Surface Texture

§1What a surface really is

A real surface carries texture at three wavelengths stacked on top of each other, plus its direction and its accidents.

Roughness is the fine, closely spaced texture the cutting process leaves — the tool’s fingerprint. Waviness is the longer-wave undulation from machine vibration, deflection and heat. Lay is the dominant direction of the pattern — along the feed on a turned part, crossed arcs on a milled face, multidirectional on a lapped one. Flaws are the accidents: scratches, pits, tears. Measurement separates roughness from waviness with a wavelength filter (§4), because they come from different causes and cure by different means — a rougher insert is a feed problem; waviness is a machine problem.

Contents

§2Ra — the arithmetic average

Draw the mean line through the profile; Ra is the average distance of the surface from it, valleys counted as positively as peaks.

Ra = 1n Σ |zᵢ| (zᵢ = profile heights from the mean line; µm)

Example 1 — eight samples by hand

Heights from the mean line: +2, −3, +1, −2, +3, −1, +2, −2 µm. Sum of magnitudes 16, so Ra = 16/8 = 2.0 µm — a healthy general-machining finish. The hero trace above carries the same arithmetic done over 800 samples, its Ra and Rz printed from the computation.

Ra’s virtues are stability and universality — averaging tames measurement noise, and everyone quotes it. Its vice is the averaging itself: §3.

Contents

§3Rz and the peaks Ra hides

Rz averages the peak-to-valley height over five sampling lengths — a measure that notices the scratch Ra shrugs off.

Example 2 — same Ra, very different surfaces

Surface A alternates ±2 µm: Ra = 2.0, total height 4 µm. Surface B runs gently at ±1 µm but carries one 6 µm scratch pair per length: Ra = (8×1 + 6 + 6)/10 = 2.0 µm — identical — yet its peak-to-valley is 12 µm, three times A’s. A seal, a plated coat or a fatigue-loaded fillet meets the 12 µm, not the average. Where extremes matter, specify Rz (or Rt, the single worst) alongside Ra; as working texture, Rz commonly runs 4–7 × Ra, and a ratio far outside that band is itself a diagnosis — isolated scratches or tears.

Contents

§4Measuring texture

The standard instrument drags a fine diamond stylus across the surface and filters what it feels.

The stylus (tip a few microns) rides the profile; electronics remove the long waves above the cutoff wavelength λc — 0.8 mm is the common default — so that only roughness remains, and evaluate over five cutoffs (a 4 mm traverse) for a stable average. Quoting a roughness number without its cutoff is quoting half a number: a wavy surface can grade “smooth” at a short cutoff and “rough” at a long one. Trace across the lay — the reading along the lay flatters. Non-contact optical instruments map whole areas without touching soft or delicate surfaces; the parameters and cutoffs carry over.

Contents

§5Typical roughness by process

Indicative Ra by process (µm — centre of each process’s usual band)
Process	Typical Ra	Process	Typical Ra
Sawing, flame cutting	12.5 – 25	Grinding	0.2 – 1.6
Rough turning / milling	3.2 – 12.5	Honing	0.1 – 0.8
Finish turning / milling	0.8 – 3.2	Lapping / polishing	0.025 – 0.4
Reaming	0.8 – 3.2	Die casting / injection moulding	0.4 – 3.2 (as-formed)
Each halving of Ra roughly means another operation or a slower one — the cost curve behind §7’s advice to specify only what the function needs.

Contents

§6The N-grade series

The roughness grades N1–N12 are a doubling ladder of Ra values, rounded to preferred numbers — one label per rung.

Roughness grade numbers N1–N12 against Ra (µm)
Grade	Ra	Grade	Ra
N1	0.025	N7	1.6
N2	0.05	N8	3.2
N3	0.1	N9	6.3
N4	0.2	N10	12.5
N5	0.4	N11	25
N6	0.8	N12	50
Each grade doubles the Ra of the one below, rounded to the preferred-number series — so N7 (1.6 µm) is a typical finish-machined face and N1 (0.025 µm) a lapped one. The grade is shorthand; the Ra value is the requirement.

Contents

§7Specifying finish on drawings

The tick symbol carries the whole requirement: the value, and only where the function pays for it.

The basic check-mark on a surface means “machined”; the Ra value (or grade) sits on the symbol’s long arm; a bar across the top means no material removal permitted (an as-cast or as-rolled face must stay); additions call the process, the lay direction and the sampling where they matter. Discipline mirrors dimensioning: a general note (“machined surfaces 3.2 Ra unless stated”) covers the bulk, and explicit ticks mark only the working faces — the bearing seat at 0.4, the seal land at 0.2, the gasket face at 1.6. Every unnecessary 0.4 on a drawing is a grinding operation bought for nothing.

Contents

§8Function and finish

Finish is not cosmetics; each function reads the mountain range differently.

