The ceramic capacitor DC bias effect means the part on your board is often not the part on the label. Class 2 ceramics shed a large fraction of their capacitance as DC voltage is applied across them, and the loss grows as that voltage approaches the rating. The value printed on the reel was measured under test conditions your circuit never reproduces.
Why Class 2 ceramics behave this way
The high volumetric efficiency of X5R, X7R and similar dielectrics comes from ferroelectric ceramics based on barium titanate. Their very large permittivity arises from domains within the ceramic that reorient in response to an applied electric field. That mechanism is what packs a useful capacitance into a tiny chip, and it is also the source of the problem. Once a steady DC field is applied, those domains progressively align and can no longer contribute the same incremental response to a small signal change. The measured capacitance, which is exactly that incremental response, falls. Push the field higher and the effect deepens, because there is less unaligned material left to respond. Class 1 dielectrics such as C0G and NP0 use a different, paraelectric ceramic with a far smaller permittivity and essentially none of this behavior, which is why they are stable but physically larger for the same value.
How much you lose depends on the part, not the value
Two capacitors with the same marked capacitance and the same dielectric code can behave very differently under bias. The dominant variables are the physical case size, the voltage rating, and the specific dielectric formulation the manufacturer chose to hit that capacitance in that package. Broadly, squeezing a large value into a very small case and a low voltage rating buys volumetric efficiency by using a thinner, more aggressive dielectric, and that part will lose the most capacitance under bias. A larger case with a higher voltage rating holding the same nominal value will keep far more of it. None of this is guesswork on your part: reputable manufacturers publish capacitance versus DC bias curves, and many provide online tools that plot effective capacitance at your actual operating voltage and temperature.
Why your bench meter does not show the problem
A handheld capacitance meter or an LCR bridge measures with a small AC test signal and, by default, no DC bias at all. Under those conditions a Class 2 ceramic reports close to its marked value, which is perfectly honest and completely unrepresentative. The same part sitting on a rail at most of its rated voltage will be delivering far less. Unless your instrument has a DC bias facility, or you build an external bias network and account for it, the bench measurement simply cannot see the effect. This is one of the more common ways an engineer talks themselves out of a correct diagnosis, because the meter agrees with the silkscreen while the circuit disagrees with both. Trust the manufacturer’s bias curve over any unbiased measurement.
Where the missing capacitance actually hurts
Bulk energy storage is the obvious casualty. A rail relying on ceramics to ride through load steps or brief input interruptions may have a fraction of the reserve the bill of materials implies. Filter corners move too: an LC output filter designed around a nominal value will have a higher corner frequency and different damping than intended, which changes ripple and can change loop stability. Timing and analog circuits suffer worse. Anything setting a time constant, an oscillator frequency, a filter response or a sample-and-hold droop rate needs a stable value, and a Class 2 ceramic under bias provides neither stability nor predictability. Supervisor delay circuits, soft-start ramps and reset timers are recurring offenders, drifting outside their intended window because the capacitor is not the size the designer assumed.
Designing so the effect cannot surprise you
Start by deciding whether the application needs a known value or merely needs enough charge. High frequency decoupling is fairly tolerant; a bulk hold-up capacitor or a filter corner is not. For anything tolerant, choose a part with a voltage rating comfortably above the rail and accept a modest loss, sizing from the bias curve rather than the marked value. For anything that needs accuracy, use C0G or NP0 ceramics, or a film capacitor, and accept the larger footprint. Where a big value is unavoidable at a small size, verify the effective capacitance from the manufacturer’s data at your rail voltage and design with that figure in the calculation. Then confirm the finished behavior on the bench, because the sum of bias loss, temperature drift and tolerance can be larger than any one of them.
What to check before you act
- Look up the capacitance versus DC bias curve for the exact part number.
- Use effective capacitance at your rail voltage in every filter calculation.
- Choose C0G, NP0 or film wherever the value must stay predictable.
- Prefer a larger case or higher voltage rating when bias loss matters.
Common questions
Does a higher voltage rating always fix the problem?
It reduces the problem but rarely eliminates it, because the effect depends on the field in the dielectric and on the formulation used. Check the actual bias curve rather than assuming a rating alone is enough.
Do C0G and NP0 capacitors lose value under bias?
Class 1 dielectrics are essentially free of the effect, which is why they are used for precision timing and filtering. The trade is a much lower capacitance for the same package size and cost.
The practical takeaway
The number on a Class 2 ceramic is a measurement made under conditions your circuit does not share. Pull the manufacturer’s bias curve for the exact part, design with the effective capacitance at your working voltage, and move to Class 1 or film wherever the value genuinely has to hold still.
References and further reading
- Department of Energy: how electronics losses turn into wasted energy and heat
- NFPA: electrical safety research and guidance on overheating equipment and fire risk
- CPSC: safety guidance for electronics and electrical products indoors
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
