
i = C × (dv/dt)
i = (100 × 10-9 F) × (2 × 106 V/s) - Calculate and track units:
i = 200 × 10-3 Amperes
i = 0.2 A, or 200 mA
Bench Insight: It seems counterintuitive that a tiny 100nF ceramic capacitor pulls 200mA of instantaneous current. This is exactly why high-speed digital ICs require local decoupling; if that 200mA had to travel through the parasitic inductance of a long PCB trace to reach the main power supply, it would cause a massive voltage spike (V = L di/dt) and corrupt the logic levels. For more on high-speed decoupling, refer to Electronics Tutorials' capacitor guides.
Worked Problem 2: Power Supply Inrush and the dt Danger
Scenario: You are building a 48V DC motor controller. The bulk filter on the DC bus consists of a single 4700μF electrolytic capacitor. When the main contactor closes, the voltage across the capacitor ramps from 0V to 48V in 2 milliseconds (limited by the wiring inductance and the contactor's bounce time). What is the average inrush current?
- Identify and convert variables to base units:
C = 4700μF = 4700 × 10-6 F = 4.7 × 10-3 F
dv = 48V - 0V = 48V
dt = 2ms = 2 × 10-3 s - Calculate the rate of voltage change:
dv/dt = 48V / (2 × 10-3 s) = 24,000 V/s - Apply the formula and solve:
i = C × (dv/dt)
i = (4.7 × 10-3 F) × 24,000 V/s
i = 112.8 Amperes
Bench Insight: 112.8A is a massive transient. Even though it only lasts for 2 milliseconds, it is enough to instantly trip a standard 10A thermal-magnetic breaker or blow a fast-acting glass fuse. This is why power supplies of this magnitude require active inrush current limiting or pre-charge circuits. The relationship between capacitance, voltage, and time is deeply explored in Wikipedia's technical breakdown of capacitor current-voltage relations.
Real-World Bench Scenario: The Blown Rectifier Diode
Formulas on a whiteboard are neat; components exploding on the bench are educational. Here is a real-world failure driven entirely by ignoring the dt variable in the capacitor current equation.
The Setup
A hobbyist was designing a 24V linear power supply for a high-torque stepper motor driver. To smooth out the 120Hz ripple from the full-wave rectifier, they installed a massive capacitor bank consisting of two 10,000μF electrolytic capacitors in parallel (Total C = 20,000μF, or 0.02F). The AC side of the transformer fed directly into a KBPC1510 bridge rectifier, which is rated for 15A continuous forward current.
The Numbers
When the power switch was flipped, the transformer output a peak DC voltage of roughly 34V (24V RMS × √2). The capacitor bank was completely discharged (0V). The voltage change (dv) was 34V. The time (dt) was limited only by the transformer's winding resistance and the PCB trace inductance, which measured out to approximately 0.5 milliseconds (0.0005s) on an oscilloscope.
Using the equation:
i = 0.02F × (34V / 0.0005s)
i = 0.02 × 68,000
i = 1,360 Amperes
The Outcome
Upon flipping the switch, there was a loud, sharp 'crack' from the power supply. The KBPC1510 bridge rectifier failed catastrophically, shorting the AC input and blowing the primary side fuse. The silicon junction inside the diode literally vaporized.
What Went Wrong
The builder looked at the 15A continuous rating of the KBPC1510 and assumed it was sufficient because the motor driver only drew 8A on average. They completely ignored the transient inrush current dictated by i = C(dv/dt). While 15A is the continuous thermal rating, silicon diodes also have an I²t rating (let-through energy) which defines how much transient energy they can survive before the wire bonds melt. 1,360A for 0.5ms vastly exceeded the I²t limit of the rectifier.
The Fix
To solve this, the builder needed to artificially increase the dt (time) variable to bring the peak current (i) down to a safe level. They installed an Ametherm MS35 10018 NTC (Negative Temperature Coefficient) inrush current limiter in series with the AC input. This component provides 10 ohms of resistance when cold, stretching the charging time (dt) of the capacitor bank from 0.5ms to roughly 40ms.
Recalculating with the NTC limiting the peak current:
Peak i ≈ V_peak / R_ntc = 34V / 10Ω = 3.4A.
This 3.4A peak is well within the safe surge rating of the KBPC1510. As the NTC heats up from the 8A continuous load, its resistance drops to near zero, allowing normal operation. For detailed sizing of these components, Ametherm's inrush current technical resources provide excellent application notes.
