The Short Answer: What Do Capacitors Do in DC?
In a purely theoretical, steady-state DC circuit, a capacitor acts as an open circuit with infinite resistance, blocking all direct current once it reaches its charge voltage. But in the real world, DC is never perfectly steady. What capacitors do in DC circuits is manage the changes and imperfections in that voltage. They filter out AC ripple from rectified power supplies, decouple high-frequency switching noise generated by digital logic, store localized energy to supply sudden current spikes, and create precise time delays in RC (resistor-capacitor) timing networks.
Think of a DC power rail like a municipal water main. The water pressure (voltage) is nominally constant, but when a factory down the street suddenly opens a massive valve (a microcontroller turning on its WiFi radio), the pressure in your pipe drops momentarily. A capacitor acts like a pressurized accumulator tank sitting right next to your faucet, instantly releasing stored water to maintain pressure until the main pump catches up. A typical 100nF (0.1µF) ceramic capacitor has an impedance of just 1.59 ohms at 1MHz, effectively shorting high-frequency digital noise to ground while completely blocking the 3.3V DC rail.
The Bench Scenario: When a Missing Decoupling Cap Bricks a Boot Sequence
To understand why capacitor selection matters in DC, let us look at a real-world failure I debugged recently on a custom carrier board for an ESP32-WROOM-32 module.
The Setup: The board was powered by a standard LM2596 buck converter stepping 12V down to 3.3V. The buck converter's output had a 220µF aluminum electrolytic capacitor, which seemed like plenty of bulk storage for a microcontroller drawing a baseline of 80mA.
The Numbers: During the initial WiFi transmission burst, the ESP32 draws a peak current of roughly 500mA for about 200 microseconds. The buck converter's feedback loop takes roughly 50µs to respond to a voltage drop. Therefore, the local capacitors must supply the deficit. If the regulator supplies 300mA, the capacitor must supply the remaining 200mA. Using the formula Q = I × t, the required charge is 0.2A × 200µs = 40µC. If we only had a 1µF ceramic capacitor on the rail, the voltage drop would be ΔV = Q / C = 40µC / 1µF = 40V. Obviously, the rail would collapse instantly.
The Outcome: The board randomly failed to boot, throwing a Guru Meditation Error: Core 1 panic'ed (Brownout detector was triggered) in the serial monitor. The 3.3V rail was dipping below the ESP32's 2.4V brownout threshold.
What Went Wrong: I had placed the 220µF electrolytic capacitor three inches away from the ESP32 module, near the power jack. At high frequencies, the PCB trace acts as an inductor. Furthermore, the cheap electrolytic capacitor had an Equivalent Series Resistance (ESR) of about 0.2 ohms and high Equivalent Series Inductance (ESL). It physically could not discharge its energy fast enough to support the 200µs transient spike. According to the Espressif ESP32 Hardware Design Guidelines, a 10µF to 100µF low-ESR capacitor must be placed as close to the VCC pin as possible.
The Fix: I added a 100µF low-ESR tantalum capacitor and a 100nF X7R ceramic capacitor directly across the ESP32's VCC and GND pins. The ceramic handled the sub-microsecond high-frequency noise, while the tantalum supplied the bulk 200µs charge, keeping the voltage drop to a safe ΔV = 40µC / 100µF = 0.4V.
Choosing the Right Dielectric: DC Capacitor Comparison
Not all capacitors are built the same. The dielectric material inside the capacitor dictates its behavior under DC bias, temperature changes, and high frequencies. Here is how to select the right type for your DC application.
| Type / Dielectric | Construction | Typical Tolerance | Tempco (Temperature Coefficient) | Best DC Use Case |
|---|---|---|---|---|
| Ceramic (C0G/NP0) | Class I ceramic layers | ±1% to ±5% | 0 ±30 ppm/°C (Ultra-stable) | RF filtering, precision analog timing, PLL loops. |
| Ceramic (X7R/X5R) | Class II ceramic layers | ±10% to ±20% | ±15% over temp range | General decoupling, bypassing, bulk storage (with DC bias derating). |
| Aluminum Electrolytic | Etched aluminum foil with liquid electrolyte | -20% to +80% | High drift, sensitive to cold | Bulk energy storage, low-frequency power supply ripple filtering. |
| Tantalum (MnO2/Polymer) | Porous tantalum pellet anode | ±10% to ±20% | Moderate stability | Compact bulk decoupling on low-voltage DC rails (3.3V/5V). |
| Film (Polypropylene) | Metallized plastic film wound or stacked | ±1% to ±5% | Very stable, low dielectric absorption | High-voltage DC bus snubbers, audio signal coupling, motor run. |
Decoding the Silkscreen: How to Read Capacitor Markings
When you are scavenging parts from an old board or sorting through a mixed bin of through-hole components, knowing how to read the markings is essential. Capacitor codes vary wildly by chemistry and package size.
