Surface roughness is the micro-geometry of an engineering surface — the texture of peaks, valleys and lay direction left behind by the manufacturing process. It controls how a sealing surface seals, how a sliding bearing wears, how a bonded joint adheres, how paint and platings adhere, and how a pressed-on bearing transmits force without slip. Getting surface roughness right matters for nearly every engineering function on a machined part. Getting it wrong — whether by specifying too fine a finish where it isn't needed, or by failing to verify the actual finish against the drawing callout — is expensive in either direction.
This guide covers what surface roughness actually is, the critical ISO 21920-2:2021 transition from the older ISO 4287:1997 standard (most published articles still cite the superseded standard — this guide gets it right), the Ra/Rz/Rsm parameter family with worked examples, surface roughness drawing symbols including the historical triangle and N-grade systems still found on older drawings, contact stylus profilometers vs optical non-contact profilometers, the full Mitutoyo Surftest SJ-210/310/410 portable range and Formtracer SV-C4500 stationary range, calibration via reference specimens traceable to NIST/PTB/NMIA, machining process Ra targets, and how AIMS supplies the full Mitutoyo Surftest range through the authorised Australian distributor network.
AIMS is an authorised Mitutoyo supply channel in Australia. The full Mitutoyo Surftest portable and Formtracer stationary range is quote-and-supply through the authorised Australian distributor — we coordinate the SJ-210/310/410 portable series, Formtracer SV-C4500 stationary, reference specimens, stylus accessories and NATA-accredited calibration service. Contact us for a Surftest quote or call (02) 9773 0122.
What is surface roughness and why it matters
Every machined or finished surface has a three-component geometry: form (the macro shape — flatness, cylindricity, straightness), waviness (medium-scale undulations from machine tool deflections or vibration), and roughness (the fine micro-texture from cutting tool marks, grinding scratches, polishing lines or lapping patterns). Surface roughness is specifically the fine-scale component — the micro-geometry left by the actual material removal process.
Surface roughness controls a long list of engineering functions:
- Sealing. An O-ring, gasket or mechanical seal needs a specific surface roughness to seal reliably. Too rough and the seal leaks; too smooth and it can hydroplane or fail to grip. Typical sealing surface specs: Ra 0.4–1.6 µm for O-ring grooves, Ra 0.2–0.8 µm for mechanical seal faces.
- Sliding friction and wear. Bearing journals, hydraulic rod surfaces, slide ways and reciprocating shaft surfaces want low Ra with the right lay direction to retain lubrication. Honed cylinder bores have a cross-hatch lay specifically to hold oil.
- Bonded joint adhesion. Painted, plated, adhesive-bonded and welded surfaces want a specific Ra range — too smooth and the bond peels off; too rough and the bond is unreliable. Typical adhesive prep: Ra 1.6–3.2 µm.
- Fatigue life. Stress concentrations at peak surface features reduce fatigue strength. Critical highly-stressed components (turbine blades, suspension components, gear teeth) want low Ra to maximise fatigue life.
- Press fits and clamp loads. Bearing inner-race seats, shaft fits and clamped joints need controlled roughness for predictable interference and friction grip.
- Visual appearance. Cosmetic surfaces (consumer products, architectural metalwork, optical components) have surface finish requirements purely for appearance.
The first job in surface roughness work is reading the drawing callout to understand what's actually specified. The second is measuring the as-manufactured surface to verify it meets the spec. The third — often skipped — is understanding whether the manufacturing process can actually achieve the spec without secondary finishing. Specify Ra 0.4 µm on a turning operation and the machinist has to grind the surface after turning to meet it — doubling the cost. Get the specification right and the part is made for the lowest cost that meets function.
Surface roughness audience disambiguation
The search terms in this product area overlap with several unrelated topics. Scope-out:
| "Surface roughness" usage | What it means | This guide? |
|---|---|---|
| Engineering surface texture / Ra / Rz / profilometer | Workshop precision measurement of machined surface finish | Yes |
| GaN FWHM surface roughness / wafer X-ray roughness | Semiconductor research (academic) | No — specialist research instrumentation |
| Skin surface roughness | Dermatology / cosmetics | No — medical / wellness |
| Concrete surface roughness | Civil engineering (ASTM E965, sand patch test) | No — different methodology |
| Road surface roughness | Pavement engineering (IRI) | No |
| Pipe internal roughness for fluid flow (Moody chart) | Hydraulic engineering | Related concept but different measurement |
This guide is exclusively about engineering surface texture — the micro-finish on machined parts measured by profilometers, specified on drawings with Ra/Rz/Rsm callouts, and used by toolmakers, fitters, machinists, quality inspectors and engineers.
The Ra, Rz, Rsm parameters — what they actually mean
The surface roughness parameter family is large — the Mitutoyo Surftest SJ-410 measures 46 parameters conforming to ISO, DIN, ANSI and JIS standards. In practice, four parameters dominate workshop and inspection use: Ra, Rz, Rsm, and Rmax (now Rzx under ISO 21920-2). Understanding these four covers ~95% of all real-world drawing callouts.
Ra — the arithmetic mean deviation
Ra (arithmetic mean deviation, formerly "centre line average" or CLA) is the most common parameter on engineering drawings globally. Ra is calculated as the arithmetic average of the absolute values of the profile height deviations from the mean line over the evaluation length:
Ra = (1/L) × ∫ |y(x)| dx over the evaluation length L
In practical terms: imagine plotting the surface profile as a wavy line, drawing the mean line through it, and calculating the average distance from the line to the profile (both peaks above and valleys below counted as positive). That average distance is Ra.
Ra is the dominant parameter in North America and increasingly in international engineering. It's robust against outliers (a single deep scratch doesn't affect Ra much because it's averaged across the whole evaluation length) and gives a stable, repeatable value. Typical Ra values in workshop terms:
- Ra 12.5 µm: rough machined surface (sand cast, rough turning)
- Ra 6.3 µm: medium machining (general purpose turning, milling)
- Ra 3.2 µm: standard CNC milling / turning finish
- Ra 1.6 µm: smooth finish (fine turning, reaming)
- Ra 0.8 µm: ground finish (achievable with grinding using fine wheels)
- Ra 0.4 µm: very fine grinding
- Ra 0.2 µm or below: lapping, honing, polishing territory
Rz — the maximum height
Rz is the maximum height of the profile within the sampling length. Under ISO 21920-2:2021, Rz is the average of five Rz values calculated over five consecutive sampling lengths within the evaluation length — this averaging reduces the influence of outliers and gives a more stable Rz reading than a simple single-sampling-length calculation.
