Within PCBs, multilayer ceramic capacitors (MLCCs) are probably the most ubiquitous components and, at the same time, the ones most likely to produce unpleasant surprises on the test bench. It is a common scene: a 10 µF capacitor that, once subjected to its operating voltage, behaves as if it only provided 3 µF. This loss of capacitance is not a manufacturing defect, but an inherent physical behavior that is often overlooked.
This phenomenon is one of the most underestimated in electronic design. Understanding why this capacitance drop occurs, what real risks it poses to system stability, and how to anticipate it from the component selection stage makes the difference between a theoretical design and reliable hardware.
This article explains, in a practical and easy-to-follow way—without assuming extensive prior knowledge—what DC Bias is, why it occurs, what real consequences it has on a PCB design, and above all, how we can anticipate and mitigate it from the component selection stage.
1. What is an MLCC and why is it used so widely?
An MLCC (Multi-Layer Ceramic Capacitor) is the most widely used passive component in modern electronics. Its structure consists of alternating layers of ceramic material (the dielectric or insulator) and metallic electrodes, stacked and compressed into SMD package formats that can be extremely small (such as 0805, 0603, 0402, 0201 and even smaller) or somewhat larger (such as 1206, 1210, 1808, 1812 and above).
Their popularity is not accidental; they offer a combination that is hard to beat: extremely small size, very low production cost, and excellent high-frequency performance thanks to their low parasitic resistance and inductance (ESR and ESL). For this reason, it is rare today to find a power rail, filter, or signal coupling stage that does not rely on them to operate.
However, that versatility comes with a technical “fine print” that is often overlooked in the early stages of design: DC Bias.
Figure 1. Four common applications of MLCCs in electronic design.
2. The DC Bias phenomenon explained
2.1. What is it exactly?
DC Bias, also known as the voltage coefficient of capacitance, is the effective reduction in the capacitance of an MLCC when a DC voltage is applied. In other words, a capacitor rated at 10 µF may behave as if it were 4 µF, 2 µF, or even less, simply because it is biased with the normal operating voltage of the circuit.
The most important thing to understand is this: it is not a defect, nor a failure, nor a tolerance issue. It is an inherent physical characteristic of the dielectric material used, and manufacturers specify it (although sometimes you have to look closely in the datasheet to find it).
Analogy for less technical profiles: imagine a sponge that, at rest, can absorb one liter of water. If you squeeze it with your hand while it is trying to absorb, it will barely hold half a liter. The capacitor behaves in the same way: at rest it provides its nominal capacitance, but when you “squeeze” it with DC voltage, its effective capacitance decreases.
Figure 2. Effect of DC Bias on effective capacitance depending on the dielectric type.
2.2. Why does it happen?
The key to this behavior lies in the dielectric material used, typically barium titanate, a ferroelectric compound. In these materials, the internal crystal structure contains electric dipoles that spontaneously align. However, when a DC voltage is applied, the resulting electric field forces these dipoles to orient in a fixed direction, effectively “locking” part of their mobility. Since a capacitor’s capacitance depends directly on the ability of these dipoles to respond to voltage variations, this internal “hardening” leads to an effective loss of the microfarads that the component can provide to the circuit as the voltage increases.
MLCCs are classified into two main families:
- Class I (for example, C0G / NP0): they use highly stable paraelectric dielectrics such as magnesium titanate. They show almost no variation with voltage, temperature, or aging (approximately 2% capacitance loss). Their drawback is that they offer relatively low capacitance values and a higher cost per unit of capacitance.
- Class II (for example, X5R, X7R, X7S, Y5V…): they use ferroelectric dielectrics mainly based on barium titanate. They allow much higher capacitance values in very small packages, but in return they exhibit a significant dependence on voltage, temperature, and aging.
Figure 3. Microscopic difference between paraelectric and ferroelectric dielectrics.
DC Bias is, essentially, a phenomenon almost exclusive to Class II capacitors. When a DC voltage is applied, the ferroelectric domains in the material become polarized and lose their ability to respond to additional changes in the electric field. The result is a decrease in effective permittivity, and therefore in capacitance.
As a practical rule of thumb, the more compact the package is for the same capacitance and voltage rating, the stronger the DC Bias effect tends to be. A 10 µF / 25 V in a 0402 package will lose much more capacitance under voltage than the same value in a 1210 package.
3. Real impact on design
The most direct consequence is obvious: the effective capacitance in the circuit is lower, sometimes much lower, than the nominal capacitance. And this has implications that, if not anticipated, translate into real problems:
- Poor power supply decoupling: if we place a 10 µF capacitor next to a microcontroller’s supply rail and, under operating voltage, those 10 µF effectively become 3 µF, the circuit will not have the intended charge reserve. This can lead to transient voltage drops, spurious resets, or erratic behavior.
- Shifted cutoff frequency in filters: an LC or RC filter designed for a given capacitance will have its cutoff frequency shifted, degrading noise filtering or dynamic performance.
- Instability in DC-DC converters: most switching regulators require a specific range of input and output capacitance to ensure control-loop stability. Severe DC Bias can push the converter out of its stable operating region, causing oscillations or even unacceptable ripple.
- Insufficient bulk capacitance: in power stages with high transient current demand (processors, power stages, switched LEDs, etc.), the actual reserve capacitance can be significantly lower than expected, leading to voltage drops that do not appear in simulation if DC Bias is not modeled.