The 3-Digit EIA Code (Ceramics)
Through-hole and small SMD ceramic capacitors use a three-digit picofarad (pF) code. The first two digits are the significant figures, and the third digit is the multiplier (number of zeros).
- 104: 10 × 104 pF = 100,000 pF = 100 nF = 0.1 µF. (This is the most common decoupling cap in electronics).
- 473: 47 × 103 pF = 47,000 pF = 47 nF.
- 221: 22 × 101 pF = 220 pF.
Sometimes a letter follows the number to indicate tolerance: J = ±5%, K = ±10%, M = ±20%.
Electrolytic and Tantalum Markings
Through-hole aluminum electrolytics are straightforward: they print the capacitance (e.g., 470µF), the maximum DC voltage rating (e.g., 25V), and a temperature rating (e.g., 105°C). A printed stripe with minus signs indicates the negative lead. For radial SMD aluminum electrolytics, the black stripe marks the negative side.
SMD Tantalum capacitors use a colored band or a thicker line on one end of the rectangular package to indicate the positive (anode) side. This is the exact opposite of aluminum electrolytics, and reversing a tantalum will result in catastrophic failure.
Safe Substitution: When the Parts Drawer is Missing the BOM Part
You are in the middle of a prototype build, and you are out of the exact capacitor specified in the schematic. Substitution is common, but you must follow strict electrical rules to avoid destroying your circuit.
- Voltage Rating: You can always substitute a higher voltage rating for a lower one (e.g., using a 50V cap instead of a 25V cap). Never substitute a lower voltage rating. However, be aware of physical size constraints on the PCB.
- Capacitance Value: For power supply decoupling and bypassing, substituting a higher capacitance is generally safe and often beneficial. For RC timing circuits, oscillators, or active filters, the capacitance must be exact, or your frequency and timing will shift.
- The DC Bias Trap (Crucial for MLCCs): When substituting Class II ceramics (X7R/X5R), you must account for DC bias derating. As Texas Instruments notes in their MLCC design guides, a 10µF X7R ceramic capacitor in a small 0805 package rated for 16V might lose up to 60% of its capacitance when 5V of DC bias is applied. If your circuit absolutely requires 10µF at 5V, you must either choose a larger physical package (like 1210) or select a capacitor with a much higher voltage rating (like 25V or 50V) to maintain the actual capacitance under load.
- Dielectric Swaps: You can swap an X7R ceramic for a C0G/NP0 in almost any application, but you cannot swap C0G for X7R in precision analog or RF circuits, as X7R introduces microphonic noise and voltage-dependent capacitance shifts.
Failure Modes: What a Dead Capacitor Looks Like on the Bench
Capacitors fail in distinct ways depending on their chemistry. Recognizing these visual and electrical symptoms will save you hours of oscilloscope probing.
Aluminum Electrolytic: The Slow Dry-Out
Electrolytics contain a liquid electrolyte that slowly evaporates through the rubber end seal over time, a process accelerated by heat. According to the Cornell Dubilier Aluminum Electrolytic Application Guide, the rule of thumb is that capacitor life doubles for every 10°C drop in operating temperature. Visual symptoms: The top aluminum vent bulges outward (the cross or K-score on top is designed to pop open to prevent explosions). You may see brown, crusty electrolyte leaking from the bottom rubber seal or the top vent. Electrical symptom: The capacitance drops slightly, but the ESR skyrockets. A cap that should measure 0.05 ohms ESR might read 5 ohms, rendering it useless for filtering high-frequency ripple.
Multilayer Ceramic (MLCC): The Invisible Flex Crack
Ceramic capacitors are brittle. If a PCB is bent during assembly, testing, or enclosure mounting, the solder joints can pull on the capacitor body, causing a microscopic crack near the terminal. Visual symptoms: Often completely invisible to the naked eye. Sometimes you will see a tiny hairline fracture on the ceramic body under a 10x loupe. Electrical symptom: The crack can cause a partial short or an open circuit. If it shorts, your DC power rail will be pulled to ground. You diagnose this by removing power and using a multimeter in continuity mode across the VCC and GND rails. If it reads near 0 ohms, you start desoldering decoupling caps one by one until the short disappears.
Tantalum: The Catastrophic Thermal Runaway
Manganese dioxide tantalum capacitors are highly sensitive to reverse polarity and high surge currents. If subjected to a voltage spike exceeding their rating or installed backward, the dielectric oxide layer breaks down. Visual symptoms: Tantalum failures are violent. The component will literally catch fire, emit a bright flash, and leave a scorched, blackened crater on the PCB. The part often cracks completely in half. To prevent this, modern designs increasingly favor polymer tantalum capacitors, which use a conductive polymer cathode that fails open (safely) rather than catching fire when abused.