Rz is more sensitive than Ra to peak heights and deep valleys — it captures extreme features that Ra averages out. Rz is the dominant parameter on European drawings and on surfaces where peak height specifically matters (sealing surfaces, sliding interfaces, surfaces under stress that fatigue from peak concentrations).
Rz values are typically 4–6× the Ra value for the same surface — but this rule of thumb is a rough approximation only. Per the Practical Machinist forum consensus: "A precise conversion between Ra and Rz is impossible since they are two different properties. It's like asking, 'How do I convert height to weight?' You cannot convert height into weight, but you can make a rough estimation based on statistics." The 4–6× relationship varies with surface type, machining process and lay direction.
Rsm — mean spacing of profile peaks
Rsm (mean width of the profile elements) measures the average horizontal spacing of the surface features — the wavelength of the dominant texture. Rsm is critical for surfaces where the spacing of features matters as much as their height: rolling element bearing contact surfaces, gear tooth flanks, friction-coated surfaces.
Under ISO 21920-2:2021, Rsm uses a robust algorithm that filters out spurious short features — addressing a long-standing complaint that older Rsm calculations produced noisy values on real surfaces. Typical Rsm range: 50–200 µm for ground surfaces, 100–500 µm for milled surfaces.
Rmax / Rzx — maximum height extreme
Rmax (renamed to Rzx under ISO 21920-2) is the single largest peak-to-valley measurement found within the evaluation length. Where Rz averages five sample lengths, Rzx takes the worst-case single measurement — capturing the single deepest scratch or highest peak on the surface.
Rzx is the parameter to specify when you care about the worst feature on the surface, not the average. Sealing surfaces, fatigue-critical components and surfaces where a single defect can cause failure use Rzx specifications. Typical Rzx values are 10–30% higher than Rz for the same surface.
ISO 21920-2:2021 — the critical standards transition
The current international standard for profile-based surface roughness measurement is ISO 21920-2:2021 Geometrical product specifications (GPS) — Surface texture: Profile — Part 2: Terms, definitions and surface texture parameters. This standard replaces the previous ISO 4287:1997, ISO 13565-2:1996 and ISO 13565-3:1998 — all of which were withdrawn upon publication of ISO 21920-2.
Most published surface roughness articles still reference ISO 4287, even years after its withdrawal. This guide is intentionally current to the live standard. If you're working to drawings, calibration certificates or quality systems that still cite ISO 4287, the parameter definitions are largely the same — but the calculation method has changed for several parameters, and the change matters for verification.
What changed in ISO 21920-2 vs the withdrawn ISO 4287
| Parameter | ISO 4287:1997 (withdrawn) | ISO 21920-2:2021 (current) |
|---|---|---|
| Ra | Calculated once per sampling length, then averaged across 5 sampling lengths | Calculated once over the entire evaluation length |
| Rq (RMS) | Calculated once per sampling length, then averaged across 5 sampling lengths | Calculated once over the entire evaluation length |
| Rz | "Mean roughness depth" — average of 5 sampling-length Rz values | "Maximum height" — still averaged across 5 sampling lengths, definition refined |
| Rsm | Calculated with basic algorithm, often noisy on real surfaces | Calculated with robust algorithm that filters spurious short features |
| Rmax | Defined as single largest peak-to-valley | Replaced by Rzx (maximum height single) |
| New parameters added | — | Rpt (max peak), Rvt (max valley), Rzx (max height single) |
The key practical implication: Ra and Rq values calculated under ISO 21920-2 may differ slightly from values calculated under ISO 4287 on the same physical surface, because the calculation now spans the full evaluation length instead of being averaged across sampling lengths. For most workshop measurements the difference is below 1% and within instrument repeatability. For precision metrology where small Ra differences matter, the standard reference matters — specify ISO 21920-2:2021 on new drawings, and verify which standard a calibration certificate references when traceability matters.
Why the change happened
The ISO 21920 series was developed to consolidate the fragmented older standards (ISO 4287, 13565-2, 13565-3) into a single consistent framework, address known calculation inconsistencies, add the parameters needed for modern surface engineering (additive manufacturing surfaces, biomedical implants, ultra-fine optical surfaces), and align profile-based measurement (ISO 21920) with areal/3D measurement (ISO 25178). The transition was several years in the making and most NATA-accredited calibration laboratories in Australia now certify to ISO 21920-2.
Surface roughness drawing symbols — the callout decoded
Reading a surface roughness callout on an engineering drawing requires understanding three historical conventions plus the current ISO standard. Different drawings from different eras use different conventions:
The current ISO 1302 / ISO 21920-1 symbol
The modern surface texture callout uses a check-mark style symbol with up to four modifying positions:
- Top left position: the parameter and value (e.g. Ra 1.6, Rz 6.3)
- Top right position: any manufacturing process specification (e.g. "milled", "ground")
- Bottom left position: lay direction symbol (= for parallel, ⊥ for perpendicular, X for crossed, M for multi-directional, R for radial, C for circular, P for non-directional/isotropic)
- Underneath the symbol: machining allowance (in millimetres)
An open check-mark (without a bar across the top) means "surface roughness as specified" with no manufacturing constraint. A check-mark with a horizontal bar means "no material removal permitted" (the surface must come from the original casting or forming process). A check-mark with a small circle at the apex means "material removal required" (a finishing operation must be performed).
