Surface Roughness Measurement: Ra, Rz, and Rq

Surface roughness measurement quantifies the fine peaks and valleys of a surface by tracing a profile and reducing it to parameters, most often Ra, Rz, and Rq, over a defined sampling length, so a finish requirement can be verified instead of judged by eye or thumbnail. The number replaces the argument.

Surface finish decides whether a seal holds, a bearing lasts, a coating sticks, or a part fatigues early, so drawings call it out with a number and a parameter. But the same surface can read very differently depending on which parameter you use and how the instrument is set up, which is why "it measured 0.8" means nothing without knowing 0.8 of what, over what length, filtered how. This guide covers what Ra, Rz, and Rq mean, how a profilometer captures them, why cutoff length changes the result, and how to read a finish callout, the metrology behind any serious dimensional inspection.

What is surface roughness measurement?

Surface roughness measurement is the quantification of a surface's texture, specifically the short-wavelength irregularities left by the manufacturing process, separated from the longer-wavelength waviness and the overall form. Every real surface has three overlapping components: form (the intended shape), waviness (broad undulations from machine deflection or vibration), and roughness (the fine tool marks). Roughness measurement isolates that third component and expresses it as a single number, or a few numbers, so a designer can specify it and an inspector can verify it.

The reason a single trace is not enough on its own is that a surface is statistical. A finish parameter summarizes thousands of height readings into one value, which is powerful and lossy at the same time: two surfaces that feel and function very differently can share the same Ra. That is exactly why more than one parameter exists, and why the parameter is part of the specification, not an afterthought.

What do Ra, Rz, and Rq mean?

Ra, Rz, and Rq all describe the roughness profile but answer different questions. Ra, the roughness average, is the arithmetic mean of the absolute distances of the profile from its mean line over the sampling length. It is the workhorse, stable, widely specified, easy to reproduce, but because it averages, it forgives the occasional deep scratch by drowning it in thousands of ordinary readings. Rq, the root-mean-square roughness, squares the deviations before averaging, so larger excursions count more; it is preferred in optics and in statistical work, and for a given surface Rq usually runs slightly higher than Ra. Rz, the maximum height of the profile, reports the average of the largest peak-to-valley height in each sampling length, which makes it sensitive to exactly the isolated defects Ra hides.

One caution worth stating plainly: the definitions of Rz have changed over the standards' history. Current ISO and ASME practice defines Rz as the average maximum peak-to-valley height across sampling lengths, while an older definition (sometimes called Rz DIN or the ten-point height) computed it from the five highest peaks and five deepest valleys. They are not interchangeable, so a drawing should reference the standard edition it means, and an inspector should set the instrument to match.

How does a profilometer measure roughness?

The most common instrument is a contact stylus profilometer. A fine diamond tip, typically a few micrometres in radius, is drawn across the surface at a controlled speed while a transducer records its vertical movement as a function of horizontal distance. That raw trace contains form, waviness, and roughness all mixed together, so the instrument applies filters to isolate the roughness band before computing the parameters. Non-contact optical profilometers, using focus variation, confocal, or interferometric methods, do the same job with light instead of a tip, which suits soft, delicate, or very fine surfaces the stylus would scratch or miss, though correlation between methods must be checked.

Whichever instrument is used, the measurement is only as trustworthy as the system behind it. A profilometer needs calibration against a reference standard and its own gage R&R study, because a roughness value from an unverified instrument is a precise-looking guess. Choosing between a stylus and an optical tool, and knowing its limits, is part of picking the right gauge for the job.

What is cutoff length and why does it change the answer?

Cutoff length is the wavelength filter that separates roughness from waviness, and it is the setting that most often explains why two people measuring the same surface get different numbers. The roughness cutoff, written λc, sets the boundary: irregularities with wavelengths shorter than λc count as roughness, longer ones are filtered out as waviness. Choose too short a cutoff and you exclude legitimate roughness, reading the surface as smoother than it is; too long, and you let waviness leak in, reading it as rougher. The evaluation length then usually spans five sampling lengths so the parameters average over a representative stretch rather than one lucky or unlucky patch.

Because the cutoff changes the result, it is not a free choice, standards tie the appropriate λc to the roughness range and the surface type. ISO 4288 and ASME B46.1 give the rules for selecting cutoff and sampling length, and a finish specification is incomplete until the cutoff is defined or defaulted to the standard. Reporting an Ra without stating the cutoff is like reporting a temperature without the scale.

How do you take a surface roughness measurement?

The procedure below applies to a stylus profilometer, the most common shop instrument.

1. Clean the surface. Wipe off oil, coolant, chips, and dust. Contamination reads as false peaks and ruins the trace; a clean surface is the cheapest accuracy you will buy.
2. Calibrate against a reference. Verify the instrument on a certified roughness standard before measuring, so you know the tip and electronics are reading true.
3. Set the parameter, cutoff, and lengths. Select Ra, Rz, or Rq to match the callout, and set the cutoff λc and evaluation length per ISO 4288 or ASME B46.1 for the expected roughness range.
4. Orient across the lay. Trace perpendicular to the machining marks (the lay), where roughness is greatest and most representative, unless the drawing specifies otherwise.
5. Take multiple traces. Measure at several locations and orientations, because a single trace can land on an unrepresentative spot. Real surfaces vary across the part.
6. Record the full result. Report the parameter, the value, the cutoff, and the number of traces, not a bare number. A result without its measurement conditions cannot be reproduced or disputed.

How do you read a surface-finish callout?

A surface-finish callout is a check-mark-shaped symbol carrying the requirement, and you read it by position. The parameter and its limit, say Ra 1.6, sit above or beside the symbol; a value alone conventionally means Ra unless another parameter is named. Added marks can specify the maximum and minimum, the machining allowance, the required production method, and the lay direction. Different processes land in characteristic roughness ranges, rough turning and milling produce coarser finishes, fine grinding a smoother one, and lapping or honing the smoothest, but treat those as expectations to verify, not values to assume; the standard and the drawing govern, and the actual number comes from measurement.

The finish requirement is one line on a drawing full of dimensions and tolerances, and like a true-position callout it means nothing until it is measured with an instrument you trust. Read the parameter first, note the cutoff, and set the profilometer to match before you touch the part, a callout read carelessly is a finish measured wrongly, and the two failures look identical on the report.

By the numbers

By the numbers. ASME B46.1 is the U.S. standard for surface texture, roughness, waviness, and lay, and defines the parameters (including Ra, Rq, and Rz), the measurement methods, and the instrument and filtering requirements used to compute them; its 2019 edition is the current revision (ASME B46.1, Surface Texture). Internationally, the roughness profile parameters are defined in the ISO surface-texture standards (historically ISO 4287, with the newer ISO 21920 series), and ISO 4288 sets the rules for selecting cutoff and sampling length so a measurement is repeatable (ISO, Geometrical Product Specifications (GPS), Surface texture). The practical takeaway from both: Ra averages and hides isolated defects, Rz and Rq react to them, and no roughness value is complete without its parameter, cutoff, and evaluation length stated alongside.

On a shop floor, the recurring failure is not the physics but the record-keeping: finish results scribbled on a traveler, cutoff conditions unrecorded, and no way to see a process drifting toward the rough end until parts come back. When roughness and dimensional results are captured digitally at the point of inspection and connected to the part they belong to, the kind of plumbing Harmony's quality intelligence and connected-systems modules handle alongside the measurement systems analysis that proves the gauge can judge the tolerance, a surface trending toward its limit shows up the same shift, not at the next audit. CLS made that shift, from measurements found the next morning to measurements visible while the parts are still on the line.