How many lamps can one impulse relay switch? There is no fixed number — the answer depends on the peak capacitive inrush current the connected LED drivers draw at switch-on, not on the relay's steady-state AC-1 amp rating printed on the nameplate. A 16 A impulse relay that comfortably carries 16 A of resistive load can still weld its contacts on a fraction of that current in LED fittings, because the driver's inrush spike, not the running current, is what the contact has to make against. The only reliable route to a lamp count is the manufacturer's published capacitive-load table for that specific relay reference, cross-checked against the driver's own inrush figure.
This is one of the more common support calls in commercial lighting retrofits: an installer replaces a corridor's old fluorescent fittings with LED, keeps the existing impulse relay because "it's rated 16 A and the new load is only 2 A", and within a few months contacts are pitted or fully welded. The relay was never overloaded in the traditional sense — it was hit thousands of times by inrush peaks it was never selected to withstand. Sizing an impulse relay for mixed lighting loads is covered in our lighting load sizing guide, and the broader mechanics behind it are in the impulse relay engineering guide. This article works through the actual lamp-count calculation.
Why the nameplate amps don't answer the question
An impulse relay's AC-1 rating describes its ability to make and break a resistive, non-inductive, non-capacitive load continuously. Incandescent and halogen lamps are close enough to resistive that the AC-1 figure is a workable guide, with AC-5b applied for the higher inrush of a cold tungsten filament. LED drivers are a different animal: the input stage is a bridge rectifier feeding a smoothing capacitor, which looks like a near short-circuit for the first few hundred microseconds after the contact closes. The result is a spike that can run into tens of amps for a single small driver, lasting a fraction of a cycle, then collapsing to a running current of well under an amp.
A relay contact does not care that the spike is short. What matters is the energy dissipated at closure, and whether the arc formed as the contact bounces has enough current behind it to locally melt the surface. Do that often enough, on enough fittings, and the surfaces weld together — covered in detail in our contact welding article. The nameplate current tells you almost nothing about this risk.
Inrush current
The brief, high-magnitude current peak drawn by a capacitive load — an LED driver, an electronic transformer — in the first fraction of a cycle after the supply is connected. It is unrelated to the steady-state running current and is the figure that determines contact stress, not the wattage on the fitting's label.
Where the manufacturer's lamp-count table comes from
Schneider Electric, ABB and other impulse relay manufacturers publish tables that state, for a given relay reference, the maximum number of LED lamps or drivers it can switch, broken down by driver inrush class. These tables exist because a single "amps" figure cannot capture the physics involved — the manufacturer has already run making-capacity tests against standardised capacitive load classes and translated the result into a usable number. If your relay's datasheet includes this table, use it directly; it supersedes any first-principles calculation, because it reflects the actual switch mechanism, contact material and spring force of that specific product.
The complication is that not every driver publishes its inrush current, and the table is built around inrush classes rather than a specific brand of fitting. When the driver datasheet is silent, common on budget fittings, you are left estimating — and that is where a working calculation earns its place as a sanity check, not a substitute for the table.
A capacitive load calculation you can actually use on site
The calculation is a straightforward capacity check: total expected inrush current against the relay's published making capacity for capacitive loads. It will not replace the manufacturer's table, but it explains why a given lamp count is or is not sensible, and it is the same logic the table was built from.
N_max = I_make(rated) / I_inrush(per driver)
Where: N_max is the maximum number of LED drivers that can be switched simultaneously by one pole, I_make(rated) is the relay's published peak making current for capacitive/LED loads (from its datasheet, not its AC-1 figure), and I_inrush(per driver) is the peak inrush current of a single driver as stated on its own datasheet, or a manufacturer's typical figure for that driver class if the datasheet omits it.
Two things make this less tidy than it looks. Inrush current does not add in a simple linear way once several drivers are in parallel, because they do not all reach peak at exactly the same instant relative to the AC waveform — the combined peak is usually somewhat lower than a straight sum, but not low enough to ignore. Cable and wiring impedance between the relay and the fittings also slightly limits the peak the relay sees, one reason manufacturers test with defined cable lengths. Treat the formula as a conservative estimate, and where the driver's inrush figure is missing, favour a smaller lamp count rather than a generous one.
Key takeaway: The relay's AC-1 nameplate rating and its LED lamp-count capacity are two different numbers derived from two different physical mechanisms. Never use the AC-1 figure as a proxy for how many LED drivers a relay can switch.
Reading the driver's inrush figure — and what to do without one
When a driver datasheet does state an inrush current, check what it is measured against. Some list a peak current at a specified pulse duration, others list an I²t figure meant for fuse coordination rather than relay contact life, and the two are not interchangeable with what a relay manufacturer's table expects. Where the units don't obviously match, contact the driver manufacturer or treat the fitting as unknown-inrush and derate accordingly. Retrofit jobs are the worst offenders, because nobody revisits the relay selection when fluorescent tubes are simply swapped for LED tubes.
When the relay datasheet has no lamp-count table
Not every relay datasheet in circulation carries an LED lamp-count table — older stock and ranges documented before LED retrofits became routine sometimes only list AC-1 and AC-5 figures. Three options exist. Contact the manufacturer's technical support with the relay reference and driver inrush figure; most can supply an unpublished figure. Choose a relay from a range that does publish the table, such as the Schneider Acti9 iTL 1P 16 A. Or, as a fallback, apply a generous derating factor and keep the fitting count per pole deliberately low rather than guessing.