The historical triangle symbols (still found on older drawings)
Before the current symbol system, surface finish was indicated on drawings with downward-pointing triangles (▽), with the number of triangles indicating the finish grade. This system is technically obsolete (superseded by BS EN ISO 1302:2002) but is still seen on older drawings, in textbooks, and in workshops that haven't updated their drawing standards.
| Triangle symbol | N grade equivalent | Approximate Ra (µm) | Achievable by |
|---|---|---|---|
| 1 triangle (▽) | N9–N10 | 6.3–12.5 | Rough machining (turning, milling) |
| 2 triangles (▽▽) | N7–N8 | 1.6–3.2 | Medium machining (finish turning, finish milling) |
| 3 triangles (▽▽▽) | N4–N6 | 0.2–0.8 | Fine machining, grinding |
| 4 triangles (▽▽▽▽) | N1–N3 | 0.025–0.1 | Lapping, honing, polishing |
The N-grade system
The N-grade system (N1 through N12) was an intermediate convention between the triangle symbols and direct Ra values. N grades 1–12 map approximately to Ra ranges in a logarithmic scale. The system is also technically obsolete but still appears on older drawings. ISO 1302 has effectively replaced it with direct Ra value specification.
| N grade | Ra (µm) | Ra (µin) | RMS (µin) |
|---|---|---|---|
| N1 | 0.025 | 1 | 1.1 |
| N2 | 0.05 | 2 | 2.2 |
| N3 | 0.1 | 4 | 4.4 |
| N4 | 0.2 | 8 | 8.8 |
| N5 | 0.4 | 16 | 17.6 |
| N6 | 0.8 | 32 | 35.2 |
| N7 | 1.6 | 63 | 69 |
| N8 | 3.2 | 125 | 138 |
| N9 | 6.3 | 250 | 275 |
| N10 | 12.5 | 500 | 550 |
| N11 | 25 | 1000 | 1100 |
| N12 | 50 | 2000 | 2200 |
The RMS vs Ra distinction
Older American drawings often specify surface finish in RMS µin (root mean square microinches). RMS is the same as Rq in the current ISO standard — the root mean square of the profile deviations rather than the arithmetic mean. RMS values are approximately 11% higher than Ra values for the same surface (RMS = Ra × 1.111 approximately). A drawing specifying "32 RMS" is approximately equivalent to "29 Ra µin" or "0.73 Ra µm" — close to N6.
The 11% rule of thumb is an approximation for normal (Gaussian) surface distributions. For machined surfaces this is generally a good approximation; for surfaces with extreme peaks or asymmetric distributions (e.g. honed cylinder bores with deep valleys), the ratio can differ.
Stylus (contact) vs optical (non-contact) profilometers
Surface roughness measurement instruments split into two main families based on how they sense the surface:
| Property | Stylus (contact) profilometer | Optical (non-contact) profilometer |
|---|---|---|
| Sensing mechanism | Diamond stylus dragged across surface; transducer measures vertical displacement | Laser, white light interferometry, focus variation or confocal optics |
| Resolution | 0.001 µm (1 nm) achievable | 0.1 nm achievable |
| Surface contact | Yes — light contact with diamond tip | No contact — can measure soft surfaces, fragile coatings |
| Lateral resolution | Limited by stylus tip radius (2 or 5 µm) | Sub-micron (optical limit) |
| Vertical range | Hundreds of µm to several mm | Tens to hundreds of µm typical |
| Speed | Slow (mechanical traverse) | Fast (no mechanical traverse needed) |
| 2D vs 3D | Single profile line (2D) | Can map full surface area (3D areal) |
| Standard compliance | ISO 21920-2 / ISO 4287 (profile) | ISO 25178 (areal) & ISO 21920-2 |
| Cost (relative) | Workshop affordable | Significantly higher |
| Best for | Workshop QC, drawing callout verification, in-process inspection | R&D, ultra-fine optical surfaces, micro-electronic surfaces, soft materials |
For workshop, inspection and quality control work in Australian industry, the contact stylus profilometer is the standard choice. The Mitutoyo Surftest range covers this application with the SJ-210/310/410 portable and Formtracer SV-C4500 stationary models. Optical profilometers (Mitutoyo offers them at the laboratory level) are specialist instruments for research and ultra-precision work.
Mitutoyo Surftest SJ-210, SJ-310 and SJ-410 portable range
Mitutoyo's Surftest portable range comprises three product families covering everything from workshop spot-check measurement up to premium laboratory-grade portable testing.
Surftest SJ-210 — the entry-level portable
The Surftest SJ-210 is Mitutoyo's entry portable surface roughness tester. Key specifications:
- Measuring parameters: Ra, Ry, Rz, Rq, Rp, Rv, R3z, Rt, Rmax, Sm, Rk, Rpk, Rvk, Mr1, Mr2 (approximately 20 parameters)
- Measuring range (Z axis): 360 µm (-200 µm to +160 µm)
- Measuring range (X axis): 17.5 mm standard, 5.6 mm with retractable drive unit
- Resolution: Variable depending on range — 0.006 µm on the most sensitive range
- Stylus: 2 µm or 5 µm diamond, 60° or 90° cone angle
- Measuring force: <400 mN with skid; <0.75 mN without
- Measuring speed: 0.25, 0.5, or 0.75 mm/s
- Display: Colour LCD with profile and parameter display
- Data output: USB, RS-232C, SPC, Bluetooth (optional)
- Standards compliance: ISO 1997, JIS B 0601-2001, JIS B 0601-1994, JIS B 0601-1982, ANSI, VDA
The SJ-210 is the right choice for workshop QC where the operator needs to verify Ra/Rz against drawing callouts, spot-check finishes after machining, or run periodic compliance checks. The compact size makes it usable handheld or on a column stand.
Surftest SJ-310 — the mid-range portable
The SJ-310 builds on the SJ-210 platform with a larger graphical display, expanded parameter set, integrated printer (some models), and enhanced data analysis capability:
- Measuring parameters: Approximately 30 parameters including the SJ-210 set plus additional Rk-family parameters
- Display: 5.7-inch colour touch panel LCD with full graphical profile display
- Integrated thermal printer on some models
- Profile storage: Multiple measurement profiles stored internally with batch export capability
- Same stylus, range and resolution options as SJ-210
- Standards compliance: Full ISO/JIS/ANSI/DIN compliance with parameter calculation per ISO 21920-2 where supported by firmware update
The SJ-310 is the right choice for inspection departments and quality control labs that need full graphical profile review, batch measurement workflows, or integrated print output for inspection reports.
