Where the Real Cost Sits in a Derating Spec
When a capacitor is specified with a voltage derating, the cost is rarely in the component itself. The cost sits in the ripple current rating, the physical size, and the dielectric thickness required to survive the surge event. A 50 V derating on a 25 V rail is not a simple safety factor; it is a request for a thicker dielectric layer, which directly increases the volume of the part. For electrolytic capacitors, that thicker oxide layer also changes the ESR and the thermal resistance, which in turn forces a larger can size to keep the core temperature within limits.
For a procurement engineer, the practical question is not whether derating is necessary, but how much margin is actually priced into the offers you receive. Two suppliers can quote the same nominal capacitance and rated voltage, yet one delivers a part with a genuine surge capability of 1.2x the rated voltage while the other delivers a part that only meets the DC rating. The difference shows up on the test bench, not on the datasheet summary page.
Comparing Offers on Equal Terms: Define the Surge Profile
The most common source of comparison errors is the assumption that “rated voltage” is the same as “surge withstand voltage.” For aluminum electrolytic capacitors, the rated voltage is the maximum continuous DC voltage at rated temperature, while the surge voltage is typically 1.1 to 1.15 times the rated voltage for a limited duration (for example, 30 seconds on, 5.5 minutes off, for a specified number of cycles). For film capacitors, the surge rating is often expressed in terms of dV/dt (volts per microsecond), which is a completely different parameter.
To put two quotes on equal footing, you need to specify the following minimum set of parameters:
- Continuous DC voltage including the maximum bus voltage ripple component.
- Surge voltage magnitude and duration (e.g., 1.15x rated for 30 seconds, or 1.5x rated for 100 ms in a hot-swap event).
- Number of surge cycles expected over the product lifetime (e.g., 1000 cycles).
- Operating ambient temperature and the allowable core or case temperature rise.
- Ripple current at the switching frequency (e.g., 1.5 A at 50 kHz) and the ESR at that frequency.
- Failure mode requirement (open vs. short) after a surge event.
If a supplier quote does not include the surge cycle count or the dV/dt limit, request it in writing. A part that can handle a single surge but cracks after 50 cycles is not a drop-in replacement. This is particularly relevant for MLCCs where mechanical flex and thermal shock during surge testing often cause microcracks that later lead to insulation resistance (IR) failures.
Incoming Inspection Checklist for Surge and Derating Verification
Incoming inspection should not attempt to replicate full qualification testing, but it should catch gross mismatches in dielectric thickness and internal construction. The following checks are practical for a receiving department with basic lab equipment:
| Check | Method | Pass / Fail Threshold |
|---|---|---|
| Capacitance at rated DC bias | LCR meter with DC bias fixture (e.g., 0 V and rated voltage) | Capacitance drop of no more than 15% from 0 V to rated DC bias for MLCC; electrolytic within 20% of nominal. |
| Surge voltage (single pulse) | Charging capacitor to 1.15x rated voltage, discharging through a low-inductance load | No short, no leakage current increase greater than 1.5x initial value after 10 cycles. |
| ESR at switching frequency | Impedance analyzer or ESR meter at 100 kHz | ESR within 20% of the datasheet maximum value at 20 °C. |
| Leakage current | Apply rated voltage for 5 minutes, measure leakage | Leakage below the datasheet limit (often 0.01 CV or 3 µA, whichever is greater, for electrolytics). |
| dI/dt capability (film and ceramic) | Measure voltage drop across a calibrated shunt during rapid discharge | No voltage collapse indicating internal breakdown at 70% of the specified dI/dt limit. |
This table is intentionally conservative. If your application requires a 10-year life at 105 °C ambient, incoming inspection cannot verify that. What it can do is establish that the parts you received are consistent with the samples that passed your qualification. If a new batch fails the surge pulse test, do not renegotiate the derating; investigate the supplier’s internal dielectric thickness and termination quality first.
Negotiation and Stocking Strategy for Derated Capacitors
When negotiating pricing for a part with a demanding derating spec, the key lever is the acceptance tolerance. Suppliers often quote a standard capacitance tolerance (e.g., ±20% for electrolytic, ±10% for MLCC) that is wider than your application requires. If your design can tolerate a tighter window (e.g., ±10% for the electrolytic), you are effectively asking the supplier to sort parts. This sorting cost is real, and it should be negotiated separately from the unit price.
A more effective approach is to consolidate the surge requirement into a single specification that the supplier can test at the end of the line. Instead of asking for a part that “meets 1.15x surge for 30 seconds,” define a surge pulse shape (e.g., 1.15x rated for 1 second with a 10 ms rise time) and a cycle count. Suppliers that have this test on their line will quote a lower price than those that have to set up a new fixture. Ask for the test report in the lot documentation.
For stocking, consider the storage conditions that affect surge withstand. Electrolytic capacitors that have been stored for more than two years may require reforming (gradual voltage application) before they can handle a full surge. Film capacitors are less sensitive, but humidity absorption can affect the dielectric strength in high-voltage parts. Your incoming inspection should include a storage age check and a reform cycle if the shelf life exceeds 24 months.
Finally, keep a functional sample of the approved part in your lab. When a new quote arrives at a lower price, run it through the same surge cycle test and the DC bias capacitance check. This simple practice prevents the common scenario where a cheaper part saves 3% on the BOM cost but requires a redesign because its true derating capability is lower than the incumbent. The cost of an engineering change order and a field retrofit will exceed any component-level savings by a wide margin.