# Surface profilometry

Contact and optical surface-height measurement, with controls for sampling, filtering, calibration and interpretation.

Reviewed: 2026-11-01

Tags: inspection, metrology, surface-texture, profilometry

Use the drawing’s parameter, trace or area, and processing settings. ISO **21920-2:2021**, corrected June 2022, defines profile texture. ISO **25178-2:2021** defines areal texture. [ISO 21920-2](https://www.iso.org/standard/72226.html), [ISO 25178-2](https://www.iso.org/standard/74591.html)

As a bounded example, NIST describes its **Taylor Hobson Form Talysurf PGI 1240** with a 60 mm stylus arm, 2 µm radius conisphere diamond tip and 1 mN force. Its capability table lists 0.125 µm X sampling interval for measurement lengths from 0.1 to 15 mm and 0.8 nm Z resolution at a 12.5 mm range. These are different quantities: lateral sample spacing does not make the probe capable of entering every feature at that spacing, and vertical resolution is not overall uncertainty. These values describe NIST's listed configuration, not every PGI system or a recommended setup for every surface. [NIST: stylus profilometer](https://www.nist.gov/laboratories/tools-instruments/stylus-profilometer)

## Define the measurand before selecting equipment

Specify the feature, units, parameter and relevant scales. A roughness parameter calculated after filtering is different from the total measured profile. Trace direction also matters on a directional surface: measuring along machining marks can give a different result from crossing them. NPL's stylus guidance illustrates both direction effects and surfaces with the same Ra but different functional behavior. One average-height parameter cannot by itself establish sealing, wear or lubrication performance. [NPL Good Practice Guide 37, issue 2](https://eprintspublications.npl.co.uk/2041/1/mgpg37.pdf)

A useful inspection request identifies whether the objective is a drawing-defined texture value, a localized step, a coating edge, a scratch or a broader contour. It also distinguishes a representative production measurement from a deliberately selected worst-looking region. Retain the location map: a precise result from the wrong surface patch does not answer the intended question.

## Contact stylus measurement

A stylus instrument traverses a tip across the surface and measures vertical displacement. The tip has finite geometry: it may bridge narrow valleys and mechanically filter the apparent profile. Contact force must maintain tracking without damaging the specimen. NPL discusses both limitations. Record the tip radius and geometry, contact force, traverse speed and direction, rather than specifying only the instrument model. Check whether the surface tolerates contact before inspection. [NPL: stylus geometry, force and measurement](https://eprintspublications.npl.co.uk/2041/1/mgpg37.pdf)

## Optical measurement

Optical profilers include interferometric, confocal and other measurement families. Coherence scanning interferometry (CSI) obtains height information from interference signals during a vertical scan. Its measurement depends on optics, signal processing and specimen interaction rather than mechanical tracing. NPL's CSI guide treats objective selection, lateral resolution, spatial sampling, slope handling and scan settings as separate considerations. A dense camera grid alone does not establish that fine surface structure is resolved. [NPL Good Practice Guide 116](https://eprintspublications.npl.co.uk/4833/1/mgpg116.pdf)

Verify that the chosen method measures the intended surface over the relevant slopes, materials and height range. Inspect the height map and signal-quality or invalid-point information before trusting a summary statistic. Missing points, steep edges and processing choices need a documented treatment, not silent replacement with plausible heights. The NPL guide discusses these acquisition and analysis limitations; optical performance must be demonstrated for the actual application. [NPL: CSI setup and analysis](https://eprintspublications.npl.co.uk/4833/1/mgpg116.pdf)

## Filtering changes what is evaluated

Acquisition and analysis settings form one measurement method. Leveling or form removal establishes a reference; filters select the spatial scales retained for texture assessment. Bruker's application note illustrates how tip geometry and electronic filtering affect stylus results and how matching scan lengths, cutoffs and filters is necessary when correlating optical and stylus measurements. Selecting the same parameter name is insufficient. [Bruker application note 558: correlation to traceable standards](https://www.bruker.com/en/products-and-solutions/test-and-measurement/3d-optical-profilers/resource-library/an-558-correlating-advanced-3d-optical-profiling-surface-measurements-to-traceable-standards.html)

Record the acquired length or area, evaluation region, sample spacing, filter type and scale settings, form-removal operation and edge treatment. Use the standard and edition required by the drawing; do not silently replace legacy settings with a newer software default. Cropping, smoothing or excluding a scratch can change the answer. If an exclusion is authorized, preserve the original data and explain its basis.

**Original simplified example:** four equally weighted, already referenced height samples are −1, −1, +1 and +1 µm. Their mean absolute height is 1 µm and their RMS height is 1 µm. Another zero-mean set, −2, 0, 0 and +2 µm, also has mean absolute height 1 µm, but RMS height √2 ≈ 1.414 µm and twice the peak-to-valley span. This arithmetic demonstrates loss of shape information in a single average. It is not a standards-compliant Ra or Sa evaluation: a real result also requires the specified acquisition, reference and filtering operations.

## Calibration, correlation and uncertainty

Verify the axes and characteristics relevant to the task. NIST's stylus system uses a calibrated sphere for roughness-height traceability, a sphere or interferometrically calibrated step for step height, and a displacement-measuring laser interferometer for lateral motion. A vertical reference alone does not establish lateral spacing accuracy. Software calculation can also be checked independently; NIST provides reference parameter-calculation tools and datasets. [NIST: traceability and parameter validation](https://www.nist.gov/laboratories/tools-instruments/stylus-profilometer)

Do not assume numerical equivalence between contact and optical methods. A 2007 NIST comparison found method-dependent differences on the particular roughness specimens studied. Its results are evidence for validating correlation, not a universal correction factor for modern equipment. Compare the same region and processing bandwidth where feasible, and investigate disagreements using the specimen's structure and each probe's response. [NIST: comparison of optical and stylus methods](https://www.nist.gov/publications/comparison-optical-and-stylus-methods-measurement-rough-surfaces)

## Inspection record and decision limits

Record component identity, surface location and orientation, preparation, instrument configuration, reference checks, environmental conditions, raw data and analysis settings. Report the parameter, units, results by location, measurement capability and applicable acceptance rule. A texture limit does not automatically specify allowable isolated defects or overall form; those need their own requirements.

Sources were reviewed November 1, 2026. NPL Guide 37 explicitly states updated February 2014; Guide 116 carries copyright 2010. Their technical guidance is used with the separately identified current parameter standards rather than treating their older standards references as current defaults. NIST's instrument page was updated September 30, 2025. Bruker's web application note did not display a revision date.

[Inspection and quality documents](../../sourcing/documents/index.md)