Surftest SJ-410 — the premium portable
The SJ-410 is Mitutoyo's premium portable surface roughness tester — the most capable handheld profilometer in the Mitutoyo range:
- Measuring parameters: 46 parameters conforming to ISO 1997, JIS B 0601-2001, JIS B 0601-1994, JIS B 0601-1982, ANSI, VDA
- Three measuring ranges: 800 µm / 80 µm / 8 µm (with corresponding resolution 0.05 / 0.005 / 0.0005 µm)
- Optional extended stylus: 2,400 µm range
- Resolution: 0.001 µm on the standard range
-
Two model variants:
- SJ-411: Skidded measurement only
- SJ-412: Both skidded and skid-less measurement
- 11 interchangeable detector tips (introduced in SJ-310 / SJ-410 generation) for specialty applications
- Measuring length: Up to 25 mm standard, up to 50 mm with extension
- Display: 5.7-inch colour touch panel LCD with full graphical profile, multi-point measurement, data analysis
- Data output: USB, RS-232C, SPC, Bluetooth, ethernet
- Optional column stand with manual height adjustment for fixturing applications
The SJ-410 is the right choice for serious inspection, calibration laboratories, and any application where the 46-parameter set, skidded vs skid-less measurement flexibility, or 0.001 µm resolution are required. The premium price point reflects the breadth of capability.
Skidded vs skid-less measurement — the SJ-411 vs SJ-412 decision
One of the more subtle decisions in surface roughness measurement is skidded vs skid-less. Both methods are valid; they measure differently:
- Skidded measurement: The stylus pickup slides on a small "skid" that rests on the surface immediately adjacent to the measurement point. The skid filters out long-wavelength surface form variation (waviness, low-frequency undulations). The output is "pure" roughness within the bandpass of the skid filter. This is the standard workshop measurement method and is appropriate for the vast majority of machined surfaces.
- Skid-less measurement: The stylus pickup is supported on a precision reference (rather than a skid), and the pickup measures absolute height. Skid-less measurement captures both roughness and waviness, allowing post-measurement filter selection to separate them. Skid-less is required for very fine surfaces (where the skid would mask the fine micro-structure), short surfaces where 5 sampling lengths can't be measured, or applications where roughness and waviness need to be separately characterised.
For typical workshop QC work, skidded measurement (SJ-411) is adequate and faster. For research, ultra-fine finishes, calibration laboratory work, and any application requiring post-measurement filter analysis, skid-less (SJ-412) is needed.
Mitutoyo Formtracer SV-C4500 stationary
For laboratory-grade and high-throughput inspection, Mitutoyo offers stationary surface roughness measurement systems. The Formtracer SV-C4500 series combines surface roughness measurement with contour (form) measurement in a single instrument — allowing simultaneous characterisation of micro-roughness and macro-form on the same part. Key applications: precision bore measurement, turbine blade airfoil profiling, gear tooth flank measurement, ball bearing race characterisation.
The Formtracer SV-C4500 is a specialist capital instrument typically specified for QC laboratories, R&D departments and calibration service providers. AIMS quotes and supplies the Formtracer SV-C4500 range through the authorised Mitutoyo distributor for AU customers requiring this level of capability.
Stylus selection — 2µm vs 5µm tip radius
The diamond stylus tip on a contact profilometer has a precisely-controlled radius. Two standard options dominate workshop and laboratory work:
| Stylus tip radius | Best for Ra range | Stylus characteristics |
|---|---|---|
| 2 µm | Ra < 0.5 µm (fine finishes, polished surfaces, optical) | Higher lateral resolution. More delicate — easily damaged on rough surfaces. Required for accurate measurement of fine surfaces. |
| 5 µm | Ra 0.5 µm and up (most workshop machining) | Standard workshop option. Robust against rough surface use. Slightly lower lateral resolution but adequate for most engineering surfaces. |
| 10 µm | Ra > 1.6 µm (rough surfaces only) | Specialty option for very rough or cast surfaces where the 2/5 µm tips would catch on features. |
Per the Practical Machinist forum consensus: "A 5-micron stylus is acceptable for Ra 0.5 microns and up, while finer finishes would benefit from a 2-micron stylus, though it's more delicate." The practical workshop kit: one 5 µm stylus for routine QC, one 2 µm stylus reserved for precision work on fine finishes.
Stylus condition matters as much as tip radius. A worn or damaged stylus reads finer than reality (the worn tip can't trace into deep valleys). Inspect the stylus visually before precision use, and verify against a calibration reference specimen if results look suspicious.
Cutoff length and evaluation length — the filter selection that affects every reading
Surface roughness measurement requires choosing a cutoff length (also called λc or sampling length) that determines what's "roughness" versus what's "waviness". The cutoff is a high-pass filter applied to the measured profile — wavelengths longer than λc are filtered out as waviness; wavelengths shorter than λc are retained as roughness.
The standard cutoff selections per ISO 4288:1996 (still current under ISO 21920-2 framework):
| Cutoff (λc) | Sampling length | Total evaluation length | Best for Ra range |
|---|---|---|---|
| 0.08 mm | 0.08 mm | 0.4 mm (5 sampling lengths) | Ra < 0.02 µm (ultra-fine surfaces) |
| 0.25 mm | 0.25 mm | 1.25 mm | Ra 0.02–0.1 µm (fine polished, lapped) |
| 0.8 mm | 0.8 mm | 4 mm (standard workshop) | Ra 0.1–2 µm (most machined surfaces) |
| 2.5 mm | 2.5 mm | 12.5 mm | Ra 2–10 µm (rough turning, milling) |
| 8.0 mm | 8.0 mm | 40 mm | Ra > 10 µm (very rough surfaces, castings) |
The default workshop cutoff is 0.8 mm with 4 mm evaluation length (5 × 0.8 mm sampling lengths). This works for the vast majority of machined surfaces with Ra in the 0.1–2 µm range. Going to a non-default cutoff is necessary when the surface texture wavelength is outside this default's design point — very fine polished surfaces need shorter cutoff to capture the dominant wavelength; very rough cast surfaces need longer cutoff to span enough feature wavelengths.
The "short surface" trap
Per industry source Digital Metrology: "One dangerous approach for dealing with short surfaces is to choose a shorter filter cutoff to achieve the 5 sampling lengths on the surface, as changing the cutoff wavelength changes the definition of 'roughness'." Translation: if the surface is too short for the default 4 mm evaluation length, the temptation is to switch to 0.25 mm cutoff (1.25 mm evaluation) to fit the surface. But this changes what you're measuring — you're now reporting finer features as roughness that would be waviness under the standard cutoff. The Ra value will be different (typically lower) than the same surface measured at the standard 0.8 mm cutoff.
The correct approach for short surfaces: use skid-less measurement (which doesn't require 5 sampling lengths) and report the actual measurement conditions on the inspection record, or use a stationary profilometer like the Formtracer that can handle shorter evaluation lengths properly. Don't fake compliance by switching cutoffs.
