Strobe versus Continuous Light

An emitter has a continuous rating and a pulsed rating, and the pulsed rating is much higher. Running a light emitting diode at ten times its continuous current for a hundredth of the time produces roughly the same average heat, but delivers ten times the light during that brief window. A strobe controller does exactly this, and it is triggered in step with the camera exposure so that the bright pulse and the open shutter coincide. Between exposures the emitter is off, or driven at a low level for aiming.

When the extra light is the point

• Fast moving parts. If the part moves far enough during the shortest exposure the camera allows to blur the feature beyond tolerance, more light is the only way to shorten the exposure further. This is the main reason strobes exist.
• Very short line periods. On a line scan station running at tens of thousands of lines per second, each line is a very short exposure, and continuous illumination at that intensity would cook the emitter and the part.
• Long working distances or large fields. Light falls off with distance and spreads over area, so a big field of view from far away may simply be beyond what a continuous source can illuminate.
• Light-hungry techniques. Dark field, low-angle and coaxial arrangements throw most of the light away by design. A technique that needs ten times the light is much easier to achieve with a strobe than with a bigger continuous source.

The counterweight is the whole apparatus around it. A strobe needs a controller unit, a trigger signal that arrives at the right time with the right width, cabling rated for the pulse current, and an emitter that tolerates being driven that way. It also puts the light source into the timing budget: the trigger has a delay and a rise time, and both have to be accounted for when the exposure begins. On a slowly moving line where a five millisecond exposure is acceptable, none of this earns its cost.

What changes when you fit one

Fitting a strobe turns the illumination into part of the timing chain, and that has consequences beyond the light. The trigger has to reach both the camera and the controller, with a delay that is stable from shot to shot — jitter in that delay shows up as brightness variation between images, which a threshold based algorithm will read as a change in the part. The pulse width has to be short enough to freeze the motion but long enough to deliver the light, and those two pull in opposite directions. And the emitter is driven hard in short bursts, so the electrical installation matters: pulse currents are large, and a supply or a cable that was fine for continuous operation may not be.

The validation requirement also grows. With continuous lighting, the station's behaviour is steady and a single image characterises it well. With a strobe, brightness depends on the pulse, and the pulse depends on the trigger, so any change in the trigger path is a change in imaging conditions. That is why strobe stations should have the trigger delay and pulse width recorded as controlled parameters alongside exposure and gain, and why changing a controller or a cable is a change that warrants re-validation.