Interferential Optical Encoder Technology: VCSEL Sources, Talbot Planes, and Nanometer Resolution

Executive Summary

Among the three major optical encoder architectures (transmissive, reflective, and interferential) interferential provides the highest achievable resolution and signal quality. 

This is not a matter of electronics sophistication; it is a consequence of the optical principle. Where conventional encoders cast a shadow of the grating on a detector, interferential encoders create an interference pattern that encodes position in the phase of a continuous wave. 

This phase measurement is inherently less susceptible to grating defects, contamination, and alignment variation than amplitude-based shadow detection.

The VCSEL Light Source

Interferential optical encoders use a VCSEL (Vertical Cavity Surface Emitting Laser) as the illumination source. VCSELs differ from conventional edge-emitting lasers and from LEDs in several relevant ways:

For interferential sensing, the illumination source must be coherent, the light must have a defined phase relationship across the illuminated area. LEDs have too short a coherence length for effective interference fringe formation. 

Edge-emitting lasers have sufficient coherence but are larger and require more complex optical coupling.

VCSELs provide sufficient coherence length for grating periods down to 20 µm while remaining compact enough to integrate into a sensor head with sub-millimeter overall dimensions. 

The circular emission pattern of the VCSEL simplifies collimation compared to the elliptical output of edge-emitting lasers.

Diffraction and Talbot Plane Formation

When a VCSEL illuminates a diffraction grating, the reflected (or transmitted) beams are diffracted at specific angles determined by the grating equation:

d × sin(θ_m) = m × λ

Where:

• d = grating pitch.
• θ_m = diffraction angle for order m.
• m = diffraction order (0, ±1, ±2, …).
• λ = wavelength.

For a 20 µm pitch grating illuminated with a VCSEL at 850 nm:

Secure Your Components Stock Now with Torquety

Reliable automation components for high-performance applications.

Check Availability Chat on WhatsApp

• First-order diffraction angle: sin(θ₁) = 850 nm / 20,000 nm = 0.0425 → θ₁ ≈ 2.44°

The +1 and -1 diffracted orders interfere constructively at specific distances behind the grating, these are the Talbot planes, also called self-image planes. At each Talbot plane, the grating is “imaged” by constructive interference; between planes, the image is blurred by destructive interference.

The Talbot distance Z_T for a grating of period d illuminated at wavelength λ is:

Z_T = 2d² / λ

For d = 20 µm and λ = 850 nm: Z_T = 2 × (20 × 10⁻⁶)² / (850 × 10⁻⁹) = 941 µm ≈ 0.94 mm

The photodetector is placed at the first Talbot plane (0.94 mm from the grating surface) or at an integer multiple of it. At this distance, the interference pattern exactly reproduces the grating period, a sharp, high-contrast fringe pattern that shifts as the grating moves relative to the sensor.

Signal Generation and Phase Detection

At the Talbot plane, the fringe pattern has a period equal to the grating pitch (20 µm). 

As the grating moves laterally by one grating period (20 µm), the fringe pattern completes one full cycle, one sine period.

To detect position within this cycle, a quadrature detector array is used. Four detector elements are positioned at 0°, 90°, 180°, and 270° relative to the fringe period:

• Detectors at 0° and 180° form a differential pair (A+ and A-).
• Detectors at 90° and 270° form a differential pair (B+ and B-).

The differential signals A (= A+ – A-) and B (= B+ – B-) are sinusoidal and in 90° quadrature. Position within the grating period is computed as:

θ = arctan(B / A)

This arctangent computation is the interpolation step. At ×4,000 interpolation, 4,000 position counts are generated per grating period (20 µm). This yields:

• Position resolution: 20 µm / 4,000 = 5 nm
• Velocity resolution at 100 mm/s: 5 nm × (100 mm/s / 5 nm) = 20 MHz count rate — within the capability of high-speed interpolation electronics

Index Mark Implementation in Interferential Systems

The index mark in an interferential optical encoder is not simply a wider opaque bar on the grating. For interferential systems, the index mark is designed to generate a focused band of light at the Talbot plane, functioning as a cylindrical lens rather than an absorber.

Multiple identical index tracks are interleaved with the position track. Position track signals and index track signals are processed separately, and push-pull signal processing is applied to eliminate the common-mode interference signal. The result:

• No false index pulses: The push-pull cancellation eliminates the signal from the position track, leaving only the true index pulse
• No missing index pulses: Multiple interleaved index tracks ensure the index is visible from multiple sensor positions simultaneously
• Index signal independent of illumination intensity: Differential processing cancels common-mode amplitude variations

This implementation contrasts with simpler index mark designs that use a single transparent slot, which can produce false pulses from edge diffraction and can be masked by contamination.

Signal Quality Metrics for Interferential Encoders

Signal quality determines the usable interpolation factor:

High-quality interferential sensors specify:

• THD < 1%
• SNR > 60 dB
• Amplitude imbalance < 2%
• Phase offset < 1°

These specifications support ×4,000 interpolation with positional accuracy within ±5 nm rms. Degraded signal quality (contamination, misalignment, beam intensity variation) produces proportionally larger interpolation errors.

Recommended articles to continue reading: 

• RF Rotary Joints: Frequency Range, Loss Specifications, and Integration with Slip Ring Assemblies, 
• in addition to our analysis of Incremental vs. Absolute Rotary Encoders: Technical Differences and Selection Criteria 
• and our read on Slip Ring Offshore Wind: Corrosion Protection, Ingress Rating, and Lifecycle Service.