AKKN ElectronicsWhen an output capacitor series is discontinued or its lead time stretches beyond the production schedule, the replacement decision for a switching power supply is never just capacitance and rated voltage. The two parameters that separate a safe substitution from a field failure are ESR and ripple current rating. Both drive self-heating, output ripple voltage, and control-loop phase margin. The sequence below starts from the circuit requirement, not from a matching part number, and ends with the bench checks that qualify a substitute for production.
Why the Original Part Becomes Unavailable
Sourcing gaps for aluminum electrolytic and polymer electrolytic capacitors usually have one of three causes: an end-of-life notice issued as a manufacturer consolidates a product family, long lead times on the preferred package that force a footprint alternative, or a single-source situation that the design team is trying to eliminate. In each case the cross-reference work is the same.
Do not match by brand alias first. A direct series swap often looks acceptable on paper and still fails in thermal or loop checks, because ripple ratings and ESR are specified at different frequencies, temperatures, and operating life points across manufacturers.
Parameters That Must Match Closely
Capacitance. The replacement must meet the original minimum capacitance at the operating DC bias and over the full temperature range. For output filters, a capacitance increase of 20–30 % is usually tolerable if the loop is re-verified. A decrease below the original tolerance band raises output ripple and degrades load-transient response.
Rated DC voltage. Do not install a substitute with a lower voltage rating than the original. For aluminum electrolytic parts, keep at least 20 % headroom above the worst-case steady-state voltage. For polymer electrolyte types, follow the manufacturer derating guidance; typical practice is to avoid sustained operation above 80–90 % of rated voltage.
ESR at the switching frequency. ESR controls internal loss, P = I_ripple² × ESR. Compare values at the converter switching frequency, usually 100 kHz, not at 120 Hz. Example: a 470 µF / 25 V aluminum electrolytic part with an ESR of about 50 mΩ at 100 kHz and 1.2 A ripple dissipates roughly 72 mW. A polymer substitute with 15 mΩ at the same current dissipates about 22 mW, but the zero created by ESR moves from 6.8 kHz to 22.6 kHz, which changes the compensation requirement. Lower ESR is not an automatic drop-in improvement.
Ripple current rating. Verify the rating at the same frequency and the same ambient temperature the original was selected for. A common trap is comparing a rating stated at 105 °C and 100 kHz with a rating stated at 85 °C or 120 Hz. Apply the frequency correction factor from the data sheet, then keep the worst-case ripple below the derated rating with a margin of at least 20 %. For aluminum electrolytic parts, the rated ripple current is tied to a specific core temperature rise and load life; running beyond it shortens life and increases the risk of the pressure-relief vent opening in sealed parts.
Cross-Reference Table for a Typical Output Capacitor
The table is a generic example of a 470 µF / 25 V output capacitor in a 12 V industrial supply. Use it as a template, and replace the figures with the values from your schematic and the series under evaluation.
| Parameter | Original Aluminum Electrolytic (typical) | Candidate Replacement | Decision Threshold |
|---|---|---|---|
| Capacitance | 470 µF ±20 % | 470–560 µF ±20 % | ≥ 470 µF at rated bias and over the operating temperature range |
| DC voltage rating | 25 V | 25 V or 35 V | ≥ 25 V, with at least 20 % steady-state headroom for electrolytic types |
| ESR @ 100 kHz | 30–80 mΩ | 10–25 mΩ (polymer) or 25–80 mΩ (electrolytic) | Re-verify loop stability when the ESR zero frequency changes by more than ±30 % |
| Ripple current @ 100 kHz / 105 °C | 1.0–1.6 A | Per data sheet with frequency and temperature correction factors applied | Corrected rating ≥ worst-case ripple × 1.2 |
| Lifetime | 5 000–10 000 h at rated ripple and 105 °C | Recalculate with actual ripple and ambient | Core temperature ≤ 105 °C; a 10 °C margin is recommended |
| Package | 10 mm × 16 mm, radial through-hole | Same or smaller within PCB keep-out | Check pad geometry and thermal-relief vias for SMD types |
Parameters that may deviate without dedicated analysis are limited to brand, product family, case finish, and packaging form. Even these have secondary effects: a larger case usually improves thermal resistance, while a smaller case with the same ripple rating may run a higher hot-spot temperature. Treat every deviation as a design input, not a free choice.
Verification Steps Before Production Release
- Measure ESR at the switching frequency. Use a bench LCR meter set to 100 kHz, or to the actual converter frequency, with leads compensated. Note that electrolytic ESR rises at low temperature; if cold-start behavior matters, verify with a cold chamber.
- Confirm ripple current at the operating point. Compute or measure the rms ripple at full load. Apply the data sheet frequency and temperature correction factors, and confirm the derated rating covers the measured value with at least 20 % margin.
- Run a steady-state thermal test. Load the prototype to full load at the maximum expected ambient temperature, in the real enclosure airflow. Measure case temperature with a thermocouple and estimate core temperature from the data sheet thermal resistance and the I²R loss. For aluminum electrolytic parts, holding core temperature 10 °C below the rated limit roughly doubles expected life compared with running at the limit.
- Re-check the control loop. If the ESR zero frequency moved by more than ±30 %, confirm gain and phase margin with a network analyzer or a step-load test. This step catches the most common problem in polymer-for-aluminum substitutions: a loop compensated for a higher ESR zero that now has insufficient phase margin.
Finally, order samples of the candidate series and run the same checks on the actual production board before issuing the cross-reference. Data sheet values and room-temperature LCR measurements agree well, but the interaction with the control loop and with the board thermal environment is visible only on the loaded unit, not in a datasheet.
