Articles

Replacing a Ceramic Capacitor? Compare Usable Capacitance

A capacitor disappears from stock, and a supplier offers another part with the same capacitance, voltage rating and footprint. That is a useful starting point, not an engineering approval. The replacement must still do the same job in the assembled product.

For a multilayer ceramic capacitor, or MLCC, the number in the bill of materials does not tell the whole story. A practical substitution review compares the exact parts under the circuit’s operating conditions, records what remains uncertain, and tests the behavior that the capacitor supports.

Start with the capacitor’s job

Identify the reference designator and the function before comparing catalog entries. Is this an output capacitor required by a regulator, a local supply bypass, a filter component, or a timing element? The acceptance criteria will differ. A successful boot is weak evidence for a part whose critical job appears only during a load change.

Write down the operating voltage range, expected temperature range and relevant load states. Identify the component datasheet or design calculation that sets the requirement. Keep minimum effective capacitance separate from nominal capacitance, and check any maximum capacitance or equivalent series resistance (ESR) limits too.

For example, the TI TPS7A20 datasheet, revision H, distinguishes its nominal output-capacitor recommendation from an effective capacitance range required for stability. Its recommended-operating-conditions table and footnotes specify 0.47 µF minimum and 200 µF maximum effective output capacitance. These are requirements for that device, not universal limits for other regulators.

Compare the exact part numbers, not just X7R

High-dielectric-constant ceramic capacitors can lose capacitance under DC bias. Murata’s DC-bias guidance explains why this must be considered in the actual equipment. A shared dielectric designation or voltage rating does not establish that two parts retain the same capacitance on your rail.

Obtain the manufacturer’s data for each complete orderable part number. Record the curve’s measurement conditions and whether its values are typical or guaranteed. A typical curve is useful for screening, but it is not automatically a worst-case production limit. If the available data cannot support the required margin, ask the manufacturer for clarification or choose a better-documented candidate.

Also distinguish an operating-condition estimate from a bench measurement taken without bias. Murata’s capacitance-measurement FAQ identifies temperature, AC and DC voltage, frequency, and time as influences on high-dielectric-constant ceramic capacitance. Record these conditions so the comparison can be repeated.

A small example that changes the decision

Consider a hypothetical circuit whose documented requirement is at least 4.7 µF of effective capacitance. Two candidate parts both carry a nominal value of 10 µF. Suppose the review produces screening estimates of 6.0 µF for candidate A and 4.0 µF for candidate B at the intended operating point.

  • Candidate A clears that screening threshold by 1.3 µF, but still needs its uncertainty and remaining requirements checked.
  • Candidate B falls below it by 0.7 µF. Matching the printed 10 µF value does not rescue the substitution.
  • If those estimates exclude tolerance, temperature or other relevant effects, neither is yet a complete lower-bound assessment.

These invented values explain the decision process. They are not manufacturer data, measurements, or approval of any real capacitor. Do not apply a generic percentage reduction to every MLCC, or multiply unrelated curve factors without checking how the manufacturer defines them.

Hypothetical capacitor comparison: two nominal 10 µF candidates screen at 6.0 and 4.0 µF against a 4.7 µF minimum.
Illustrative screening comparison, not real part data. Both candidates are nominally 10 µF; only candidate A clears the hypothetical 4.7 µF threshold. Final approval still requires uncertainty and circuit checks.

Make the approval record useful to purchasing

A short comparison sheet should let another engineer reconstruct the decision and let purchasing order the exact approved item. Include:

  • Identity: original and proposed manufacturer part numbers, reference designators and affected board revision.
  • Fit: package dimensions, height, termination and assembly requirements, rather than footprint name alone.
  • Electrical evidence: voltage and temperature limits, nominal tolerance, effective-capacitance evidence, and application-specific impedance or ESR requirements.
  • Traceability: source documents, revisions, access date, assumptions, reviewer and unresolved questions.
  • Release scope: the builds or board revisions covered by the approval, plus the tests that must pass.

A supplier’s proposed alternative should remain identifiable as a proposal until that review is complete. Keep the decision with the hardware revision record, rather than leaving it only in an email conversation.

Retest the behavior, not just the component

Define pass limits before testing. Use representative boards, the intended supply and relevant operating corners. Compare the original and replacement under the same setup, recording the board revision and exact fitted part. For a supply capacitor, useful observations may include rail excursions during load transitions, startup and shutdown behavior, and the product’s response during its demanding operating modes.

Changing capacitance can also change startup behavior. Treat the power-rail startup and inrush check as a separate test rather than assuming that more nominal capacitance is always better. Preserve traces and test conditions, including probe connection and load stimulus, so a later reviewer can distinguish an actual improvement from a changed measurement setup.

This is a system-level decision. MakersGround’s Gomicron development work spans mechanical design, PCB electronics, firmware and control, slicing software, prototyping and manufacturing. That multidisciplinary context explains why a component change belongs in the product’s engineering record; it does not imply that the hypothetical capacitor example came from that project.

The release question is simple: does this exact replacement meet the documented requirements, with enough evidence to repeat the decision? Approve that conclusion, not a matching line in a distributor’s search results.