Parasitics & Capacitor Selection, Part 4B: Dissipation Factor

Since capacitance bridges measure capacitance using a small AC signal voltage, it is possible to measure DF simultaneously. So, when capacitors are 100% tested, both capacitance and DF are measured at the same frequency.

For aluminum electrolytic, tantalum, and niobium solid electrolytic and high-capacitance Class II ceramic capacitors (e.g., X5R > 10µF) testing is performed at 120 Hz. The resultant DF will range from 1% to 10% for tantalum and ceramic dielectrics, due to their smaller capacitance and lower voltage ratings, and from 10%+ for aluminum electrolytic dielectrics. The DF for these technologies will also gradually increase with frequency due to their self-resonance, which is typically in the 10 kHz to 10 Mz range (depending on capacitance value), and then increase more rapidly after that as the inductive reactance of the capacitor dominates. For Class II ceramics, there is an additional voltage coefficient effect, so DF will also increase as the ratio of applied to rated voltage increases.

For other technologies, including Class I and Class II ceramics < 10µF, film, and glass, the capacitance and DF measurement frequency is 1 kHz. DF is an especially important consideration for these technologies, as their particular combinations of capacitance and application frequencies suit them to a vast number of AC/DC power supply output applications in which capacitor smoothing is used to both minimize output ripple current and the amount of heat generated in the capacitor due to power dissipation.

One of the lowest DF technologies is Class I (e.g., NP0) ceramic, which exhibits a DF of approximately 0.1%. Class I ceramics also have a low dielectric constant (K), ranging from roughly 15–100 depending on the exact dielectric formulation. In SMD form, Class I ceramics have a low available capacitance and high self-resonant frequency, making them ideal for use in high-frequency applications. However, larger versions of Class I ceramic capacitors (e.g., those arranged in multiple stacks) are ideal for switch mode input filtering up to the 10µF/100V range since their high dielectric strength allows them to filter the spikes and transients associated with industrial applications.

Polypropylene film also has low DF (~0.2%) and is very stable over temperature and frequencies up to the MHz range, but has a lower dielectric constant (K) than NP0 (~2.2). However, the ability to wind large sheets of thin polymer material into large capacitor elements makes this an ideal technology for high-capacitance power filtering with minimal power loss and self-heating from ripple current. Consequently, film capacitors are being designed into an increasing number of DC link applications in which switching outputs can generate high transients, as they effectively keep these transients from radiating back to the input. Prime examples of such applications are the inverter circuits in hybrid drive automobiles, windmills, solar panels, and trains.

DF is an important diagnostic parameter for all capacitor technologies, but it's a particularly important parameter for power applications. Since resistive loss is such an important factor in a given capacitor’s DF reading, it provides a good indication of the health of the insulating dielectric. For example, capacitors that develop high DC leakage in a circuit can often be identified by DF measurements that are outside of specification, even if direct DC measurements cannot be taken. In fact, this aspect of DF also makes it an important diagnostic tool for checking insulation integrity in the cable industry.

In sum, DF is most useful as a design parameter in high-power, low-frequency applications (e.g., DC links in inverter applications) in which the power losses become critical for the thermal management of the system. At low frequencies (< 10kHz), DF is also a useful metric because it’s a summation of the total power inefficiency of all of a capacitor’s parasitics: ESR, ESL, and (to some extent) DCL. However, these parasitics are often better addressed individually in digital or low-power analog applications.

1 comment on “Parasitics & Capacitor Selection, Part 4B: Dissipation Factor

  1. Davidled
    July 8, 2014

    Lower value of DF might be cost the circuit and might improve the circuit performance. If the blog provides the example with a different capacitor and output performance comparison, at least, we could observe how much DF influences the circuit behavior. Then, engineers could select an appropriate CAP for their application.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.