Six-step deterministic magnetorheological finishing closed loop, from residual error measurement and data registration through removal-function calibration, dwell solution, correction and re-measurement.
Magnetorheological finishing (MRF) is a deterministic corrective process used after grinding, conventional polishing or ultra-precision machining. A magnetically stiffened polishing ribbon produces a calibrated removal function; measured surface-error data are then converted into a dwell-time and path plan for controlled local material removal.
MRF is normally used to reduce residual form error and improve process convergence. It is not intended to replace blank generation or uncontrolled bulk removal. Successful implementation depends on a stable removal function, correct registration between metrology and machine coordinates, appropriate edge strategy, material/slurry compatibility and a repeatable measurement chain.
• Plano, spherical, aspheric, and general freeform optical components
• Convex, concave and high-local-slope regions after accessibility and collision review
• Fused silica, optical glass, selected crystals and other qualified substrates or coatings after process validation
• Residual form correction after grinding, conventional polishing, diamond machining or another qualified pre-process
• Removal-function stability over the required process time and material/slurry condition
• Registration of measured error maps to the workpiece and machine coordinate systems
• Dwell-time solution, path planning and control of edge behavior or excluded zones
• Head, fixture and probe accessibility on concave, steep or non-circular surfaces
• Comparable before/after metrology with fixed aperture, fit, removed terms, filtering and uncertainty
Acquire a repeatable surface map and confirm that the residual error is within the correction range of the selected head, material system and available stock.
Apply the agreed aperture or mask, nominal-surface fit, removed terms, filtering and stitching. Transform the error map into the workpiece and machine coordinate systems.
Generate and measure a representative spot or ribbon under controlled fluid temperature, wheel speed, immersion, pressure and material conditions. Define the stability metric used for acceptance.
Calculate dwell distribution and tool path, then check motion limits, head and fixture interference, edge treatment, excluded regions and expected residual.
Run the controlled polishing process with parameter and fluid-condition monitoring. Record the process revision, head/slurry configuration and actual execution time.
Measure using the same evaluation rules, compare predicted and actual removal, and repeat only where required by the agreed convergence and acceptance plan.
• MRF machine, polishing head, fluid-circulation unit and process software
• Removal-function calibration, dwell-time calculation, path generation and process parameter management
• In-machine non-contact measurement for posture identification, program verification, and error analysis where included in the selected configuration
• Workholding, support, measurement/data interface and material-specific process development
• Sample validation, FAT/SAT, operator training, maintenance and consumable planning as agreed
Model | Nominal XY envelope | Standard axes | Standard wheel | Typical applications |
PKC300-Q | 300 × 300 mm | X/Y/Z/B/C | Ø50 mm | Small-/medium-aperture, steep-surface and multi-product work |
PKC500-Q | 500 × 500 mm | X/Y/Z/A/B/C | Ø50 mm | General 500 mm-class platform; Ø50/Ø150 mm common heads |
PKC700-Q | 700 × 700 mm | X/Y/Z/A/B; C optional | Ø150 mm | Medium-/large-aperture, multi-head and automated preparation |
PKC1200-Q | 1200 × 1200 mm | X/Y/Z/A/B; C optional | Ø300 mm | Large-aperture gantry; project-specific support and metrology |
Note: Nominal XY envelope does not equal guaranteed effective optical aperture; actual range depends on geometry, fixture, head/probe access and the project configuration.
Published Typical Capability
Published PKC datasheets list a typical surface-form result of RMS ≤ 4.2 nm (approximately λ/150 at a wavelength of 632.8 nm) and a typical profile roughness of Ra ≤ 0.5 nm. These are conditional, typical process results—not unconditional guarantees. Applicability depends on material, initial form, effective aperture, head/MRF-fluid configuration, support, thermal environment, measurement bandwidth, and data-processing rules.
The acceptance plan should define the nominal surface, effective aperture or mask, wavelength where relevant, fit and removed terms, spatial filtering, stitching, repeated measurements, uncertainty and decision rule. Before/after values are comparable only when these conditions are consistent.
ISO 10110-5:2026 and ISO 14999-4:2026 may be used for surface-form specification and data interpretation; ISO 10110-8:2019 may be used for optical surface texture. Removal-function stability is not a single ready-made ISO metric and must be defined by the project, for example by peak depth, removal volume or another stated characteristic and statistic.
Machine-axis acceptance may reference ISO 230-1 and ISO 230-2, but axis positioning values do not directly equal finished-optic surface form. The optical sample and metrology chain require a separate acceptance plan.
• Nominal XY envelope is not the same as guaranteed effective optical aperture. Curvature, local slope, fixture, head size, probe access and collision clearance can reduce the usable range.
• Published local-slope or head-size capabilities are conditional. They do not mean that every workpiece can be processed over its full aperture without a geometry and interference review.
• Final guarantees are project-specific. Typical values and representative case results become contractual only when the workpiece, process configuration, measurement method and acceptance criteria are written into the signed technical agreement and FAT/SAT plan.
• Workpiece drawing or model, material, coating, size, mass and mounting references
• Surface type, radius/asphere equation, clear aperture and maximum local normal angle
• Initial and target PV/RMS/texture, including fit, removed terms, filtering and edge exclusions
• Available metrology, wavelength, stitching/null method and data format
• Pre-process, expected correction stock, acceptable number of cycles and target throughput
• Required wheel/head, slurry family, rotational axis, fixture, posture measurement and site utilities
Q: Is MRF a replacement for grinding or bulk polishing?
A: Normally no. MRF is a deterministic corrective process for controlled residual errors. The initial surface must be within the removal range and texture/defect condition agreed for the selected process.
Q: How is a PKC model selected?
A: Selection is based on nominal size and mass, curvature and local slope, concave/convex access, head size and removal function, fixture and probe clearance, required throughput, support strategy and metrology—not on diameter alone.
Q: How many MRF cycles are required?
A: There is no universal number. It depends on initial error, removal-function stability, convergence prediction, metrology repeatability and the acceptance target. The project plan should define the maximum or expected correction cycles.
