Author: Site Editor Publish Time: 2026-07-22 Origin: Site
A low RMS value alone does not prove process capability. This guide explains how to make MRF results reproducible, comparable and suitable for a technical agreement.
Magnetorheological finishing (MRF) is commonly used to correct residual surface-form errors after grinding, conventional polishing or ultra-precision machining. Because the process is driven by measured error data, a calibrated removal function and controlled dwell time, it is generally described as a deterministic finishing process.
Deterministic, however, does not mean that one low RMS value proves identical capability for every material, aperture and surface geometry. A technically meaningful result must be interpreted together with the workpiece definition, evaluated aperture, fit and removed terms, spatial filtering, metrology chain, process configuration and acceptance rule.
MRF is a sub-aperture, controlled material-removal process. A magnetic field stiffens an abrasive-bearing magnetorheological fluid near the polishing wheel, creating a polishing ribbon with a measurable local removal response. Dwell time is then distributed across the surface according to the target error map.
MRF is normally positioned near the end of the manufacturing chain for final form correction and precision finishing. It is not normally a replacement for blank generation, deep grinding or bulk stock removal. If a component still has large geometric errors, significant subsurface damage, an unqualified material or coating system, or no reliable metrology, the upstream process must be addressed first.
The relevant question is therefore not simply “What final RMS can the machine achieve?” It is “On which workpiece, over which evaluated aperture, under which measurement and data-processing conditions, with how many correction cycles and which process configuration was the result obtained?”
1. Surface measurement: acquire the pre-process error map and confirm the coordinate system, evaluated aperture and nominal surface definition.
2. Data processing: register coordinates, apply the agreed mask, handle invalid points, fit the specified reference and remove only the agreed low-order terms.
3. Removal-function calibration: measure the local removal response for the defined wheel, fluid, wheel speed, magnetic condition and immersion state.
4. Dwell-time solution: calculate the dwell distribution from the target removal map, removal function and accessible tool path.
5. Deterministic correction: execute the verified trajectory while monitoring fluid condition, temperature and other critical process parameters.
6. Re-measurement and convergence: repeat the measurement using the same evaluation rules, quantify convergence and perform another correction cycle if required.
A change in any of these steps can change the reported result. In particular, before-and-after PV/RMS values may no longer be directly comparable when the aperture, removed terms or spatial filter is changed.
Field | What Must Be Stated | Risk When Omitted |
Workpiece & Surface | Material/coating, dimensions, surface type, radius or asphere equation | The result is incorrectly extrapolated to another material or geometry |
Evaluation Area | Clear aperture, mask, edge exclusion and data region | PV/RMS values from different apertures are compared as if equivalent |
Metric & Unit | PV/RMS units; base wavelength when λ is used | Wavelength-based data are converted to nanometres incorrectly |
Fit & Removed Terms | Nominal surface, tilt, power/defocus, astigmatism and other terms | Different low-order treatment produces substantially different values |
Spatial Filter | Sampling, cut-offs and analysis bandwidth | Low-, mid- and high-frequency errors are mixed or hidden |
Metrology Chain | Instrument, wavelength, reference, stitching and repeat measurements | The result lacks traceability and an uncertainty basis |
Process Conditions | Wheel, fluid, removal function, cycles and definition of process time | Process stability and production time cannot be assessed |
A publishable, comparable and contract-ready MRF result should state eight basic information groups.
Case | Processing results | Inference That Should Not Be Made |
Ø82 mm sphere-to-asphere correction | Three MRF cycles; published polishing time approximately 20 h; PV from approximately 10.992 µm to 142.538 nm and RMS from approximately 3.093 µm to 7.224 nm. If the before/after evaluation settings are confirmed as identical, the calculated reductions are about 98.70% PV and 99.77% RMS. | Do not present this as a guarantee for every Ø82 mm component or material. Approximately 20 h should not automatically be treated as the complete project cycle time. |
Ø170 mm steep spherical surface | The published record shows a pre-process RMS of 17.136 nm and a post-process result in the approximately 2 nm RMS class. | Multiple post-process values and incompletely aligned evaluation settings mean an improvement percentage should not be published. |
400 × 50 mm microcrystalline straightedge | The published result is PV 0.072 λ and RMS 0.008 λ. | The base wavelength is not disclosed, so the result must not be converted to nanometres or used to calculate a reduction. |
482 × 482 mm data region | The published result is PV 34.085 nm and RMS 4.048 nm over a data region of approximately 482 × 482 mm. | Do not describe this as a stable 4.048 nm RMS result over the full 1200 mm nominal machine envelope. |
These examples demonstrate correct data interpretation; they do not establish a universal guarantee across materials, apertures or surface geometries.
A nominal XY processing envelope describes the basic working space of a machine platform. It is not an unconditional full-aperture optical capability for every component. The usable aperture is also affected by curvature, local normal slope, workpiece thickness and mass, fixture geometry, polishing-head diameter, probe accessibility, rotary-axis configuration, cable and hose constraints, and collision-safe tool paths.
The published PKC-series nominal XY envelopes are 300 × 300 mm, 500 × 500 mm, 700 × 700 mm and 1200 × 1200 mm. For convex, concave, steep-slope and large aspheric components, the effective processing region must be confirmed through the workpiece model, fixture concept, polishing-head selection and trajectory simulation.
