Choosing the Right Capacitor: The Importance of Accurate Measurements

This presentation from Würth Elektronik by Jon Izkue provides guidelines How capacitor measurements help you to select the correct capacitor technology.

Introduction

Selecting the right capacitor for electronic applications is critical, yet nuanced, due to varying measurement conditions, data sheet specifics, and simulation models.

While capacitance is the initial go-to parameter, real-world performance depends on understanding ripple current limitations, voltage derating, ESR, and the effects of aging and operating conditions. This white paper synthesizes advanced content from an expert webinar to guide engineers on optimizing capacitor choices.

Measurement Essentials for Capacitor Selection

Capacitance as listed in data sheets is typically measured at specific reference conditions: 1V AC, 1kHz frequency, 0V DC bias, and 25°C ambient temperature. These conditions ensure repeatability but do not reflect real circuit scenarios. For example, MLCC class 2 capacitors can lose 30–90% of their rated capacitance when operated at their maximum DC voltage, especially in smaller package sizes like 0805 compared to larger 1812 sizes.

Preconditioning and Memory Effects

Ceramic capacitors, especially MLCC class 2 types, require thermal preconditioning (e.g. heating to 150°C for 1 hour, then cooling for a day) to reset the dielectric material before measurement. Failing to precondition results in permanent polarization effects, causing non-reproducible capacitance readings due to a memory effect in the ceramic domains.

Impact of AC and DC Test Signals

Manufacturers use LCR meters with specific AC excitation (often 1V) to test capacitance. Real-world signals can be lower—meaning less measured capacitance—or higher, but may exceed practical device limits. DC bias dramatically reduces capacitance, more so than AC level. When DC bias approaches 50% or more of rated voltage, remaining capacitance may drop below one third of nominal value. Temperature can cause further divergence, but under high DC bias, the effect converges — temperature “saturates” rather than adds losses.

Parameter	Test Condition	Remark
Capacitance	1V/1kHz at 25°C	Preconditioning resets material, prevents memory effect
AC Level	1V excitation	Lower signals can show lower capacitance
Ripple current limit	Rated temp/freq	Higher value possible at lower temp or higher freq (see multipliers below)

Temperature increases ripple current capability (due to increased permissible self-heating), but shortens operating life. Ripple current multipliers increase with switching frequency (e.g., from 900mA RMS up to nearly 2A RMS at higher frequency/low temp), but this must be balanced against thermal endurance—exceeding rated endurance causes rapid aging or outright failure.

Parameter	Test Condition	Typical Value	Importance
Capacitance	1V/1kHz at 25°C	10 µF (MLCC)	Reference only; expect major losses in real use
Ripple Current	max temp/frequency	900 mA RMS (Electrolytic)	Limit for safe continuous operation
Multipliers	Frequency, Temp	up to 2x ripple (higher freq/lower temp)	Boost permissible ripple with operation tweaks
AC Level	1V excitation	—	Lower signals can show lower capacitance
Endurance	max stress	2,000 hours	Benchmark for max rated conditions
Useful Life	relaxed limits	5,000 hours	Statistical failure probability & extended use

Time-dependent Aging Effects

Capacitance can decrease further with time under bias in MLCC class 2 capacitors. After 10 seconds, losses may reach 50%; after an hour, extra 15% loss, and after 24 hours, total losses may exceed 70%. After about 100 hours, losses tend to saturate. Applications with continuous power-on cycles experience ongoing degradation, emphasizing need for worst-case design consideration.

Manufacturers measure capacitance at reference conditions (1V, 1kHz, 0V bias, 25°C). Real-world operation often shifts these values. Especially in MLCC class 2 capacitors, capacitance can drop by 30–90% under rated DC voltage, with more loss in smaller package sizes. These critical curves are not always published in datasheets.

Aging effects (DC bias/time) are not captured in most simulation models; designers of mission critical applications shall empirically measure circuit capacitance over extended periods (days).

Capacitor Technologies Core Performance Dependencies

• DC Bias Curve: Capacitance drops with applied DC voltage, especially in MLCC Class 2. The effect depends heavily on size and material profile. Simulation platforms provide real bias curves for correct selection.
• Temperature Impact: Elevated temperature shifts capacitance and increases permissible ripple, but decreases lifetime. Electrolytic capacitors rated for higher temperature offer extended operational life even at moderate temperatures.
• Voltage Impact: For supercapacitors, higher voltage at higher temperatures drastically shortens life due to oxide/electrolyte breakdown. Lower voltage improves longevity.
• Electrolytic (and hybrid polymer) capacitors actually benefit from continuous DC voltage, which reforms the oxide layer, reduces leakage, and extends life—provided rated conditions are observed. A small voltage derating (e.g., use 300V part at 250V) gives modest life extension (~125%).
• Ripple & Lifetime Calculators: Modern tools allow sorting and selection based on not just rated specs but adjusted values under real-world use—frequency, voltage, and temperature.

Capacitor Type	DC Bias Impact	Temperature Impact	Ripple Current Multiplier
MLCC Class 2	Up to 90% drop at rated voltage	Loss converges at high bias	N/A
Aluminum Electrolytic	Minimal, positive for lifetime	Strong influence (lifetime halves every 10°C rise)	2x at lower temp, higher freq
Supercapacitors	Decreases lifetime at higher voltage	High temp + voltage accelerates aging	N/A
Film Capacitors	Permissible voltage derates with temp	Lifetime and voltage interdependent	N/A

Simulation Models & RedExpert Platform

Simulation models (LTSpice, PSI, etc.) from advanced suppliers include DC bias effects but commonly reflect bias curves at short time intervals (~10 seconds loading); aging effects over longer time frames are typically not modeled. Designers should build in extra safety margins in simulations, especially for critical or high-voltage MLCC circuits.

Platforms like RedExpert let users input real application conditions: voltage, ripple, temperature, expected switching frequency. Results are displayed in sortable tables, showing true operating lifetime for selected part numbers. For example, a capacitor rated for 105°C offers a multiyear lifespan at 70°C, but operating at maximum voltage, ripple and temp can drop this to under two years.

Simulation tools such as RedExpert goes beyond standard datasheets, offering:

• Bias-dependent capacitance curves
• Lifetime calculators factoring voltage, frequency, and temperature
• Downloadable simulation models (LTSpice, PSI), but DC bias aging is not modeled; only initial bias curves are included

Access: Model libraries downloadable from manufacturer homepage, GitHub, or within RedExpert. Always verify inclusion of DC bias curves.

Practical Selection & Worst-Case Design

• For MLCC Class 2, derating by 50% is grossly inadequate: losses up to 90% are documented. Always use manufacturer bias curves and worst-case estimation based on both bias and temperature.
• Couple all effects—bias, temperature, aging—in calculation for robust design. Instability, noise, and power delivery issues otherwise result.
• Plan for uncertainty: Each measurement is affected by operational history, memory effect, temperature, and bias.
• Use platform tools for worst-case estimation and confirm with in-circuit measurement if possible.

Conclusion

Advanced capacitor selection is more than picking a rated value from a datasheet. It requires understanding real measurement conditions, temperature and bias effects, and leveraging tools that incorporate these variables for safe and reliable electronic design. Always read the complete data sheet, utilize specialized selection platforms, and design to handle worst-case scenarios for mission-critical applications.