Sensor Alignment for Small-Diameter Rotary Encoder Scales: Tolerance and Mounting Guidance

Optical encoder manufacturers publish alignment tolerance tables that define acceptable ranges of sensor position relative to the scale. 

These tables are generally derived from linear scale or large-diameter rotary scale testing. Applying those tolerances directly to small-diameter rotary scales (specifically scales with diameters at or below 20 mm) introduces a systematic error: the radial alignment tolerance for small rotary scales is substantially tighter than the published tolerance for linear scales. 

Designs that rely on benching surfaces to achieve alignment for small-diameter discs without independent sensor adjustment will typically produce inadequate signal quality.

Why Small Rotary Scales Have Different Alignment Requirements

Grating Geometry: Wedge vs. Parallel Lines

On a linear scale, grating lines are straight and parallel. The period (pitch) of the lines is constant at every point on the scale. Consequently, moving the sensor laterally across the scale face (Y direction) does not change the period the sensor reads. Alignment tolerance in the Y direction is generous.

On a rotary scale, grating features are wedge-shaped — thinner at the inner diameter, thicker at the outer diameter. The period of the grating changes continuously as radius changes. A standard optical encoder scale uses a grating with a period of exactly 20 µm at the nominal optical diameter — the specific radius at which the sensor must be positioned for correct operation.

If the sensor is positioned at a radius different from the nominal optical diameter:

• The grating period at the sensor position no longer matches 20 µm.
• The interpolation electronics process a signal with incorrect period, producing systematic position errors.

The Effect of Diameter on Wedge Severity

For a large-diameter rotary scale (e.g., 100 mm diameter), the fractional change in grating period per unit of radial displacement is small. Moving the sensor ±200 µm radially changes the effective grating period by a small fraction, and signal quality remains acceptable.

For a small-diameter scale (e.g., 12 mm diameter), the same ±200 µm radial displacement represents a much larger fractional change in radius. The grating period at ±200 µm from nominal deviates significantly from 20 µm. Signal amplitude can drop to only 20% of the ideal value at this displacement.

Practical consequence: For a 12 mm rotary scale:

• Published tolerance (linear scale / large rotary): ±200 µm radial
• Required tolerance for adequate signal quality: ±100 µm or less

This 2:1 tightening of tolerance in the Y (radial) direction applies to scales at or below the 20 mm diameter threshold.

Secure Your Components Stock Now with Torquety

Reliable automation components for high-performance applications.

Check Availability Chat on WhatsApp

Alignment Axes and Terminology

Standard encoder alignment uses a six-axis coordinate system:

For small rotary scales, the Y axis (radial direction) is the axis with the critical tightened tolerance. All other axes retain their standard tolerances or are less sensitive.

Mounting Design Requirements for Small Rotary Scales

Adjustable Sensor Mount

The critical design requirement: the sensor mount must provide fine adjustment in the Y (radial) direction with resolution adequate to position the sensor within ±100 µm of nominal.

Implications for mechanical design:

• Fixed benching surfaces alone cannot reliably achieve ±100 µm due to manufacturing tolerances and tolerance stack-up across the encoder hub, shaft, bearing, and sensor mount.
• The sensor head bracket must incorporate a radial adjustment mechanism — slotted mounting holes are the minimum, but a fine adjustment screw provides better control.
• Adjustment should be made while monitoring encoder signal amplitude (using manufacturer-provided alignment tools or oscilloscope).

Thermal Expansion Considerations

The nominal optical diameter is specified at a reference temperature (typically 20°C). For applications with significant temperature variation, differential thermal expansion between the sensor mount material and the scale disc will shift the radial sensor position:

Radial shift = ΔT × (CTE_mount – CTE_scale) × nominal_radius

For a 10 mm nominal radius mount (20 mm disc) with steel mount (11 ppm/°C) and soda lime glass disc (8 ppm/°C):

At 30°C temperature rise: Radial shift = 30 × (11-8) × 10⁻⁶ × 10 mm = 0.9 µm

This thermal shift is small relative to the ±100 µm tolerance for a 20 mm disc. However, for an aluminum mount (23 ppm/°C):

Radial shift = 30 × (23-8) × 10⁻⁶ × 10 mm = 4.5 µm — still acceptable.

Thermal effects become more significant at smaller diameters: a 5 mm disc at ±100 µm tolerance with an aluminum mount over 100°C range would see approximately 15 µm thermal drift — still within tolerance, but consuming 15% of the margin.

Alignment Verification Protocol

Best practice for small-diameter scale installations:

1. Mount the encoder sensor head with the adjustment mechanism at mid-travel.
2. Connect to the manufacturer’s alignment tool or oscilloscope.
3. Monitor the sin/cos amplitude (or quadrature signal edge quality) while adjusting radially (Y direction).
4. Set the radial position at the peak amplitude reading — this is the nominal 20 µm grating period position.
5. Verify amplitude stability through a full rotation of the disc — excessive amplitude variation indicates eccentricity or scale runout.
6. Lock the sensor mount and re-verify amplitude after tightening.

Do not rely on benching surfaces for small-diameter rotary scales, tolerance stack-up from the hub, disc, and mount will typically place the sensor outside the ±100 µm radial tolerance.

Impact on System Accuracy

Signal amplitude directly affects interpolation accuracy. The interpolation electronics compute position from the ratio of the sin and cos channels. When amplitude is reduced by radial misalignment:

• The ratio computation becomes more susceptible to noise.
• Gain differences between the sin and cos channels amplify periodic position errors.
• At 20% amplitude, interpolation errors are typically an order of magnitude larger than at nominal amplitude.

For a sensor specified at ±2 µm accuracy with nominal signal amplitude, operating at 20% amplitude may increase effective position error to ±10–20 µm or more, negating the high-resolution capability of the encoder.

Before you go, you might want to dive deeper into 

• Fiber, Friction, and Flow: Navigating Slip Ring Technologies, 
• discover more about From Fear to Function: How Robotics Is Augmenting, Not Replacing, Humans., 
• or check out our guide on Are Magnetic Encoders the Same as Inductive Encoders? A Technical Analysis.