Overdrive strobing without visible flashes: hidden strobe technique

A strobed inspection light solves the camera's problem and creates the operators'. The sensor gets a bright, motion-freezing burst; everyone else on the line watches flashes all shift. Hidden strobing — pulsing the LED at overdrive current only while the shutter is open — keeps the image benefit and removes the flash. Below is the arithmetic that makes a strobe invisible: the duty-cycle limit, the overdrive factor a pulse rating allows, the flicker-fusion threshold that decides whether anyone sees anything, and a worked example at 120 fps. It ends with the controller and light characteristics to look for, and the cases where a plain continuous light still wins.

What is hidden strobe lighting?

Hidden strobe lighting fires the LED at many times its continuous-rated current, typically 5 to 20 times, but only during the few hundred microseconds the sensor exposes: the camera records a bright frozen frame while the room sees an average brightness of a few percent of continuous output.

A conventional strobe gets the image right and the room wrong: fire a 10 A pulse every frame and the station flashes like a camera flash. The sensor cannot tell what the light does between frames, though. Photon count, motion freezing and ambient rejection all happen inside the pulse, so the dark intervals cost nothing optically — make them dark. The trigger discipline is identical to any strobed station, and the overdrive, duty cycle and timing math carries over unchanged; what changes is the peak current you pick and the duty cycle you accept.

Why does the camera see a bright flash that people do not?

The sensor adds light only while its shutter is open, so it receives the full peak intensity of every pulse; the eye integrates over roughly 16 ms — flicker fusion sits near 60 Hz — and judges a repeating pulse train by its time average, which is peak current times duty cycle.

Above the fusion frequency the eye follows the Talbot-Plateau law: apparent brightness depends only on the time-averaged light. Two rules drop out of that. The thermal rule: LED heating follows average power, so peak current times duty cycle must stay at or below the continuous rating. The visibility rule: that same product, divided by the continuous rating, is how bright the light looks to a person. Run the product at the thermal limit and a strobed station looks like a continuously lit one; run it at 0.1 to 0.3 and it reads as a dim indicator lamp.

Both rules live and die by the duty cycle, so match controller and light on that number first; the duty cycle matching guide walks the comparison.

How much overdrive current can the light take in a short pulse?

Take the overdrive factor from the light's pulse-current table, which derates with pulse width — commonly around 10 times continuous current at 100 µs and around 5 times at 1 ms — then confirm that peak current times duty cycle stays at or below the continuous rating.

Two limits set that table. The peak limit sits in the die: bond wires and current crowding cap what the emitter survives in one pulse, so shorter widths allow higher factors. The average limit is thermal, the same I_peak × d inequality as above, and a table that lists a maximum duty cycle at each current has already folded it in. High-luminance bar and spot lights, with dense arrays on metal-core boards, usually tolerate the larger factors; large diffused panels tolerate smaller ones. Phosphor-converted white LEDs also lose efficacy as current climbs, so a 10× current pulse delivers somewhat less than 10× the photons; the current droop diagnosis shows how to measure that on your own light. The factor is a property of the light, printed in its strobe table — that table, not a rule of thumb, sets your M.

What duty cycle keeps the flash invisible?

Keep the pulse rate at or above about 60 Hz — the camera frame rate — so the eye fuses the flashes, and keep overdrive factor times duty cycle at or below roughly 0.3 so the fused glow reads as a dim lamp rather than a working light.

Fusion thresholds are typical values, not constants: about 50 to 90 Hz, lower for dim sources, higher for bright ones. Below roughly 50 flashes per second the eye starts counting pulses, so a 30 fps line shows discrete flashes unless the average brightness drops below about 0.05 of continuous output — a budget the current math may not allow. Run the numbers before promising anyone a dark station.

Worked example — conveyor inspection at 120 fps:

• Part speed 200 mm/s, allowed blur 0.05 mm → exposure t ≤ 0.05 ÷ 200 s = 250 µs.
• Frame period T = 1 ÷ 120 s = 8.33 ms, so duty cycle d = 0.25 ÷ 8.33 = 0.030 = 3.0 %.
• Example light: 1000 mA continuous rating, pulse table allows 10 A at widths up to 1 ms → overdrive M = 10 A ÷ 1 A = 10.
• Average current: 10 A × 0.030 = 0.3 A, which is 30 % of the continuous rating, a thermal margin of about 3×.
• Apparent brightness ∝ M × d = 10 × 0.03 = 0.30: the station glows at about 30 % of continuous brightness, fused at 120 Hz into a steady faint light with no visible flashes.
• Quieter still: cut the exposure to 100 µs (blur 0.02 mm) → d = 0.0001 × 120 = 1.2 % and apparent brightness 0.12, barely noticed, while the camera still receives 10 A of peak current in every exposure.

