Sizing a 400 Hz Source for Inrush: kVA Is the Last Number to Pick

Aviation & 400 Hz · 4 min read · 22 cited facts

The first question asked about a 400 Hz source is how many kVA, and it is the wrong first question. Avionics loads are rectifier-fed: they draw their current in narrow peaks at the crest of the waveform, not smoothly across the cycle. A source sized by matching nameplate kVA to nameplate VA can hit current limit on the very first cycle — and a source in current limit is distorting the exact waveform your test depends on.

What the standard says the waveform must survive

Those three rows are voltage rows, but they size the current path. Holding distortion at ≤0.05 while a nonlinear load pulls its current in peaks means the source has to deliver that peak current without flattening the crest of the wave — the 1.31–1.51 crest factor window is gone the moment the output stage clips. Whether it clips is decided by the source's peak-current capability, not by its kVA rating.

The datasheet line that actually answers the question

For inrush and rectifier-fed loads, the figure to find is the peak-to-RMS output current ratio. The AN61(F) series provides peak current at four times its rated RMS output. Per model, on the low gear, that cashes out as:

• AN61500S-350(F) — 5 A effective, 20 A peak
• AN61501S-350(F) — 10 A effective, 40 A peak
• AN61502S-350(F) — 20 A effective, 80 A peak
• AN61504S-350(F) — 40 A effective, 160 A peak

Read the list backwards from your load: the peak column is the number your DUT's front end actually meets on its first cycle, and it is what decides whether the start succeeds. A larger frame with a weaker peak ratio can fail an inrush event that a smaller frame at four times clears. The same series programs switch-on and switch-off angles across 0-359.9 degrees, which is what lets it place the start on the part of the cycle the test calls for rather than wherever the contactor happens to close.

Inrush is not the only overcurrent. Motor loads pull sustained overcurrent measured in seconds and minutes, and that is what an overload rating covers — a separate line of the datasheet from the peak ratio, and one worth asking for by name. Overload ratings and peak ratios describe different phenomena — a per-cycle rectifier peak is untouched by the 60-minute rating, and a motor start cares more about seconds of sustained current than about a single-cycle peak.

When a bigger source is not the answer

• The failure is the transient envelope, not the load. Reproducing the 180 V for 10 ms surge and 80 V sag of Figure 3 takes voltage headroom and millisecond sequencing; extra kVA buys neither.
• The waveform is distorting under a nonlinear load. The fix is a source that holds the Table I distortion factor of ≤0.05 at your load's peak draw — a capability question that a larger frame with the same output stage does not resolve.
• The load returns energy. A motor-type EUT pushes energy back at the source, so confirm that the source is specified to absorb it and to operate into an unbalanced load. A bigger source without that behavior is a bigger problem, not a smaller one.
• The work is component-level. The AN61(F) provides three to four times peak current for surge measurement, with programmable switch-on and switch-off angles, which covers bench inrush work that would otherwise get argued into a cabinet-sized purchase it does not need.

When the answer genuinely is more power, the catalog steps are wide. AN61(F) sources run 500VA to 120kVA on a 15~1000Hz band that spans 400 Hz outright. AFC converters run 0.5kVA to 2MVA, single- or three-phase, with 400 Hz available either as a fixed output or on the 350~450Hz adjustable band. But sizing lands on a specific model only after the peak current, the overload duration and the waveform-quality requirement are each known.