Derating factor
A safety margin applied to the calculated maximum lamp count to account for inrush variability between drivers, ageing contacts, and incomplete inrush data. Roughly 20-30% below N_max is a reasonable starting point on mixed or unknown driver populations; well-documented, single-brand runs can be sized closer to the calculated figure.
Load type comparison at a glance
| Lamp / load type | Switch-on behaviour | Limiting factor for relay sizing | Typical utilisation category |
|---|---|---|---|
| Incandescent / halogen | Brief inrush from cold filament resistance | Cold-filament inrush multiple of running current | AC-5b |
| Fluorescent, magnetic ballast | Inductive, moderate inrush | Ballast inrush and power factor | AC-5a |
| Fluorescent, electronic ballast | Capacitive front end, short spike | Ballast capacitor inrush | AC-5a (manufacturer table preferred) |
| LED driver | Sharp capacitive inrush, sub-millisecond | Driver inrush peak, not running current | Manufacturer LED lamp-count table |
Splitting large lighting groups across poles and relays
When a calculated or published lamp count is lower than a circuit actually needs, the fix is not to force the group onto one relay and hope — split it across additional poles or relays, each within its own capacity. A multi-pole impulse relay lets you divide a large lighting zone into sub-groups on separate poles, controlled by the same push-buttons, without duplicating control wiring; both the ABB E290-32-20/230 and higher-pole Acti9 variants are built for this. Where the pole count isn't the constraint but the total load is, a second relay wired in parallel from the same buttons splits the inrush burden instead.
A 4-pole 16 A relay carrying four separate LED-lit zones off one set of corridor buttons is a common configuration, provided each pole's fitting count stays inside the manufacturer's table — a single-pole 16 A E290 and a 4-pole 16 A E290 do not necessarily share the same per-pole ceiling, so check the table for the specific reference, not just the amp figure in the part number.
Key takeaway: If a lighting zone exceeds one pole's LED lamp-count capacity, split it across additional poles or relays rather than accepting a relay that is undersized for the inrush it will actually see.
Key takeaway: A missing lamp-count table is not permission to estimate loosely — it's a reason to either get the figure from the manufacturer directly or choose a documented relay range instead.
For silent, high-cycle applications such as hotel corridors, the Acti9 iTLs 32 A combines the higher making capacity of a 32 A frame with quiet operation, useful when the ceiling void sits above guest rooms. See the full range in the relay collection.
Frequently Asked Questions
Can I use the AC-1 amp rating to work out how many LED lamps an impulse relay can switch?
No. The AC-1 rating describes continuous resistive current handling, not the brief capacitive inrush spike an LED driver draws at switch-on. A relay rated well above the running wattage of your load can still be undersized for the inrush, and can weld its contacts without ever seeing a traditional overload. Use the manufacturer's LED lamp-count table for the specific relay reference instead.
Where do I find the LED lamp-count table for a specific impulse relay?
It's published in the technical datasheet for that relay reference, usually alongside the AC-1 and AC-5 tables, listed against driver inrush classes or a maximum driver count. If it isn't on the public datasheet, the manufacturer's technical support can usually provide it, or confirm the reference wasn't tested for LED loads and should be treated conservatively.
What happens if I exceed the rated lamp count on one relay pole?
Nothing happens on day one — the relay switches the load correctly the first several hundred times. Over time, repeated inrush stress pits and eventually welds the contact surfaces, so the relay fails to release or fuses in the closed position. It's a cumulative wear failure, not an instant overload trip, which is why it's often missed during commissioning.
Does splitting a lighting zone across more relay poles solve an inrush capacity problem?
Yes, provided each pole's fitting count stays within its own published capacity. A 4-pole relay switching four separate lighting sub-groups from the same push-buttons is a normal way to handle a large zone, as long as no single pole carries more drivers than its lamp-count table allows. Splitting only helps if the split is real, not just convenient wiring.
Is fluorescent lighting subject to the same inrush limits as LED?
Fluorescent fittings with electronic ballasts have a capacitive front end similar in principle to an LED driver, so the same inrush logic applies, though usually at a different magnitude. Magnetic ballast fluorescent is inductive rather than capacitive and is governed more by AC-5a figures than a lamp-count table. Mixed retrofits, where some fittings are original fluorescent and others are replacement LED, should be sized against the worst-case driver type present, not the average.
Do halogen and incandescent lamps need a lamp-count calculation too?
Not in the same way. Their inrush comes from cold filament resistance, not a capacitive front end, and is covered by the AC-5b utilisation rating rather than a separate LED-style table. The calculation in this article is specifically for capacitive, driver-based loads — halogen and incandescent circuits are sized from the standard AC-1/AC-5b figures instead.
Getting the number right before you order
The practical sequence is short: identify the driver's inrush behaviour, check the manufacturer's published lamp-count table for the exact reference, and split the load across poles or relays when the count exceeds what one pole can take. Skipping straight from "the wattage is low" to "the relay is fine" is how corridor lighting circuits end up back on a maintenance ticket, contacts welded shut, with nobody able to say why a 16 A relay failed on a load that never drew more than 3 A running. The relay was never wrong for the running current — it was wrong for the inrush, and that's the number that should have driven the selection.