Ra and Rz on a Parts Report: Read the Measurement Conditions
Share
A surface-roughness value is the output of a defined measurement and processing chain. The parameter, filter, evaluation length, trace direction, stylus and standard edition all shape what the number means. Two reports can display the same Ra value yet provide different evidence, while an Ra result cannot be turned into Rz through a fixed multiplier.
This review concerns profile surface-texture report comparability. The existing guide to track-roller floating seals and lubrication provides product context. It does not create a universal seal finish, and this article supplies no finish value. The controlled drawing, specification operator and measurement plan govern the actual surface.
Confirm which roughness parameter was specified
Start with the part, drawing number and revision and identify the exact feature or surface. A report for a convenient coupon or nearby flat face does not automatically represent a finished curved journal, seal interface or machined flank. Mark the trace locations on a controlled view so each value can be connected to the specified feature.
Determine whether the drawing specifies a profile parameter or an areal parameter. This article addresses profile Ra and Rz. Retain the exact symbol, including any qualification defined by the invoked notation and standard. Similar names in a three-dimensional areal report belong to another measurement framework and should not be substituted by visual resemblance.
Ra and Rz describe different aspects of a profile. Ra expresses an arithmetic mean of profile deviations under its defined operator, while Rz represents a height-related parameter under its applicable definition. Their relationship changes with profile shape. A report containing only Ra cannot prove an Rz requirement through a fixed ratio, and an Rz value should not be relabeled as Ra.
Record the parameter-definition standard and edition. The official page for ISO 21920-2:2021 states that it defines profile surface-texture terms and parameters and replaced withdrawn ISO 4287:1997 and related documents. Parameter-name continuity does not guarantee that every definition, default or operator remained identical across editions.
Transcribe the limit type and value and the units. Preserve whether the callout establishes an upper, lower or bilateral requirement and any specification operator attached to it. Record the number of traces and whether the decision uses every result, a maximum, an average or another defined statistic. Do not assume the reporting convention from a single summary row.
Include the manufacturing condition: as machined, treated, coated, lapped or another controlled stage. Identify the acceptance authority. A bare or ambiguous callout needs technical resolution before the report is interpreted; it is not a reason to choose familiar default settings after the measurement.
Read the filter and measurement lengths
Profile roughness is separated from shorter- and longer-wavelength content through a defined specification operator. Record the nesting index or cutoff terminology required by the invoked edition, along with the filter type and settings. Do not copy a familiar cutoff from an older chart or instrument default without showing that it applies to the drawing's standard and surface.
Keep the length terms separate. A sampling or section length, the evaluation length and the instrument's traversing length can be different quantities. The stylus may travel beyond the evaluation region to allow run-up, run-down or filtering behavior. A report that lists only total travel does not necessarily state the length over which Ra or Rz was evaluated.
Record the number of sections or sampling lengths combined in the evaluation and any form-removal operation. Identify the profile type delivered to the parameter calculation. Where a legacy method explicitly distinguishes periodic and aperiodic profiles for a selection step, preserve that decision and its evidence; do not carry it forward as an unstated current ISO 21920 rule.
A Mahr metrology guide explains, in a legacy ISO 4288 context, why cutoff selection and measuring, evaluation and traversing lengths affect a result. It is useful for understanding the distinction, but its legacy table is not the current universal ISO 21920 operator.
Capture instrument software settings and state whether defaults were used. A default can be valid only when the invoked requirement allows it and the report identifies it. Preserve the raw or primary profile where the contract requires it or when a borderline result needs technical review. A recalculation from raw data must retain the software version, operator settings and revision trail.
Review exported reports for hidden normalization. Some templates show the final parameter while omitting the selected filter, evaluation region or form-removal choice. A screenshot of the setup screen can support the audit when controlled by the inspection record, but it should link to the trace and report rather than stand alone. If software was updated or a file was reprocessed, identify which version produced each displayed result.
Review measurement uncertainty near a specified limit using the agreed acceptance rule. Do not create a guard band or adjust the value informally. Identical Ra values produced with different filters or evaluation lengths are not automatically equivalent evidence, even if rounding makes them look the same.
