Surface Finish Demystified

Ra, Rq (RMS), Rz—how to spec what you actually need (and avoid paying for what you don’t).

TL;DR

• Ra (average roughness) is the most common callout, but it doesn’t capture peaks/valleys or texture direction (lay).
• Rq (RMS) is always ≥ Ra; on many machined surfaces Rq ≈ 1.1–1.3 × Ra. Don’t assume a fixed conversion.
• Rz (peak-to-valley) tells you about worst-case asperities; great for sealing or coatings—but varies by standard/method.
• Specify where it matters (seals, sliding, optics, thermal interfaces). Keep everything else “as-milled/turned” to save time and cost.
• If you care about function (leak-tight seal, friction, reflectivity), pair Ra with notes on lay, cutoff/filter, and measurement method.

1) What “surface finish” actually measures

Surface finish parameters summarize tiny height variations on a surface:

• Ra: arithmetic mean of absolute heights from a center line. Simple, robust, cheap to measure.
• Rq (a.k.a. RMS, root-mean-square): more sensitive to outliers than Ra; useful for optical/tribology work.
• Rz / Rt: peak-to-valley metrics. Helpful for seals, coatings, paint, and any application where a single tall peak or deep pit is the problem.
• Waviness vs roughness: low-frequency form undulations (waviness) are not the same as micro-roughness. Cutoff filters separate them.
• 3D analogs (for areal maps): Sa, Sq, Sz. Increasingly common with optical metrology.

Key idea: Ra is a good “first-order” number. If a tall burr or a deep pit would kill performance, add Rz or a max peak height limit, not just Ra.

2) How it’s measured (and why the method matters)

• Contact stylus profilometer: a diamond tip drags across the surface to trace a 2D profile.
  • Pros: fast, cheap, widely available, standards-based.
  • Watch-outs: stylus radius can “bridge” narrow valleys; direction matters (measure across the lay for conservative results).
• Optical methods (confocal/white-light interferometry): produce 3D maps; great for micro-features/optics.
  • Pros: non-contact, detailed areal data.
  • Watch-outs: highly reflective or transparent surfaces can be tricky; settings (filtering, thresholds) change numbers.

Cutoff/Filtering matters. Standards define default cutoff lengths and filters (ASME/ISO). If you don’t specify them, two shops can measure the same part and get different results—both “correct.”

3) Typical finish ranges by process (ballpark)

• Rough milling/turning: ~125–250 µin Ra (3.2–6.4 µm)
• Finish milling/turning: ~32–63 µin Ra (0.8–1.6 µm)
• Fine turning / precision pass: ~16–32 µin Ra (0.4–0.8 µm)
• Grinding: ~8–16 µin Ra (0.2–0.4 µm)
• Honing / Lapping / Superfinishing: ~2–8 µin Ra (0.05–0.2 µm)
• Bead-blast: hides tool marks but raises Ra (often > 100 µin Ra). Great cosmetically; not for seals.

Cost signal: Every step down the “finish ladder” usually adds tool changes, slower feeds, or post-processing. Tighten finish only where it pays for itself.

4) When finish truly matters (and what to spec)

Sealing surfaces (O-ring / metal-to-metal)

• Goal: prevent leak paths.
• Suggestion: Ra 16–32 µin max, Rz limit (to catch spikes), lay ⟂ to seal direction, no bead-blast on sealing land.
• Add: flatness and scratch-free note.

Sliding/tribology (bushings, rails, pistons)

• Goal: reduce wear and friction, retain lubricant.
• Suggestion: Ra 8–16 µin with a preferred lay; consider bearing ratio (Abbott curve) or plateau-honed note.

Optics/laser/reflective parts

• Goal: minimize scatter; sometimes anisotropic textures are harmful.
• Suggestion: Ra ≤ 8–16 µin (or lower), Rq report, and specify measurement method (optical) if critical.

Thermal interfaces

• Goal: better contact area with TIM or bare contact.
• Suggestion: Ra 16–32 µin with tight flatness; extremely low Ra without flatness does not help.

Coatings & plating (anodize, Ni, Au)

• Plating can partly smooth or amplify texture; anodize often increases Ra slightly.
• Suggestion: Call out finish before coating, and if the coated surface is functional, specify post-coat finish or a masking plan.

5) Spec tips that save time and money

1. Call out finish only on functional surfaces. Everywhere else: “as-machined”.
2. Include lay. Example: “Lay ⟂ to axis” or “parallel to axis.”
3. Mention cutoff/filter when it matters. “Evaluate per ASME/ISO defaults” or specify the λc.
4. Avoid conflicts. “Bead-blast + Ra 32 µin max on the same land” fights itself—split the surfaces.
5. Pair finish with geometry. A super-smooth but non-flat sealing face still leaks.
6. If spikes are deadly, add Rz. Ra alone can hide a burr.
7. Note the measurement direction (across the lay) to make results consistent.

6) Quick conversion helper (µin ↔ µm)

(1 µin = 0.0254 µm)

About grit numbers: “Grit ↔ Ra” charts are rough heuristics that vary with media, pressure, and technique. Use them for ballpark only—verify on your actual parts.

7) Ready-to-paste drawing notes (examples)

A) Sealing face (O-ring land)
Surface finish: 16 µin Ra max, Rz ≤ 100 µin Lay: perpendicular to seal path Measure across lay; cutoff per standard defaults No bead-blast on sealing land; no scratches > 0.002"

B) Sliding bore
Surface finish: 8–16 µin Ra, plateau preferred Report Rq and bearing ratio if outside nominal performance Hone final pass; measure across lay

C) Cosmetic chassis
Bead-blast cosmetic surfaces only (callouts shaded) Functional surfaces “as-machined” unless otherwise noted Edges: break 0.2–0.4 mm; remove loose media

8) Common pitfalls (and how to avoid them)

• Specifying a single global Ra across the whole part → cost spike for no gain.
• Measuring along the lay → artificially low numbers; be explicit.
• Ignoring cutoff/filter → two “good” measurements that don’t match.
• Chasing Ra while ignoring burrs → add Rz or peak height limit where spikes matter.
• Finishing after tight tolerances without compensation → you’ll lose material; plan the order of ops.

9) Ziqual’s practical approach

• Function first. We push Ra/Rq/Rz only where performance needs them; everything else stays “as-machined.”
• Clarity beats magic. We match measurement direction, cutoff, and lay to your spec and include the method on FAIRs when needed.
• Repeatability. Each part’s traceability record (“DNA”) keeps the recipe—so the next run hits the same finish without a fresh learning curve.

10) Copy-and-keep checklist

• Which surfaces actually need a finish? (circle them)
• Pick the metric(s): Ra only, or Ra + Rz / Rq?
• Note lay and measurement direction.
• Confirm cutoff/filter (defaults or specified).
• Check for conflicts (bead-blast vs seal, finish vs tolerance).
• If coated: pre- vs post-coat finish? Any masking?
• Add the inspection method if it’s critical (contact vs optical).
• Verify cost/lead-time impact before locking spec.