AKKN ElectronicsWhen a capacitor bank in a power converter fails, the standard reaction is to match capacitance, rated voltage, and case size from the original bill of materials. For distributors and design engineers sourcing cross-brand replacements, however, a frequent cause of field returns is not a wrong capacitance value. It is a thermal mismatch that appears months later as bulged vent disks, dry electrolyte, or cracked film. A replacement that passes every electrical check at the bench can still run hot because of a different ESR, a changed internal construction, or a longer can that blocks airflow through a tightly packed bank.
The original part may be unavailable for several reasons. The manufacturer discontinued the series, the OEM specified a custom-inscribed body for a single production run, or the supplier changed the internal construction while keeping the same series name. Any of these situations forces a substitution. The thermal duty must be evaluated as part of the selection, because the temperature of the capacitor core—not the nominal voltage rating—determines whether the alternate survives the expected service life.
Why the Original Part Disappears and What That Changes Thermally
Discontinued aluminum electrolytic series are the least predictable. An original capacitor may never have been thermally stressed because the design had generous airflow. The replacement candidate from another manufacturer can have the same capacitance and can diameter yet a higher ESR, a lower ripple current rating, or a different thermal resistance between the core and the case. Film capacitors, common in DC-link and snubber positions, tolerate ripple differently, but they are equally sensitive to the hot spot at the center of the winding.
Cooling changes when the bank layout changes. A 10 mm longer can may block forced air from an adjacent fan; terminal styles that differ from the original alter busbar spacing; mounting clamps with higher thermal resistance can impair the path from the can to the chassis. These mechanical differences change the thermal result even when all electrical parameters match.
Parameters to Match Exactly vs Parameters That May Deviate
Before searching for a cross-reference, separate the specification into hard constraints and acceptable deviations. Use the table below with the original series name and its datasheet values in hand.
| Parameter | Representative Original Series | Alternate Acceptance Criteria | Why It Matters |
|---|---|---|---|
| Capacitance | 470 µF ±20% at 120 Hz | Within ±10% of the original nominal and inside the circuit tolerance | PFC tuning steps must stay accurate; DC-link ripple voltage changes with capacitance |
| Rated DC voltage | 400 Vdc | Equal to or higher than the original | Line transients require the same voltage derating margin |
| ESR at switching frequency | ≤ 350 mΩ max at 10 kHz, 20°C | ≤ 1.25 × original maximum | ESR controls the I²R power dissipated as heat inside the can |
| Ripple current rating | 1.8 A rms at 105°C, 10 kHz | ≥ 1.6 A rms; if lower, add forced airflow or reduce ripple | Sets the maximum permissible internal temperature rise |
| Case-to-core thermal resistance | Not published in most datasheets | ≤ 5 K/W for snap-in; ≤ 7 K/W for screw-terminal | Converts a measured case temperature into a core temperature estimate |
| Case diameter × height | 35 × 50 mm | Diameter within ±1 mm; height within +5 / −3 mm | Retains the cooling surface and spacing between capacitors |
| Maximum operating temperature | 105°C | 85°C only if ambient and ripple are derated accordingly | Changes the rated hot-spot lifetime |
Capacitance tolerance is the value most often over-specified in a substitution. For DC-link smoothing, a ±10% deviation is usually invisible as long as the ripple voltage remains within the original limits. For power factor correction banks, however, the per-step reactive power target requires a closer match, so verify the compensation table before selecting. Impedance at 100 kHz is a weak proxy for ESR at the switching frequency; for a converter running at 10 kHz, match ESR at 10 kHz instead. A higher ESR at the operating frequency creates additional loss that must be rejected to ambient, and it raises both the case temperature and the hot-spot temperature.
Verification Steps After Installation
Electrical bench testing alone will not reveal a thermal problem. Run the full bank at nominal load and measure the steady-state temperatures before the repair is closed out.
- Measure capacitance at 120 Hz and ESR at the switching frequency, for example 10 kHz, with an LCR meter after the candidate parts have reached 20°C. Record the values on the incoming-inspection report.
- Verify case dimensions, terminal pitch, and vent orientation against the existing clamp and busbar layout. Select the correct clamp for the alternate diameter; do not modify the clamp at installation.
- Attach a thermocouple to the bottom center of the can with thermal grease or a thin insulating washer, away from direct airflow from the fan.
- Run the bank at the worst-case ripple current, typically 105% of nominal output power, for four hours in a 25°C ambient, and record the case temperature. If the bottom of the can is inaccessible, measure the top of the can and add 1–2 K.
- Estimate the hot-spot temperature by adding the case-to-core gradient from the alternate’s datasheet, or the generic values above, to the measured case temperature. Acceptance threshold: the estimated hot spot at rated load must be at least 5 K below the rated maximum core temperature of the alternate.
- Record the ambient temperature near the capacitor bank, not at the cabinet intake. The thermal measurement is valid only if airflow over the bank matches the field installation, including blanking plates and dust filters.
A cross-brand replacement is acceptable when it passes both the electrical match and the thermal check. For aluminum electrolytic capacitors, a useful approximation is that lifetime doubles for each 10 K reduction in hot-spot temperature below the rated maximum; for film capacitors, the same margin reduces the risk of premature dielectric degradation. When the original series is obsolete and the field data is limited, keeping the hot-spot estimate 10 K below the rated maximum provides a conservative design margin.