4. Consequences in the circuit when operating near the rated voltage
Choosing a capacitor whose rated voltage is very close to its operating voltage has a direct impact on system performance. In Class II materials such as X7R or X5R, this proximity drastically reduces the available effective capacitance, which particularly affects rails supplying digital electronics. As this charge reserve decreases, the rail may exhibit higher-than-expected ripple, compromising the stability of the voltage supplied to the integrated circuits’ power pins (VDD).
The most challenging aspect of this phenomenon is that it often goes unnoticed during debugging phases. Since it is not an obvious catastrophic failure, it is easy to attribute erratic behavior to other causes, such as external noise or logic errors, when in reality the root cause is a component selection that does not provide the required capacitance under actual operating conditions.
Figure 4. Degradation of decoupling when effective capacitance drops due to DC Bias.
5. How to mitigate or avoid the problem
The good news is that DC Bias is predictable and manageable. It is not a random phenomenon, and with proper component selection its impact can be drastically minimized.
5.1. Proper selection of the rated voltage
The most widely used rule of thumb is to never operate above 50% of the capacitor’s rated voltage, and ideally to stay below 30–40%. This means selecting a voltage rating that is two to three times higher than the operating voltage. In critical applications, this margin should be even greater. In fact, it is standard industry practice to use 100 nF capacitors with a 50 V rating for decoupling low-voltage rails, ensuring that the effective capacitance remains as close as possible to the nominal value.
5.2. Oversizing the capacitance
If it is known that DC Bias will reduce the effective capacitance, a straightforward solution is to specify a higher nominal value. For example, if a design requires 10 µF effective capacitance at a 5 V bias, it may be reasonable to specify 22 µF or even 47 µF nominal, depending on the dielectric and package size.
5.3. Choosing more stable dielectrics
When stability is critical (precision filters, feedback loops, references, etc.), it is advisable to use Class I dielectrics (C0G/NP0), which are essentially unaffected by DC Bias. Their available capacitance is lower, but they offer extremely stable behavior.
Within Class II, there are also differences: dielectrics with better thermal behavior (such as X7R, X7S) typically also exhibit a more favorable DC Bias characteristic than less stable types (such as Y5V).
5.4. Using larger package sizes
For the same capacitance and voltage rating, a larger package typically exhibits less DC Bias degradation, since it provides more dielectric volume. Moving from a 0402 to a 0603 or 0805 can make a noticeable difference.
5.5. Using a combination of capacitors
In power rails, it is common to combine capacitors of different values and technologies: an electrolytic or tantalum capacitor for bulk storage, several Class II MLCCs for mid-range capacitance, and small-value Class I MLCCs for high-frequency decoupling. This strategy distributes the load and reduces the overall sensitivity of the design to DC Bias.
5.6. Reviewing manufacturer curves
This is perhaps the most important advice: it is not enough to look at the nominal capacitance in the datasheet. It is important to consult the curves of Capacitance vs. DC Voltage (and also vs. Temperature and vs. Frequency ) that manufacturers publish for each part number.
These curves directly show the actual capacitance that can be expected under the specific operating conditions of the design. Two components with the same nominal capacitance, voltage rating, and tolerance can behave in radically different ways under DC bias.
6. Best design practices
As a summary, here are some practical recommendations to minimize surprises:
- Design based on effective capacitance, not nominal capacitance. The capacitance that matters is the actual one under operating conditions.
- Leave ample voltage margin. Select MLCCs whose rated voltage is at least twice the operating voltage.
- Use Class I when precision is required. In filters, control loops, and sensitive analog circuits, the stability of C0G/NP0 compensates for their higher cost and lower capacitance density.
- Prefer larger packages when PCB space allows and DC Bias is a concern.
- Validate in the laboratory the effective capacitance and the transient response of the design, especially in critical power stages.
- Document component selection decisions: this helps procurement and sourcing teams understand why not all “equivalent” capacitors in catalogs are truly equivalent in the circuit.
7. Conclusion
DC Bias is not a defect: it is an inherent property of high-permittivity ceramic dielectrics that allows MLCCs to achieve such high capacitance values in such small sizes.
Understanding the phenomenon, reviewing manufacturer curves, leaving voltage margin, choosing the right dielectric, and combining technologies enables the design of robust, stable, and predictable circuits. Ignoring it, on the other hand, leads to subtle issues—instability, sporadic resets, filters that do not filter as they should—that often appear late, during validation phases or, worse, in the field.
For hardware teams, the key is to design with effective capacitance in mind. For procurement teams, the key is to understand that two capacitors with the same nominal capacitance, voltage rating, and tolerance may not be functionally equivalent in the final circuit. Selecting an MLCC correctly is not just a matter of specifications, lead time, and price: it is also a matter of dielectric type, voltage margin, and performance curves.
In a sector where more and more is demanded in less space and at lower cost, understanding these details makes the difference between a design that works on paper and one that actually works in practice.
Eurotronix, experts in MLCCs since 1986
At Eurotronix, we offer high-quality MLCC ceramic capacitors manufactured by leading brands in the market. In addition, our team of Product Managers and FAEs will support you in your projects, providing reliable, safe solutions tailored to today’s electronic design challenges.
Authors
Álvaro Guerris
MLCC Product Manager
Néstor Galera
Electronic compontents Field Application Engineer