Fatigue starts cracks in valleys — smoother fillets live longer, and a ground-then-polished radius on a cycled shaft is cheap life insurance. Sealing needs peaks low enough for the elastomer to follow: dynamic seal lands typically want Ra around 0.2–0.4, and Rz watched, because one scratch is one leak path. Bearings and slides want an apparent paradox: too rough tears the oil film, but a mirror finish holds no oil at all — the classic answer is the plateau surface (honed cylinder bores): flat load-bearing lands with deliberate valleys left as oil reservoirs, a surface two roughness parameters describe better than one. Coating and bonding want controlled roughness as a key. State the function; the number follows.

Contents

§9Quick reference

The working core of the page on one card rack.

Ra

mean of |z| from the mean line

the universal number

Rz

mean peak-to-valley, 5 lengths

≈ 4–7 × Ra when healthy

Measurement

stylus · cutoff λc 0.8 mm

trace across the lay

Processes

turn 0.8–3.2 · grind 0.2–1.6

lap 0.025–0.4

Specify

general note + ticks where it pays

halving Ra ≈ one more operation

Contents

Handbook application: from concept to controlled practice

Purpose. This expanded section turns the original page into a practical handbook. It preserves the supplied material and adds a repeatable way to apply, check and review Surface Texture. It does not replace a contract, legislation, a controlled standard, competent engineering judgement or specialist advice.

The operating aim is to carry the subject from function and assumptions through design evidence, verification and controlled release. Read the original explanation first, then use the workflow and checks below to convert knowledge into evidence.

Apply Surface Texture by beginning with the duty, not the component or software command. Convert the key ideas—finish, surface, roughness, texture, average—into measurable requirements and interfaces. Record operating and non-operating environments, duty cycle, expected life, loads, energy sources, human interaction and reasonably foreseeable abnormal conditions. When a value is not a project requirement or verified supplier datum, identify it as an assumption or illustrative value.

Create a calculation and evidence trail that another competent person can audit. Every input should carry a source, unit, revision and uncertainty or tolerance where relevant. Every model should state its boundary conditions and limitations. Keep nominal capacity separate from design capacity, and keep verification margin separate from an arbitrary safety factor. If a code or standard governs the work, confirm the applicable edition and contractual status rather than copying a number from a secondary summary.

Design for manufacture, assembly, inspection, operation and maintenance at the same time. A technically valid geometry can still fail because it cannot be fixtured, measured, cleaned, guarded, reached or replaced. Review process capability, datum or reference strategy, tolerance accumulation, access, error-proofing and changeover. Where people interact with plant, apply the hierarchy of controls and consult those who will operate, clean, maintain and recover the equipment.

Plan verification before release. Define the characteristic, method, equipment, sample or test condition, acceptance criterion, record and responsible person. Validation then asks a different question: whether the resulting system is effective and suitable in the intended use context. A passed drawing check or analysis does not by itself validate usability, maintainability or production performance.

Step-by-step operating method

Define the duty. Capture the required function, interfaces, operating environment, life, loads and unacceptable outcomes.
Establish the model. Identify governing principles, units, material or process data, assumptions and uncertainty.
Develop alternatives. Compare feasible concepts against performance, manufacturability, safety, maintainability and cost.
Verify the design. Use analysis, test, inspection or demonstration with acceptance criteria defined before execution.
Release and learn. Baseline the design, control changes, retain evidence and feed operating results into the next revision.

Illustrative design review record

Illustrative values only. Build a one-page record with the required function, input sources, assumptions, governing load or process condition, failure consequences, selected concept, verification method and acceptance criterion. Mark every numerical input as project requirement, verified supplier data, measured value, calculation output or assumption. Review the weakest evidence first. If an assumption can change safety, compliance, interchangeability or capacity, it must be resolved before release rather than buried in a calculation note.

Evidence class	Question	Release expectation
Requirement	What must the design do and under which conditions?	Approved and traceable
Input	Where did the load, property, tolerance or process limit come from?	Source, unit and revision recorded
Analysis	Which model and assumptions connect input to result?	Checkable calculation or simulation
Verification	How will conformity be demonstrated?	Method and acceptance criterion agreed
Validation	Will the solution work for intended users and conditions?	Representative use evidence

Common failure modes and recovery actions

1. Watch for

Starting detailed design before interfaces and operating limits are agreed.

Recovery: Return to the governing definition or requirement and restate the decision in one sentence.

2. Watch for

Using catalogue or typical values as though they were certified project inputs.

Recovery: Separate evidence from assumption, assign an owner and set a date for validation.

3. Watch for

Checking nominal performance while ignoring tolerances, degradation and foreseeable misuse.

Recovery: Run a small counterexample, boundary test, pilot or independent check before proceeding.

4. Watch for

Confusing verification of requirements with validation of user need.

Recovery: Record the consequence, decision and rationale, then update the controlled baseline.

5. Watch for

Releasing drawings or procedures without configuration, inspection and change controls.

Recovery: Escalate when the issue affects safety, compliance, acceptance, material value or an agreed tolerance.