Optical Encoder Air Gap Setup: Setting Readhead Clearance That Survives Motion, Heat and Debris

Why the gap is an optical parameter, not just a clearance

An optical readhead does not measure the scale by touching it; it projects illumination onto the grating and collects the diffracted or reflected light across the air gap. The phase and intensity relationships that become sine and cosine signals are functions of that distance. Inside the window the readhead was designed for, amplitude and contrast are flat and forgiving. Outside it, they fall away—gradually at first, which is what makes gap errors insidious, because a slightly-too-large gap still produces a signal good enough to count.

The two failure directions are not symmetric. Too small, and the risk is physical: debris trapped between scale and readhead, a dust particle scored across the grating by a moving head, or contact during a thermal excursion or a stage pitch event. Too large, and the risk is electrical: the modulation depth of the received signal falls, the interpolator works closer to its amplitude floor, and sub-divisional error grows while the output still looks healthy on a counter. Neither mode shows up on a slow 'does it move and count' commissioning pass, which is why both keep appearing in the field months after handover.

The gap window itself is not a universal number. It depends on the optical principle—coarse reflective scales tolerate generous gaps, while fine-pitch and diffractive designs use tighter, more sensitive windows—and on the readhead's rated tilt tolerances, because tilt consumes gap margin by varying the effective distance across the sensing field. The only number that matters is the one in the datasheet for the specific readhead and scale combination.

Setting the gap correctly on the machine

The gap is the last mechanical parameter to be set, not the first. Scale alignment, readhead parallelism and angular tolerances must be finished and fastened to final torque first, because every correction to alignment changes the gap. With the axis stationary at mid-travel, the supplier's gap gauge, shim stock of the specified thickness, or the follower tool supplied with the readhead is inserted between scale and sensing face, the readhead is brought into light contact and clamped, and the gauge is removed. The value is recorded together with the gauging method, the temperature and the axis position.

Target the middle of the window, not the edge closest to a bigger signal. A gap sitting just above the minimum may produce the strongest amplitude on the bench, but it has reserved nothing for the contamination film that will inevitably form, for the bracket creep and thermal expansion that will slowly change the distance, or for the debris particle that must pass through the gap without scoring anything. Mid-window placement converts the datasheet tolerance into real lifetime margin.

Where the readhead exposes a diagnostic amplitude or gap indication, the mechanical gauging and the electrical reading should be closed on each other: set the gap by gauge, confirm the signal level sits where the supplier's documentation says it should for that gap, and note both. A discrepancy between the physical gap and the optical response is itself diagnostic—it points to tilt, contamination on installation, or a scale-reflectivity issue worth resolving before the machine ships.

• Alignment final and torqued before any gap measurement
• Gauge or shim at the mid-window value, never 'as close as it looks right'
• Record gap, method, temperature and axis position
• Cross-check physical gap against the diagnostic signal level

What moves the gap after commissioning

The gap you set is not the gap the machine keeps. Thermal growth changes the distance between the scale mounting surface and the readhead bracket, especially when the two sit on different structures with different expansion coefficients—steel frame against aluminium bracket is the classic combination. Bracket compliance lets acceleration loads deflect the readhead transiently, so the gap modulates with every move; a bracket that is stiff enough on paper can still ring at the axis's motion frequencies. Vibration from adjacent machinery adds a continuous, smaller modulation. Each of these mechanisms alone sits inside the window; together they can push the instantaneous gap out of it.

Verification therefore has to be dynamic. After the static gap is set, the amplitude or diagnostic reading is checked at both ends and the middle of travel, where bracket loading and scale straightness differ. The axis is then run through its warm-up to operating temperature and the readings repeated, separating cold-machine values from production values. Finally a maximum-speed traverse is captured; amplitude dips during motion reveal the gap modulation that a static check cannot see. Any reading outside the window at any state means the mid-window target was not actually achieved in the installed geometry.

Writing gap control into installation and purchase documents

A complete installation specification names the datasheet gap window for the exact readhead and scale pair, prescribes the gauging method and the mid-window target, and requires the measured value, temperature and signal levels to be recorded at commissioning. Maintenance documents then reference that baseline: a periodic diagnostic check that compares current amplitude to the recorded gap-related levels catches slow drift early, and any re-set of the gap goes through the same gauging procedure rather than being adjusted 'until the alarm stops'.

For purchasing, the questions that separate a qualified encoder from a catalogue part are: does the supplier state the gap window and tilt tolerance for the specific configuration, does the readhead expose signal level or gap status for monitoring, and can the supplier provide a gauge or fixture for setting the gap? An encoder whose gap cannot be measured, monitored and recorded cannot be commissioned in the engineering sense—it can only be installed. The distinction shows up within the first year of service.