Check the surface, direction and stylus conditions
Map each trace to the feature and record its direction relative to lay. A trace perpendicular to machining lay can encounter a different profile from one taken parallel to it. If the drawing or plan requires multiple directions, keep the individual results and their orientation labels. Do not combine them into an unlabeled average.
Inspect and document the surface condition. It should be clean and dry under the approved procedure, without handling residue that distorts the stylus path. Identify whether coating or treatment was present. Cleaning must not polish away the texture under review, and a pre-coating measurement cannot silently represent the coated final surface.
Record curvature, access and the contact-instrument configuration. A measurement on a flat surrogate does not automatically represent a curved product feature. Identify skid or skidless operation, stylus tip geometry and, where required, stylus force and instrument range. These settings affect what the system follows and resolves, but this article does not prescribe one configuration for every geometry.
Identify the instrument manufacturer, model and serial number and its applicable calibration or verification status. Document leveling and alignment, trace speed and relevant environment or vibration controls under the instrument procedure. A vendor's stated capability does not prove that a specific trace was valid.
A Keyence stylus measurement procedure illustrates surface cleaning, trace direction and selection concepts under cited ISO 4288-era rules. Treat it as a limited educational example. It does not demonstrate current ISO 21920 compliance or establish settings for the buyer's feature.
Flag a trace interrupted by a hole, edge, damage, steep geometry or loss of contact. Record repeatability by location and direction and keep all individual trace values. Repeating until a preferred value appears is not a repeatability study. The plan should say when a damaged trace is replaced and how the replacement is identified.
One valid trace does not automatically represent the complete feature. Surface texture can vary around a circumference, across machining passes or near process transitions. The drawing or approved plan should define the number and distribution of locations and the statistic used for acceptance. Where only a sample was measured, state that coverage plainly instead of describing the whole surface as conforming.
| Comparison state | Drawing and parameter | Measurement conditions | Review action |
|---|---|---|---|
| Normal: comparable | Part revision, feature, native Ra or Rz parameter, definition edition, limit, units and statistic are identified. | Filter, nesting index, evaluation length, location, direction, stylus, instrument and valid individual traces match the requirement. | Send the matched record to the authorized drawing-conformance review. |
| Missing: conditions incomplete | The callout is controlled, but the report omits its parameter definition or trace-to-feature link. | Filter, evaluation length, location, direction, stylus detail, calibration or raw trace needed for review is absent. | Request the missing evidence or a repeat measurement under an approved plan. |
| Conflict: operators differ | Ra is offered for an Rz requirement, or the drawing and report invoke different editions without an equivalence decision. | Filters, evaluation lengths, trace directions, surface states or locations differ between results being compared. | Preserve each native result and route the technical difference to the authorized specification owner. |
The normal state makes a result comparable; it does not release a lot. Missing and conflict states show why a bare number is insufficient. Neither state can be resolved with a fixed Ra-to-Rz conversion or a copied legacy cutoff.
Handle reports based on different standard editions
Record the edition invoked by the drawing and the edition used by the report. The ISO technical committee catalogue lists the published ISO 21920-1, ISO 21920-2 and ISO 21920-3 profile surface-texture family and shows their lifecycle. Those documents form the current published family at this research cutoff, even though their catalogue status indicates revision activity.
ISO 4287 and ISO 4288 have been withdrawn in the ISO catalogue context addressed by the newer family. A result made under an older edition is not automatically invalid when a legacy contract expressly invokes it. Contract precedence matters. At the same time, the report cannot be silently relabeled as ISO 21920 evidence simply because both systems use the names Ra or Rz.
For a legacy drawing, ask the authorized design or metrology owner whether to inspect under the invoked historical operator, revise the drawing, or approve a documented technical equivalence for the specific feature and decision. That review should compare definitions, filters, lengths, defaults and acceptance treatment, not only parameter names.
If repeat measurement is required, identify the part and sample, location, direction, instrument configuration and standard edition. Retain the original report and link the new result rather than overwriting it. A deviation or concession should state its scope, affected lot and approvers. Corrected reports need version control and a clear change record.
The final disposition stays with the authorized owner using the drawing, contract and complete measurement evidence. This article does not supply a universal finish for seals or undercarriage parts and does not establish a supplier measurement capability. Its endpoint is a native comparison panel that tells the reviewer whether parameter, edition, operator and physical trace conditions align, are missing or conflict.