RESEARCH / BLOG

Research,
Explained.

Read the research question, follow the authors’ methods and explore the evidence behind the article.

Research Blog

Electronics research · Research insights

Correcting encoder signals improves tiny-motion measurements

A 2026 FPGA-based experiment improves grating-encoder measurements through signal correction. Understand the difference between resolution and accuracy.

By NPT Research3 min read

An optical encoder turns movement into a changing light signal. The position estimate is only as good as both the optical pattern and the electronics that interpret it. Unequal channel amplitudes, offsets and electrical noise can turn a regular grating into an irregular measurement.

Latest breakthrough

A 2026 Measurement paper combines signal correction and noise suppression for self-imaging grating encoders on an FPGA platform. The experiment addresses amplitude imbalance, phase error and direct-current offset. Using a four-micrometre grating period, the authors report measurements at the ten-nanometre precision level with an accuracy of 0.15 micrometres, alongside a 66.88% improvement in their standard-deviation-based stability measure. These quantities describe different aspects of performance. The advance is a measured combination of correction and hardware noise control, not a claim that every reported position is accurate to ten nanometres.

Xin and colleagues. Measurement (2026). Read the original study.

How the method works

The encoder uses a grating-generated optical signal to infer position. Amplitude imbalance, phase error and a DC offset can distort that inference, so the method combines correction with noise suppression on an FPGA. An FPGA is configurable digital hardware that can perform a chosen processing pipeline. Precision concerns repeatability, accuracy concerns closeness to the reference and the reported stability statistic describes variation under the study's conditions. Correcting one does not automatically give the same numerical improvement in all the others; each result must retain its own definition.

A simple analogy

A ruler that starts at the wrong zero and a shaky hand create different measurement problems. Correcting the zero does not remove the shake, and repeated readings can agree while remaining offset. Encoder signals have more complicated errors than a ruler, but the analogy separates systematic correction, noise and the difference between accuracy and precision.

Why two channels are useful

Two appropriately offset signals can reveal both movement and direction. Their relative phase provides a way to estimate position within one grating period. If one channel has an offset or a different amplitude, the calculated phase can become distorted. Correcting that relationship helps the electronics extract more useful information from the same basic optical pattern.

A practical lesson for precision machinery

Our interpretation is that improving signal quality can be as important as specifying a finer grating. Machine tools and measuring systems need repeatable readings across their operating environment. A well-designed correction method could make existing optical hardware more useful, provided calibration remains stable. The broader benefit would come from better process control and fewer measurement-driven errors, which need application-level evidence.

Resolution is not a certificate of truth

A display can show many decimal places while still having systematic error. Resolution, repeatability and accuracy answer different questions, and the paper reports them separately. Thermal expansion, alignment and reference calibration remain important. The study concerns linear grating measurements; applying the approach to another encoder architecture requires a fresh error analysis. Better position information can support control, but it is distinct from the higher-level task choices explored in the modular robot-arm experiment.

Compare linear sensing with a rotary model

The gallery rotary encoder has a 24-period disc, two optical channels, a shaft and interface board. Follow how rotation changes the light received by each channel. This illustrates the phase-and-direction principle shared with linear encoders. Its disc is not the paper's micrograting, and the viewer does not implement the FPGA correction algorithm or reproduce its metrology results.

Sources and further reading

  1. Improving measuring accuracy of self-imaging-based grating encoder by signal correction and noise suppression