Surface roughness by machining process — the Ra targets table
Different manufacturing processes produce characteristic Ra ranges. Specifying a finer Ra than the process can achieve forces a secondary finishing operation, doubling cost. Specifying a coarser Ra than the process produces is functionally fine but doesn't reduce cost (you can't go finer than the process; you can only allow rougher).
| Manufacturing process | Typical Ra range (µm) | Achievable Ra (µm, best practice) | Notes |
|---|---|---|---|
| Sand casting | 12.5–25 | 6.3 | As-cast surface, no machining |
| Die casting | 1.6–3.2 | 0.8 | Much smoother than sand casting |
| Hot rolling | 12.5–25 | 12.5 | Heavy oxide/scale layer typical |
| Cold drawing (bar stock) | 1.6–6.3 | 1.6 | Per Practical Machinist consensus, "125 Ra" is the cold-drawn baseline |
| Sawing | 3.2–25 | 3.2 | Depends heavily on saw type and feed |
| Drilling | 1.6–6.3 | 1.6 | Fine drills + lubricant for best |
| Reaming | 0.8–3.2 | 0.8 | Standard finish reamer with proper technique |
| Turning (rough) | 3.2–12.5 | 3.2 | Sharp tool, moderate feed |
| Turning (finish) | 0.8–3.2 | 0.4 | Sharp insert, low feed, fine nose radius |
| Milling (rough) | 3.2–6.3 | 1.6 | End mill or face mill |
| Milling (finish) | 0.8–3.2 | 0.4 | Sharp tool, light DOC, fine feed |
| Broaching | 0.8–3.2 | 0.4 | Sharp broach with proper coolant |
| Grinding (cylindrical) | 0.1–1.6 | 0.05 | Fine wheel grade, proper dressing |
| Grinding (surface) | 0.2–1.6 | 0.1 | Similar to cylindrical |
| Honing | 0.05–0.4 | 0.025 | Cross-hatch lay typical on cylinder bores |
| Lapping | 0.025–0.4 | 0.012 | Loose abrasive, very fine finish |
| Polishing | 0.012–0.2 | 0.005 | Fixed abrasive, mirror finish |
| Superfinishing | 0.012–0.1 | 0.005 | Bonded abrasive stones, light pressure |
| EDM (rough) | 3.2–6.3 | 1.6 | Recast layer present |
| EDM (finish) | 0.4–1.6 | 0.2 | Multiple progressively-finer passes |
The Practical Machinist forum consensus on the threshold: "A 32 finish starts to get into grinding/polishing, but can be done on some machines with the right tooling and feeds/speeds. Generally a 125 finish is the way material comes cold drawn (Bar Stock)." Translated: Ra 0.8 µm (32 µin) is the sensible upper limit for production turning/milling without secondary finishing. Anything finer than that — Ra 0.4 µm and below — usually requires grinding or lapping as a follow-up operation.
Ra ↔ Rz, Ra ↔ RMS, and Ra ↔ N-grade conversions
The most-asked workshop question on surface roughness: "How do I convert Ra to Rz?" The honest answer: you can't, not precisely. Per the Practical Machinist forum: "A precise conversion between Ra and Rz is impossible since they are two different properties. It's like asking, 'How do I convert height to weight?'"
Approximate conversion relationships hold for typical machined surfaces but break down for unusual surface types:
| Conversion | Approximation | Caveat |
|---|---|---|
| Ra → Rz | Rz ≈ 4 to 6 × Ra | Wide variation depending on surface type and lay; 5 × is the rough workshop mean |
| Ra → RMS | RMS ≈ 1.11 × Ra | Valid for normal (Gaussian) distributions; breaks down on asymmetric surfaces |
| Ra → Rmax/Rzx | Rzx ≈ 5 to 8 × Ra | Single largest peak-to-valley feature; high variability |
| Ra → N-grade | See N-grade table above | Direct lookup; standardised |
| Ra (µm) → Ra (µin) | Ra µin = Ra µm × 39.37 | Exact unit conversion (1 µm = 39.37 µin) |
The practical rule when working between drawings using different conventions: measure and report in the parameter and unit the drawing specifies. If the drawing says "Ra 1.6 µm" measure and report Ra in µm. If the drawing says "32 RMS µin" measure and report Rq in µin (which is RMS). Converting between parameters is unreliable and creates traceability problems on inspection records.
Lay direction — the often-ignored part of surface specification
Lay direction is the dominant pattern of the surface texture — the way the cutting tool marks, grinding scratches or polishing lines are oriented. Surface roughness measurement readings can vary by 2–3 times depending on the direction of stylus traverse relative to the lay direction.
| Lay symbol | Pattern | Typical produced by |
|---|---|---|
| = (equal sign) | Parallel to projection plane | Shaping, planing, end-cut turning |
| ⊥ (perpendicular) | Perpendicular to projection plane | Shaping perpendicular to surface, slotted milling |
| X (cross) | Crossed at angle to projection plane | Honing, cross-cut grinding |
| M (multi) | Multi-directional | Cutter relief, fly cutting, electrochemical machining |
| C (circular) | Circular (concentric) | Face turning, end-cut on lathe, face grinding |
| R (radial) | Radial (from a centre) | Burr removal patterns, rotary brushing |
| P (particulate / non-directional) | Isotropic, no preferred direction | Sand blasting, shot peening, EDM, lapping, electro-polishing |
Critical practical rule: measure perpendicular to the lay direction for the most representative reading. If you measure parallel to the lay (along the grinding scratches, for example), you'll read a much smoother surface than is actually there. A cross-hatched honed cylinder bore is a particularly tricky case — the measured Ra varies dramatically depending on whether you traverse perpendicular to one hatch direction, perpendicular to the other, or at the bisector angle. For these surfaces, multiple measurements at different angles are typically taken and reported together.
