I Thought I Was Ordering Capacitors. Turns Out I Was Buying Hopes.
In early 2024, I placed an order for 10,000 pieces of Kemet C0805C222K5RACTU. That's a 2,200 nF, 50V, X7R ceramic capacitor. Standard part. Nothing exotic.
They arrived. I sent a sample to our quality team for verification. The measured capacitance? 1,850 nF. Almost 16% below the nominal value. And within spec per the datasheet (which allows -20%/+80% for this series).
So. The 2,200 nF issue. That's the surface problem.
What You Think the Problem Is (and Why It's Not)
Most engineers and buyers I talk to assume the problem is 'bad parts' or 'counterfeit goods.' That's what I assumed too, after the first batch in 2020 showed similar deviations. I spent weeks auditing our supplier, testing samples, and even arguing with the manufacturer's rep.
Turns out, I was looking in the wrong direction.
The real issue isn't that the parts are wrong. It's that our expectations—and our testing methods—aren't aligned with how ceramic capacitors actually work.
I'm not a design engineer, so I can't speak to the physics of dielectric materials in detail. What I can tell you from a procurement perspective is that the gap between 'what we order' and 'what we get' is mostly a case of assumption failure.
We assume '2,200 nF' means exactly that capacitance under all conditions. In reality, it's a specific value measured at a specific DC bias, temperature, and frequency. Change any one parameter, and the capacitance can drop by 50% or more.
Learned never to assume 'same specifications' means identical performance across vendors after comparing Kemet MLCCs with a cheaper alternative in 2022. Both marked 2,200 nF. The Kemet parts held 1,950 nF at 10V DC bias. The generic ones dropped to 1,200 nF.
That's a 40% difference hidden behind the same part number.
The Deeper Issue: We Don't Test for What We Actually Need
We didn't have a formal process for verifying DC bias characteristics at the time. Cost us when a prototype failed during a high-voltage stress test—the capacitors couldn't maintain rated capacitance above 30% of rated voltage.
The third time a similar issue surfaced—this time with a customer complaint about a power supply ripple exceeding spec—I finally created a pre-qualification checklist that includes voltage coefficient testing. Should have done it after the second time.
The deeper cause is this: Most procurement teams treat datasheets as truth rather than guidelines. A datasheet is not a guarantee. It's a range of possibilities, with a marketing-friendly nominal value at the top.
Why does this matter? Because when you're ordering an MLCC like the Kemet C0805, the real question isn't 'Is it 2,200 nF?' It's 'Will it be 2,200 nF under my actual operating conditions?'
The answer is almost always no. Sometimes it doesn't matter. Sometimes it costs you a lot.
The Price of Ignorance: What These Assumptions Cost
Let me give you a concrete example from Q2 2024. Our engineering team needed 1,000 pieces of a specific X7R capacitor for a prototype run. We ordered from a new vendor to save $0.03 per piece—a total saving of $30. The parts arrived, looked identical, and were marked with the same capacitance code.
We didn't verify the DC bias characteristics. That was a mistake.
The prototype failed during testing. The root cause: the capacitors couldn't maintain rated capacitance at 80% of rated voltage. We lost two weeks of development time and $5,800 in expedited shipping for replacement parts from our regular supplier.
The $30 saving cost us $5,800 and a missed deadline. The vendor who couldn't provide proper invoicing cost us $2,400 in rejected expenses back in 2022. That unreliable supplier—who promised 'probably on time' delivery—made me look bad to my VP when materials arrived late.
To be fair, the alternative wasn't always clear. We didn't have a formal process for verifying voltage coefficient data before that incident. Cost us the project timeline.
The cost of getting this wrong isn't just the price of the parts. It's the rework, the delays, the lost customer confidence. In March 2024, we paid $400 extra for rush delivery from Kemet's distribution channel. The alternative was missing a $15,000 event. The math was simple.
After getting burned twice by 'probably on time' promises, we now budget for guaranteed delivery from authorized distributors. It's not the cheapest option—but the uncertainty premium is worth it when you're dealing with tight production schedules.
Granted, this requires more upfront work—vetting suppliers, requesting DC bias data, and testing samples. But it saves time later. That relationship saved us when we needed an urgent batch of 10,000 MLCCs in Q3 2024. The supplier who had maintained a consistent internal quality record could prioritize our order.
The Modest Fix: What I Do Differently Now
This gets into some territory that's not entirely in my control—the physics of MLCCs is a design engineering domain. What I can tell you from a procurement perspective is how to build a buffer against this problem.
First, I always request DC bias curves from the manufacturer before placing a large order. Most reputable brands like Kemet provide this data on request. If a supplier can't produce it, I treat that as a red flag.
Second, I order a small sample batch—100 pieces—and send it to our quality lab for verification under actual use conditions. This costs maybe $200 in testing time. It's saved us thousands.
Third, I maintain a shortlist of vendors with proven track records. For MLCCs, that means authorized distributors who can trace the supply chain back to the original manufacturer. We've narrowed down to about 3 main vendors for critical components, including Kemet for specialty ceramics and high-reliability tantalum parts.
I'm not 100% sure this approach would work for every component type. But for ceramic capacitors, it's reduced our rejection rate from about 12% to under 2% over the past two years.
Take this with a grain of salt: my experience is specific to passive components in industrial electronics. Your mileage may vary.
The short version: Don't take a datasheet at face value. Verify under your conditions. And build relationships with suppliers who can provide the data you need—not just the lowest price.
That's it. Simple.
"The question isn't whether the capacitor is 2,200 nF. It's whether it will be 2,200 nF when you need it to be."
Prices as of January 2025; verify current rates. This is based on personal procurement experience; consult your engineering team for specific design decisions.
