How to Verify Capacitor Quality: A 6-Step Checklist for Procurement and Engineering Teams

When This Checklist Matters

If you're sourcing ceramic capacitors, tantalum capacitors, MLCCs, or relays for a new product launch or a high-volume production run, the first batch you receive sets the tone. One bad lot can delay a project by weeks and cost thousands in rework. I'm a quality compliance manager at an electronics distributor — I review roughly 200 unique component batches each year, and in 2024 I rejected about 12% of first deliveries due to spec deviations or hidden defects.

This checklist is for engineers and procurement teams who need a repeatable method to verify that parts from any supplier — whether it's Kemet, another authorized distributor, or a second-tier source — actually match the datasheet and your manufacturing tolerances. It's not theoretical; I've used these steps on everything from T491 tantalum caps to 3210-series MLCCs and Duravx Extreme relays.

Below are six steps. Skip one at your own risk (I have, and I paid for it).

Step 1: Cross-Check Part Markings Against the Datasheet

This sounds obvious, but you'd be surprised how often markings are misread or omitted. Take a sample from the shipment and compare the printed codes with the manufacturer's datasheet. For Kemet T491 series tantalum capacitors, for example, the marking includes capacitance, voltage rating, and case size code. I once received a reel marked "106" (10 µF) that was actually "226" (22 µF) — the vendor had swapped reels (note to self: never trust a reel label without verification).

Pro tip: use a best multimeter with capacitance measurement mode to quickly spot-check a few units. If the reading deviates more than the tolerance spec (e.g., ±10% for X7R), flag the batch.

Step 2: Visual Inspection Under Good Lighting

Look for cracks, chips, discoloration, or bent terminations. This is where the quality reputation of a brand like Kemet really shows. In my experience, authorized distributors of well-known brands rarely ship physically damaged parts (thankfully). But when they do — often due to poor packing — it's a red flag. Inspect at least 5% of the quantity or 20 units, whichever is larger.

Pay special attention to the edges of MLCCs (like the 3210 size). Hairline cracks are hard to see but will cause failures during soldering. If you spot even one cracked unit, reject the entire batch — in my opinion, it's not worth the risk. I learned this the hard way when I skipped the visual check on a rush order (ugh).

Step 3: Measure Electrical Parameters on a Sample Set

Don't just trust the datasheet. Use an LCR meter or a high-quality multimeter to verify capacitance, ESR, and dissipation factor. For tantalum capacitors like Kemet T491, ESR is critical — a higher-than-specified ESR can cause thermal runaway in power circuits.

I typically test 10 units from different reels or trays. Record the values and look for consistency. A batch where one unit is 9.8 µF and another is 10.2 µF is fine; a batch where the range is 8.5 µF to 11.5 µF suggests poor process control (which, honestly, is a deal‐breaker for any production environment).

Step 4: Check Temperature Coefficient (for Ceramics)

This step is often ignored by procurement teams who assume "ceramic is ceramic." For Kemet Duravx Extreme series and other high‐reliability MLCCs, the temperature coefficient (X7R, X8R, C0G, etc.) dictates how capacitance changes with temperature. If your circuit requires stable capacitance from -55°C to +125°C, don't accept a Y5V substitute (surprise, surprise — it'll drop 80% capacitance at 85°C).

How to verify: heat a few caps with a hot air station (carefully) to 85°C and measure the capacitance. Compare the drift to the datasheet curve. It takes 10 minutes and can save you from a full board failure in the field.

Step 5: Test Under Realistic Voltage Bias

MLCCs lose capacitance when DC voltage is applied — this is called DC bias effect. A 10 µF rated capacitor might drop to 4 µF at 50% of its rated voltage. This matters a lot for power supply decoupling. I ran a blind test with our engineering team: same nominal cap from two different brands under 10V bias. One delivered 8.2 µF, the other 5.1 µF. The cost difference was $0.003 per piece, but on a 50,000‐unit run that's a $150 savings for an inferior component. Not worth it (in my opinion).

Use a bias‐capable LCR meter (most modern ones have it) or build a simple test circuit. For Kemet parts, their datasheets usually provide bias curves — but verify anyway. I've seen batches where actual bias performance was 15% lower than the spec.

Step 6: Assess Supplier Documentation and Traceability

This step ties directly into quality perception of your brand. When a customer sees that your product uses authentic, traceable components from a reputable manufacturer, their trust increases. Request the certificate of conformance (CoC) and check that the lot number matches the shipped reels. For Kemet parts, the lot code can be traced back to the production date and factory.

If the supplier can't provide a CoC or tries to give you a generic one, that's a major red flag (note to self: walk away). The total cost of a counterfeit or substandard component isn't just the replacement cost — it includes field failures, warranty claims, and brand damage. As the value proposition anchor states, "Total cost of ownership includes base product price, setup fees, shipping, potential reprint (or rework) costs — the lowest quoted price often isn't the lowest total cost." So apply that same thinking here.

Common Mistakes to Avoid

I've made most of these mistakes myself, so consider this a free lesson:

• Skipping Step 1 because “the supplier is authorized.” Authorized distributors can make errors too. Always verify markings.
• Testing only one unit. Sample size matters — test at least 10 pieces from different positions in the shipment (top, middle, bottom of the reel).
• Ignoring lead time uncertainty. The value of guaranteed delivery isn't just speed — it's certainty. If you rush a quality check to meet a deadline, you may end up with a defective batch that delays you even more.
• Assuming “better brand = no need to test.” Kemet makes excellent tantalum caps, but even the best manufacturers have occasional yield issues. Test anyway.
• Using a cheap multimeter for capacitance. Not all multimeters are equal. For accurate readings, use a meter with a capacitance range under 1% basic accuracy — something you'd call a best multimeter for electronics work.

I'm not 100% sure this checklist covers every possible failure mode (take it with a grain of salt), but following these six steps has saved my team from at least three major production delays in the last two years. If you implement even half of them, you'll significantly reduce the risk of getting burned by a bad batch.

(Funny enough, the most common reaction I get from engineers who start doing these checks is: "Why didn't we do this sooner?")