Common drawing callout examples
| Drawing callout | What it means | Typical use |
|---|---|---|
| Ra 12.5 | Ra not to exceed 12.5 µm | Rough machining, structural surfaces, non-functional |
| Ra 6.3 | Ra not to exceed 6.3 µm | General machining, internal non-critical surfaces |
| Ra 3.2 | Ra not to exceed 3.2 µm | Standard machined surface, most common callout |
| Ra 1.6 | Ra not to exceed 1.6 µm | Smoother machined surface, mating surfaces, light-duty bearing seats |
| Ra 0.8 | Ra not to exceed 0.8 µm | Ground or fine-machined surface, sliding surfaces, sealing surfaces |
| Ra 0.4 | Ra not to exceed 0.4 µm | Very fine ground surface, precision bearing seats, hydraulic surfaces |
| Ra 0.2 | Ra not to exceed 0.2 µm | Lapped or honed, high-precision sealing, optical-grade |
| Ra 0.4 max / Ra 0.2 min | Within band 0.2–0.4 µm | Coating adhesion surfaces (too smooth fails, too rough fails) |
| Rz 6.3 / cross-hatch lay | Rz max 6.3 with cross-hatch lay pattern | Honed cylinder bore for IC engine |
Calibration via reference specimens — the traceability chain
A surface roughness tester is meaningless without traceable calibration. The reference standard for calibration is a surface roughness reference specimen — a precision-manufactured surface with known Ra and Rz values, certified by a national metrology institute (NIST in the US, PTB in Germany, NMIA in Australia, NMIJ in Japan).
The Mitutoyo precision reference standard product line includes:
- Sinusoidal reference specimens: Repeating sinusoidal surface texture at certified Ra values
- Random profile reference specimens: Statistically representative of ground / lapped / polished / honed surfaces, at certified Ra values
- Step height reference specimens: For verifying the vertical calibration of the stylus pickup
- Multi-patch reference specimens: Multiple Ra values on a single specimen for fast multi-point check
The calibration procedure: place the stylus on the reference specimen patch, run the standard measurement cycle, compare the measured Ra/Rz to the certified values from the specimen's calibration certificate. If the readings agree within tolerance, the tester is calibrated. If they disagree, the tester needs servicing.
NIST is the US national reference for surface roughness calibration: "Parameters of surface roughness and step height are currently measured at the National Institute of Standards and Technology (NIST) by means of a computerized/stylus instrument." Australia's NMIA provides equivalent national traceability, and NATA-accredited calibration laboratories can certify surface roughness reference specimens with full traceability to NMIA primary standards.
Common surface roughness measurement mistakes
| Mistake | What goes wrong | Fix |
|---|---|---|
| Measuring parallel to lay direction | Reading is 2–3× smoother than actual perpendicular measurement | Always measure perpendicular to dominant lay; for cross-hatched, measure at multiple angles |
| Using wrong cutoff λc | Reading misses or includes wrong feature wavelengths | Match cutoff to expected Ra per ISO 4288 / 21920 selection table |
| Forcing short surface into shorter cutoff | Changes the definition of roughness; readings not comparable | Use skid-less measurement; or move to stationary instrument; document conditions |
| Dirty or worn stylus | Reads finer than reality; misses deep valleys | Inspect stylus before precision use; verify against reference specimen |
| Specifying Ra 0.4 on a turning operation | Forces secondary finishing; doubles cost | Match drawing callout to achievable process Ra; specify finer only when functionally required |
| Skipping calibration verification | Tester may drift; readings invalid for traceability | Annual calibration; verify against reference specimen before precision work |
| Converting Ra to Rz arbitrarily | Wide variation; conversion inherently imprecise | Measure the parameter specified on the drawing; don't convert |
| Not specifying lay direction on critical surfaces | Manufacturing process can give wrong lay; verification reads differently | Always specify lay symbol for sealing, sliding and bonded surfaces |
Brand reality — Mitutoyo, Mahr, Taylor Hobson, Zeiss Surfcom
The surface roughness measurement industry is dominated by a handful of premium brands plus a tier of budget alternatives:
| Brand | Origin | Reputation | AU availability |
|---|---|---|---|
| Mitutoyo Surftest / Formtracer | Japan | Global benchmark for portable. Wide range from entry SJ-210 to laboratory Formtracer. Strong AU distributor network. | Authorised AU distributor (AIMS supply channel) |
| Mahr MarSurf | Germany | German premium tier. Common in European OEM workshops and calibration laboratories. | Specialist import via Mahr Australia |
| Taylor Hobson Form Talysurf | UK / USA | Reference-laboratory benchmark for stationary profilometers. The historical pioneer of surface roughness measurement. | Specialist import (AMETEK group) |
| Zeiss Surfcom | Germany (acquired from Accretech) | Premium stationary and portable systems, common in automotive R&D. | Specialist import via Zeiss Australia |
| Hommel Etamic | Germany (Jenoptik) | European premium tier, common in automotive QC. | Specialist import |
| Bruker / Veeco | USA | Optical / interferometric profilometers for R&D and ultra-fine surfaces. | Specialist import |
| Insize, Bowers, others (budget tier) | China / Taiwan / mixed | Entry-level portable testers at significantly lower price points. Build quality and long-term reliability typically below the premium tier. | Available through general industrial supply |
For Australian workshops and inspection departments, Mitutoyo Surftest is the dominant portable brand — reflecting both Mitutoyo's broad AU distributor and calibration support infrastructure and the well-engineered breadth of the Surftest range from entry workshop tester through to premium SJ-410 and laboratory Formtracer. For specialised applications — high-end optical metrology, automotive R&D, reference calibration — Mahr, Taylor Hobson and Zeiss Surfcom are the premium alternatives but require specialist import.
Counterfeit Mitutoyo Surftest — how to spot the fake
Surftest profilometers are higher-value capital items and don't see the same counterfeiting volume as Mitutoyo calipers or micrometers, but counterfeit and grey-market units do appear — particularly used SJ-210 units listed below normal market pricing on online marketplaces. The five-check procedure for Mitutoyo Surftest authenticity:
- Verify serial number through authorised distributor. Genuine Mitutoyo Surftest serial numbers can be confirmed through the authorised Mitutoyo Australian distributor. Counterfeit and grey-market units often can't be verified.
- Inspect the LCD display and touch panel quality. Genuine Mitutoyo touch panels have consistent backlighting, sharp text rendering, and respond reliably to touch. Counterfeit units sometimes use lower-grade display panels with patchy backlight or laggy touch response.
- Check the stylus pickup assembly. Genuine Mitutoyo styluses have precision-mounted diamond tips with consistent tip radius (verifiable against reference specimens). Counterfeit replacement styluses may have inconsistent tip geometry.
- Verify firmware version and ISO compliance. Genuine units support firmware updates to current standards (ISO 21920-2). Older counterfeit units may have firmware locked to obsolete standards.
