Condensator MLCC (Ceramic Multistrat) OEM China

Selecția Clasei Dielectrice: C0G vs X7R vs X5R

The dielectric class is the single most consequential specification on an MLCC. It determines temperature stability, capacitance change under applied DC voltage, and aging drift. Many BOM spec sheets underspecify this, leading to production surprises when a factory substitutes within the same nominal value.

C0G / NP0 (Class I — stable dielectric). Capacitance change across the full operating temperature range: ±30ppm/°C maximum. Effectively zero DC bias effect. Zero piezoelectric noise. Ideal for timing circuits, oscillator load capacitors, RF matching networks, and any application where capacitance accuracy affects circuit function. Trade-off: C0G is only practical for values up to approximately 100nF in standard case sizes — the ceramic formulation has lower permittivity than Class II materials, so higher values require physically larger cases or are simply unavailable. Price premium over X7R of the same nominal value is typically 2–4×.

X7R (Class II — stable-ish). Temperature coefficient: capacitance changes within ±15% of the nominal value across –55°C to +125°C. Additionally, X7R capacitors exhibit significant DC bias derating: at the rated voltage, capacitance can drop 30–60% below the nominal value depending on manufacturer and case size. For decoupling applications where a 10µF capacitor might actually deliver 4–6µF under operating bias, this is a real design constraint, not a datasheet footnote. X7R is the standard choice for bypass and decoupling capacitors, coupling capacitors, and filtering where ±15% variation is acceptable. At 0402 and 0603 case sizes, X7R is the most commercially available and competitively priced dielectric class from Chinese manufacturers.

X5R (Class II — reduced temp range). Capacitance change within ±15% across only –55°C to +85°C — narrower than X7R. Often chosen for cost reasons in consumer electronics targeting ambient use only. DC bias derating is similar to X7R and can be worse in high-capacitance small-body parts. For any product that will experience elevated board temperatures (above 70°C ambient, or near power stages), X5R is a risk that outweighs its modest cost savings.

X8R and Y5V. X8R extends stability to +150°C and is relevant for automotive and industrial applications. Y5V (–82%/+22% change across temperature) should be treated as a filter-only or noise-suppression component — any design that depends on Y5V capacitance being close to nominal will behave unpredictably in production.

When sourcing through our sourcing service, we verify that the dielectric code on delivered reels matches the approved vendor part numbers, not just the nominal capacitance value.

Deratarea la Polarizare DC și Dimensiunea Carcasei: Regula de 50%

MLCC capacitance under applied DC voltage can be dramatically lower than the nominal (zero-bias) measurement. This is a materials property of Class II ceramics, not a quality defect — but ignoring it leads to under-capacitance in production circuits.

The 50% voltage derating rule. For X5R and X7R capacitors used in any power supply bypassing or decoupling role, select a component rated for at least twice the operating voltage. A 10µF / 6.3V 0805 X5R at 3.3V operating voltage (52% of rated voltage) will deliver approximately 6–7µF of effective capacitance at room temperature. The same 10µF / 10V 0805 X5R at 3.3V (33% of rated voltage) delivers 8–9µF. The marginal cost difference per reel between 6.3V and 10V rated parts at 0805 is typically under $5 per 4,000-piece reel. The difference in effective capacitance at operating voltage is material.

Case size interactions. High-capacitance values in small case sizes (e.g., 10µF in 0402) use thin dielectric layers to achieve the capacitance — which simultaneously pushes the DC bias derating curve steeper. A 10µF 0402 at 50% rated voltage may deliver under 2µF effective capacitance. The same capacitance in 0603 or 0805 will have a less aggressive bias curve. For power delivery network decoupling, it is generally better to use larger case sizes at modest capacitance values than to squeeze maximum capacitance into the smallest footprint.

Acoustic noise (piezoelectric effect). Class II ceramics have a piezoelectric response — voltage variations across the capacitor cause microscopic mechanical deformation, generating audible buzz in the 1–20kHz range. This is relevant for audio equipment PCBs and any board where the capacitors are mounted on a thin PCB near a speaker or microphone. Mitigation options: use C0G for audio-path decoupling (no piezo effect), use polymer or electrolytic capacitors on bulk supply rails, or specify soft-termination MLCCs (available from Murata and Samsung SEMCO) which mechanically decouple the ceramic body from the PCB pad.

