LED driver inrush current — why a circuit with many LED fixtures can still trip on a curve-C breaker

The guide on MCB trip curves B/C/D already briefly mentions LED drivers as a load type with inrush current peaks that can call for a curve C. This article goes deeper into why that is: the specific behaviour of an LED driver's inrush current, and why this phenomenon — unlike a regular overload — can cause an unwanted trip in a fraction of a millisecond, right at switch-on itself.

The phenomenon: the driver's input capacitor

Almost every LED driver has an input-side capacitor (for rectification and/or power-factor correction) that, on switch-on from a fully discharged state, charges very quickly. During that very short charging time — typically well under one millisecond — the driver draws a current peak that per unit can reach tens of times its rated operating current. For a single fixture this peak is too short and too small to be a problem; the risk arises when many fixtures are switched on at once on the same circuit — for example via a single light switch, relay, or impulse relay (see the guide on impulse relays) — causing the individual peaks to add up almost simultaneously into a combined inrush current that can exceed the miniature circuit-breaker's magnetic instantaneous trip threshold.

Why this isn't an overload

This phenomenon must be distinguished from a regular overload: the circuit's continuous load (the number of fixtures times their rated power) can remain well within the breaker's rated current, while the instantaneous switch-on peak — lasting only a fraction of a millisecond — still exceeds the magnetic trip threshold (see the [guide on MCB curves](/guides/nen-3140/automaat-b-c-d) for the thresholds per curve: B = 3-5×In, C = 5-10×In, D = 10-20×In). The breaker's thermal (overload) protection plays no role here; it is exclusively the magnetic, instantaneous trip that is triggered by the combined inrush peak.

Measures to prevent this

Choosing a higher curve (for example curve C instead of B) — this raises the magnetic trip threshold and gives more margin for the combined inrush peak, without materially reducing the continuous overload protection.
Limiting the number of fixtures per circuit to a number whose combined inrush peak (per the driver manufacturer's data) stays well below the trip threshold of the chosen breaker.
Staggered switching: in large installations (for example a warehouse or car park with hundreds of fixtures), switching the fixtures on not all at once but in small groups with a short time delay between them, so the individual inrush peaks don't overlap.
Using drivers with built-in inrush-current limiting (for example with an NTC resistor or an active limiting circuit on the input side) — this reduces the peak at the source, per driver, rather than adapting the installation around it.

Practical relevance

When designing a lighting circuit with many LED fixtures, or when investigating a circuit that trips unexplainably on switch-on without an identifiable overload, the combined inrush current must be considered as a separate factor, distinct from the continuous load. Fixture or driver manufacturers typically specify a peak value and duration of the inrush current per unit, which serves as the basis for determining the maximum number of fixtures per circuit and/or the breaker curve to choose.

Common mistakes

Sizing a circuit with many LED fixtures based only on the continuous (thermal) load, without checking the combined switch-on current.
Attributing an unexplained trip when switching on lighting to a faulty breaker or an overload, while the cause may be a combined inrush peak that only occurs at switch-on.
Switching on all fixtures of a large building at a single switching moment (or a single impulse-relay pulse) without considering staggered switching when a large number of fixtures is involved.
Simply switching to curve D as a "safe" solution without checking whether the circuit's short-circuit protection and Zs value still allow it — a higher curve also reduces sensitivity to an actual fault at the far end of the circuit.