Surface Finish Measurement: How Ra and Rz Are Actually Checked

Q: What is the difference between Ra and Rz?

Ra is the arithmetic mean of the absolute profile deviations from the centre line over the sampling length — an average that smooths out individual spikes. Rz is the mean height between the highest peak and lowest valley within each of five sampling segments — a value dominated by the extremes. The same surface can measure Ra 0.8 µm while carrying Rz 5–6 µm, because a few deep tool marks barely move the average but fully set Rz. They measure different things and cannot be exactly converted; the ratio Rz/Ra typically falls between 4 and 7 for turned and milled surfaces but depends on the process.

Q: Which instrument should be used to verify a finish callout?

A stylus profilometer is the reference method for specified Ra or Rz values — it traces the surface and computes the parameter over the correct sampling length. Surface comparison plates are acceptable for coarse in-process checks but are subjective and not evidence-grade. Optical methods (interferometry, focus variation) work well on delicate or very fine surfaces where a stylus could scratch. For dimensional plus finish verification, profilometer results are recorded in the inspection documentation.

Q: How many measurements does a finish check need?

One trace is not a verdict. Good practice is several traces at different locations and orientations on the face in question — turning marks run circumferentially, milling marks in the feed direction, so the stylus is traversed across the lay — and reporting the average with the worst single reading noted. Where a drawing specifies sampling and evaluation lengths, those settings are used exactly; otherwise standard 0.8 mm cut-off is applied for the usual 0.2–3.2 µm machined range.

Measurement practice from the quality room · Boyang Hardware, Zhongshan, Guangdong, China

Direct answer Surface finish is measured by tracing the profile and computing a parameter over a defined sampling length. Ra is the average absolute deviation from the centre line; Rz is the mean peak-to-valley height over five sampling segments — Rz reacts to single deep marks that Ra averages away, and the two are not interchangeable (typical Rz ≈ 4–7 × Ra on machined faces). The reference instrument is a stylus profilometer; comparison plates serve only for coarse shop checks. A credible result needs several traces across the lay, standard 0.8 mm cut-off for the usual machined range, and the readings recorded in the inspection report.

Ramean deviation from centre line
Rz ≈ 4–7 × Ratypical on milled/turned faces
0.8 mmstandard cut-off, 0.2–3.2 µm range
100%inspection before shipment

Ra, Rz and related parameters

Machined-surface callouts are overwhelmingly Ra, with Rz appearing on sealing and bearing work where single deep scratches matter. Rsk and other higher-order parameters appear on specialist drawings but are rare in general manufacturing practice. For reference values by application, see the Ra surface finish chart and the aerospace-specific bands in aerospace surface finish requirements.

Instrument comparison

Table 1 — Instrument options for verifying a machined surface, with honest suitability notes.
Instrument	What it gives	Suitable for	Caution
Stylus profilometer	Traced profile; Ra, Rz computed over set sampling length	The reference for specified callouts; evidence-grade numbers	Stylus radius filters very fine detail; skid-type units average more
Surface comparison plates	Visual/tactile match against graded samples	Fast shop-floor checks of coarse finishes	Subjective; not defensible against a disputed callout
Optical profilometry (interferometer / focus variation)	Full 3D topography, area parameters	Delicate, very fine or curved surfaces; research-grade work	Cost; setting area vs line parameters confuses results
CMM with scanning probe	Profile along a scanned line	Convenient when the part is already on the CMM	Probe filtering differs from a dedicated profilometer — align settings

A measurement procedure that holds up

Read the callout: parameter (Ra or Rz), value, and any lay or machining-direction note.
Set cut-off and evaluation length: 0.8 mm cut-off covers Ra 0.2–3.2 µm, the normal machined band; the drawing may state its own.
Traverse across the lay, not along it — turning marks run circumferentially, milling marks in the feed direction.
Take 3–5 traces at different spots on the face; report the average and flag the worst reading.
Record results against the drawing in the inspection documentation, alongside the dimensional data.

What appears on the report

Where a finish is dimensioned, our inspection documentation records: instrument used, parameter and cut-off, number of traces, measured values, and pass/fail against the callout — filed with the 100% dimensional inspection report and material certificate that travel with every shipment. The full documentation set is described in 100% inspection and reporting.

Honest limits: two traps we see regularly. First, Rz↔Ra "conversions" are approximations from process statistics, not identities — if your customer spec says Rz, measure Rz. Second, measurement settings change results: a 2.5 mm cut-off will report a different Ra than 0.8 mm on the same face, and skidless vs skidded stylus units disagree at the fine end. If a dispute arises, the tie-breaker is a defined method on a defined instrument — agree it before production, not after. We measure dimensioned callouts with calibrated equipment and report what we measured, including settings.

FAQs

What is the difference between Ra and Rz?

Ra averages all deviations from the centre line; Rz is mean peak-to-valley over five segments and is dominated by extremes. The same face can read Ra 0.8 µm with Rz 5–6 µm. They are different questions about one surface and cannot be exactly converted.

Which instrument should be used to verify a finish callout?

A stylus profilometer for specified values; comparison plates only for coarse shop checks; optical methods for delicate or very fine surfaces. Results are recorded in the inspection documentation.

How many measurements does a finish check need?

Several traces at different locations, traversed across the lay, averaged with the worst reading noted — one trace is not a verdict. Standard 0.8 mm cut-off applies unless the drawing states otherwise.