Free Trial

← Back to blog

Surface qualityReading time 13 min

What that little tick symbol on the drawing is actually asking for

Surface roughness is one of those rare specifications that looks almost identical and measures very differently. It is also the one most often reduced to “Ra 1.6” on the drawing, and the one people argue about at incoming inspection.

Ra

01What each field of the symbol says

The body of the symbol is a tick. A plain tick means the surface condition is required but the method is not specified. A tick with a bar means the surface must be produced by removing material (turning, milling, grinding). A tick with a circle means material must not be removed (as-cast, as-forged, as-drawn).

PositionContentsExample
a (left)Roughness parameter, limits and valueRa 1.6、Rz 6.3
b (above the bar)Machining methodGround, turned, EDM
c (above the bar, right)Sampling length / cut-off0.8
d (inside the tick)Lay direction symbol=、⊥、X、M、C、R
e (below the bar)Machining allowance0.5
Field layout of the roughness symbol (per ISO 21920-1). Most drawings fill in only a and leave the rest at default.

The one near the title block governs the whole drawing

A roughness symbol placed near the title block is the global default: it applies to every surface not called out individually. If it is followed by a bracket and a tick, it reads as “all other surfaces are this value; exceptions are marked individually”. An inspection sheet built only from the symbols on the drawing body will miss that whole category, and it usually covers more than eighty per cent of the part.

02Ra is the average, Rz is the extreme

Ra, the arithmetic mean roughness, takes the profile deviation from the mean line over a sampling length, in absolute value, and averages it. It is stable, repeatable and forgiving of operator error, which is why it is the most-specified parameter in the world. Its weakness is the same thing: a single deep scratch gets averaged away.

Rz, the maximum height, is the sum of the highest peak and the deepest valley within the sampling length. It is extremely sensitive to a single defect, which is exactly why sealing faces that must not leak, and shaft shoulders under cyclic stress, are usually given both Ra and Rz.

ParameterDefinitionGood for controlling
RaArithmetic mean of the absolute profile deviationsOverall process stability, general fitting surfaces
RzMaximum peak height plus maximum valley depthSealing faces, fatigue-sensitive faces, single defects
Rmax / Rz1maxThe largest Rz among the sampling lengthsExtremes that are simply not allowed
RsmMean spacing of profile elementsTexture pitch, paint adhesion
Rmr(c)Material ratio at a given heightBearing surfaces, sliding fits
RskSkewness, whether the distribution leans to peaks or valleysThe surface's ability to retain an oil film
Common profile parameters. Ra and Rz account for more than ninety per cent of what appears on real drawings.

03From profile to area: the S parameters of ISO 25178

Traditional contact measurement draws a line: the stylus travels a set distance across the surface and produces a two-dimensional profile. The trouble is that moving or rotating the trace changes the number, especially on milled surfaces with a strong lay, where measuring along the tool marks and across them give very different answers.

Areal roughness measures a whole region instead. The parameters begin with S and are defined by ISO 25178. Sa corresponds to Ra and Sz to Rz, but the set adds information a line trace cannot capture: Sdr, the developed interfacial area ratio, describes how much larger the real surface is than its projected area and relates directly to coating adhesion and contact heat transfer, while Spk and Svk split the surface into peak, core and valley zones, particularly useful for cylinder bores and bearings, which need a smooth core with valleys that hold oil.

A machined surface under magnification: what looks flat to the eye turns into peaks and valleys, and those are what roughness actually describes.
A machined surface under magnification: what looks flat to the eye turns into peaks and valleys, and those are what roughness actually describes.Wikimedia Commons・Silicon chip under microscope・CC BY-SA 4.0

04Why the same face measures differently twice

Roughness is not simply whatever the instrument reads. The surface is first separated by wavelength, and only the short-wave part is kept. How it is separated is decided by the filters.

NameSymbolWhat it does
λs filterλsRemoves components shorter than itself (noise, artefacts from the stylus radius)
λc cut-offλcThe boundary between roughness and waviness, the single most critical setting
λf filterλfSeparates waviness from form error
Three filters divide the surface signal into roughness, waviness and form error.

Change λc and a different amount of the signal is classified as waviness and filtered out, so Ra changes with it. ISO recommends choosing λc from the expected Ra (for example, Ra 0.1–2 μm pairs with λc 0.8 mm), with the evaluation length usually five sampling lengths. Strictly speaking, then, “Ra 1.6” with no measurement conditions is an incomplete specification, two parties using different λc values will reach different conclusions.

Contact and non-contact

  • Stylus profilometer. Mature, traceable and inexpensive. The limits are the stylus radius (typically 2 or 5 μm), which cannot enter narrower valleys, and the risk of scratching soft or coated surfaces.
  • Confocal, white-light interferometry, laser microscopy. Non-contact, capable of areal measurement, with vertical resolution down to sub-nanometre. They struggle on highly reflective or steeply sloped surfaces.
  • Atomic force microscopy. The highest resolution available, but a tiny measurement range and slow acquisition make it a research tool rather than a production one.