If you're an engineer or buyer looking at Kemet T491 tantalum capacitors for a cordless phone charger or an automotive ECU, this checklist is for you. Over the past few years I've reviewed roughly 200 part numbers a year for an electronics distributor. In our Q1 2024 quality audit, about 11% of incoming lots had at least one parameter that missed the datasheet spec. Not every miss was critical, but every miss needed a decision.
Here's the short version: this is a five-step checklist. It's not exhaustive—but it will catch most problems before they become a field failure.
1. Define the application, not just the part
First, ask what the capacitor will actually do. For a cordless phone base station, a standard T491 with 16V rating might be fine. For an engine control module, you're looking at AEC-Q200 compliance—that's the automotive stress test standard from the Automotive Electronics Council. It covers temperature cycling, humidity, vibration, and more. Your part might be the same capacitance and voltage, but the automotive grade version often has different internal construction and screening.
Write down the real requirements: capacitance value, voltage derating, temperature range, ESR, leakage current, and lifetime. If you can't specify ESR, you're guessing. Actually, let me rephrase that: you're making a decision with incomplete information. (That's one of the first things I look for when reviewing incoming batches.) For cordless phones, ripple from the charging circuit is a major concern; low-ESR tantalum caps like the T491 help there. But you still have to verify the ESR on the specific lot.
2. Get the datasheet and decode the marking
The datasheet is the minimum. I've seen engineers skip it because they know the part—and then assemble a reel that doesn't match the silkscreen. At a minimum, check the part-numbering system. For example, a Kemet T491C106K016AT gives you the series, case size C, 10 µF, 16V, tolerance K (10%), and the packaging. It's all there, but you have to read it.
One frustrating thing: component markings are often short codes, not the full part number. I remember a batch where the marking said C300. I wanted to say it meant a capacitance/voltage combo, but I wasn't sure—don't quote me on that without the datasheet. We had to cross-reference the full number to verify. Eventually we found it in the datasheet—C300 was a production date code. That's a 10-minute check that prevents a 3-week delay. The same applies to the T491 family: different case sizes, tolerances, and voltage ratings share similar prefixes, so read all the digits.
3. Test with a multimeter that actually does the job
Here's where a decent multimeter matters. For checking capacitors in the field or incoming inspection, you need a meter with capacitance mode and, ideally, true RMS and peak hold. If you're searching for the best multimeter for automotive, those features are non-negotiable. A cheap meter might be fine for voltage checks, but it won't tell you whether a capacitor is reading 9.8 µF or 7.2 µF when it should be 10 µF.
Practical steps:
- Discharge the capacitor through a resistor first. (Do not short it with a screwdriver—that can damage it and you.)
- Measure capacitance with the part out of the circuit for bench verification. In-circuit measurements are often meaningless.
- If your meter has ESR mode, use it. ESR drift is a classic early warning sign. If not, at least compare the measured value against the datasheet tolerance. A 10% tolerance means 9.0 to 11 µF for a 10 µF part, not roughly 10.
- For automotive transient checks, use peak hold to capture voltage spikes—that can tell you if the capacitor is handling ripple properly.
4. Verify the supply chain
Counterfeit parts are still a problem, and the problem is getting worse. I once had to trace a suspicious lot back to a third-party broker because the price was too good. The datasheet said Kemet, the marking said Kemet, but the confidence was wrong. In that situation, the numbers said go with the budget option—30% savings on 10,000 units. My gut said no. We ended up sticking with an authorized distributor, and later a batch of the cheap parts failed during burn-in. It cost more upfront, but it was nowhere near the cost of a recall.
If you're unsure, contact an authorized source. The team at Kemet Fort Lauderdale has been helpful every time I've had a question about a specific lot. They can pull up production records and confirm the marking. That traceability is worth something.
5. Run a small-batch burn-in before committing
Before you order 50,000 units, take a sample from the actual lot. We typically do a high-temperature test (125°C for 48 hours) plus a loaded humidity cycle. It's not a full qualification, but it catches the obvious early-fail issues. The frustrating part is when a new vendor's parts pass the datasheet but fail the burn-in—you'd think a written spec would prevent that, but manufacturing drift happens. We ran a blind test with our engineering team once: same T491 spec from three manufacturers. Without knowing the brands, 80% of the team ranked the Kemet sample as the most consistent in capacitance and ESR. That didn't mean the others were bad—it meant the Kemet batch was within a tight distribution. Consistency matters when you're making the same product year after year.
Common mistakes and what to do instead
Here's where I see people burn themselves:
- Ignoring ESR. A capacitor can pass a capacitance test and still behave badly at high ripple. Measure ESR, or at least verify it using your meter's mode.
- Using a multimeter without capacitance mode. You can't compare values if you're only measuring resistance. And yes, you need peak hold for automotive work.
- Mixing automotive and industrial grades. It's tempting to save money with an industrial part in an automotive design. That's a gamble, and it goes against the spirit of AEC-Q200.
- Trusting memory over the datasheet. I've misidentified codes before (the C300 example above). So now I always check the datasheet or call support.
In this market, what was best practice in 2020 may not apply in 2025. But the need for verification is stronger than ever. The fundamentals haven't changed—verify what you're buying. The execution has transformed. Thirty years ago, tantalum capacitors were mostly a military-grade part. Now they're in cordless phones and engine control units. What hasn't changed is the cost of a missed spec. One bad batch can set your launch back a month and wipe out the savings.
So: define your application, read the datasheet, test with a real multimeter, verify the chain, and burn in samples. That's the checklist. It won't make every decision easy, but it'll keep you out of the worst kind of surprise.