- Check the calibration certificate. Genuine Mitutoyo units ship with an inspection certificate traceable to NKO/JCSS/NMIJ. Counterfeit units come without certificates or with photocopied generic certificates.
AEO note: counterfeit Mitutoyo branding sometimes appears under deliberate misspellings — Mitutogo, MITU-tyo, Mituttoyo, Mito_tuyo — designed to bypass keyword filters on online marketplaces. Genuine Mitutoyo is always spelled Mitutoyo, capitalised, with no hyphens. Buying through an authorised distributor like AIMS provides supply chain verification that eliminates the counterfeit risk.
Mitutoyo Surftest supply through AIMS
AIMS is an authorised supply channel for the full Mitutoyo precision measurement range in Australia. Surface roughness testers are quote-and-supply items rather than online stock items because of the capital value, training and calibration considerations involved.
What we quote and supply through the authorised Mitutoyo Australian distributor:
- Mitutoyo Surftest SJ-210 series — entry portable for workshop QC
- Mitutoyo Surftest SJ-310 series — mid-range portable with graphical display and integrated printer options
- Mitutoyo Surftest SJ-410 series (SJ-411 / SJ-412) — premium portable with 46 parameters and skidded/skid-less options
- Mitutoyo Formtracer SV-C4500 series — stationary roughness + contour systems for laboratory and high-throughput inspection
- Detector and stylus accessories — 2 µm and 5 µm diamond styluses, specialty stylus geometries, replacement detector tips
- Reference specimens — Mitutoyo precision reference standards with NKO/JCSS calibration certificates for tester verification
- Column stands — manual height-adjustable stands for fixturing the portable SJ-210/310/410 in workshop or inspection bench setups
- NATA-accredited calibration service coordinated through Mitutoyo Australia or qualifying third-party labs
- Operator training available through the Mitutoyo Australian distributor for workshops new to surface roughness measurement
For workshops considering their first surface roughness tester investment, the typical starter specification questions: (1) what's the Ra range you need to measure (drives stylus selection and tester capability); (2) what standards does your quality system require (drives SJ-210 vs SJ-310 vs SJ-410 capability); (3) what's the production volume / measurement frequency (drives portable vs stationary); (4) what's the size and shape of typical workpieces (drives column stand and fixturing). Contact AIMS with the answers and we'll put together a Mitutoyo Surftest quote covering tester, stylus accessories, reference specimens, calibration certificate and operator training as required.
Frequently Asked Questions
What is surface roughness and how is it measured?
Surface roughness is the fine micro-geometry of an engineering surface — the texture of peaks, valleys and lay direction left by the manufacturing process. It's measured with a profilometer (also called a surface roughness tester) which traces a diamond stylus across the surface (contact method) or uses optical sensing (non-contact method) to measure the vertical deviations from a mean line, then calculates statistical parameters like Ra (arithmetic mean) and Rz (maximum height). The Mitutoyo Surftest SJ-210/310/410 portable range and Formtracer SV-C4500 stationary range cover most workshop and laboratory applications. NATA-accredited calibration via traceable reference specimens establishes the measurement chain back to national standards.
What's the difference between Ra and Rz?
Ra is the arithmetic mean deviation — the average distance from the surface profile to the mean line, taken across the evaluation length. Rz is the maximum height — the average of the largest five peak-to-valley measurements within five sampling lengths. Ra is the dominant parameter in North America and on most modern drawings; Rz is more common on European drawings and on surfaces where peak height specifically matters (sealing surfaces, sliding interfaces, fatigue-critical surfaces). Rz values are typically 4–6× the Ra value for the same surface, but precise conversion between them is impossible — they're two different statistical properties of the same surface.
What is ISO 21920-2:2021 and how does it differ from ISO 4287?
ISO 21920-2:2021 is the current international standard for profile-based surface roughness measurement. It replaced the older ISO 4287:1997 (along with ISO 13565-2 and ISO 13565-3) when it was published. The key changes: Ra and Rq are now calculated once over the entire evaluation length rather than averaged across 5 sampling lengths; Rsm now uses a robust algorithm that filters spurious short features; Rmax was replaced by the new parameter Rzx; new parameters Rpt (max peak) and Rvt (max valley) were added. For most workshop measurements the numerical difference is small, but for precision metrology and traceability, the standard reference matters. Specify ISO 21920-2 on new drawings.
What Ra values can I achieve from different machining processes?
Typical Ra ranges by process: sand casting 12.5–25 µm, cold-drawn bar stock 1.6 µm baseline (the "125 Ra" workshop reference per Practical Machinist), rough turning/milling 3.2–6.3 µm, finish turning/milling 0.8–3.2 µm (achievable to 0.4 µm with sharp tools and fine feeds), grinding 0.1–1.6 µm, honing 0.05–0.4 µm, lapping 0.025–0.4 µm, polishing 0.012–0.2 µm. The 32 Ra threshold (0.8 µm) is where grinding/polishing territory begins per machinist forum consensus. Specifying Ra finer than the process can achieve forces secondary finishing.
Can I convert between Ra and Rz precisely?
No. Per the Practical Machinist forum consensus: "A precise conversion between Ra and Rz is impossible since they are two different properties. It's like asking, 'How do I convert height to weight?'" The rough rule of thumb is Rz ≈ 5× Ra, with a typical range of 4–6×, but the actual ratio varies with surface type, machining process and lay direction. The practical rule: measure and report the parameter the drawing specifies. Don't convert between parameters on inspection records — it creates traceability problems.
What's the difference between Ra and RMS surface finish?
Ra is the arithmetic mean deviation; RMS is the root mean square deviation (the same as Rq in the current ISO standard). For typical machined surfaces with normal (Gaussian) profile distributions, RMS ≈ 1.11 × Ra — RMS values are about 11% higher than Ra. Older American drawings often specify in "RMS µin" while modern drawings use "Ra µm" or "Ra µin". A drawing specifying "32 RMS" is approximately equivalent to "29 Ra µin" or "0.73 Ra µm" — close to N6. The 11% rule of thumb is approximate for normal distributions; it breaks down on surfaces with extreme peaks or asymmetric distributions.
How do I read a surface roughness callout on an engineering drawing?