For boards destined for PCB assembly in China, we recommend locking down the approved vendor list and dielectric class in the BOM before factory handoff to avoid equivalent substitutions that look identical in the pick-and-place machine but behave differently in circuit.

Lanțul de Aprovizionare: Murata/TDK vs CCTC/Walsin

MLCCs have experienced two significant supply disruptions in the past decade: the 2018 global shortage (driven by 5G infrastructure build-out) and component allocation tightening in 2021–2022. Chinese buyers now maintain safety stock, but lead times on factory-direct orders still extend to 6–14 weeks for non-stocked values.

Japanese and Korean tier-1 manufacturers (Murata, TDK, Yageo/KEMET, Samsung SEMCO). The default specification reference for most Western OEM designs. Murata’s GRT and GRM series are the benchmark for bias derating performance, and Samsung SEMCO’s CL series offers competitive pricing at 0402/0603 with consistent quality. The trade-off: lead times from authorized distributors stretch to 16–26 weeks for high-demand values; pricing is 40–80% higher than equivalent Chinese-manufactured parts.

Chinese-manufactured alternatives (CCTC, Walsin, FengHua, TDK China / Shenzhen Megasun). CCTC (China Zhenhua Group) and Walsin Technology produce X5R and X7R MLCCs across the standard case size range. Quality at the factory-direct level is significantly higher than what appears through spot-market gray channels. For consumer electronics targeting cost parity with Chinese-made end products, CCTC and Walsin represent a practical trade-off: 20–40% lower cost than Murata equivalents with REACH/RoHS certification, at the expense of slightly broader distribution on capacitance tolerance and higher individual-part variance on the DC bias curve.

What to verify when evaluating Chinese MLCC suppliers:

• Request the dielectric code documentation (not just nominal capacitance and voltage) — genuine C0G vs X7R must be documented in the part number system.
• Test DC bias derating against the published curve using an LCR meter with DC bias capability (Keysight E4980A, Hioki IM3533, or equivalent). Measure capacitance at 0V, 25% rated voltage, and 50% rated voltage. Compare against the manufacturer’s published curve. Significant deviation indicates either a wrong dielectric class or a substituted part.
• For AEC-Q200 grade parts (automotive applications), verify the test report — AEC-Q200 is a qualification test protocol, not a factory-level mark. The report should reference the specific part number, test house, and test conditions.

For short-run procurement and shortage mitigation, our PCB assembly guide covers BOM management strategies and approved vendor list setup.

Detectarea Contrafacerilor

The MLCC counterfeit risk is lower than for ICs or battery cells, but relabeling does occur — primarily, lower-grade Class II parts (X5R or even Y5V) relabeled as C0G, or parts with degraded dielectric (returns, rework salvage) repackaged as new. Counterfeit MLCCs are more common in spot market procurement than in direct factory or authorized distributor channels.

Capacitance measurement at operating frequency. An LCR meter at 1kHz (the standard test frequency) will confirm nominal capacitance and broadly identify wrong dielectric class, but it won’t catch degraded parts that happen to measure correctly at zero bias. Measure at the actual operating frequency of your application (100kHz for switching supply decoupling, 1MHz+ for RF bypass) and at the operating DC voltage. A degraded X7R part may measure close to nominal at 1kHz/0V bias but fall significantly short at 100kHz/3.3V.

Temperature sweep test. For C0G parts, perform a capacitance measurement across –40°C to +85°C (achievable with a bench-top thermal chamber or dry ice + oven sequence). Genuine C0G parts hold within ±30ppm/°C. X7R parts sold as C0G will show capacitance drift of several percent across this range — unambiguous identification.

Visual inspection under microscopy. The ceramic body color is not a reliable indicator of dielectric class, but the termination plating is visible under 40× optical microscopy. Genuine AEC-Q200 or automotive-grade parts use a Ni/Sn or Cu/Ag/Pd termination stack with defined layer thickness. Salvaged or reworked parts often show solder wicking, residue, or uneven termination edges.

Buy from accountable supply chains. Parts purchased from gray-market spot brokers at significant discounts vs authorized distributor pricing are the primary counterfeit vector. For MLCC procurement in volume, our sourcing service sources from factory-direct channels and Murata/Samsung SEMCO authorized distributors, with incoming inspection including LCR measurement at operating frequency and DC bias derating verification. The inspection step is documented and reported before parts reach your assembly line. See our inspection service for pre-shipment component verification protocols.