The same arithmetic pushed to its short-pulse extreme sizes the strobes in barcode tunnels; the high-speed barcode strobe sizing guide works through pulses of a millisecond and less.

How do you time the pulse to the camera exposure?

Fire the controller's strobe input from the camera's exposure-active output and set the pulse width equal to the exposure time; the LED adds a few microseconds of switching delay, which costs about 2 % of the pulse at a 250 µs exposure and only dominates below roughly 20 µs.

The chain is short: exposure-active signal out of the camera, into the controller's strobe input, rated pulse current out for the set width. Cameras that only offer a frame trigger run the pulse after a fixed delay, and a black first frame usually means the pulse landed outside the exposure — the black-image timing fix walks that fault. At 10 µs exposures the delay and rise budgets decide how much of the pulse is usable light, which the optical rise-time measurement quantifies. A frame that is dark at its start can also trace to exposure settling rather than wiring, as in the first-frame-dark diagnosis. One trap: free-running PWM dimming is not hidden strobing, because an unsynchronized pulse train beats against the exposure and produces banding instead of one full-brightness pulse per frame. Multi-light and multi-camera stations can share one strobe signal, as in the multi-host synchronization set-up.

Which parts make up a hidden strobe station?

Three parts: a controller that delivers the peak pulse current — 10 A in the worked example — while carrying the 0.3 A average, a light whose pulse table allows that current at your pulse width, and the trigger wiring that brings the exposure-active signal to the controller's strobe input.

The light controllers range splits into constant-current and constant-voltage output types, and a dedicated controller earns its place here because it regulates current through the pulse; the five reasons for a lighting controller apply doubly at overdrive. When one station needs pulse power per channel, the constant-current family goes biggest: the LS-CTL-0054 is a two-channel high-power constant-current controller rated 240 W per channel, 320 W total.

Where several lights should stay under one digital control box, the LS-CTL-0070 gives four channels and 200 W with a strobe feature.

On the light side, high-luminance types concentrate the emitters so their pulse tables allow the large factors: the LS-BAR-0400 is a 312 x 30 mm high luminance bar light sized for conveyor fields, and the LS-SPT-0175 is a 6 mm, 5 W high luminance spot light that mounts beside small-field lenses. Size light and controller together in the lighting configurator.

The exposure-active signal leaves the camera through its GPIO connector; if your camera's output set-up is unclear, the GPIO cable guide covers when a cable is needed. Both light ranges also include 850 nm and 940 nm infrared versions, and infrared is the most complete version of the technique: the pulses are invisible at any frame rate, which is why IR strobes are standard around dock doors and daylight-leaky cells.

When is hidden strobing the wrong choice?

Skip it when the camera runs below about 60 fps at high pulse brightness — people resolve the individual flashes — when operators need continuous light for changeovers or manual work at the station, or when the exposure grows so long that the pulse becomes most of the frame.

The boundary cases, in order:

• Below roughly 50 fps: either dim the average to 0.05 or less of continuous output, or move to infrared, where fusion stops mattering.
• Exposures beyond a few milliseconds: duty cycle climbs, the pulse table derates at long widths, and the average current approaches the continuous rating — a dim continuous light then does the same job with no trigger wiring at all.
• If ambient rejection is the real goal, strobing still helps, because the signal peak rises against fixed ambient; measure it on your bench rather than assuming it, with the strobed ambient rejection method.
• Lights whose tables stop at a 2× or 3× factor add timing complexity for little image gain; spend the budget on a higher-luminance light instead.

The decision in one line: frame rate at or above 60 fps, exposure under about 1 ms, and a light whose pulse table allows 5× or more — then hidden strobing buys a brighter image and a darker room at the same time. Check your numbers against the worked example above, pick a controller in the lighting controllers finder or the lights finder, or step through both in the controller configurator. More lighting technique articles sit on the lighting topic page, and for a second opinion on a specific light's pulse table, contact us.