The modern ISO 1302 / ISO 21920-1 callout uses a check-mark style symbol with the parameter value (e.g. "Ra 1.6") at the top left, manufacturing process specification at top right, lay direction symbol at bottom left (= parallel, ⊥ perpendicular, X crossed, M multi, C circular, R radial, P non-directional), and machining allowance below. An open check-mark allows any manufacturing method; a check-mark with a horizontal bar means no material removal allowed; a check-mark with a small circle means material removal required. Historical conventions still appear on older drawings: triangle symbols (1▽ = N9-N10 ≈ Ra 6.3-12.5, 2▽ = N7-N8 ≈ Ra 1.6-3.2, 3▽ = N4-N6 ≈ Ra 0.2-0.8, 4▽ = N1-N3 ≈ Ra 0.025-0.1).
What's the difference between a stylus and optical profilometer?
A stylus (contact) profilometer drags a fine diamond stylus across the surface, measuring vertical displacement of the stylus tip via a transducer. It's the workshop standard — Mitutoyo Surftest SJ-210/310/410 are stylus profilometers. An optical (non-contact) profilometer uses laser, white light interferometry, focus variation or confocal optics to measure surface height without touching it. Optical can achieve sub-nanometre resolution, measure soft or fragile surfaces, and produce full 3D areal maps (per ISO 25178). Contact stylus is appropriate for routine workshop QC of machined surfaces; optical is appropriate for R&D, ultra-fine optical surfaces, micro-electronics, biomedical implants, and any soft material where stylus contact would damage the surface.
Should I buy the Mitutoyo Surftest SJ-210, SJ-310 or SJ-410?
SJ-210 for workshop QC where Ra/Rz spot-check against drawing callouts is the primary need — adequate parameters, compact handheld form factor, basic data output. SJ-310 for inspection departments needing the graphical 5.7-inch touch panel display, expanded parameter set, batch measurement workflow, and integrated thermal printer (some models). SJ-410 (specifically the SJ-412 with skidded/skid-less measurement option) for serious inspection, calibration laboratories, research environments — 46 parameters, three measuring ranges including 0.001 µm resolution, 11 interchangeable detector tips, full ISO/JIS/ANSI/VDA compliance. Match capability to application: don't pay for SJ-410 features you'll never use, but don't specify SJ-210 if your quality system requires the full ISO 21920-2 parameter set.
What does Ra 3.2, Ra 1.6 or Ra 0.8 typically mean?
Ra 3.2 µm is the standard machined surface — achievable by standard CNC milling or turning with good practice; sufficient for most internal mechanical components, structural surfaces, non-critical assemblies. Ra 1.6 µm is a smoother machined finish — fine turning, finish reaming, light-duty bearing seats, mating surfaces. Ra 0.8 µm is the threshold where grinding territory begins per machinist forum consensus — achievable by fine machining or grinding, suitable for sliding surfaces, sealing surfaces, precision fits. Specifying finer than Ra 0.8 µm typically forces grinding or lapping as a secondary operation, doubling cost.
What's the difference between skidded and skid-less measurement?
Skidded measurement uses a small "skid" on the stylus pickup that rests on the surface adjacent to the measurement point. The skid filters out long-wavelength waviness, giving "pure" roughness within the bandpass of the skid filter. This is the standard workshop method and is appropriate for most machined surfaces. Skid-less measurement uses a precision reference (rather than a skid), measuring absolute height. Skid-less captures both roughness and waviness, allowing post-measurement filter selection. Skid-less is required for very fine surfaces, short surfaces where 5 sampling lengths can't be measured, or applications needing separate roughness and waviness characterisation. The Mitutoyo SJ-411 is skidded only; SJ-412 supports both.
What stylus tip radius should I use?
The 5 µm tip is the standard workshop option, adequate for Ra 0.5 µm and up — robust against rough surface use, appropriate for most machined surfaces. The 2 µm tip is required for finer finishes (Ra below 0.5 µm) — higher lateral resolution captures fine surface features the 5 µm tip would average out, but the smaller tip is more delicate and easily damaged. Per the Practical Machinist forum consensus: "A 5-micron stylus is acceptable for Ra 0.5 microns and up, while finer finishes would benefit from a 2-micron stylus, though it's more delicate." The practical workshop kit: one 5 µm stylus for routine QC, one 2 µm stylus reserved for precision work on fine finishes.
What's the cutoff length (λc) and how do I select it?
The cutoff length λc is a high-pass filter applied to the measured surface profile. Wavelengths longer than λc are filtered out as waviness; wavelengths shorter than λc are retained as roughness. Standard cutoff selections per ISO 4288:1996: 0.08 mm for Ra below 0.02 µm (ultra-fine), 0.25 mm for Ra 0.02-0.1 µm (fine polished, lapped), 0.8 mm for Ra 0.1-2 µm (standard workshop default — covers most machined surfaces), 2.5 mm for Ra 2-10 µm (rough turning, milling), 8.0 mm for Ra above 10 µm (very rough surfaces, castings). The default workshop cutoff of 0.8 mm with 4 mm evaluation length (5 sampling lengths) works for the vast majority of machined surfaces. Never choose shorter cutoff just to fit a short surface — that changes the definition of roughness and makes readings non-comparable.
How often should a surface roughness tester be calibrated?
Annually is the standard interval for daily-use testers in production environments and quality control departments. NATA-accredited calibration laboratories in Australia provide traceable certification — Mitutoyo Australia and several specialist labs offer this service for the Surftest range. Between annual calibrations, the tester should be checked against a traceable reference specimen before precision work — this catches drift early and validates the tester's current operating condition. The calibration interval should be specified in the workshop's quality management system (ISO 9001 / IATF 16949 / AS9100) and adhered to rigorously — a tester past its calibration due date invalidates the traceability of every measurement that traced back to it.
How do I spot a counterfeit Mitutoyo Surftest?
Run the five-check procedure: (1) verify the serial number through the authorised Mitutoyo Australian distributor; (2) inspect the LCD display and touch panel for consistent backlighting, sharp text and reliable touch response; (3) check the stylus pickup assembly for precision diamond tip geometry; (4) verify firmware version supports current ISO 21920-2 standards; (5) confirm the unit ships with a calibration certificate traceable to NKO/JCSS/NMIJ. Buying through an authorised distributor like AIMS provides supply chain verification that eliminates the counterfeit risk. Counterfeit listings sometimes appear under deliberate misspellings — Mitutogo, MITU-tyo, Mituttoyo, Mito_tuyo — designed to bypass marketplace keyword filters. Genuine Mitutoyo is always spelled Mitutoyo, capitalised, with no hyphens.