Industrial website copy should therefore use the term “nominal processing envelope,” not “guaranteed maximum optical aperture,” unless the latter has been validated for the specific project.
Platform | Published Nominal XY Envelope | Standard Axes | Standard Wheel | Correct Interpretation |
PKC300-Q | 300 × 300 mm | X/Y/Z/B/C | Ø50 mm | Small-to-medium aperture platform; usable range requires curvature, slope, fixture and head-access review |
PKC500-Q | 500 × 500 mm | X/Y/Z/A/B/C | Ø50 mm | 500 mm-class general platform; other wheels and steep-slope capability are configuration dependent |
PKC700-Q | 700 × 700 mm | X/Y/Z/A/B; optional C | Ø150 mm | Medium-to-large aperture platform; C axis and multi-head options are project specific |
PKC1200-Q | 1200 × 1200 mm | X/Y/Z/A/B; optional C | Ø300 mm | Large-aperture platform; support, gravity, thermal stability and large-aperture metrology must be defined together |
PV and RMS are commonly used for low-spatial-frequency surface-form deviation relative to a nominal or fitted reference surface. Ra and Rq are profile texture parameters, while Sa and Sq are areal texture parameters. They describe different characteristics and depend on different sampling and filtering conditions.
An optic may have very low roughness but unacceptable form error, or good form with problematic mid- or high-spatial-frequency texture. “Ra 0.5 nm” must not be interpreted as “surface-form RMS 0.5 nm,” and a single profile trace must not replace a full-aperture form evaluation.
Published results should report surface form, surface texture and any relevant mid-spatial-frequency metric separately, together with the instrument, objective, scan area, sampling and filtering conditions.
An executable MRF acceptance plan should define at least the drawing revision; material and coating; component dimensions, mass and datums; surface type, radius or asphere equation; clear aperture and mask; initial condition and target PV/RMS; surface-texture requirement; instrument and wavelength; reference fit, removed terms and spatial filter; stitching and reference calibration; definition of correction cycles and process time; measurement uncertainty and decision rule; raw-data retention, report identification and FAT/SAT responsibilities.
When a measured value is close to the acceptance limit, the decision rule should account for measurement uncertainty. Published specifications should also distinguish typical capability, representative case data and project-specific guaranteed values. Typical capability describes observed process performance; a representative case applies only to its recorded conditions; a project guarantee is established through sample validation and the signed technical agreement.
MRF is a strong candidate for high-value optical components that already have a controlled upstream surface and require deterministic correction of residual form error; plano, spherical, aspheric and accessibility-validated freeform optics; projects with local, edge or steep-slope correction requirements; and research or production tasks that need a traceable measurement–calculation–machining loop.
Additional process development is required when the project involves bulk stock removal, an unqualified material or coating chemistry, unstable support or thermal conditions, an unmeasurable clear aperture, or target regions that cannot be reached safely by the polishing head and metrology probe.
APEX begins with the workpiece rather than a standalone machine model. The recommended input package includes a 2D drawing or 3D model, material and coating, external dimensions and mass, surface definition and radius/asphere equation, clear aperture, maximum local normal slope, pre-process metrology, target PV/RMS and texture, upstream process, batch size and throughput, available measurement equipment and data format.
From these inputs, APEX evaluates the machine envelope, axis configuration, polishing head, workpiece support, removal function, metrology interface, trajectory accessibility and expected correction strategy. Final configuration and guaranteed values are established through drawing review, sample validation, the signed technical agreement and FAT/SAT records.
Tell us the material, approximate dimensions, surface geometry, incoming condition and target result. A short, non-sensitive summary is enough to start the equipment discussion.
MRF uses a calibrated local removal function and a calculated dwell-time distribution to correct a measured residual form-error map. Conventional polishing can also produce excellent surfaces, but local removal and convergence are often more dependent on process experience, tool condition and manual control.
No. RMS must be interpreted with the evaluated aperture, reference fit, removed terms, filtering and measurement uncertainty. RMS values generated with different analysis settings are not directly comparable.
No. The nominal processing envelope is the platform working space. The usable optical aperture depends on workpiece geometry, fixture, polishing-head and probe access, rotary-axis configuration, gravity and thermal stability, and the metrology plan.
Only when the base wavelength and the quantity definition are explicitly stated. If the wavelength is not disclosed, the result should remain in λ and should not be converted by assuming 632.8 nm or another value.
No. Ra is a profile surface-texture parameter. Surface-form RMS describes deviation from the nominal or fitted surface. They have different measurement scales, bandwidths and physical meanings.
· ISO 10110-5:2026: drawing indication of optical surface-form tolerances; nanometres are the preferred standard unit, while fringe or wavelength-based statements require the base wavelength to be stated.
· ISO 14999-4:2026: interpretation and evaluation of measured surface-form and wavefront-deformation data.
· ISO 10110-8:2019: drawing indication of optical surface texture; it does not replace low-spatial-frequency surface-form evaluation.
· ISO 14253-1:2017: conformity and nonconformity decision rules that account for measurement uncertainty.
· These standards are referenced as terminology, reporting and acceptance frameworks. Their mention does not claim certification or that every published case was executed under every clause of the standards.
