Assembling a Medical Power-Input Bench: The Generator Is the Specification

Medical equipment · 4 min read · 34 cited facts

Strip away the medical overlay and the power-input bench under IEC 60601-1-2 is an IEC 61000-4-11 rig. The dips themselves are unremarkable; what defines the bench is the generator clause behind them and the operating points in front of them: the profile runs at both the minimum and the maximum rated input voltage, so the wide-range 100 to 240 V supply that most medical equipment uses needs a source that sits accurately at both ends of that range. A source specified only at a single nominal point has been specified for neither.

The generator clause is the bench

Read the second and fourth rows together and the shape of the generator problem appears. The equipment inside this standard's scope draws at most 16 A per phase (61000-4-11 §1), yet the generator must drive 500 A of peak inrush on a 220–240 V supply and 250 A on a 100–120 V one (61000-4-11 §6). The difference is the recovery instant: when the supply returns after a dip, the EUT's front end takes what it takes, and a generator that cannot deliver it re-shapes the recovery edge — the test that was run stops being the test the report claims. The 23 A at 70% U_nom and 40 A at 40% U_nom rows police the same physics at reduced voltage.

One row is deliberately absent. IEC 61000-4-13 defines per-class harmonic percentage tables, and they sit behind the IEC paywall — they are not reproduced here and should not be reconstructed from vendor summaries. The synthesis range above is enough for scoping the source; the programme that tests to the clause buys the document.

Accurate at 100 V and at 240 V, not on average

The two-ends requirement is where range architecture earns its keep. The AFV-P runs two user-selectable output voltage ranges at 0.1Vrms resolution, and the wide-range input splits across them naturally — 100 V sits on the lower range, 240 V on the upper — so each end of the 100 to 240 V window lands inside a range built for it rather than at the far corner of a single span sized for the middle.

The same datasheet is candid about where the AFV-P stops, and the candour is worth copying into the test plan. Its slew specification is under 300 µs from 0 to 90% of output voltage, and the sheet positions the unit for pre-compliance tests such as IEC 61000-4-11; the certification generator's edge specification is 1–5 µs (61000-4-11 §6), a different class of edge. Inrush tells the same story: the largest model, the AFV-P-5000, delivers 180 A / 90 A peak against 40 A / 20 A rms — strong for a bench source, still short of the 500 A the generator clause demands on a 220–240 V supply. A pre-compliance bench finds the design problem before the certification lab does. It does not produce the certification edge, and a quotation that pretends otherwise fails at the lab, not at the purchase order.

The parts of the quote that are not a source

• Proximity magnetic fields — added at Edition 4.1 and run to IEC 61000-4-39:2017 — need a field generator. No programmable AC source produces this test.
• Conducted immunity now triggers on signal cables longer than 1 m, down from 3 m. It is injected through a coupling network onto cables, not through the power port, and not by this bench.
• Equipment drawing more than 16 A per phase leaves IEC 61000-4-11's scope entirely (61000-4-11 §1). IEC 61000-4-34 takes over, and none of the sizing above carries across.